ebooksclub.org encyclopedia of physical science and technology 3e organic chemistry

P1: FPP 2nd Revised Pages Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology EN002C-64 May 19, 2001 20:39

Table of Contents (Subject Area: Organic Chemistry)

Article Authors Pages in the Encyclopedia

Acetylene Robert J. Tedeschi Pages 55-89

Alkaloids Armin Guggisberg and Manfred Hesse Pages 477-493

Bioconjugate Chemistry Claude F. Meares Pages 93-98

Carbohydrates Hassan S. El Khadem Pages 369-416

Catalysis, Homogeneous Piet W. N. M. van Leeuwen Pages 457-490

Fuel Chemistry Sarma V. Pisupadti Pages 253-274

Heterocyclic Chemistry Charles M. Marson Pages 321-343

Organic Chemical Systems, Theory Josef Michl Pages 435-457

Organic Chemistry, Synthesis John Welch Pages 497-515

Organic Macrocycles J. Ty Redd, Reed M. Izatt and Jerald S. Bradshaw Pages 517-528

Organometallic Chemistry Robert H. Crabtree Pages 529-538

Pharmaceuticals, Controlled Release of

Giancarlo Santus and Richard W. Baker Pages 791-803

Physical Organic Chemistry Charles L. Perrin Pages 211-243

Stereochemistry Ernest L. Eliel Pages 79-93

P1: GLQ Revised Pages Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology EN001F-4 May 25, 2001 16:2

AcetyleneRobert J. TedeschiTedeschi and Associates, Inc.

I. IntroductionII. Acetylene and Commodity ChemicalsIII. Production of Acetylene and Commodity

ChemicalsIV. Acetylene-Based Processes for Large-Volume

ChemicalsV. Important Chemical Uses for Acetylene

VI. Reppe ProductsVII. Specialty Acetylenics and Derivatives

VIII. Processes for Acetylene ProductionIX. Chemistry of Specialty ProductsX. Vinyl EthersXI. Flavor and Fragrance Compounds and Vitamins

A and EXII. Acetylenic PesticidesXIII. Acetylenic Reactions with Research and

Commercial PotentialXIV. Acetylene Research in Russia

GLOSSARY

Acetylenic Pertaining to organic compounds containinga triple bond ( C C ) or acetylene group in themolecule.

Adjuvant Acetylenic diol used with pesticides to en-hance activity, lower the rate of application, and in-crease safety.

Alkynol Primary, secondary, or tertiary acetylenic alco-hol with the hydroxyl group generally adjacent or α tothe triple bond.

Ammoxidation Acrylonitrile manufactured by oxidationof propylene in the presence of ammonia.

Aprotic Pertaining to highly polar solvents such asdimethyl sulfoxide (DMSO), hexamethylphospho-ramide (HMPA), N -methyl-pyrrolidone (NMP), andacetonitrile that can hydrogen bond or complex withacetylene and acetylenic compounds; used to dissolveand activate acetylene.

Carbonylation Reaction of acetylene with carbonmonoxide and alcohols to form acrylate esters; Reppeprocess.

Commodity chemicals Large-volume, multitonnagechemicals, some of which are derived from acetylene.

Ethynylation Reaction of acetylene with aldehydes andketones to form acetylenic alcohols and diols.

Grignard Organomagnesium halide used in acetylenicand other syntheses.

Oil-well acidizing Use of acetylenic alcohols (alkynols)as corrosion inhibitors to protect steel pipe duringacidizing operations undertaken to free oil from lime-stone formations.

Pesticides Agricultural chemicals including herbicides,insecticides, miticides, fungicides, and bacterial con-trol agents.

Reppe products and technology Pioneered by Dr.Walter Reppe of the I. G. Farben; various prod-ucts derived from the reaction of acetylene with

55

P1: GLQ Revised Pages


56 Acetylene

formaldehyde to yield butyne-1,4-diol and propargylalcohol.

Surfynols Acetylenic hydroxyl compounds used as high-speed, low-foam wetting–dispersing agents and asagricultural adjuvants and formulation aids.

Trofimov reaction Formation of substituted pyrrolesand N -vinylpyrroles by reaction of acetylene withketoximes.

Vinylation Reaction of acetylene with alcohols or pyrol-lidone to form vinyl ethers and vinylpyrrolidone.

ACETYLENE CHEMISTRY is the chemistry of thecarbon–carbon triple bond ( C C ). This functionalitydefines the unique chemistry of this reactive group, in ad-dition to its diverse and important applications. The highelectron density of the triple bond with its circular, sym-metrical π field makes acetylene and its derivatives reac-tive and useful intermediates for synthesizing a wide va-riety of organic compounds. These organic products findwide use in the synthesis of flavors and fragrances, vita-mins A, E, and K, β-carotene, pesticides, surfactants, cor-rosion inhibitors, and specially intermediates. This articledescribes the technology and applications of acetylene,acetylenic compounds, and the chemicals derived fromthem.

In the mid-1960s more than 1 billion pounds ofacetylene were used annually for the production oflarge-volume (commodity) chemicals. Since then, acety-lene has been gradually supplanted by less expensiveolefin feedstocks. However, acetylene is still used inmultimillion-pound levels to produce Reppe chemicals(butynediol, propargyl alcohol, butanediol, butyrolactone,N -methylpyrrolidone, polyvinylpyrrolidone, and vinylether copolymers) and specialty acetylenic chemicals andtheir derivatives. Large volumes of butanediol are used inthe manufacture of engineering plastics such as polybuty-lene terephthalate. Other significant uses for acetylene inspecialty areas include acetylene black, vitamins A and E,flavor–fragrance (F & F) compounds, corrosion inhibitors,acetylenic surfactants, and pesticides. Acetylenic chemi-cals, polymers, and derivatives of potential value in re-search and commerce are also discussed. Some specialaspects of acetylene chemistry research in Russia are alsosummarized.

I. INTRODUCTION

Acetylene has always been an important raw material formaking chemical products. In the early years of its history(circa 1890–1900) it was used extensively as an illuminantfor trains and city streets. It was soon realized that acety-

lene burned in the presence of oxygen provided a very hotflame, useful in the joining of metals. The welding indus-try today still uses significant amounts of acetylene in spiteof the availability of less expensive fuels such as propane.

Initially, acetylene was handled in industry as the undi-luted liquefied gas below its critical temperature of 36◦Cat a pressure greater than 600 psig. This appeared to bea safe procedure until a series of industrial explosionsin the early 1900s eliminated the practice. Today acety-lene is shipped in cylinders under pressure that contain amixture of diatomaceous earth or asbestos, acetone, andstabilizers. Another safe method of transport is via granu-lar calcium carbide in sealed containers, free of water. Thelarge-volume use of acetylene for the manufacture of com-modity chemicals has led to the building of plants, eitherpetrochemical or calcium carbide, “across the fence” fromacetylene producers. Today the factors leading to acety-lene hazards are well understood and have been well doc-umented. Acetylene now poses a minimum risk in well-operated processes and plants.

II. ACETYLENE AND COMMODITYCHEMICALS

From 1965 to 1970 more than 1 billion pounds of acetylenewere used annually to manufacture a variety of chemi-cals. Welding applications constituted ∼10% of this to-tal. After 1970, the use of acetylene for the manufac-ture of commodity chemicals began to decline markedly,and by 1979–1983 only 269 million pounds were em-ployed. Less expensive petrochemical raw materials suchas ethylene, propylene, butadiene, amylenes, and methanewere replacing acetylene. Table I summarizes chemicalsmanufactured mainly from actylene before 1965, types ofprocesses, and replacement raw materials. The principaluse of the monomers listed in the table was in diversepolymer applications spanning plastics, latex emulsions,rubbers, and resins. Chlorinated solvents were importantin vapor degreasing, but their use in more recent yearshas been gradually limited as a result of toxicity and airpollution.

The applications of the polymer products derived fromthe above monomers are not within the scope of this article.

III. PRODUCTION OF ACETYLENE ANDCOMMODITY CHEMICALS

The following tabulation shows the gradual decline in U.S.acetylene production from its high point of 1.23 billionpounds in 1965 to ∼269 million pounds in 1979. In 1965bulk acetylene was valued at 7–12c//lb, while ethylene was



Acetylene 57

TABLE I Early Acetylene-Based Chemicals

Product C2H2 process Replacement raw material Replacement process

Acrylates and acrylic acid Reppe carbonylation (CO + C2H2) Propylene (C3H6) Two-stage oxidation

Acrylonitrile C2H2 + HCN C3H6 Ammoxidation (C3H6–O2–NH3)

Chloroprene C2H2-Vinylacetylene-HCl Butadiene Chlorination and dehydrochlorination

Chlorinated hydrocarbons C2H2 + Cl2 C1–C3 feedstocks; C2H4 Chlorination–dehydrochlorination

Vinyl acetate C2H2 + acetic acid Ethylene (C2H4) Oxyacetylation

Vinyl chloride C2H2 + HCl Ethylene Oxychlorination

C2H2 + HCl C2H2 + C2H4 Balanced ethylene–acetylene

3–4c//lb. By 1983–1984 the cost ratio was approximatelythe same, with acetylene valued at about 55–75c//lb andethylene at 23–29c//lb.

Year C2H2 used (106 lb)

1965 1230

1967 1065

1969 1195

1971 852

1973 571

1976 490

1979 269

1984 286

From 1967 to 1974, 23 plants making such acetylene-based products as acrylonitrile, chlorinated hydrocarbons,chloroprene (neoprene), vinyl acetate, and vinyl chloridewere shut down. Sixteen of these plants manufacturedvinyl acetate and vinyl chloride. Table II presents acety-lene usage for various products in 1970, 1979, and 1984.

In 1984 total U.S. capacity for the production of acety-lene was estimated to be 384 million pounds. This produc-

TABLE II U.S Acetylene Usage: Large-VolumeChemicals

Product 1970 1979 1984

Acrylic acid and acrylates 70 16–45a 0

Acrylonitrile 42 0 0

Chloroprene (neoprene) 242 0 0

Chlorinated solvents 91 0 0

Vinyl chloride 268 100–110a 146

Vinyl acetate 158 37–52a 10

Reppe chemicalsb 41 73–80a 114

Other acetylenics and 10 14 26derivativesc

a Mainly butane-1,4-diol plus other Reppe products.b Acetylene black, vinyl fluoride, specialty acetylenics.c Estimated value.

tion is a mix from calcium carbide, by-product acetylenefrom cracking, and partial oxidation processes.

The nine U.S. acetylene producers, with their capacityin millions of pounds, were AIRCO-BOC, Calvert City,KY, and Louisville, KY (75); Dow, Freeport, TX (16);Hoffmann–La Roche, Nutley, NJ (5); Monochem,Geismar, LA (180); Rohm and Haas, Deer Park, TX (55);Union Carbide, Ponce, P. R. (12); Union Carbide; Seadrift,TX (12); Taft, LA (10); Texas City, TX (16).

The 1984 demand for acetylene was 286 millionpounds, and it was estimated to be 292 million pounds in1988. Growth from 1974 to 1983 was negative at −6.9%per year, while through 1988 it was slightly positive at0.5% per year. Hoffmann–La Roche generates acetylenefrom calcium carbide for use in the manufacture of vita-mins A and E and β-carotene.

Of all the commodity chemicals listed in Table II,vinyl chloride showed the least decline from 1970 to1984. In 1984 acetylene converted to vinyl chloride rep-resented 51% of total acetylene consumption. However,this production was from only one site, Monochem atGeismar, LA, and may be vulnerable in the future ifMonochem decides to convert completely to ethylene asraw material. The most promising growth areas for acety-lene in the near term are Reppe chemicals, particularlybutane-1,4-diol, used extensively in engineering plasticsand polyurethanes. Acetylene black and vinyl fluoride arealso specialty growth areas, as are acetylenic surfactantsand corrosion inhibitors. These acetylene-drived productsare discussed in greater detail in Section V.

A. Acetylene Production on a World Basis

Table III shows that, although U.S. acetylene productionis modest, acetylene usage worldwide is still significant,amounting to ∼1.9 billion pounds.

In the longer term, it is believed that worldwide acety-lene capacity and usage will gradually increase as oilprices escalate. Acetylene usage for such products as vinylacetate, vinyl chloride, Reppe chemicals, and specialty



58 Acetylene

TABLE III World Acetylene Usagea

Location Chemical Industrial Total

United States 282 114 396

Western Europe 814 200b 1014b

Japan 106b 100b 206

Others 200b 150b 350b

Total 1402b 564b 1966b

a In millions of pounds.b Estimated value.

acetylenics is expected to increase. Worldwide acetylenecapacity is spread over a wide geographic area, as shownin Table IV. The calcium carbide (CaC2) process, basedon coal and limestone, is still extensively practiced or ispresent as a backup capacity. In Russia there is probably alarge calcium carbide capacity that has not been reported.

The large-scale use of acetylene for the manufactureof commodity chemicals will be dependent on the costdifference between coal and oil and natural gas. It is certainthat sometime in the future oil and natural gas reserves willbecome limited and more expensive than coal. The targetdate is the early 21st century. Coal-based technologiessuch as the calcium carbide and AVCO (coal–hydrogenplasma arc) processes are prime candidates for large-scaleacetylene production. The AVCO process (Section VIII)has been studied successfully at the pilot-plant level.

IV. ACETYLENE-BASED PROCESSESFOR LARGE-VOLUME CHEMICALS

The processes summarized in the equations below wereimportant in 1940–1965 for producing commodity chem-icals. Below each acetylene-based process is shown itsreplacement process.

TABLE IV Worldwide Acetylene Capacity

CapacityCountry (Company, Location) (millions of pounds) Acetylene process

West Germany (BASF, Ludwigshaven) 176 Partial oxidation of natural gas

West Germany (Chem. Werke Huels, Marl) 264 Arc process–refinery gas

West Germany (BASF, Ludwigshaven) 13 By-product C2H2 from ethylene

Italy (Anic, Ravenna) 132 Naphtha cracking

Italy (Montedison, Porto Marghera) 154 Partial oxidation of natural gas

Japan (Denki Kagaku Kagyo, Ohmi) 200 Calcium carbide process

Japan (Igegana Electric Co., Ogaki) 616 Calcium carbide process

South Africa (African Explosives) 110 Calcium carbide process

Russia (Lissit Chansk) 77 Partial oxidation of natural gas

United States (Kentucky, Texas) 384 Calcium carbide, cracking by-product,partial oxidation

A. Acrylates and Acrylic Acid

1. Reppe Carbonylation

4C2H2 + Ni(CO)4 + 4C2H5OH + 2HCl −→CH2 CHCO2C2H5 + H2 + NiCl2

�H2O

CH2 CHCO2H + C2H5OH

t = 30–40◦C; atmospheric pressure.

2. Replacement Process: Two-Stage PropyleneOxidation

CH3 CH CH2 + 12 O2

A−−−−−→300−450◦C

CH2 CHCHOB−−−−−→

275−365◦CCH2 CHCO2H

↙ROH

Acrylate esters

A and B = fixed or fluidized-bed reactors.

B. Acrylonitrile

1. Acetylene–Hydrogen Cyanide

C2H2 + HCN −−−−−→40−600◦C

H2C CHCN

Fixed-bed process; catalyst, Ca(CN)2.

2. Replacement Process: Ammoxidationof Propylene

H3C CH CH2 + 32 O2 + NH3 → H2C CHCN + 3H2O



Acetylene 59

Fixed-bed or fluidized-bed reactors; t = 240–460◦C; typ-ical catalysts, Bi–P–Mo (Sohio), Mo–Vo–Bi (Snam),Mo–Te–Ce (Montedison–UOP), Se–CuO (BP-Dist.)

C. Chloroprene (Neoprene,2-Chloro-1,3-Butadiene)

1. Acetylene–Vinylacetylene–Hydrogen Chloride(Nieuwland–Carothers Process)

2C2H2CuCl−−−→

NH4ClH2C CH C CH −−→

HCI

H2C CH CCl CH2 + HCl

Liquid-phase process; t = 50–75◦C; atmosphericpressure.

2. Replacement Process: High-TemperatureChlorination of Butadiene

Cl2�→ 2Cl.

H2C CH CH CH2 + 2Cl.600◦C−−−→

H2C CH CCl CH2 + HCl

D. Chlorinated Hydrocarbons (Solvents)

1. Chlorination of Acetylene Followed byDehydrochlorination

C2H2 + Cl2 → CHCl2CHCl2−HCl−−−→

CHCl CCl2 −→Cl2

CHCl2CCl3↓�

Cl2C CCl2

2. Replacement Process

This includes the chlorination of hydrocarbon feedstocks(methane, ethane, propane, ethylene, propylene), whichrequires multiproduct technology.

E. Vinyl Acetate

1. Acetylene–Acetic Acid

C2H2 + CH3CO2H −−−−−→180−210◦C

H2C CHOCOCH3 + H2O

Fixed-bed reactor; catalyst, zinc acetate on carbon.

2. Replacement Process: Oxyacetylationof Ethylene

H2C CH2 + CH3CO2H + 12 O2 −−−−−→

175−200◦C

H2C CHOCOCH3 + H2O

Fixed-bed or fluidized-bed reactors; catalyst, palladium ormixture of noble-metal salts deposited on supports.

F. Vinyl Chloride

1. Acetylene–Hydrogen Chloride(Hydrochlorination)

C2H2 + HCl −−−−→90−140◦C

H2C CHCl

Fixed-bed reactor; catalyst, 10% mercuric chloride oncarbon.

2. Balanced Ethylene–Acetylene (Chlorination–Dehydrochlorination)

C2H4 + Cl2 → ClCH2CH2Cl → H2C CHCl + HCl

C2H2 + HCl → H2C CHCl

Overall: C2H4 + C2H2 + Cl2 → 2H2C CHCl

3. Replacement Process: Oxychlorinationof Ethylene

C2H4 + 2HCl + 12 O2 → ClCH2CH2Cl + H2O

C2H4 + Cl2 → ClCH2CH2Cl

ClCH2CH2Cl�→ H2C CHCl + HCl

2HCl + 12 O2 → Cl2 + H2O

Overall: 2C2H4 + Cl2 + 12 O2 → 2H2C CHCl + H2O

Fixed-bed or fluidized-bed reactors; oxychlorination,t = 450–500◦C (atmospheric pressure); catalyst, CuCl2–KCl on Kieselguhr; chlorination of ethylene, t = 50–140◦C (P = 4–10 atm); pyrolysis of ClCH2CH2Cl,t = 470–540◦C (P = 24–25 atm).

These acetylene-based processes are well-tested, high-yield, and high-selectivity operations, averaging well over90% of theory. The extensive use of these processes islikely once oil-based feedstocks become more expensivethan coal, which will be the preferred raw material foracetylene production.



60 Acetylene

V. IMPORTANT CHEMICALUSES FOR

As shown in Tables II and III, there has been steady growth

ACETYLENE

in the production of acetylene-based specialty chemicals.These products comprise acetylene black, vinyl fluo-ride, vinyl ethers, Reppe chemicals, and miscellaneousacetylenic alcohols and diols. Reppe products consistmainly of butyne-1,4-diol, propargyl alcohol (1-propyn-3-ol), butane-1,4-diol, butyrolactone, tetrahydrofuran,pyrrolidone, vinylpyrrolidone, N -methylpyrrolidone,polyvinylpyrrolidone, polyvinyl ethers, and copolymers.Butane-1,4-diol has emerged as the most importantuser of acetylene for acetylenic chemicals and derivedproducts. Its usage in specialty polymers and chemicalsincreased steadily during the 1977–1979 period from82 to 92% of total acetylene consumption (74 millionpounds in 1979). The principal uses for butanediol are inpolybutylene terphthalate and other engineering plastics,besides polyurethanes. By 1984 the total acetyleneusage for acetylenic chemicals had been estimated to be115 million pounds.

A. Acetylene Black

This unique, conductive form of carbon is made by theexothermic decomposition of acetylene:

C2H2800◦C−−→ 2C + H2.

The reaction is highly exothermic and the evolved heat(∼55,000 cal/g) is used to maintain the pyrolysis temper-ature. Preferred processes involve (1) continuous explo-sion of acetylene at 1–2 atm using an electric spark and(2) decomposition via the electric arc.

The principal uses of acetylene black are in the man-ufacture of dry cell batteries and the fabrication ofconducting rubber and plastic sheeting and molded plas-tics. Although acetylene black is too expensive to com-pete with petrochemical carbons in large-volume polymercompounding, it has carved out for itself a growing spe-cialty market in conductive applications. Acetylene con-sumption (in estimated millions of pounds) for acetyleneblack grew steadily from 1967 to 1984 in the United States:1967 (10); 1979 (19); 1980 (20); 1984 (23).

Producers of acetylene black are Shawinigan (Canada),Union Carbide (Ashtabula, OH, and Ponce, P. R.), andGulf (Cedar Bayou, TX). The Gulf production facilitywent on stream in 1979 and is rated at 15–20 millionpounds/year. The markets for batteries and conductingpolymer composites grew significantly from the mid-1980s, and acetylene black participated in this growth.The total production of acetylene black from all sources(millions of pounds) from 1981 to 1987 is estimated as

follows: 1981 (50); 1984 (55); 1987 (60). Union Carbideproduced ∼65% of this total, which was used internallyin its battery division.

B. Vinyl Fluoride

The principal use for this specialty monomer is the produc-tion of polyvinyl fluoride (PVF), a polymer considerablymore stable than polyvinyl chloride (PVC). The use ofPVF in exterior coatings for aluminum siding, steel build-ing panels, hardboard siding, and asbestos-impregnatedfelt roofing continues to grow. PVF was introduced byDu Pont in 1963 under the registered trademark Tedlar.In 1966 acetylene consumption for PVF production wasestimated to be ∼1 million pounds and by 1984 it was6–9 million pounds, showing steady growth.

Vinyl fluoride is produced by the addition of hydrogenfluoride to acetylene in the presence of a mercuric saltcatalyst deposited on carbon. The intermediate vinyl flu-oride is not isolated due to the difficulty of separating itfrom acetylene. The resulting difluoroethane (DFE) is inturn cracked to vinyl fluoride and the evolved hydrogenfluoride is recycled:

C2H2 + 2HF → [H2C CHF]Vinyl fluoride

→ CH3CHF2DFE

DFE�→ H2C CHF + HF.

The principal U.S. producers of vinyl fluoride and PVFare Du Pont (Louisville, KY) and Diamond Shamrock(Houston, TX). The intermediate DFE is also importantfor the manufacture of vinylidene fluoride, as shown inthe next section.

C. Vinylidene Fluoride (Vinylidene Difluoride)

Vinylidene fluoride (VDF), a monomer, can be producedfrom DFE via chlorination and cracking or from 1,1,1-trifluoroethane:

C2H2 + 2HF → CH3CHF2Cl2→

H3C—C(Cl)F2CDFE

�→ H2C CF2VDF

(1)

CH3CF3�→ H2C CF2 + HF (2)

At present the preferred starting material for VDFproduction is 1-chloro-1,1-difluoroethane (CDFE), alsoknown as refrigerant 142b. Numerous other intermediateshave been described for the preparation of VDF. The prin-cipal use for this monomer is the production of polyviny-lene difluoride (PVDF, PVF2). The polymer is semicrys-talline and has the following desirable properties; highmechanical and impact strength, resistance to most chemi-cals, very high dielectric constant, and excellent resistance



Acetylene 61

to ultraviolet and nuclear radiation and general weather-ing. It can be cross-linked via electron-beam radiation togive further improved properties.

It is used extensively in the interior coating of pipesand as a base for long-lasting decorative finishes on alu-minum, galvanized steel, and other architectural metals.Other applications include wire insulation and connectorsfor computer boards, telecommunications uses, and resis-tant coatings for harsh environments.

The U.S. suppliers of PVDF are Pennwalt, Solvay, andDynamit Nobel. Du Pont produces PVDF for its own in-ternal use, primarily to produce Viton plastic. The totalU.S. production of this polymer is significantly greaterthan that of PVF.

VI. REPPE PRODUCTS

Reppe products, a line of large-volume, acetylene-basedspecialty chemicals, bear the name of their discoverer,

FIGURE 1 Reaction flow diagram: Reppe chemicals. (1) 2-Butyne-1,4-diol; (1a) propargyl alcohol; (2) 2-butene-1,4-diol; (3) butane-1,4-diol; (4) tetrahydrofuran; (5) γ -butyrolactone; (6) 2-pyrrolidone; (7) N-vinylpyrrolidone;(8) polyvinylpyrrolidone; (9) N-methylpyrrolidone.

Walter Reppe of the I. G. Farben, who is one of the leg-endary figures of acetylene chemistry and its applicationto the production of valuable chemicals for Germany dur-ing World War II. Reppe’s research spanned the 1930sand reached commercial importance during the war yearsin the 1940s. Ethynylation technology to produce the keystarting materials 2-butyne-1,4-diol and propargyl alcohol(1-propyn-3-ol) is summarized in Section IX.B.

C2H2 + HCHOCuC2H−−−→

HOCH2 C CCH2OH2-Butyne-1,4-diol

+ HOCH2C CHPropargyl alcohol

Products derived from butynediol are 2-butene-1,4-diol,1,4-butanediol, γ -butyrolactone, N -methyl-2-pyrroli-done, 2-pyrrolidone, N -vinyl-2-pyrrolidone, polyvinyl-pyrrolidone (PVP), and polypyrrolidone. The reactionflow diagram in Fig. 1 shows the interrelationship ofthese products. Other Reppe chemicals such as vinylethers, polyvinyl ethers, vinyl ether–maleic anhydride



62 Acetylene

TABLE V Applications of Reppe Chemicalsa

Product Applicationsb

Propargyl alcohol Metal treatment; corrosion inhibition; oil-well acidizing; electroplating; intermediate for vitamin A and pesticides

Butyne-1,4-diol Important Reppe starting material; metal treatment; acid pickling; electroplating additive; stabilization ofchlorinated solvents; intermediate for pesticides

Butene-1,4-diol Intermediate for pesticides, pharmaceuticals, fungicides, and bacteriacides; polyurethane intermediate

Butane-1,4-diol Important Reppe intermediate for THF, butyrolactone, and PVP polymers; manufacture of polybutyleneterphthalate; polyurethanes, spandex fibers; specialty plasticizers

γ -Butyrolactone Intermediate for pyrrolidone and PVP; acetylene solvent (Sacchse); specialty solvent for polymers, lacquers,paint removers, and petroleum processing; intermediate for herbicides, azo dyes, and methionine

Tetrahydrofuran Continuous top coating of automotive vinyl upholstery, general solvent for resins; coating of cellophane withvinylidene polymers; polymers; polyurethane polymers; spandex fibers

Pyrrolidone Intermediate for N -vinyl and PVP polymers; powerful solvent for resins; acetylene solvent; polar reaction solvent;formulating agent; floor waxes; specialty inks; reaction intermediate

N -Methylpyrrolidone Commercial acetylene solvent; uses same as pyrrolidone; selective extraction solvent for butadiene purification(cracked naphtha); spinning synthetic fibers; surface coatings; pigment dispersant; formulating agent

N -Vinylpyrrolidone Intermediate for PVP polymers; functional monomer; lube oil manufacture; copolymer applications; paints,paper-coating adhesives, cosmetics; increased dye receptivity; pigment dispersant

a Reprinted from Tedeschi, R. J. (1982). “Acetylene-Based Chemicals from Coal and Other Natural Resources,” pp. 101–102, courtesy ofDekker, New York.

b THF, tetrahydrofuran; PVP, polyvinylpyrrolidone.

copolymers, acetylenic alcohols and diols, and poly-butylene terphthalate are discussed in subsequentsections.

A. Applications of Reppe Products

The uses of both butynediol and propargyl alcohol andtheir derivatives are quite diverse (Table V). They vary

TABLE VI Applications of Polyvinylpyrrolidone Polymersa

Product Compositionb Applications

Polyvinylpyrrolidone — Approved by the Food and Drug Administration for food and drug use; bloodplasma expander; nontoxic, biodegradable; varied uses

Polyclar AT Standard PVP grade Beverage clarification aid (fruit juices, wines, beer); chill proofing of beer;(white powder) complexing agent

Plasdone Pharmaceutical-grade PVP Tablet manufacture (granulating agent); tablet coating; liquid dosages; topicalpreparations; stabilizer, dispersant, drainage aid (syringes); skin creams, hairsprays, shampoos, general cosmetic use

Kolima adhesive Modified, clear high-solids Superior glue-line strength; sealing of smooth hard surfaces; grease andpolymers (Kolima PVP solutions chemical resistance; superior adhesion at temperature extremes; superior35, 55, 75) surface wetting; residual tack and viscosity control applications

Ganex V polymers Proprietary PVP polymers Coatings, detergents, pigment dispersants, plastic additives, textiles, petroleumapplications, protective colloids for vinyl lattices, improved freeze-thawstability and scrub resistance

Polectron (P) emulsion PVP copolymer lattices: P-130 Adhesives: remoistenables, pressure sensitise, heat-sealing applicationscopolymers (ethyl acrylate); P-230 Coatings: leather sizes and finishes, metal primers

(2-ethylhexylacrylate); P-430, Paper: board sizing, pigment binding, precoating, heat sealing450 (styrene); P825L, Textiles: fabric laminates, oil repellent finishes, permanent pressing, polyester845L (vinyl acetate) sizing

a Reprinted from Tedeschi, R. J. (1982). “Acetylene-Based Chemicals from Coal and other Natural Resources,” pp. 134–135, courtesy ofDekker, New York.

b

from metal treatment and corrosion inhibition to inter-mediates, specialty solvents for acetylene and polymers,top coating and spinning of fibers, preparation of speci-alty polymers and engineering plastics, vitamins A andE, and pesticides. The most important products in termsof production and sales are butanediol and tetrahydrofu-ran (THF), which are key building blocks for engineer-ing plastics, polyurethanes, and spandex fibers. Table VI



Acetylene 63

TABLE VII Applications of Vinyl Ethers and Polymersa

Product Composition Applications

Methyl vinyl ether CH3O CH CH2 Specialty monomer; intermediate; Gantrez resins

Ethyl, n-butyl, isobutyl, CH2H2 + alcohol → RO CH CH2 Specialty monomers; copolymers (adhesives); lacquers, paints,decyl vinyl ethers plasticizers, pour-point depressants; reaction intermediates

Gantrez M Polyvinyl methyl ether grades M-55, Adhesive formulations with high wet tack; adhesionM-154 (50% H2O) M-574 to plastic and metal surfaces; uses: adhesives, coatings,(70% solids in toluene) printing inks, textiles

Gantrez AN Copolymer of methyl vinyl ether and Polyelectrolyte in aqueous media; uses: hair sprays, shampoors,maleic anhydride grades AN-139, detergents, leather, paper coating, pigment grinding,AN-149, AN-169 nonwoven fabrics, topical remedies, tablet coating,

photoreproduction

Gantrez VC Copolymers of alkyl vinyl ethers Traffic, marine, and maintenance surface coatings; foundation(e.g., isobutyl vinyl ether) and sealants; corrosion-resistant paints; flame retardancy for vinylvinyl chloride substrates; primer coat for metals, paper coating

a Reprinted from Tedeschi, R. J. (1982). “Acetylene-Based Chemicals from Coal and Other Natural Resources,” p. 142, courtesy ofDekker, New York.

summarizes typical uses of polymers and copolymers de-rived from N -vinylpyrrolidone.

Propargyl alcohol is used extensively in the UnitedStates as a corrosion inhibitor in acid media for such appli-cations as oil-well acidizing and acid pickling of steel andelectroplating and as an intermediate for vitamin A andpesticides. These combined uses are estimated to com-prise about 3–5 million pounds annually.

Other Reppe derivatives such as γ -butyrolactone,pyrrolidone, N -methylpyrrolidone, and N -vinylpyrrolidone are used mainly as intermediates in the manu-facture of PVP. However, butyrolactone and N -methyl-pyrrolidone are used extensively as solvents for polymersand acetylene. Butynediol has minor uses in metal treat-ment and as a pesticide intermediate (Barban). However, itis very important as a starting point in the manufacture ofTHF and butanediol, the two largest volume Reppe prod-ucts. Both of these products have large-volume uses inthe manufacture of engineering plastics and polyurethanepolymers (see Table V and Sections VI.C and VI.D). PVPhas important uses in the clarification of beer, wine, andfruit juices, in pharmaceutical tablet manufacture, andin cosmetic manufacture. Its high degree of solubilityand compatibility makes it useful in numerous adhesiveapplications.

Vinyl ethers and derived polymers are summarized inTable VII. These products are particularly useful in formu-Table VII. These products are particularly useful in formu-lating adhesives and coating additives. The copolymer ofmethyl vinyl ether and maleic anhydride is sold by GeneralAniline and Film Corporation (GAF) under the nameGantrez AN series. These copolymers are important in hairsprays and shampoos and compete with the PVP polymersas versatile specialty polymers. Polyvinyl methyl ether and

copolymers have various uses as adhesives, particularly inhigh-tack formulations.

Vinyl ether monomers are prepared by the base-catalyzed reaction of alcohols with acetylene. With reac-tive alcohols such as methanol, ethanol, and isopropanol,the reaction can be carried out at atmospheric pressure:

C2H2 + ROH → RO CH CH2.

The original Reppe process involved heating thealcohol–acetylene charge at 150–180◦C using 1–2% KOHor sodium as catalyst under pressure, with a partial ni-trogen atmosphere. GAF has produced a variety of alkylvinyl ethers, but the most important is methyl vinyl ether,probably due to the commercial success of its copolymerseries with maleic anhydride. The vinyl ethers polymer-ize readily via free-radical catalysts to yield both homo-and copolymers. They are reactive molecules and canalso be used as intermediates to make a large variety ofproducts.

B. Butane-1,4-Diol and Tetrahydrofuran:Large-Volume Polymer Uses

Butanediol has evolved into the most important chemicalin the Reppe line on the basis of its uses in the manufac-ture of polybutylene terephthalate (PBT) and THF. It isalso important as a chain extender for polyurethanes. Iftotal acetylene usage for acetylenic chemicals in 1984 isestimated to be 115 million pounds and butanediol pro-duction is estimated conservatively to be 75% of this total,then at least 270 million pounds of butanediol were thenproduced.



64 Acetylene

C. Polybutylene Terephthalate

Although this engineering plastic competes with polyethy-lene terephthalate (PET), PBT is considered a supe-rior polymer because of its outstanding properties. It isproduced by ester interchange between dimethyl tereph-thalate and butanediol, while PET is derived from ethyleneglycol as the diol component. PBT is most widely used asa 30% glass-filled blend.

The desirable properties of PBT include excellentdimensional stability, superior strength and stiffness,exceptional moldability and low mold shrinkage, lowcreep at elevated temperatures, excellent electrical (dielec-tric) properties, and good chemical resistance.

It has end uses in the automotive industry: dielec-tric and heat-resistant uses to 300◦F; ignition systemcomponents—distributor caps, coil bobbins, rotors; lightsockets; emission control parts; body panels and louvers;transmission components. It is also used in electronicsfor connectors, switches, relays, cases, and covers. In themanufacture of appliances, PBT is a component of han-dles, small housings, impellors, and general engineeringparts; it also has high-voltage applications (resistance toarc tracking).

The annual growth rate of PBT from 1974 to 1977 was∼40%, and in 1984 it was estimated to be 10–15%. Ifacetylene usage between butanediol and THF productionis about equal, then PBT derived from acetylene could beof the order of 200 million pounds/year.

D. Tetrahydrofuran: Commercial Uses

Du Pont built a THF production facility in Houston in 1969based on Reppe technology. The capacity of the plant wasincreased from 50 million to 90 million pounds annuallyby 1973. This plant made Du Pont the largest producer ofTHF in the United States. By virtue of the uses summa-rized below, THF has been growing at an annual rate of7%; by 1979 U.S. production was ∼112 million pounds.

1. Continuous top coating of automotive vinylupholstery via THF–PVC solutions

2. Coating cellophane with vinylidene copolymers3. Coating of PVC and acrylic polymer sheets4. Protective coatings, film casting, printing inks,

adhesives5. Activating solvent for tetraethyllead manufacture,

specialty solvent6. Polymer intermediate for polyethers, polyurethane

elastomers, spandex fibers

To date the Reppe process for both butanediol and THFremains dominant.

VII. SPECIALTY ACETYLENICSAND DERIVATIVES

Specialty acetylenics and their derivatives have a varietyof uses which are described in the subsections that fol-low. Further details are provided in Sections IX–XII. Theproduction volumes and dollar sales of such products areoften proprietary and seldom exceed 2–15 million pounds.

A. Vitamins A, E, and K and β-Carotene

Acetylene is an important building block in the lengthysynthesis of vitamins A, E, and K and β-carotene, all ofwhich are important polyterpene compounds. Syntheticvitamin E, also known as dl-α-tocopherol, has emergedas one of the most important vitamins in the world to-day. Before 1950 its role in nutrition was in dispute, andit was not considered an essential vitamin. The synthesis,development, and manufacture of vitamins A and E werepioneered by Hoffmann–La Roche in Basel, Switzerlandand Nutley, NJ. Both vitamins are important in animal,human, and pet nutrition. The largest commercial marketsare animal feeds and supplements. In 1983 the productionof both synthetic and natural vitamin E was estimated tobe 11 million pounds with a value of about $164 million.Vitamin A production is about one-tenth that of vitamin E,and that of vitamin K is even less. Vitamin K is used almostentirely as an antihemorrhagic agent, while β-carotenemade from vitamin A (trans-retinol) is used as a safe foodcolorant. Since β-carotene is metabolized in mammalianorganisms partly to vitamin A, it is far more effective thanordinary yellow food dyes. The technology of vitamins Aand E and flavor–fragrance (F & F) compounds is sum-marized in Section XI.

B. Flavors and Fragrance Compounds

Although it is possible to produce a variety of F & F com-pounds by the reaction of acetylene with C5 to C8 ketonesfollowed by hydrogenation, only a small number of thesehave attained commercial importance, as shown in the fol-lowing subsections.

R1R2 C O + C2H2KOH−−−−−→

solvent

R1R2 C C CHH2→ R1R2 C CH CH2| |

OH OH

1. Linalool (3,7-Dimethylocta-1,6-dien-3-ol)

The fragrance compound linalool is prepared via theroute shown above by the reaction of methylheptenone



Acetylene 65

(6-methylhept-5-en-2-one) with acetylene to yield de-hydrolinalool (DHL, 3,7-dimethylocta-6-en-1-yn-3-ol),which is semihydrogenated to give the final product.

DHL is an important intermediate in the commercialsynthesis of vitamin A, while linalool is a key intermediatein the production of vitamin E (dl-α-tocopherol). Linaloolis one of the most important fragrance compounds in usetoday. It is used in many perfume compositions where adesirable floral-rose note is desired. In the past an impor-tant natural source of linalool was Brazilian rosewood. Ithas also had minor uses as a flavor ingredient. Besidesthe acetylenic route, large amounts of linalool are pro-duced by a turpentine-based intermediate, myrcene. TheGlidden–SCM Company produces a commercial grade oflinalool that is used extensively as a vitamin E intermedi-ate. The chemistry of DHL and linalool is summarized inSection XI.

2. Citral (3,6-Dimethylocta-2,6-dienal)

The isomerization of the alkynol DHL with vanadium cat-alysts yields citral, an important lemon F & F compoundused in many F & F compositions. Citral is also used asa vitamin A intermediate. Its condensation with acetoneyields pseudoionone, which on cyclization (H2SO4) formsβ-ionone, a key compound with the vitamin A ring systemand a valuable perfumery compound (Section XI).

CH3 CH3| |CH3 C CHCH2CH2 C CHCHO

3. β-lonone [4-(2,6,6-Trimethyl-1-cyclohexen-1-yl)-3-buten-2-one]

Besides the acetone–citral route just described, β-iononeis made by condensing the acetylenic alcohol DHL withdiketene or methylisopropenyl ether. Both routes yieldpseudoionone in comparable yields. The diketene routedescribed in Section XI yields a tertiary acetylenic ace-toacetate ester as the primary product, which is then rear-ranged with the loss of CO2 to yield pseudoionone. Thelatter compound is readily converted to β-ionone, withdilute sulfuric acid as catalyst. This terpenic ketone, inaddition to its use in the production of vitamin A, is animportant ingredient of floral (iris) perfumes with woodynotes. It also has minor uses as an ingredient in fruitflavors.

CHCH3

CHCOCH3

CH3

β-Ionone

CH3

C. Acetylenic Surfactants

The reaction of methyl ketones such as methyl ethyl andmethyl isobutyl ketones with acetylene via base-catalyzedethynylation yields tertiary acetylenic diols. These com-pounds containing the 1,4-dihydroxyacetylenic groupingat carbon chain lengths of eight or higher are superior

2R COCH3 + C2H2KOH−−−→

solvent

CH3 CH3| |R C C C C R| |

OH OH

surfactants. They are produced by Air Products and Chem-icals, Inc., and are sold under the trademark SurfynolTM.Typical products are listed in the tabulation below:

Trademark Generic name or description

Surfynol 82 3,6-Dimethyl-4-octyne-3,6-diol

Surfynol 104 2,4,7,9-Tetramethyl-5-decyne-4,7-diol

Surfynol 104-A Surfynol 104 + 50% 2-ethylhexanol

Surfynol TG Proprietary Surfynol surfactant

Surfynol 440 Ethoxylated (40%) Surfynol 104

These surfactants, particularly Surfynols 104 and 440,are superior wetting–dispersing agents with very low foamcharacteristics. They are used extensively in water-basedindustrial primers and finishes, in rinse-aid applications, inpigment grinding, and in agricultural formulations. Theygive synergistic performance with other surfactants interms of wetting and defoaming action, and in coatingformulations they yield latex paints that can be painteddirectly on oily surfaces without the finished paint crawl-ing or developing “fish eyes” on drying. These specialtysurfactants have enjoyed steady growth since 1974, anda 1984–1985 production estimate was 5 million poundsannually.

D. Corrosion Inhibitors

Acetylenic alcohols have been used extensively sincethe mid-1950s in oil-well acidizing operations to freetrapped oil in limestone formations. The acetylenic al-cohol (alkynol) is formulated with 15 to 28% HCl andpumped downhole via steel tubing (N-80, J-55) to thelimestone formation. Some acidizing operations attaindepths of more than 20,000 ft with temperatures greaterthan 300–400◦F. The alkynols, by virtue of complexing orreacting with the iron surface, minimize HCl corrosion and

Alkynol Chemical structure

Propargyl alcohol HO CH2 C CH

1-Hexyn-3-ol CH3(CH2)2 CH(OH) C CH

4-Ethyl-1-octyn-3-ol CH3(CH2)3 CH(C2H5) CH(OH) CH CH

OW-1 Proprietary acetylenic inhibitor



66 Acetylene

TABLE VIII Acetylenic Pesticides

Name Uses Starting material Pesticide structure

Kerb (pronamide) Control of grasses in lettuce production; other 3-Methyl-1-butyn-3-olCONH

Cl

Cl

C(CH3)2 C CHcrops: alfalfa, trefoil, endive, escarole

Barban (carbyne) Control of wild oats in production of spring 2-Butyne-1,4-diol NHCO2CH2 C CCH2Cl

wheat, durum wheat, barley, sugar beets,soybeans, lentils, flax

Omite (propargite) Acaricide with residual killing action; control Proporgyl alcohol(CH3)3C

O

OSOCH2 C CH

of many varieties of mites; very effectiveagainst spider mites; used on fruit (apples,apricots, oranges, peaches, plums) and fieldcrops (corn, potatoes)

extend the life of the tubing for years. These products havebeen recommended for the acid pickling of steel to removemill scale. This use is smaller than oil-well acidizing.

The alkynols used in these applications are propargyl al-cohol (1-propyn-3-ol), hexynol, ethyl octynol, and OW-1(proprietary alkynol-based product) (see table below).

The most effective acetylenic inhibitors are secondaryacetylenic alcohols of C6–C8 chain length, with the hy-droxyl group α to the terminal ethynyl group. Propargylalcohol, a primary alkynol, is somewhat less effective thanthe best inhibitors, but since it is a by-product of the Reppeprocess it is significantly less expensive than hexynol orethyl octynol. The total 1984–1985 usage of acetyleniccorrosion inhibitors was estimated to be about 3–4 mil-lion pounds with propargyl alcohol accounting for abouthalf of this total.

E. Acetylenic Pesticides and Adjuvants

Acetylenic compounds have found selective uses as her-bicides and miticides in the control of weeds and pests.It is likely that if more research and development effortwere extended to triple-bonded compounds, superior pes-ticides would be developed. Three acetylenic compoundsused commercially are Kerb (pronamide), Barban (car-byne), and Omite (propargite). They are summarized inTable VIII and Sections XII.A–XII.C in terms of theacetylenic starting material, pesticide structure, and uses.

It is believed that the use of Omite is considerablygreater than that of either Kerb or Barban. Surfynol104 (tetramethyldecynediol) and the Surfynol 400 series(ethoxylated S-104), while not pesticides, have been ex-tensively evaluated as adjuvants for use with crop herbi-cides such as atrazine, alachlor, basagran, sencor, treflan,and kerb. In most cases they have been found superior tostandard commercial surfactants of the nonionic and an-ionic types, if used with the above herbicides. They are

also effective when used with bacterial control agents forthe control of larva-worm infestations. They are used insignificant amounts as agricultural formulation aids.

VIII. PROCESSES FOR ACETYLENEPRODUCTION

Common raw material sources for the production of acety-lene are coal, natural gas, and liquid petroleum feedstockssuch as naphthas and light oils. Calcium carbide tech-nology using coal (coke) and limestone was the earliestmeans of producing acetylene. It is still used, but on alimited scale.

Petrochemical acetylene, based on natural gas and liq-uid feedstocks, began to emerge as a competitor of the car-bide process in 1940–1950. Until the mid-1960s acetylenewas still the favored raw material for such products as acry-lates, vinyl chloride, vinyl acetate, chloroprene, acryloni-trile, and chlorinated hydrocarbons. However, as steamcracking of C1–C3 alkane feedstocks to olefins (ethylene,propylene) became important, the role of both petrochem-ical and carbide acetylene declined markedly by the mid-1970s. These olefin cracking processes also produced by-product acetylene, which could be recovered by selectivesolvent extraction and sold or utilized for chemical pro-cesses by nearby chemical companies. Companies suchas Union Carbide, Tenneco, and Monochem have suchby-product acetylene facilities in Texas and Louisiana.

Table IX summarizes processes that were or are prac-ticed in the United States and elsewhere, with the excep-tion of the newer AVCO coal-based plasma are process.Coal-based processes such as calcium carbide or AVCOwill probably become important after the year 2000 as oilreserves become depleted.

Petrochemical acetylene processes are characterizedby a common principle. A hydrocarbon feedstock is



Acetylene 67

TABLE IX Acetylene from Natural Gas, Petroleum, and Coal Sourcesa

Typical process Principal feedstock Technology Typical companies

Electric arc Methane, gas oils Arc or plasma Huels, Du Pont

Sachsse Methane, natural gas mixtures Partial combustion (one stage) BASF, Dow, Monsanto

SBA Methane, natural gas mixtures (first stage); Partial combustion SBA, M. W. Kelloggnaphthas, heavier feedstocks (one and two stage)(second stage)

Wulff Natural gas, naphthas, heavier feedstocks Regenerative furnace pyrolysis Union Carbide, Wulff(four cycles)

Montecatini Natural gas, naphthas Partial combustion under pressure Montecatini, Diamond Alkali

By-product acetylene Ethane, hydrocarbons, naphthas, oil Steam cracking Major oil and chemical companies(EXXON, Shell, Dow, Union Carbide)

Calcium carbide Limestone (CaCO3) and coke CaC2 from C + CaCO3; C2H2 AIRCO-BOC, Union Carbidefrom CaC2+ H2O

AVCO Coal and hydrogen Hydrogen plasma AVCO (pilot and demonstration plants)

a Reprinted from Tedeschi, R. J. (1982). “Acetylene-Based Chemicals from Coal and Other Natural Resources,” p. 11, courtesy of Dekker, NewYork.

subjected to an intense energy source and thereby heatedto 1200–1500 K. By the use of very short residence times(0.01–0.001 sec) and quick quenching of the cracked gasto 550 K, acetylene and the starting feedstock can be re-covered. The recovery and purification of petrochemicalacetylene is a lengthy operation compared with the simplecalcium carbide process, which readily yields high-purityacetylene.

The petrochemical acetylene processes most likely tobe practiced today are the partial oxidation types. Promi-nent among these are Sachsse, SBA, and Montecatini pro-cesses. The SBA and Montecatini processes utilize eithernatural gas or naphtha feedstocks, while the Sachsse pro-cess is designed primarily for natural gas or methane, butcan be modified for naphtha. However, Sachsse technol-ogy has been widely practiced both in the United Statesand in Europe, showing it to be reliable and trouble free.

In contrast, the electric are process is more sensitiveto process variables, which can lead to the formation oflarge amounts of by-product carbon. The Wulff processwas once widely used in Europe, South America, Japan,and the United States. By controlling the feedstock andoperating conditions, the four-cycle regenerative processcould be made to deliver mainly acetylene or ethylene,making it more versatile than other petrochemical pro-cesses. The fact that the Wulff process is now seldom usedmay be due to the high efficiency of stream cracking ofalkanes to form ethylene and propylene at lower energyusage, thereby making the Wulff mixed product streamless attractive.

A process diagram of the BASF Sachsse (burner) reac-tor is shown in Fig. 2. Normally, with natural gas the burneris a one-stage reactor. However, it can be modified to usenaphtha by extension of the combustion chamber (second

stage) and an oxygen–steam off-gas mixture to provide amoist flame zone in the lower end of the burner to cracknaphtha to acetylene. About two-thirds of the hydrocar-bon feed is burned in the reactor to provide the thermalenergy needed to crack the remaining feed to acetylene.

The AVCO coal–hdrogen plasma process has not yetbeen scaled up to commercial production. However ina joint AVCO–DOE project (1980) at Wilmington, MA,a coal-fed hydrogen–plasma reactor capable of produc-ing 2 million pounds/year of acetylene was successfullydemonstrated, The pilot unit gave an acetylene yield of35% based on coal with low electrical usage. AVCO claims

FIGURE 2 BASF (Sachsse) burner. (1) Oxygen; (2) inert gas(safety purge); (3) methane or naphtha feed; (4) neck and mix-ing chamber; (5) auxiliary oxygen; (6) burner block; (7) reac-tion chamber; (8) water quench. [Reprinted from Tedeschi, R. J.(1982). “Acetylene-Based Chemicals from Coal and Other NaturalResources,” p. 21, courtesy of Dekker, New York.]



68 Acetylene

FIGURE 3 Coal–hydrogen (AVCO) plasma arc reactor. (1) Cath-ode and drive rolls; (2) hydrogen feed; (3) coal feed; (4) plenum;(5) reactor body; (6) anode; (7) field coils; (8) hydrogen quenchports; (9) rotating arc; (10) char and gas to cyclone. [Reprintedfrom Tedeschi, R. J. (1982). “Acetylene-Based Chemicals fromCoal and Other Natural Resources,” p. 30, courtesy of Dekker,New York.]

that this process can be scaled rto commercial size to giveacetylene commpetitive with or less expensive than ethy-lene used for commodity chemicals such as vinyl chlorideand vinyl acetate.

Figure 3 is a schematic diagram of the AVCO plasmareactor. By means of an external magnetic field, the arcis made to rotate and spread out radially in the reactor.The finely divided coal in contact with the arc area isactivated and reacts with the hydrogen plasma to formacetylene at a temperature gradient of 8000–15,000 K.Acetylene concentrations in the off gas of up to 16% (yield33 wt%) and energy consumption of 9.5 kWh/kg havebeen claimed. About 67% of the coal is consumed, andthe remainder is converted to char, which can be used asfuel for the process.

The use of AVCO coal-based technology for the large-scale production of acetylene will depend on future oilreserves and prices. At present, it does not seem likelythat coal will be a cheaper raw material than oil in theUnited States for this purpose. Also, the calcium carbideprocess, a competitive coal-based process for acetylene,has worldwide capacity with some of it in use. Its tech-nology and economics are well understood and have beenpublished, so that increased acetylene production can beimplemented if needed without excessive developmentand capital costs. An improved electric furnace process

for the production of 300 million pounds of acetylene peryear has been reported by SRI International, together withdetailed economics and process diagrams.

IX. CHEMISTRY OF SPECIALTYPRODUCTS

The chemical compounds discussed in this section areeither acetylenics or their derivatives. Often, acetyleneis a building block in the synthesis of a more complexmolecule and is not recognizable as such in the final com-pound. Products such as vitamins A and E examplify this.Specialty chemicals derived from acetylene have variedand selective uses, and their market size can vary fromthousands of pounds to millions of pounds per year.

A. Ethynylation Reaction

The introduction of an acetylenic group ( CH CH orC C ) was called ethynylation in Reppe days, and the

term has persisted. The reaction on a commercial scale isapplied to the reaction of acetylene with aldehydes and ke-tones to yield acetylenic alcohols (alkynols). The ethyny-lation of carbonyl compounds occurs via the followingroutes:

1. Reppe system: Reaction of acetylene withformaldehyde in the presence of a copper acetylidecatalyst under pressure

2. Lithium or sodium acetylide–liquid ammonia(atmospheric pressure)

3. Potassium hydroxide–acetylene–organic solvents(atmospheric to low pressure, 5 psig)

4. Catalytic potassium (sodium) hydroxide–acetylene–liquid ammonia (pressure, 50–400 psig)

5. Catalytic quaternary ammonium hydroxide catalyst,ethynylation of acetone under pressure (1000 psighydrostatic)

6. Liquid acetylene as solvent and reactant (pressure300–650 psig)

7. Grignard route: reaction of carbonyl compounds withacetylenic Grignard compounds (R C C MgBr)(atmospheric pressure)

These ethynylation methods are discussed in greaterdetail in Sections IX.B–IX.I.

B. Reppe Process for 2-Butyne-1,4-Diol andPropargyl Alcohol (1-Propyn-3-ol)

Cuprous acetylide deposited on silica-based carriers orsupports has proved to be a unique catalyst for the reactionof aqueous formaldehyde with acetylene under pressure



Acetylene 69

FIGURE 4 Butynediol–propargyl alcohol process. (1) Acetylene storage; (2) compressor; (3) recycle compressor;(4) pressure reactor; (5) vent; (6) separator; (7) pumps; (8) stripping still; (8a) recycle formaldehyde; (9) formaldehydestorage; (10) NaOH feed; (11) settling tank; (12) waste effluent; (13) butynediol storage; (14) recycle C2H2 streamfor vinylpyrrolidone production. [Reprinted from Tedeschi, R. J. (1982). “Acetylene-Based Chemicals from Coal andOther Natural Resources,” p. 105, courtesy of Dekker, New York.]

at 90–130◦C. The products resulting from this reaction,butynediol and propargyl alcohol, are formed in approxi-mately 9:1 ratio, with butynediol the predominant product.The reaction proceeds readily (exothermic) in total con-versions and selectivity of well over 90%.

HCHO + C2H2CuC2H−−−→ H C C CH2OH

Propargyl alcohol

+ HO CH2 C CButynediol

CH2 OH

Higher aldehydes or reactive methyl ketones fail to re-act in the Reppe system. Even a reactive aldehyde suchas acetaldehyde reacts very slowly and gives only an 18%conversion in 22 hr at 125◦C. In contrast, formaldehydereacts exothermically and is the basis for the continuousoperation in current Reppe plants. Flow diagrams of thebutynediol–propargyl alcohol process and the continuousprocess for the production of all Reppe products are shownin Figs. 4 and 5, respectively. The commercial importanceof these acetylene-based products and their estimated pro-duction volumes are discussed in Sections VI.A–VI.D.

C. Sodium and Lithium Acetylidesin Liquid Ammonia

Although alkali metal acetylides can be prepared in polarorganic solvents by the reaction of acetylene with alkalimetal dispersions, the preparation and synthetic use ofeither sodium or lithium acetylides in liquid ammonia hasproved to be the preferred route both in the laboratory and

in commercial practice. Below is shown the formation ofsodium acetylide in liquid ammonia and its reaction witha vitamin E intermediate, geranylacetone (GA), to yieldthe important terpene alkynol, dehydronerolidol (DHN):

1. Na + Liquid NH3

Fe(NO3)3−−−→−33◦C

NaNH2 + 12 H2 (3)

2. NaNH2 + C2H2NH3−→ NaC2H + NH3

3.

CH3|(CH3)2 C CH (CH2)2 C CH (CH2)2

GACOCH3 + C2H(NH3)

�

1.NaC2H2.H2O−NH4Cl (4)

CH3 CH3| |(CH3)2 C CH (CH2)2 C CH (CH2)2 C C CH|

OHDHN

(5)

This technology is used by Hoffmann–La Roche to pre-pare intermediates for the manufacture of vitamins A andE. The process has been in operation since the early 1950s.The use of lithium acetylide is preferred for more sensitiveor less reactive carbonyl compounds. Reactions (3)–(5)are run consecutively in liquid ammonia without isola-tion. The formation of the alkali metal acetylide (sodium,



70 Acetylene

FIGURE 5 Continuous process for the production of Reppe chemicals (GAF). (1) Crude acetylene; (2) scrubber;(3) compressor; (4) ethynylation reactor; (4a) formaldehyde; (5) separator; (6) vent; (7) still; (8) butynediol; (9) crudepropargyl alcohol; (10) propargyl alcohol still; (10a) to waste; (11) propargyl alcohol still; (12) batch hydrogenation;(13) hydrogen feed; (14) catalyst feed; (15) catalyst filter; (16) halogenation reactor; (16a) catalyst; (16b) halogen;(17) propargyl halide still; (18) butenediol batch still; (19) butanediol hydrogenator; (20) vent; (21) butanediol still; (21a)to butanol recovery; (22) butanediol still; (22a) waste; (23) final butanediol still; (23a) waste; (24) to dehydrogenationreactor; (25) dehydrogenation reactor; (26) separator; (27) to hydrogen recovery; (28) butyrolactone stills; (28a) waste;(29) pump; (30) pyrrolidone product reactor; (31) liquid NH3 or methylamine feed; (32) product stills; (32a) waste; (33)pyrrolidone feed; (34) catalyst feed; (35) catalyst preparation; (36) vinylation reactor; (37) compressor; (38) purifiedacetylene; (39) vinylpyrrolidone stills; (39a) waste; (40) polymerizer; (41) catalyst and distilled water; (42) purified air;(43) spray dryer; (44) vent; (45) iodine feed; (46) reactor. [Reprinted from Tedeschi, R. J. (1982). “Acetylene-BasedChemicals from Coal and Other Natural Resources,” p. 106, courtesy of Dekker, New York.]

lithium) is rapid and quantitative, and the formation of thealkynol averages more than 90%.

D. Potassium Hydroxide–Acetylene–OrganicSolvent Route

Favorsky, the Russian chemist, discovered that potassiumhydroxide ground in solvents such as ethers, acetals, andamines was an effective medium for reacting acetylenewith ketones and certain aldhydes to yield acetylenic al-cohols and glycols. This reaction system yields secondaryand tertiary acetylenic hydroxy compounds and is partic-ularly useful for the tertiary series. This technology is nowpracticed by Air Products and Chemicals via a continuousprocess whereby both acetylenic alcohols and diols areproduced in organic media.

The ethynylation reaction, in general, can be controlledvia acetylene concentration, reaction temperature, and the

ratio of base to ketone and acetylene. Alkynol formationis favored at lower temperatures, generally below 10◦C,while alkynediol formation is favored at 30–50◦C. Whatfollows is the formation of methylbutynol and dimethyl-hexynediol from acetone and acetylene:

(CH3)2 CO + C2H21. KOH−solvent−−−−−−−→2. H2O

(CH3)2 C C CH + (CH3)2 C C C C (CH3)2| | |OH OH OH

Acetylenic alcohols and diols produced commerciallyby this process are shown in the tabulation below.

Products with the (S-) designation are Surfynol sur-factants extensively used as nonfoaming wetting and dis-persing agents. The applications for these products arediscussed in Section VII.C.



Acetylene 71

E. Ethynylation in Liquid AmmoniaUsing Catalytic Amounts of KOHor Sodium Acetylide

Tedeschi showed that catalytic amounts of either NaOH orKOH in liquid ammonia–acetylene medium could affectthe ethynylation of aldehydes and ketones in high cat-alytic efficiency. The reaction was carried out at 20–40◦Cat acetylene–ammonia pressures of 150–300 psig. Themole ratio of reactants for methylbutynol formation aver-ages 6–18 mol acetone, 18–24 mol acetylene, 0.5–1.5 molKOH, and 18–24 mol ammonia. The conversion to methyl-butynol based on acetone averages 75–100%, with a cat-alytic yield based on KOH averaging 380–900%. Longerchain aldehydes and ketones react in the ammonia systemwith equal ease. The ethynylation of methylheptenone toDHL proceeds catalytically in more than 90% conversion.When stoichiometric to excess amounts of acetylene areused in the ammonia system, acetylenic diol formation isquite low (0.5–1.0%).

The reaction proceeds via the formation of an alkynol–KOH complex. the isolated crystalline complex can besubstituted for KOH in the ethynylation reaction withequal results. However, lithium hydroxide, unlike lithiumacetylide in liquid ammonia, fails to yield any acetylenichydroxy compound. The mechanism of catalytic ethyny-lation in liquid ammonia and aprotic solvents based onkey intermediates has been published by Tedeschi. Alkalimetal acetylides can be used as catalytic species in placeof KOH. Although acetylenic glycols can be formed in

Carbonyl compound Alkynol Alkynediol

Acetone Methylbutynol Dimethylhexynediol

Methyl ethyl ketone Methylpentynol Dimethyloctynediol (S-82)

Methyl isobutyl Dimethylhexynol Tetramethyldecynediolketone (S-61) (S-104)

Butyraldehyde Hexynol

2-Ethylhexaldehyde Ethyloctynol

liquid ammonia using KOH, the reaction is not catalytic,requiring stoichiometric to excess amounts of base. Thecatalytic liquid ammonia–KOH process has great poten-tial for economically producing secondary and tertiaryalkynols. Long-chain terpenoid alkynols used as interme-diates in the synthesis of vitamins A and E can be madein high conversions and catalytic efficiency in the am-monia system. Secondary alkynols, such as 1-hexyn-3-oland 4-ethyl-1-octyn-3-ol used commercially as corrosioninhibitors, are also readily formed. Methylbutynol, madefrom acetone and acetylene, is important commercially asa starting point for vitamins A and E, as a stabilizer forchlorinated solvents, as an intermediate for the herbicide

Kerb, and for the large-scale production of isoprene. In thelast-named use it is made catalytically in liquid ammonia,as described in Section IX.F.

F. Isoprene from Acetylene and Acetone

Anic (an affiliate of Ente Nazionale Idocarburi) atRavenna, Italy, has been operating a large production facil-ity for the manufacture of isoprene, which is used to makecis-polyisoprene. The route developed by Anic involvesthe catalytic formation of methylbutynol (MB), semihy-drogenation to methylbutenol (MBe), and dehydration ofthe latter to isoprene:

1. (CH3)2 CO + C2H2NH3−−−−−→

50% KOH(CH3)2 C C CH|

OHMB (6)

2. MB + H2

Pd catalyst−−−−−→ (CH3)2 C CH CH2|OHMBe

(7)

3. MBe−H2O−−→ H2C C(CH3) CH CH2 (8)

Methylbutynol is formed continuously in liquid am-monia under pressure using 50% aqueous KOH as cat-alyst. The conversions to both MB and MBe are closeto 100%, and the overall yield to isoprene is well over90%. The process delivers high-purity isoprene, readilypurified for use in the isotactic polymerization of iso-prene to cis-polyisoprene. A process flow diagram of theAnic methylbutynol–isoprene process (SNAM Progetti)is shown in Fig. 6. The plant originally had a capacity of66 million pounds of isoprene per year but is believed tohave been expanded.

G. Grignard Route

The use of Grignard reagents to form acetylenic productsis a relatively expensive route used primarily in the phar-maceutical industry to produce drugs or vitamin interme-diates. The acetylenic Grignard reagent is readily formedin 100% yield by reacting ethylmagnesium bromide witha terminal acetylenic compound in THF media:

H5C2 MgBr + R C CH → R G C MgBr + C2H6

The acetylenic Grignard is used in situ and is best em-ployed with sensitive aldehydes or ketones, which are af-fected adversely by basic reagents. An intermediate stepin the manufacture of vitamin A involves the formationof the 1,5-acetylenic di-Grignard, which is then reacted



72 Acetylene

FIGURE 6 Anic (SNAM Progetti) process: isoprene from acety-lene and acetone. (A) Methylbutynol; (B) methylbutenol andisoprene; (C) isoprene purification. (1) Acetylene feed; (2) re-frigerant; (3) ammonia condenser; (4) acetone feed; (5) KOHcatalyst feed; (6) stopper; (7) refrigerant; (8) acetylene con-denser; (9) ethynylation reactor; (10) flash tank; (11) steam;(12) NH3 + unreacted C2H2; (13) unreacted acetone; (14) re-frigerant; (15) acetone recovery column; (16) water; (17) heavyends column; (18) cooling water; (19) aqueous methylbutynol tohydrogenation; (20) hydrogen feed; (21) inhibitor feed; (22) hy-drogenation reactors; (23) fresh catalyst and catalyst feed;(24) unreacted hydrogen; (25) catalyst separator; (26) methyl-butenol (MBe) receiver; (27) MBe evaporator; (28) dehydratiorreactor; (29) fuel gas; (30) raw isoprene to purification; (31) wash-ing tower; (32) water to washing tower; (33) water distillation;(34) extractive distillation column; (35) extraction solvent;(36) purge line; (37) isoprene stripper; (38) pure isoprene; (39) re-ject isoprene from polymerization plant. (Reprinted from Tedeschi,R. J. (1982). “Acetylene-Based Chemicals from Coal and OtherNatural Resources,” pp. 164–165, courtesy of Dekker, New York.)

with a β-ionylidene (C14) aldehyde to yield an acetylenicdihydroxy vitamin A precursor, as illustrated below. Ingeneral, the yields are high and the conditions mild forthe Grignard route. When acetylene is reacted with ethyl-magnesium bromide, both hydrogen atoms are replacedand the resulting acetylenedimagnesium bromide is ex-clusively formed. This di-Grignard can be used to formsensitive acetylenic diols (see Scheme 1).

H. Use of Basic Quaternary AmmonHiumHydroxide Resins

Considerable effort has been expended on the develop-ment of basic resins as replacements for KOH or NaOH inthe catalytic ethynylation of carbonyl compounds. The ob-jective in the past has been to develop a fixed-bed continu-ous process in which neutralizaion of the reaction mixtureis not required and product isolation is much simplified.This type of process has been extensively studied in thecatalytic ethynylation of acetone to form methylbutynol.

The quaternary ammonium hydroxide resin is preparedfrom the chloride form of the resin by elution with sodiumhydroxide solution, followed by washing with water, thenmethanol washing, and finally vacuum drying. A seriousproblem associated with these resins is their degradationand loss of activity on continued use. Furthermore, alde-hydes react very poorly and give almost entirely aldol by-products. Ketones higher than methyl ethyl ketone react soslowly that the process is not economical. The use of liq-uid ammonia has been employed to activate the resin andreactants. However, due to the limited application of thisprocess, it is doubtful whether this technology is widelyused or whether it is used at all.

I. Use of Liquid Acetylene as Solventand Reactant

Liquid acetylene employed below its critical pressure(650 psi) and temperature (36◦C) was used at the bench-scale level, in batch reactors, to prepare such productsas alkali metal acetylides (lithium, sodium, potassium),methylbutynol, aminobutynes, and transition metal com-plexes. The work was carried out under 600 psig acetylenepressure in a special reactor system and pressure cubicle.The research showed that liquid acetylene could be han-dled safely in a variety of syntheses, with no exothermicdecompositions or explosions. Pure, white alkali metalacetylides were easily formed at 10–25◦C, and the forma-tion of methylbutynol took place in high conversions (80–85%) using only catalytic amounts of KOH and ammonia.This work by Tedeschi and Moore was never scaled up tocommercial practice.



Acetylene 73

CH2CH C

CH3

CH

OH

C C C CHCH2OH

HO CH2CH C

CH3

C CH2C2H5 MgBr

BrMgCH2CH C C CMgBr

CH2CH C CHO

CH3

CH3

CH3

SCHEME 1

X. VINYL ETHERS

The reaction of alcohols with acetylene to form vinylethers, commonly termed vinylation, takes place readilyat 150–180◦C using basic catalysts such as alkali metalhydroxides or alkoxides:

ROH + C2H2 → RO CH CH2

The vinylation process can be carried out at both at-mospheric pressure and under elevated pressures with anacetylen–nitrogen atmosphere. The formation of methylvinyl ether from methanol and acetylene can be carriedout continuously at atmospheric pressure and 180◦C. TheGAF Corporation has produced a variety of vinyl ethersderived from methoanol, ethanol, n-butanol, isobutanol,and decanol. The vinyl ethers find their most importantuse as the copolymers derived from methyl vinyl etherand maleic anhydride, known as the Gantrez AN series.Applications of these products and other polyvinyl ethersare summarized in Section VI.A.

XI. FLAVOR AND FRAGRANCECOMPOUNDS AND VITAMINSA AND E

The reaction flow sheets in Figs. 7–13 summarize therather lengthy chemistry and technology developed since1955 for the production of polyterpene compounds. Thischemistry is characterized by change and is highly propri-etary. Prominent companies in this field are Hoffmann–LaRoche and BASF. The choice of a given synthetic route de-pends on cost and availability of raw materials in additionto manufacturing sites.

The synthesis sequence pioneered by Roche andused for more than 30 years involves (1) ethynylation,(2) semihydrogenation, (3) reaction with diketene or

methylisopropenyl ether, and (4) allylic isomerizationwith loss of carbon dioxide. This four-step sequence isrepeated a number of times in the synthesis of both vi-tamins A (trans-retinol) and E (dl-α-tocopherol). Someof the intermediate products such as methylheptenone,linalool, pseudoionone, β-ionone, citral, geranylacetone,and nerolidol, are valuable flavor–perfumery chemicals.This circumstance increases the overall versatility andprofitability of the process by pro

The reaction sequence in Fig. 7 illustrates the followingviding diversification.

transformations:

Acetone + acetylene → methylbutynol

→ methylbutenol → methylheptenone

→ dehydrolinalool → linalool

This route and related routes are high-yield, high-selectivity steps of basic importance to the final economicsof the processes for vitamins A and E. The second se-quence, shown in Fig. 8, illustrates two processes for theproduction of citral from the alkynol DHL. The DHL–acetic anhydride route involves earlier technology, whichmay have been replaced by the direct isomerization ofDHL, the second process. Large amounts of citral arealso produced by SCM-Glidden via the oxidative dehy-drogenation of geraniol.

The flow sheet shown in Fig. 9 illustrates the followingprocess sequences:

DHL or citral → pseudoionone → α- and β-ionones;

linalool → geranylacetone → dehydronerolidol

→ nerolidol

Citral and DHL compete with one another in the prepa-ration of pseudoionone, the precursor of the ionones.Lemongrass oil and myrcene (from turpentine) are nat-ural sources of citral. β-Ionone is a very important



74 Acetylene

FIGURE 7 Linalool and dehydrolinalool from acetylene–acetone (Roche, BASF). Process conditions (average process yields underformulas): (1,5) catalytic ethynylation in polar solvents using KOH or sodium under pressure (t = 30–50◦C); (2,6) partial hydrogenationusing a palladium catalyst modified (t = 25–45◦C); (3) reaction of methylbutenol with diketene (DK)

CH2 C O

H2C CO

(t = 20–40◦C): catalyst, tertiary amine; MBe, acetocetate ester not isolated [see (4)]; (4) allylic rearrangement and cleavage of product(3) using aluminum isopropoxide under fractionation column.

FIGURE 8 Processes for the production of citral. Roche: Catalysts, siloxyvanadium moieties; excess tris(trimethylsiloxy) vanadium oxide;conditions, liquid phase, 2–20 hr at 125–140◦C; conversions 72–77%; yield 96%. Rhone Poulenc (Rhodia): Catalysts, vanadium alkoxides,excess cyclohexyl orthovanadate; conditions, liquid phase, 15 min to 4 hr at 125–160◦C; conversions 37–40%; yield 77–82%; OxidativeDehydrogenation of Geraniol (SCM-Glidden).

C

CH3

CH(CH2)2(CH3)2 C CHCH2OH Citral H2O−H

O�

Possibly vapor-phase process; copper-type catalyst; estimated yield 70–80%.



FIG

UR

E9

Hig

her

terp

ene

and

frag

ranc

eco

mpo

unds

deriv

edfr

omci

tral

,de

hydr

olin

aloo

l,an

dlin

aloo

l.D

K,

Dik

eten

e;E

AA

,eth

ylac

etoa

ceta

te.

avor

fl



76 Acetylene

FIGURE 10 Vitamin A production. Reaction flow sheet (steps 10–14).

intermediate, since it contains the correct isomeric ringsystem for vitamin A. Linalool, made from DHL via par-tial hydrogenation, is a key intermediate in the synthesisof vitamin E (dl-α-tocopherol) via the above compounds,leading to isophytol, the precursor of dl-α-tocopherol.

Vitamin A production is shown by the reaction flowsheets in Figs. 10 and 11: β-ionone (βI) → C14 aldehyde,and C14 aldehyde → vitamin A, respectively. β-Ionone re-acted with chloroethyl acetate via the Darzens reaction isconverted to the intermediate epoxy ester, which on rear-rangement yield C14 aldehyde.

The di-Grignard reagent (Fig. 10) of 3-methyl-pent-2-en-4-yn-1-ol (cis-1-ol) on reaction with C14 aldehyde(Fig. 11) yields the C20 eneynediol (C20 yn-D), whichcontains the correct ring system and carbon chain ofvitamin A. The important cis-1-ol compound is derivedfrom methyl vinly ketone and lithium acetylide, followed

by isomerization. The use of lithium acetylide is preferredsince sodium acetylide gives considerably lower yields.The remaining four steps of the process [(16)–(18)] areconcerned with the synthesis of the all-trans vitamin A(retinol).

The flow sheet in Fig. 12 shows the 12-step processleading to the C20 polyterpene alcohol isophytol (IP). Theformation of vitamin E is shown in Fig. 13, as are synthesesfor the second component, trimethylbenzoquinone(TMHQ). The condensation of IP and TMHQ yields dl-α-tocopherol (vitamin E).

Vitamin A is also manufactured by Wittig technol-ogy, starting with β-ionone. This route was pioneeredby G. Wittig and H. Pommer at BASF. A general syn-thesis step involves the condensation of terpene alcohols,such as vinyl-β-ionol, with triphenylphosphine in the pres-ence of HCl to yield quaternary phosphonium salts. These



Acetylene 77

FIGURE 11 Vitamin A production. Reaction flow sheet (steps 15–18): C14 aldehyde → vitamin A sequence.

salts in the presence of base undergo rapid elimination tophosphoranes (ylides), which react readily with carbonylcompounds to form olefins to high yield. By a repetitionof this sequence the carbon chain is built up to that ofretinol.

XII. ACETYLENIC PESTICIDES

Acetylenic compounds that have attained commercial sta-tus as pesticides are Barban (carbyne), Kerb (pronamide),and Omite (propargite). The last-named compound is amiticide, while the other two are herbicides. Their usesare outlined in Section VII.E. Their synthesis is summa-rized as follows:

A. Barban

HO CH2 C C CH2 OH

Cl CH2 C C CH2 OH

Cl

NH2 � COCl2�HCl

Cl

NCO

Cl

NHCO2 CH2 C C CH2 Cl

Barban

SOCl2 or PCl3

(A)

(B) A � B



78 Acetylene

P1: GLQ Final Pages


Acetylene 79

B. Kerb

(CH3)2 C C CHHCl

(CH3)2 C

Cl

C CHNH3

(CH3)2 C C CH

NH2OH

Cl

Cl

CONH C C

CH3

CH3

CH

Kerb

Base

Cl

Cl

COCl

C. Omite

(CH3)3 C OH � ONaOH

catalyst (CH3)3 CO

OH

(CH3)3 CO

OSO2 CH2 C CH (CH3)3 CO

OSOCl

SOCl2�C5H5N

omite

OC CH2 C CH

XIII. ACETYLENIC REACTIONS WITHRESEARCH AND COMMERCIALPOTENTIAL

The products and reactions described in this sectionhave not been extensively developed on a commercialscale but are believed to have potential in fine-chemicalsapplications.←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−

FIGURE 12 Vitamin E: reaction sequence leading to isophytol.

Abbreviation Common name Process step and reaction

MB Methylbutynol (0) Ethynylation of acetone

MBe Methylbutenol (1) Semihydrogenation (Pd)

DMAAA Dimethylallyl acetoacetate (2) Diketene (DK), ketenization

MH Methylheptenone (3) Allylic isomerization (Carrol)

DHL Dehydrolinalool (4) Ethynylation of MH

L Linalool (5) Same as (1)

GA Geranylacetone (6) Same as (2) and (3)

DHN Dehydronerolidiol (7) Same as (4)

N Nerolidol (8) Same as (1) and (5)

FA Farnesylacetone (9) Same as (2), (3), and (6)

HHFA Hexahydrofarnesylacetone (10) Complete hydrogenation, side chain

IP Isophytol (11) Same as (4) and (5)

A. Cyclic Carbonates from Alkynol–Alkynediol–Carbon Dioxide Reactions

An interesting reaction discovered by Dimroth andPasedach at BASF involves the reaction of tertiaryacetylenic alchols and diols with carbon dioxide underpressure to form substituted cyclic carbonates (5-methy-lene-4,4-di-alkyldioxolan-2-ones). The reaction is cat-alyzed by the combination of a cuprous salt and a tertiary



FIG

UR

E13

Fin

alco

nver

sion

step

tovi

tam

inE

(α-t

ocop

hero

l),T

MH

Q,t

rimet

hylh

ydro

quin

one;

TM

BQ

,trim

ethy

lben

zoqu

inon

e;P

SC

,pse

udod

ocum

ene.



Acetylene 81

amine and is generally carried out at 30–50 atm and tem-peratures of 80–120◦C.

(CH3)2 C C

OH

CH � CO2CuCl

N(C2H5)3

(CH3)2 C CH2

OC

O

C

OMB(CO2)

In the presence of excess, gaseous CO2 at 80◦C and800 psig, methylbutynol reacts in 12 hr to form the cycliccarbonate MB(CO2) in 90% yield. Liquid CO2 has alsobeen used successfully as both solvent and reactant be-low its critical temperature (35◦C) with secondary andtertiary acetylenic alcohols, 1-haloalkynols, and tertiaryacetylenic glycols. The mass action and solvent effect ofliquid CO2 makes it possible for these reactions to proceedreadily at lower temperature and pressures, at faster reac-tion rates, and with fewer by-products, particularly withsecondary alkynols. However, the reaction is unsuccessfulwith propargyl alcohol or secondary acetylenic glycols,illustrating an unusual reaction specificity. The reactionof dimethyloctadiynediol (oxidative coupling product ofmethylbutynol) and CO2 leads to the formation of a bi-functional carbonate with interesting potential in polymerchemistry. Acetylenic alcohols and diols reacted success-fully with liquid CO2 at 20–30◦C under pressure (500 psig)are listed in Table X. Compounds (a), (b), (c), (e), and (f)are tertiary alkynols or alkynediols, while (d) is a sec-ondary alkynol. The reaction gives highest yields with thetertiary series.

The versatility of the cyclic carbonates is illustrated bytypical reactions carried out by Pasedach and co-workers.The alkynol (MB) is reacted in situ to form the cycliccarbonate MB(CO2), which then undergoes further reac-tion in the presence of an active hydrogen reactant.

Hydroxy ketones can be formed via reaction with wa-ter, followed by loss of CO2. This is an alternative methodof hydrating a triple bond without the use of mercury

TABLE X Acetylenic Compounds Reacted withLiquid Carbon Dioxide

Yield of cyclicAcetylenic compound carbonate (%)

(a) 3-Methyl-1-butyn-3-ol 90

(b) 1-Chloro-3-methyl-1-butyn-3-ol 76

(c) 1-Ethynylcyclohexanol 87

(d) 4-Methyl-1-pentyn-3-ol 66

(e) 2,5-Dimethyl-3-hexyne-2,5-diol 93

(f) 2,7-Dimethylocta-3,5-diyne-2,7-diol 76

salts or expensive transition metal catalysts. Reactionwith alcohols in turn yields the corresponding methylketoopen-chain carbonate in high yield (88–90%). Ethynyl-diols result in good yield through the sequential conversionof the carbonate to the hydroxy ketone via water presentin the KOH, followed by ethynylation of the carbonyl pre-cursor. The substitution of excess primary amine for thetertiary amine cocatalyst results in good yields of disub-stituted methylene oxazolidones. The reactions are car-ried out at temperatures of 80–150◦C at pressures below40 atm.

MB(CO2)H2O−−−→

−CO2

(CH3)2 C COCH3(85%)|OH

(a)

(9)

MB(CO2)KOH−H2O−−−−→

THF(a)

C2H2−→CH3|

(CH3)2 C C C CH| |OH OH (10)

MB(CO2) + ROH −→ RO CO2 C(CH3)2 CO CH3

(11)

MB(CO2) + R NH2 −→ (CH3)2 C C CH2

O N R❙ �C

O

(12)

The cyclic carbonate from methylbutynol MB(CO2)can also be copolymerized with such monomers as methylacrylate, styrene, and acrylonitrile. The acrylate copoly-mer was considerably harder than the homopolymer ofmethyl acrylate and had a glass transition temperature of90◦C vs 9◦C for polymethylacrylate. By virtue of its car-bonate functionality, MB(CO2) is an interesting reactivemonomer that not only provides increased hardness andstrength to the polymer matrix, but can also be used tomake modified polymer structures by further reaction withthe reactive ring system.

B. Arylacetylenes from Methylbutynoland Aromatic Compounds

Selwitz and co-workers at Gulf Research and Devel-opment Company have disclosed interesting technol-ogy for the production of substituted arylacetylenes,where the ethynyl group is directly attached to the aro-matic nucleus. Arylacetylenes are normally quite diffi-cult to prepare by direct substitution. The important fea-tures of this process are the reaction of an active aryl



82 Acetylene

such as m-nitrobromobenzene with 3-methyl-1-butyn-3-ol (MB), using a catalyst composed of palladiouschloride and triphenylphosphine. The catalyst complex[PdCl2–P–(C6H5)3] is further activated with a catalyticamount of cuprous iodide, used as promoter. The reactionis carried out in an amine solvent at 50–100◦C. The re-sulting nitrophenylhydroxyacetylene is then cleaved withbase to the m-nitrophenylacetylene and acetone. The basecleavage can be carried out in situ, and overall yields are,on average, greater than 80%.

Br � HC C C

OH

(CH3)2catalyst

C C C

OH

(CH3)2

NaOH∆

C CH � (C CO( )

C CH

Η2�catalyst

NO2

NO2

NH2

NO2

H3)2

The arylnitroacetylene can be selectively hydrogenatedto the amino derivative, as shown above, using an un-supported ruthenium sulfide (RuS2) catalyst. The yieldsand selectivities of the above reactions are greater than90%. Other useful products made by this technology arem-bromophenylacetylene and m-diethynyl-benzene:

Br � HC C C

OH

(CH3)2

C C C

OH

(CH3)2

C C C (CH3)2CCC(CH3)2

OH

�

Br

Br

OH

Base cleavage of the above compounds yields m-bromophenylacetylene and m-diethynylbenzene. Thesearylacetylene intermediates are used in the preparationof stable, high-temperature polymers, as discussed inSection XIII.C. The m-bromophenylacetylene can be used

as a precursor to form m-aminophenylacetylene, whichis an important acetylenic capping agent for polyimidepolymers. This is an alternative method of preparing them-amino derivative in place of catalytic hydrogenation,cited above.

C. Acetylene-Capped Polyimide Polymers

The work of N. Bilow and co-workers under sponsor-ship of the U.S. Air Force and Hughes Aircraft Com-pany was responsible for novel, acetylene-terminatedoligomers that could be cured in the absence of cata-lysts at 200◦C to high-strength polymers. These acetylene-terminated polyimides cured readily without objectionalevolution of volatiles such as water or alcohol to yieldnonporous, void-free, high-strength polymers stable up to370◦C. Graphitereinforced composites cured with theseacetylenic polyimides and subjected to aging in air at320◦C for 1000 hr had 75% of their flexural strengthretained. The polymers were also observed to be supe-rior laminating resins for 6 A1-4V titanium in formingtitanium–titanium bonds stable at 500◦F. The initial appli-cation of these resins was directed primarily at the aircraftand aerospace industries.

The basic polymer structure of some of these uncuredacetylene-capped polyimides is as follows:

OC

CN

C

X

O

O

HC

CN

C O O CN

C

O

O

O

O

X

CN

C OC

O

O

CH

m

n

m

Polymer n X m

HR-600 1 CO 0

HR-602 2 CO 0

HR-650 1 CO 1

The uncured polymers are made by reacting aromaticdianhydrides with aromatic diamines to form the polyamic



Acetylene 83

acid precursor, which is then reacted with m-aminophenyl-acetylene to end-cap the prepolymer. Azeotropic removalof water (refluxing benzene or toluene) converts thepolyamic to polyimide acetylenic prepolymer. The latteris then cured at 200◦C.

Gulf Chemical Company purchased the technology forthese polymers and marketed the HR-600 product underthe trade name Thermid 600. The applications recom-mended for the product included aircraft and missile struc-tures, structural adhesives, low-friction bearings, nuclear-radiation-resistant parts, and circuit boards. The Thermid600 line included a polyimide molding powder (MC-600),an N -methylpyrrolidone solution of the polyamic polymer(50% solids, LR-600), and a fast-drying alcohol solutionof the polyamic ester form (Thermid AL-600). Gulf hassold this specialty acetylenic polymer business to NationalStarch.

D. Polyacetylene ( CH )x

McDiarmid and Shirakawa observed that the polymeriza-tion of acetylene in the vapor state on glass surfaces withZiegler–Natta catalysts, followed by doping with a vari-ety of additives such as halogens, antimony pentafluoride,perchloric acid, sodium, and lithium, resulted in greatlyincreased conductivity. A typical increase in conductivityfrom that of the original undoped polyacetylene to that ofthe doped polymer was of the order of 10−9 to 103 (doped)−1 cm−1.

Polymerization of acetylene at room temperature yieldsa mixed cis–trans product, approximately 70–80% cis iso-mer. Polymerization at low temperature gives an all-cispolymer. Heating the cis isomer to 200◦C for 1 hr con-verts it completely to the trans form. The conductivities ofthe undoped cis and trans isomers are essentially identical(cis, trans, 10−9), while both doped isomers give conduc-tivities up to 103 −1 cm−1 in resistivity.

Polyacetylene film is a fibrillar mat of randomly ori-ented strands. The dopants used are classified as elec-tron acceptors (bromine, iodine, arsenic pentafluoride)or electron donors (lithium, sodium, potassium). A poly-mer treated with donor or acceptor dopants yields n-typeor p-type semiconductors, respectively. Polyacetylenedoped with SbF5 or AsF5 is a better conductor of electricityon a weight basis than mercury. The doped, random filmmat can be stretched oriented by thermal and mechani-cal treatment to give conductivities as high as 2000 −1

cm−1. The conjugated polyene chain of polyacetylene andits formation of charge-transfer complexes with dopantsare responsible for the greatly enhanced conductivitiesobserved.

Polyacetylene has been given much publicity in pub-lications, trade magazines, and the press. Applications

for which it is claimed to have great potential are rechar-gable, lightweight automobile batteries and rechargable,portable energy-storage devices, solar batteries, radio andEMI shielding, fuel cells, Schottky devices, and wire-cableapplications. Companies that have carried out research andcommercial development activities with polyacetylene areRohm and 2Haas, Allied, BASF, Xerox, IBM, GTE, andShowa Denko. Rohm and Haas has actually produced pilotquantities of undoped polyacetylene film for commercialsampling.

To date, no significant markets have developed forpolyacetylene. The following problems are associatedwith its production and the utilization of doped polyace-tylene:

1. Difficulty in melt processing and extruding2. Deterioration and loss of conductivity in air and in the

presence of moisture3. Need for expensive processing equipment4. Lack of compatibility with other polymeric materials

Polyacetylene, on a weight basis, is an outstanding con-ductive material, but its chemical environments and con-ditions of use must be carefully controlled. New dopantssuch as hexafluorophosphide ion (PF−

6 ) render the poly-mer more stable to oxidation and hydrolysis, and furtherresearch along this line may be fruitful. Also, copolymer-ization with new acetylenic monomers or the formationof lower molecular weight, linear polyacetylenes may im-prove the melt processability of the polymer.

Polyacetylene doped with lithium (as the anode) isa highly conductive system, which because of poly-acetylene’s highly unsaturated (polyene) structure has theunique property of storing and releasing electrons. Thisproperty, together with the low unit weight of the cell,makes rechargable, portable batteries a commercial pos-sibility. However, such a cell is highly vulnerable to airoxidation and traces of moisture. Also, the other difficul-ties cited above for doped polyacetylene have not yet beenresolved. At some future time in new space–age applica-tions using the inertness of outer space, polyacetylene mayprove its worth.

The year 2000 Nobel Prize in Chemistry was awardedto Alan J. Heeger, Alan G. MacDiarmid, and HidekiShirakawa for the discovery of and development of con-ductive polymers, which among the first were the poly-actylenes.

XIV. ACETYLENE RESEARCH IN RUSSIA

Russia has maintained a strong interest in acetylene chem-istry since the discovery of the Favorsky reaction, in which



84 Acetylene

aldehydes and ketones in the presence of potassium hy-droxide and polar solvents yield acetylenic alcohols anddiols. A current site of Russian research is the Instituteof Organic Chemistry of the Russia Academy of Sci-ences, at Irkutsk, Siberia. B. A. Trofimov and co-workershave utilized KOH–dimethyl sulfoxide (KOH–DMSO) asa superbase medium for a series of novel, acetylene-basedreactions. Some of the reaction systems developed byTrofimov are described in this section.

A. Pyrroles and Vinylpyrroles from Acetyleneand Ketoximes

The reaction of excess acetylene with ketoximes in KOH–DMSO medium under pressure at 70–140◦C yields a mix-ture of pyrroles and N -vinylpyrroles. The reaction hasbeen studied intensively with aliphatic, alicyclic, and aro-matic ketoximes and has become known as the Trofimovreaction.

NR1

R2

�

CH CH2

CR1 CH2R2

NOH

� C2H2DMSO-KOH

NR1

R2

H

This method makes available many substituted pyrrolesnot readily synthesized. By using a limited or stoichio-metric amount of acetylene, the pyrrole can be made thepredominant product. The N -vinylpyrroles readily poly-merize by a free-radical mechanism using azobisisobuty-ronitrile as the initiator.

B. Oligomers from Acetylenein KOH–NH3–DMSO Media

Trofimov observed that low molecular weight oligomersof the polyene type could be formed from acetylene ina KOH–NH3–DMSO system at 80–120◦C under pres-sure (12–18 atm). The oligomers were deep brown toyellow solids, readily soluble in acetone, and insolublein hexane and melted with decomposition at 300–400◦C.The products by infrared examination contained C O,C O, and OH functionalities, but no nitrogen. Yieldsbased on acetylene consumed averaged 75–88%. Theoligomers were paramagnetic and conductive. Moderatelypolar amines such as triethylamine could be substituted forammonia with equivalent results. However, the use of ani-

line required the addition of water to suppress nucleophilicaddition of the amine across the triple bond.

Trofimov views oligomerization as proceeding simul-taneously via the following routes:

+ −✭N: + HC CH −→ ✭N CH CH (13)+ −✭N CH CH + C2H2 −→

+|✭N CH CH (CH CH)n−1 (14)

CH CH2|OH−

(CH3)2 SO + KOH −→←−+K + H2O∣

∣∣∣∣

CH3 S CH−2 ←→ CH3 S CH2

↓O O −

∣∣∣∣∣

(15)

I

−I + C2H2 ⇐⇒ K+|(CH3)2 SO|C CH2

II (16)

II + nC2H2 −→−

HC C |CH CH|n−1 CH CH |(CH3)2SO|K+

(17)

−OH− + CH CH −→ HO CH CH2

C2H2−−−→−CH CH (CH CH)n OH (18)

The products are interesting because of their polyene,linear structure, solvent solubility, and conductiveproperties.

C. Reactions of Triads: S8–KOH–DMSO,Se8–KOH–DMSO, Te–KOH–HMPA

A number of interesting transformations reported byTrofimov and co-workers were based on the reaction ofS8, Se8, and tellurium with acetylene in DMSO–KOHor HMPA (hexamethylphosphoramide)–KOH medium toyield the corresponding divinyl compounds:

C2H2 + 12 S8 + KOH + H2O

DMSO−−−−→80−120◦C

2(H2C CH)2 S

The reaction requires aprotic solvents such as DMSOand HMPA for satisfactory yields, with both solvents be-ing approximately equal in performance (DMSO, 80%;HMPA, 76% yields). A small amount of water is requiredin the reaction to function as a proton transfer agent.



Acetylene 85

The reaction of the selenium triad (Se–KOH–DMSO)with acetylene yields a significant amount of 2-vinyloxy-1,3-butadiene, besides the expected divinyl selenide:

C2H2 + Se8 + KOH + H2O −→(CH2 CH)2 Se + CH2 C CH CH2

O CH CH2

26% 20%

This by-product results from the competitive hydration–trimerization of acetylene vs vinylation. Also, the loweryield of divinyl selenide is believed to be due to the rapidoxidation of selenide ions by DMSO.

The tellurium triad gives a yield (54%) intermediatebetween the sulfur- and selenium-based reactions:

C2H2 + Te + KOH + H2OHMPA−−−−→

110−120◦C(H2C CH)2 Te

54%

The same reaction carried out in DMSO gives only a30% yield. The most important commercial aspect of thesereactions is the process for divinyl sulfide, since sulfur isan inexpensive abundant raw material.

(CH2 CH)2 S O

CH2 CH S CH2CH2 NR2

(R2N CH2CH2)2 S

O

30

NHR2 � EtOH

2NHR2

–50 C° O

D. New Reactions and ChemicalsBased on Sulfur and Acetylene

Trofimov’s research in acetylene–sulfur chemistry in-cludes the activation of anions and triple bonds with su-perbase systems, the reaction of sulfur nucleophiles withacetylene in superbase media, the reaction of substitutedacetylenes with chalcogen nucleophiles (sulfur, selenium,tellurium) in superbasic media, and prospective applica-tions. He predicts that in the future both acetylene and sul-fur will be sources of inexpensive raw materials for makingimportant acetylene–sulfur chemicals. Aprotic solventssuch as DMSO and HMPA are required for the high yieldsobtained.

Divinyl sulfide (DVS) is one of the most important prod-ucts discussed, since it can be formed in 98% yield fromhydrogen sulfide and acetylene:

H2S + C2H2KOH−DMSO−−−−−−→

80–100◦C(H2C CH)2 S

DVS (98%)

The reaction involving the addition of sulfide ion toacetylene may involve a one-step concerted mechanism,since all attempts to isolate monovinyl sulfide have failed.Either sodium sulfide or sulfur (cyclic S8) can be used asthe starting material, with the yields being somewhat lower(80%) than with H2S. However, the use of elementarysulfur has an economic advantage.

DVS has been studied extensively as a starting ma-terial for other products via additions across its doublebonds, cycloadditions, free-radical reactions, and elec-trophilic additions. A potentially important derivative isdivinyl sulfoxide (DVSO) made by the oxidation of DVSwith hydrogen peroxide:

DVSH2O2/H2O−−−−→

50◦C(CH2 CH)2 S → O

While DVS is inert to nucleophiles, DVSO is quite ac-tive in Michael-type additions, leading to numerous mono-and di-adducts based on reactions with amines, alcohols,thiols, and active C H compounds.

DVS readily polymerizes with free-radical catalystsand, as a cross-linking agent, yields macroreticularcopolymers, which have excellent ion-exchange proper-ties. DVSO has potential as a monomer in various copoly-mer systems and as a synthon for making new monovinylderivatives.

E. Vinylox (2-Vinyloxyethoxymethyl Oxirane)

Vinylox is a monomer with interesting commercial po-tential. It is produced by partial vinylation of ethyleneglycol followed by condensation of the resulting 2-vinyl-oxyethanol with epichlorohydrin.



86 Acetylene

FIGURE 14 Vinylox-based derivatives. X = O, S, C||O

O; X1 = O, S, N, C||O

O; n = 0–7.



Acetylene 87

OH(CH2)2OH � C2H2KOH

CH2 CHO(CH2)2OH

CH2 CHO(CH2)2OH � ClCH2CH CH2

O

NaOHCH2 CHO(CH2)2OCH2CH CH2

OVinylox

FIGURE 15 Hydroxyacetylenic esters and nitriles and their derivatives. Y = COOMe; R1 = R2 = alkyl, R1–R2–cycloaklylidene.

A process for 2-vinyloxyethoxymethyl oxirane hasbeen developed and piloted in Russia by Trofimov et al.This proprietary process for Vinylox production is nowbeing scaled up in Russia, since Vinylox has been shownby Russian scientists to be a highly active bifunctionalmonomer, reactive intermediate, and building block fordiverse applications. It provides direct and reliable routesto numerous compounds with commercial promise suchas epoxyacetals, vinyloxyethoxymethyl cyclocarbonates,and vinyloxyethoxymethyl thiranes. Some important



88 Acetylene

reactions and derivatives of Vinylox are shown inFig. 14.

New epoxy resins of unique structures and high puritywhich are nontoxic, noncorrosive, and possess lower thanexpected viscosities are readily formed from this versatilemonomer. The cured epoxy resin, in turn, exhibits higherstrength and greater flexibility in a variety of composite ap-plications. Polyols, glycols, carbohydrates, dicarboxylicand hydroxycarboxylic acids, polythiols, and other sulfurhydroxy compounds react readily with Vinylox to formthe corresponding polyglycidyl derivatives. Vinylox andits derived epoxides are used as adhesives, active diluents,plasticizers, and modifiers for diverse epoxide materials.Small additions of Vinylox to synthetic rubbers enhancetheir strength, elasticity, and water-freeze resistance. Thismonomer and some of its derivatives exert a thermostabi-lizing effect on PVC formulations. Vinylox can be copoly-merized with both radical initiators and anionic catalyststo give curable copolymers with n-butylvinyl ether, vinylacetate, N -vinylpyrrolidone, and other typical monomers.With cationic catalysts, Vinylox forms both soluble andcross-linked polymers, depending on conditions.

Potentially important derivatives of Vinylox are2-vinyloxyethoxy-methyl thirane (trade name Vinylox-S)and 3-(2-vinyloxyethoxy)-propylene-1,2-carbonate (tradename Cyclovin).

F. Hydroxyacetylenic Esters and Nitriles

Trofimov and co-workers have explored the chemistry ofγ -hydroxy-α,β-acetylenic esters and acetylenic nitriles,prepared by the following routes.

1. R2 C C CH + CO + MeOHCuCl2−−−→PdCl2|

OHR2 C C C CO2Me

|OH

2. R2 C C CHKOBr−−−→

OH

R2 C C C BrCuCN−−−→DMSO

R2 C C C CN

OH OH

Esters and nitriles derived from tertiary acetylenic al-cohols add smoothly to diverse O-, N-, and S-containingnucleophiles to yield hetrovinyl derivatives which undergointramolecular cyclization to dihydrofurans and other het-erocycles. From this chemistry, novel spirocyclicimines,lactones, and 1,3-oxathiolanes have been obtained fromsulfur containing nucleophiles (sulfide and rhodanide

ions), while acetylenic nitriles give new 1,3-thiazine sys-tems and polyconjugated iminodihydrofurans. Figure 15summarizes some of this novel chemistry, some deriva-tives of which exhibit pesticide and other bioactivity.

SEE ALSO THE FOLLOWING ARTICLES

CATALYSIS, INDUSTRIAL • PHARMACEUTICALS • PHY-SICAL ORGANIC CHEMISTRY

BIBLIOGRAPHY

Bilow, N., Landis, A. L., and Boschan, R. H. (1978). U.S. Patents4,098,767 and 4,100,138.

Bilow, N., Landis, A. L., and Boschan, R.H. (1982). SAMPE J.Brandsma, L., Vasilevsky, S. F., and Verkruijsse, H. D. (1997). “Appli-

cation of Transition Metal Catalysts in Organic Synthesis,” Springer-Verlag, Berlin/New York.

Copenhaver, J. W., and Bigelow, M. H. (1949). “Acetylene and CarbonMonoxide Chemistry,” pp. 246–294, Reinhold, New York.

Kirk, R. E., and Othmer, D. F. (1978). “Encyclopedia of Chemical Tech-nology,” 3rd ed., Vol. 1, pp. 192–276, Wiley-Interscience, New York.

McDiarmid, A. G., Park, Y. W., Heeger, A. J., and Shirakawa, H. (1977).J. Chem. Soc., Chem. Commun. p. 578.

McDiarmid, A. G., Park, Y. W., Heeger, A. J., and Shirakawa, H. (1978).Appl. Phys. Lett. 33, 180.

McDiarmid, A. G., Park, Y. W., Heeger, A. J., and Shirakawa, H. (1978).J. Am. Chem. Soc. 100, 1013.

McDiarmid, A. G., Park, Y. W., Heeger, A. J., and Shirakawa, H. (1979).Coat. Plast. Prepr. Pap. Meet. Am. Chem. Soc., Div. Org. Coat. Plast.Chem.

McDiarmid, A. G., Park, Y. W., Heeger, A. J., and Shirakawa, H. (1979–1980). Synth. Met. 1, 101–118.

McDiarmid, A. G., Park, Y. W., Heeger, A. J., and Shirakawa, H. (1982).Phys. Rev. B: Condens. Matter [3] 26, 2327–2330.

McDiarmid, A. G., Park, Y. W., Heeger, A. J., and Shirakawa, H. (1984).Mol. Cryst. Liq. Cryst. 105, 89–107.

McDiarmid, A. G., Park, Y. W., Heeger, A. J., and Shirakawa, H. (1984).Polym. Prepr. Am. Chem. Soc., Div. Polym. Chem.

Miller, S. A. (1965–1966). “Acetylene, Its Properties, Manufacture andUses,” Vols. 1, 2, Academic Press, New York.

Pasedach, H., Dimroth, P., and Schneider, K. (1961). German Patent1,098,953.

Pasedach, H., Dimroth, P., and Schneider, K. (1962). German Patents1,129,941 and 1,130,803.

Pasedach, H., Dimroth, P., and Schneider, K. (1963). German Patents1,135,894 and 1,145,632.

Pasedach, H., Dimroth, P., and Schneider, K. (1963). U.S. Patent3,082,216.

Roth, S. (1995). “One-Dimensional Metals,” VCH, Weinbeim/NewYork.

Salaneck, W. R., Lundstrom, I., and Ra❡

nby, B., eds. (1993). “Nobel Sym-posium in Chemistry: Conjugated Polymers and Related Materials:The Interconnection of Chemical and Electrical Structure,” OxfordScientific, Oxford.

Selwitz, C. M., and Sabourn, E. T. (1978). U.S. Patents 4,128,588.Selwitz, C. M., and Sabourn, E. T. (1980). U.S. Patents 4,204,078,

4,215,226, 4,216,341, 4,219,679, and 4,223,172.



Acetylene 89

Skvortsov, Yu. M., Fartysheva, O. M., Mal’kina, A. G., Sigalov, M. V.,and Trofimov, B. A., Zh. Org. Khim. 22, 255.

Skvortsov, Yu. M., Gritsa, A. I., Mal’kina, A. G., Sokolyanskaya, L. V.,and Sigalov, M. V. Sulfur Lett. 6(3), 87–92.

Skvortosov, Yu. M., Abramova, N. D., Andrijankova, L. V., Mal’kina,A. G., Kosinzina, E. I., Albanov, A. I., and Skvortsova, G. G. (1986).Khim. Hetherocycl. Soed. 10, 1412–1415.

Skvortsov, Yu. M., Trofimov, B. A., Mal’kina, A. G., and Fartysheva,O. M. (1985). Khim. Hetherocycl. Soed. 12, 1689–1690.

Tedeschi, R. J. (1963). J. Org. Chem. 28, 1740.Tedeschi, R. J. (1965). Ind. Eng. Chem. Prod. Res. Dev. 4, 236.Tedeschi, R. J. (1965). J. Org. Chem. 30, 3045.Tedeschi, R. J. (1968). Ind. Eng. Chem. Process Des. Dev. 7, 303.Tedeschi, R. J. (1969). J. Org. Chem. 34, 435.Tedeschi, R. J. (1969). U.S. Patents 3,441,621, 3,474,117, and

3,474,118.Tedeschi, R. J. (1970). Ind. Eng. Chem. Process Des. Dev. 9, 83.Tedeschi, R. J. (1970). U.S. Patent 3,541,087.Tedeschi, R. J. (1973). Ann. N.Y. Acad. Sci. 214, 40–61.Tedeschi, R. J. (1982). “Acetylene-Based Chemicals from Coal and Other

Natural Resources,” Dekker, New York.Trofimov, B. A. (1980). “Heteroatomic Derivatives of Acetylene—New

Polyfunctional Monomers, Reagents and Semi-Products,” Nauka,Moscow.

Trofimov, B. A. (1980). J. Polym. Sci. Polym. Chem. Ed. 18, 1547–1550.Trofimov, B. A. (1982). “New Reactions and Chemicals Based on Sulfur

and Acetylene,” 10th Int. Symp. Org. Chem. Sulfur.Trofimov, B. A. (1982). Tetrahedron Lett. 23(48), 5063–5066.Trofimov, B. A. (1982). J. Appl. Chem. Russia 51(1), Pt. 1, 1091.Trofimov, B. A. (1982). Tetrahedron 38(5), 713–718.Trofimov, B. A. (1984). Sulfur Lett. 2(2), 55–58.Trofimov, B. A., Stankevitch, V. K., and Belozerov, L. E. (1984). Russia

Patent 1,129,208.Trofimov, B. A., Nedolya, N. A., Khil’ko, M. Ya., Vyalykh, E. P.,

Mironov, G. S., and Moskvitchev, Yu. A. (1980, 1984), Russia Patents751,811 and 1,074,881.

Trofimov, B. A., Nedolya, N. A., Vyalykh, E. P., and Figovskii, O. L.(1980). Rusia Patent 761,490.

P1: LDK Revised Pages Qu: 00, 00, 00, 00


AlkaloidsArmin GuggisbergManfred HesseThe University of Zurich

I. Alkaloids in HistoryII. History of AlkaloidsIII. OccurrenceIV. NomenclatureV. Extraction, Isolation, Purification,

and AnalysisVI. Structure ElucidationVII. Classification

VIII. BiosynthesisIX. ChemotaxonomyX. Role

GLOSSARY

Chemotaxonomy (chemosystematics) System of clas-sifying plant species by their alkaloid (or other naturalproduct) content.

Chromatography Method of separating the componentsof a mixture by distribution between a mobile phase(gas or liquid) and a stationary phase (e.g., silica gel,alumina).

Chromophore Functional group (nitro, carbon–carbondouble bond, etc.) in an organic molecule that causesabsorption of ultraviolet or visible light.

Configuration Spatial arrangement of atoms or groups ofatoms around a central atom. (In structural formulas,C—H means the H is below, C—H means the H isabove the paper plane).

Heterocyclic ring Ring in which one or more of the ringatoms are noncarbon atoms (e.g., oxygen, nitrogen).

Optical activity Capacity of some (chiral) compoundsto rotate the plane of polarized light passing through.In the terminology of chiral compounds, the follow-ing symbols are used: (+), dextrorotatory, clockwise;(−), levorotatory, counterclockwise; ( ), indeterminaterotation or not to be found in literature.

ALKALOIDS are naturally occurring, nitrogen-containing organic compounds with the exception ofamino acids, peptides, pterines and derivatives, purinesand derivatives, amino sugars, and antibiotics. Clearly,the division of naturally occurring, nitrogen-containingsubstances into alkaloids and nonalkaloids is somewhat

477

P1: LDK Revised Pages


478 Alkaloids

arbitrary, and the boundary line is drawn in differentpositions by different authors.

I. ALKALOIDS IN HISTORY

Among the most notorious poisons in the history ofcivilization are the alkaloids. Very early in history, hu-man beings learned to exploit these organic compoundsfor different purposes such as therapeutic treatment, de-fense, cosmetics, and getting food. They also used alka-loids to murder disagreeable fellow beings. Because theywere not aware of the toxic effects of some special alka-loids, the use of these compounds led to small and greatdisasters.

During the course of civilization, knowledge of naturalplant poisons and, accordingly, knowledge of poisonousplants was lost, especially in towns, which is documentedby the increasing number of modern cases of poisoning.

The use of alkaloid-containing plants or plant parts withtherapeutic or stimulating effects is mentioned in the old-est written documents. About 2700 BC, the Chinese ShenLung described the drug ma huang and its use in medicine.The active component of this drug prepared from the plantEphedra chinensis is the alkaloid ephedrine. In the “Pa-pyrus Ebers” (1600 BC, Egypt), among the more than80 medicinal plants or drugs described are the follow-ing alkaloid-containing drugs: hemlock (Conium macu-latum), opium (from Papaver somniferum), Ricinus, andStrychnos. The European discovery of America madenative American medicinal knowledge available in Eu-rope and Asia. Drugs such as ipecacuanha (Psychotriaipecacuanha) and the Peruvian bark of Cinchona spp.stimulated medical interest because of their potent phar-macological activities. At about the same time, Europebecame familiar with tabacco (Nicotiana tabacum). Inthe 17th and 18th centuries, symptoms of diseases suchas fever (Peruvian bark), pain (opium), and constipation(Ricinus) were treated with plant products. Even diarrhea(ipecacuanha) and malaria (Peruvian bark) became cur-able. Tobacco and ephedra (against sleep) were used asstimulants.

Even in ancient times alkaloid-containing drugs suchas opium, mandragora (Mandragora officinalis), and henbane (Hyoscyamus niger) were used as anodynes in op-erations. Synthetic chemicals such as narcotics were notintroduced until the 19th century.

Some of the drugs that were well known in the past aretoday being misused and abused and their use is there-fore restricted. They are prepared mainly from alkaloid-containing plants, for example, opium and morphine fromPapaver somniferum, cocaine from Erythroxylum coca,and mescaline from Lophophora williamsii. Repeatedconsumption of these drugs leads to mania and addiction(morphinism, cocainism, etc.).

In medieval Italy, extract of the deadly nightshade (At-ropa belladonna) was once used to make women appearmore beautiful (Italian: bella donna) by dilating the pupilsof the eyes. Since prehistory, poisoned arrows and dartshave been used to enhance the effect of a shot. The ancientGreek word toxon means “bow and arrow,” and the Latinword toxicum for the poison used on arrows demonstratesthe wide use of poisons in ancient times, for both warand hunting. In many regions of the world (East and WestAfrica, South America, southern Asia, Borneo, Java), ex-tracts of alkaloid-containing plants were prepared to ob-tain the poison. In the northern part of South America,especially the upper basins of the Amazon and Orinocorivers, the arrow poison (calabash and tubo curare) pre-pared from different species of the genus Strychnos causedrespiratory arrest.

Another historical aspect of alkaloid use involves ex-ecution. For example, in 399 BC the Greek philosopherSocrates was sentenced to die by drinking a cup of hem-lock (a liquid extract of the poisonous plant Conium mac-ulatum). This was a common sentence at the time.

The biological effect of alkaloids also played a ma-jor role in the so-called Opium War (1840–1842). TheBritish East India Company cultivated and monopolizedopium preparation from P. somniferum in Bengal (India),and, since 1773, this company had exported an increasingamount of the drug to China. The importation of opiumwas forbidden in 1820 by the Chinese government. Thisact precipitated the Opium War in 1840. China was van-quished by England and, as a consequence, Hong Kongwas annexed to England (Treaty of Nanjing). Furthermore,China was forced to open its ports to Western Europeannations and to the British opium trade.

Until the 19th century, an epidemic disease caused bybread poisoning was known in Western, Central, and East-ern Europe. It was called raphania or St. Anthony’s fire.Its spread between the ninth and thirteenth centuries inFrance and Central Europe reached dramatic proportionssince, especially for poor people, no substitute for the sta-ple food, corn, existed. Today, we know that alkaloids ofthe fungus Claviceps purpurea were responsible for thesepharmacological effects. The fungus grows mainly on ryeand barley and sometimes on wheat and wild-growinggrass, producing several alkaloids, the so-called ergot al-kaloids, which contaminate the food by way of the flour.Consuming the poisoned food led to this disease, to multi-lations, and in many cases to death. Nevertheless, ergot hasbeen used in medicine for many centuries (1582, herbalbook of Lonizer) to enhance labor in women.

II. HISTORY OF ALKALOIDS

The year 1817 is regarded as the birth of alkaloid chem-istry. In this year, the German pharmaceutical chemist



Alkaloids 479

TABLE I Historical Data on Isolation, Structure Elucidation, and Synthesis of Some Well-Known Al-kaloids

Isolation of the Elucidation of the Determination of theAlkaloid Source pure alkaloid correct structure absolute configuration Synthesis

Morphine Opium 1805 1925 1955 1952

Emetine Psychotria ipecacuanha 1817 1948 1959 1950

Strychnine Strychnos nux-vomica 1818 1946 1956 1954

Atropine Atropa belladonna 1819 1901 1933 1903

Quinine Cinchona bark(s) 1820 1907 1950 1944

Coniine Conium maculatum 1827 1881 1932 1886

F. W. A. Serturner isolated in pure state the compoundmorphine (or morphia), which he had already describedin 1805. He identified morphine as the basic and activeprinciple of the drug opium.

It was now possible to measure an exact dose of mor-phine owing to this first preparation of an active drugprinciple in the crystalline state. Before this time, thedosing of drugs by varying the composition was verydangerous. Serturner’s success stimulated others to iso-late other alkaloids. Initially, the well-established medic-inal plants or drugs were investigated, for example,Psychotria ipecacuanha, S. nuxvomica, A. belladonna,Cinchona bark, and Conium maculatum, to extract eme-tine, strychnine, atropine, quinine, and coniine, respec-tively (Table I). At that time pharmacists and chemistswere concerned mainly with the purity, elemental com-position, and subsequently the structures of these com-pounds. But in most cases, the structures of the alkaloidswere too complicated to be elucidated because the neces-sary physical and chemical methods were in their infancy.As a result, structure elucidation in some cases requiredmore than 100 years. Some of the methods for structureelucidation developed in the 19th century are still used(see Section VI).

After the elucidation of its structure, an alkaloid issubsequently synthesized. Synthesis may offer proof ofwhether a structure is correct. The first synthesis of an al-kaloid was that of coniine in 1886, 59 years after its firstisolation from the plant.

The structure of a natural product is considered to beestablished as soon as its absolute configuration (i.e.,the three-dimensional orientation of all atoms of themolecule) has been determined.

III. OCCURRENCE

The number of structurally different alkaloids has beenestimated to be 6000. Most of them occur in flora, ∼1%in animals, and not more than 0.5% in fungi and bacteria.

Coccinelline (from the European ladybird beetle Coc-cinella septempunctata) is an example of an animal al-kaloid, whereas the deep-blue pyocyanine (from Pseu-domonas aeruginosa) exemplifies bacterial alkaloids. Itshould be mentioned that, to date, not all living organ-isms have been investigated with respect to their alkaloidcontent. Therefore, the percentages mentioned are notdefinitive.

N

H

H H

CH3

N

NO

CH3

O

Coccinelline Pyocyanine

Within the botanical classification system, it is inter-esting to consider the occurrence of alkaloid-containingplants. Botanists estimate the number of plant generato be more than 20,000. (A genus is one order higherthan a species.) Only 9% of all genera have alkaloid-containing species. These alkaloid-containing plants arenot statistically distributed over the genera; rather, theyoccur most abundantly in genera belonging to the Di-cotyledones and Monocotyledones of the Angiospermae(flowering plants). Table II shows the increasing number

TABLE II Alkaloid-Containing Plant Sections and FamiliesKnown in 1950 and 1985

Number ofalkaloid-bearing plant

families known

Section Class 1950 1985

Angiospermae Dicotyledones 28 120

Monocotyledones 7 14

Gymnospermae 2 5

Pteridophytae 2 3



480 Alkaloids

of plant families known to contain alkaloids and the dom-inance of the flowering plants. Alkaloids have rarely beenfound in the Gymnospermae and Pteridophytae.

It has been suggested that in 40% of all plant families,at least one alkaloid-containing species is known. On theother hand, there are plant families (e.g., Papaveraceae)in which all species contain alkaloids. Consequently, thedistribution of the alkaloids in flora is very dissimilar.

Alkaloids may occur in several parts of a plant (e.g.,roots, stem, bark, leaves, fruits, seeds). In some cases (e.g.,Papaver somniferum, A. belladonna) alkaloids are isolatedfrom all parts, whereas in other cases they are found inonly one part (e.g., the alkaloid of Aphelandra squarrosais found only in the roots). It should be noted that thealkaloid-containing part of the plant may not necessarilybe the site of alkaloid formation.

The number of structurally different alkaloids variesfrom plant to plant. For example, Catharanthus roseuscontains more than 100 alkaloids, whereas only 1 alka-loid has been detected in A. squarrosa. Quite often theratio of the components of an alkaloid mixture is differ-ent in different parts of the plant. Other factors that in-fluence the level and diversity of the alkaloid content arethe age of the plant (e.g., investigated with Adhatoda va-sica), the season (e.g., P. somniferum), the gathering timeduring the day (e.g., Conium maculatum), and the habitat(e.g., Maytenus buxifolia). The total amount of alkaloidsin plants fluctuates between ∼10% of quinine (Cinchonasp.) and ∼5 × 10−6% of triabunnine in Aristotelia pedun-cularis, based on the dried weight of the drug.

IV. NOMENCLATURE

The system of nomenclature of alkaloids is similar tothat of other natural products such as flavonoids and ter-penoids. To date, no unique system has been developed.One of the major reasons is the vast number of differ-ent skeletal types. In most cases alkaloids are denotedby trivial names, which are derived from the (systematic)botanical name of the source plant. The compound nameis derived either from the genus name (e.g., papaverine isisolated from different Papaver spp.) or the species name(harmaline is isolated from Peganum harmala). Some ex-ceptions include pelletierine (referring to the name of thefamous alkaloid chemist Pierre Joseph Pelletier), mor-phine (referring to the main physiological effect of analkaloid; the Latin deity of dreams was Morpheus), andemetine (from Greek emetikos, “an emetic”).

A common feature of nearly all alkaloid names is theending “-ine” (English, French) or “-in” (German). In thecase of several different alkaloids occurring in the sameplant, the suffixes “-idine,” “-anine,” “-aline,” “-inine,”and so on are added to the common principal form.

V. EXTRACTION, ISOLATION,PURIFICATION, AND ANALYSIS

A. Physical and Chemical Featuresof Alkaloids

As with other organic molecules, the physical propertiesof alkaloids strongly depend on their molecular structures.Low molecular weight amines such as coniine (from Co-nium maculatum) and pelletierine (from Punica grana-tum) are colorless liquids. Other examples of liquid al-kaloids are the Nicotiana alkaloids nicotine (from N.tabacum) and actinidine (from Actinidia polygama). Alka-loids of higher molecular weight—sometimes with addi-tional oxygen functions—are usually colorless crystallinecompounds, for example, papaverine (from Papaver som-niferum). Several colored crystalline alkaloids are alsoknown, such as the red-violet cryptolepine (from Cryp-tolepis triangularis).

N CH3

HH

N CH3

HH

O

(�)-Coniine (�)-Pelletierine

N N(�)-Nicotine (�)-Actinidine

NH H

CH3

CH3

H3C

The majority of alkaloids are basic compounds.Whereas an alkaloid is usually soluble only in an organicsolvent such as ether, ethanol, toluene, or chloroform andmostly insoluble in water, alkaloid salts (e.g., hydrochlo-rides, quaternary methochlorides) are soluble in water.Solubility in water is important for the therapeutic useof alkaloids.

Most isolated alkaloids are optically active. The planeof polarized light passing through a solution of an opti-cally active compound is rotated either to the left (coun-terclockwise, levorotatory) or to the right (clockwise,dextrorotatory). The value of deflection is given, for

N

OCH3

OCH3

H3CO

H3CO N

N

CH3

Papaverine(m.p. 147–148°C)

Cryptolepine(m.p. 175–178°C)



Alkaloids 481

example, as the [α]D value. Optical activity is displayed byany molecule containing a carbon atom with four differentsubstituents (asymmetric carbon atom).

ce

d

a

b

ce

d b

a

a, b, d, e = different substituents

Occasionally, the same alkaloid occurs as an equimolarmixture of levo- and dextrorotatory forms in the sameplant. In this case the [α]D value is zero. Such mixturesare called racemates. Vincadifformine was isolated fromdifferent plants in (−), (+), and (±) forms.

N

COOCH3

N

H

CH3

(�)-Vincadifformine

N

COOCH3

N

HCH3

(�)-Vincadifformine

H H

Source [α]D

Amsonia tabernaemontana +600◦

Vinca difformis ±0◦

V. minor −540◦

Hydrogenation of −600◦(−)-Tabersonine

B. Extraction and Isolation

Several methods have been established for extracting ba-sic alkaloids from plant material. As an example a methodfrequently used for labscale extraction will be described.In plants, alkaloids usually occur as salts of commonplant acids. The powdered drug is treated several timeswith a mixture of methanol and acetic acid, transformingthe salts to alkaloid hydroacetates, which are soluble inmethanol. The combined filtrates are evaporated, and theresidue is taken up in 1% hydrochloric acid. Chlorophylland other neutral or acidic compounds and neutral alka-loids can be extracted with an organic solvent such as etheror methylene chloride. The remaining acid solution is ren-dered basic with potassium carbonate, liberating the freealkaloids, which can be extracted with other or methylenechloride to yield the crude alkaloids. Alkaloids contain-ing a quaternary nitrogen atom remain saltlike in the basicaqueous solution. They can be isolated by precipitationas water-insoluble salts such as picrates (with picric acid)

or as Reineckates (with ammonium Reineckate) followedby displacement of the complex anion to chloride (e.g.,through ion exchange). An example of a quaternary alka-loid is toxiferine dichloride from Calebash curare.

NCH3

N

H

CH3H

Cl

N

N

CH3 H

H3C

Cl

H

H

(�)-Toxiferine dichloride

H

C. Purification

Before preparative separation, it is necessary to exam-ine the crude alkaloid extract by analytical thin-layerchromatography or by high-performance liquid chro-matography to obtain preliminary information about thecomposition of the mixture. To detect alkaloids chro-matographically, chromogenic reagents are used—forexample, Schlittler’s reagent (potassium iodoplatinate),Dragendorff’s reagent (potassium bismuth iodide), orcerium(IV) sulfate/sulfuric acid.

If the crude alkaloid mixture consists mainly of onecomponent, repeated crystallization from a suitable sol-vent may furnish the pure alkaloid. A more complex mix-ture must be separated by chromatography using silica gel,alumina, or cellulose powder as adsorbents and mixturesof organic solvents depending on the behavior of the in-dividual alkaloids. Ion exchange and gel chromatographyhave also been successfully used. Other separation meth-ods based on differences in partition between two immis-cible liquids of functionally diverse alkaloids are some-times preferred; among these are (Craig) countercurrentdistribution and droplet countercurrent chromatography.

VI. STRUCTURE ELUCIDATION

The elucidation of the structure of an alkaloid requiresthe determination of its elemental composition and thebonded relationships of all constituent atoms in space.The starting point is a homogenous compound that cannot



482 Alkaloids

be separated into components by physical methods. Theelemental composition of an alkaloid is determined bycombustion analyses or by high-resolution mass spec-trometry. Structure elucidation begins in earnest with thespectral investigation of the substance. Ultraviolet (UV)spectroscopy (light absorption in the UV region) yieldsinformation on the presence of chromophoric units suchas aromatic and other unsaturated systems. Infrared spec-troscopy identifies functional groups such as ketones, es-ters, amides, and hydroxyl groups and the degree of sub-stitution of double bonds. Nuclear magnetic resonance(NMR) spectroscopy (especially of 1H and 13C nuclei) re-veals information on the number and the electronic andstereochemical environment of the hydrogen and carbonatoms. Mass spectrometry provides the accurate molecularweight and the molecular formula through accurate massmeasurement of the molecular ion and may provide struc-tural information for the molecule by comparison of thefragmentation pattern with those of analogous systems.

In some cases it is possible to deduce the structure ofan alkaloid using only one of these techniques, especiallyNMR and mass spectroscopy, but usually a combination ofall methods is necessary. Chiroptical measurements (opti-cal rotatory dispersion and circular dichroism) assist in de-termining the optical activity of the alkaloid. This reveals,by comparison with known compounds, the molecule inits three-dimensional orientation and may lead to the ab-solute configuration.

Finally, and independently of knowledge about elemen-tal composition and spectral data, it is possible to deter-mine the structure of a crystalline compound by X-raydiffraction analysis. An example of the results of such ananalysis is the strychnine stereoprojection below. There isa possibility that the absolute configuration can be deter-mined in this way.

N

NH

O

H

H

O

H

(�)-Strychnine (absolute configuration)

Stereoprojection of a (�)-strychnine skeletonwithout hydrogen atoms (absolute configuration)

H

Only a small amount of a substance is required for spec-troscopic analysis, particularly UV spectroscopy, massspectroscopy, and X-ray diffraction (approx 10−5, 10−9,and 10−3 g, respectively).

Chemical methods are also applied to the determina-tion of structure. The major structural difference betweenalkaloids and most of the other natural products such asflavonoids and terpenoids is the presence of at least onenitrogen atom. Since 1820, chemically related structureelucidation of alkaloids has been based on specific reac-tions occurring due to the special reactivity of the nitrogenatom. The original reactions are of no value today; theseinclude destructive distillation, potash melt, and zinc dustdistillation. In many cases the destructive operations led toheterocyclic systems that were part of the alkaloids. Someof these heterocyclic systems were hitherto unknown, andthe names given to them reflect the name of the parentalkaloid. Some examples illustrate this point.

The nomenclature of such compounds as quinoline,isoquinoline, and quinolizidine is derived from the firstpreparation of quinoline from quinine-type alkaloids(Scheme I)hhh. The name piperidine reflects its first prepa-ration from piperine, a pepper alkaloid (Piper nigrum).The zinc dust distillation of the atropine-degradation prod-uct tropane furnished 2-ethylpyridine. Destructive reac-tions of this type are only of historical interest, since anumber of alkaloid degradations led to heterocyclic sys-tems that exhibited no structural relation to the alkaloiditself.

The most important reactions used in alkaloid chem-istry are those that yield information about the neighbor-hood of the nitrogen atom(s). One of the most extensivelyused reactions is the Hofmann degradation of exhaustivemethylation. Treatment of an alkaloid containing a ba-sic nitrogen atom with methyl iodide leads to a quaternaryammonium salt. With base, the salt undergoes eliminationof a β-proton and scission of the N—Cα bond to yield anolefin and a compound containing a tertiary nitrogen atom.Through repetition, until the nitrogen atom is extruded asN(CH3)3, this reaction permits the determination of thenumber of bonds between the nitrogen atom and the car-bon skeleton of an alkaloid. In Table III the basic stages ofthe Hofmann degradation of three different alkaloid typesare given. Other specific C—N degradation reactions arealso occasionally used in structure elucidation of the alka-loids. Depending on the variation of alkaloid structures,oxidative, hydrolytic, reductive, and other reactions canalso be applied.

VII. CLASSIFICATION

In the following survey of alkaloids, we consider, first, thenitrogen atom and its direct environment, and accordingly



Alkaloids 483

SCHEME 1 Classical degradation reactions of alkaloids.

we classify the alkaloids on the basis of their nitrogen-containing structural features. In cases of alkaloids withmore than one nitrogen atom, preference will be given tothe most characteristic nitrogen function. Exceptions tothis criterion will be allowed only if other structural ele-ments are more typical, as in the case of steroidal, terpene,spermidine, spermine, and peptide alkaloids. Therefore,with respect to the location of the nitrogen atom, the alka-loids can be classified as follows:

1. Heterocyclic alkaloids2. Alkaloids with an exocyclic nitrogen atom and

aliphatic amines3. Putrescine, spermidine, and spermine alkaloids4. Peptide alkaloids5. Terpene and steroid alkaloids

According to this classification, the majority of alkaloidscomprise the heterocyclic group, whereas the putrescine,

spermidine, and spermine alkaloids form the smallestgroup.

A. Heterocyclic Alkaloids

The representatives of this class have the nitrogen atomin a heterocyclic ring. Alkaloids containing one nitro-gen atom as part of a five-membered ring are the so-called pyrrolidine alkaloids; an example is (+)-hygrine(from Erythroxylum sp.). The indole alkaloids form a largegroup within this class. The indole chromophore con-tains a five-membered ring and a benzene ring fused onone side. Some examples are vincadifformine, toxiferinedichloride (a bisindole alkaloid), strychnine, strictosidine,ajmalicine, and vincamine. Harmaline (Peganum har-mala), (−)-physostigmine or (−)-eserine (Physostigmavenenosum), and rutaecarpine (Evodia rutaecarpa) aswell as cryptolepine and triabunnine are of differ-ent types. In pilocarpine (Pilocarpus pennatifolius), the



484 Alkaloids

TABLE III Basic Stages of the Hofmann Degradation of Alkaloids

N-Methylation product First degradation product Second degradation product Third degradation productof an alkaloid Repetition Repetition−−−−−−−−→ −−−−−−−−→

CH CH2

H5C6N

CH3CH3CH3

H

OH

αβCH

H5C6CH2

� H2O

� N(CH3)3

Result: one bond betweenN atom and C skeleton

CH2

CH2

NCH3H3C

H2C

H2C

OH

β βCH2

CH2

NCH3H3C

H2C

HC

� H2O

βCH

H2C

HC� H2O

� N(CH3)3CH2

Result: two bonds betweenN atom and C skeleton

CH3

N

CH2

C

H

H2CC

H

CH2

CH

OH

β β

β

CH3

N

CH2

C

H

H2CC

H

H2C

� H2O

β

β

CH3

NCH3

H2CC

H

H2C

� H2OCH2

β H2CCH2

CH2� H2O

� N(CH3)3

Result: three bonds betweenN atom and C skeleton

five-membered ring contains two nitrogen atoms, as inthe amino acid histidine.

NH

NH3CO

CH3

Harmaline

N

( )-Triabunnine

HON

H

H3C O

H

CH3

CH3

H

The peptide ergot alkaloids consist of one molecule oflysergic acid and a peptide unit. Typical examples are er-gotamine and ergocornine, both isolated from Clavicepspurpurea.

In a similar variation of alkaloids with a five-memberedring, the nitrogen atom is part of a six-membered ring. Ex-amples are coniine, piperine, sedridine (Sedum acre), β-skytanthine (Skytanthus acutus), pelletierine, and actini-dine. Nicotine consists of both a five- and a six-memberedring, each containing a nitrogen atom.

Through fusion of a benzene to a six-membered ringwith one nitrogen atom, two ring systems can be formed:

N

N

H

CH3HOOC

(�)-Lysergic acid

N

N

HN

H

O

O

N

NO

H

R2

CH3

CH

R

R1

O

(�)-Ergotamine, R = R′ = H; R2 = CH2 C6H5

(�)-Ergocornine, R = R′ = CH3; R2 = CH(CH3)2

HO

H

H



Alkaloids 485

the quinoline and the isoquinoline system. An example ofa quinoline alkaloid is the cinchona alkaloid cinchonine.

NQuinoline

N

Isoquinoline

Natural isoquinoline alkaloids are very abundant; ex-amples are papaverine, norlaudanosoline, emetine, andmorphine. Although morphine cannot be classified as anisoquinoline alkaloid on the basis of its structure, biosyn-

N

HN

H3CO

H3CO

OCH3

OCH3

CH3H

(�)-Emetine

O

HO

H

HO

H

NCH3

(�)-Morphine

thetically it is derived from an isoquinoline alkaloid. Twonitrogen atoms in one six-membered ring are present inpyocyanine.

In quite a number of heterocyclic alkaloids, the nitrogenatom is common to two rings; some basic skeletal typesand corresponding examples are the following:

N N N

Pyrrolizidine Indolizidine Quinolizidine

N

(�)-Monocrotaline

HO O

O

CH3HO

H3C

CH3

O

HO

(from Crotolariasp.)

( )-Slaframine(from Rhizoctonia

leguminicola)

NH2N

H O CH3

O

N

HCH2OH

(�)-Lupinine(from Anabasis

and Lupinussp.)

HN N

H3C

O

H

C

O

C6H5

CH2OH

H

Tropane (�)-Atropine(racemate of hyoscyamine)

The nitrogen atom can also be common to three ringsas in coccinelline. In addition to the heterocyclic systemsmentioned above, some others occur as parts of alkaloids.

B. Alkaloids with an Exocyclic Nitrogen Atomand Aliphatic Amines

Ephedrine and mescaline are two well-known representa-tives of this class. Capsaicine was isolated from Piper andCapsicum species.

C C

CH3

N

HH

CH3HOH

NH2

H3CO

H3CO

H3CO

N (CH2)4

O

CH CH CH(CH3)2H3CO

HOCapsaicine

(�)-Ephedrine Mescaline

H

C. Putrescine, Spermidine,and Spermine Alkaloids

The alkaloids in this class contain putrescine, spermi-dine, and spermine as the basic backbone; typical ex-amples are paucine, inandenin-12-one, and aphelandrine,respectively.

NH2H2N N

NH2H2N

Putrescine Spermidine

H

NN NH2

H2N

Spermine

H

H



486 Alkaloids

NH

O

HO

H2N

N N

O

ONH2

Paucine(from Pentaclethra sp.)

Inandenin-12-one(from Oncinotis inandensis)

H

O

H

N

NH

H O

ON HH

OH

N

(�)-Aphelandrine(from Aphelandra squarrosa)

H

D. Peptide Alkaloids

Besides a (cyclic) peptide, the peptide alkaloids contain anadditional basic nitrogen atom, as demonstrated by mau-ritine A built up from the L-amino acids valine, trans-3-hydroxyproline, N ,N -dimethylalanine, and phenylala-nine, together with p-hydroxystyrylamine as the aminocomponent. This group of alkaloids are also called rham-naceous alkaloids owing to their predominant occurrencein the plant family of Rhamnaceae.

N

O

NN

NH

H

H

HO

NCH3

H

O

H3C

H3C

HH

H3C

CH3

OO

H

(�)-Mauritine A(from Zizyphus mauritiana)

E. Terpene and Steroidal Alkaloids

Terpenes and steroids can be part of an alkaloid skeletonthat contains in addition at least one nitrogen atom. Be-

sides monoterpene (e.g., β-skytanthine), sesquiterpene,diterpene (e.g., aconitine, the poison of monk’s hood),triterpene alkaloids, and steroidal alkaloids (e.g., cones-sine) are known.

H

O

H2C

HO

H3CO

OCH3OCH3

CH3

O

OH

OH

OCH3

O CH3

O

N

CH3

(�)-Aconitine(from Aconitum napellum)

H

N

N

CH3

HH

H3CH

CH3

CH3

H3C

(�)-Conessine(From Holarrhena species)

Other classification systems for alkaloids have beenproposed. In the biogenetic classification, the class des-ignations are derived from the amino acids or other fun-damental units that are considered precursors of alkaloidformation in plants. This system has some advantages. Asdemonstrated in this article, even the chemical classifica-tion systems usually require biogenetic arguments in thecase of special alkaloids (e.g., morphine). However, it isknown from various organisms that they can biosynthesizethe same alkaloid from different amino acids. In Scheme 2,the biogenesis of nicotinic acid is presented. The precur-sor in the first pathway is tryptophan; that in the second isaspartic acid. Although many biosynthetic pathways havebeen established experimentally, some biogenetic propos-als are only speculative. Considering these disadvantages,it seems more plausible to use the classification systembased on the alkaloid structures.

VIII. BIOSYNTHESIS

The alkaloid chemist is interested in the identity of thenatural building blocks of alkaloids. On the basis ofmany experiments it has been demonstrated that for the



Alkaloids 487

SCHEME 2 Two different pathways in the biosynthesis of nico-tinic acid.

large variety of alkaloids only a few building blocks areneeded. Structural variety is generated by special enzy-matic systems in plant and animal cells. As expected,amino acids play the most important role among thenitrogen-containing precursors. But only a few aminoacids act as precursors in alkaloid biosynthesis. Scheme 3lists the important amino acids (without specification oftheir absolute configuration) together with their corre-sponding alkaloid derivatives. In the case of rutaecarpine,tryptophan is required in addition to anthranilic acid as asecond amino acid precursor.

The same amino acids that are the precursors of the al-kaloid examples given in Scheme 3 act as precursors formost of the other alkaloids. The remaining plant bases arederived by the addition of ammonia or an alkylamine tononbasic precursors. Many of the nitrogen-building blocksof alkaloids are terpenoids, fatty acids, and cinnamic acidderivatives. Small carbon fragments, such as methyl ac-etate, and carbonate complete the peripheral functional-ities. Carbon–carbon bond formation, oxidation, and re-duction reactions are typical enzyme reactions.

It is possible to establish a biogenetic pathway in a liv-ing organism by using isotopic tracers. In feeding experi-ments, specifically labeled precursors are introduced into aplant or microbial broth. The labeling may be with eitherradioactive (e.g., 3H, 14C) or stable (e.g., 13C) isotopes.Some difficulties are associated with this procedure. Thelabeled precursor must be synthesized with the labels inthe known positions, and the amount of the label at eachposition established. Once labeled precursor has been in-corporated, the product must be isolated in a pure state,the amount of label present in the whole molecule deter-mined, and the precise location of the label establishedby chemical degradation of the molecule to pure prod-ucts, the radioactivity of which is then measured. In thecase of stable-isotope labeling, 13C-NMR spectroscopy is

SCHEME 3 Examples of alkaloids and their corrsponding aminoacid precursors.



488 Alkaloids

SCHEME 4 Main pathways in the biosynthesis of papaverine from tyrosine in Papaver somniferum L ( �and � = 14C).

necessary for the detection of the marked positions. Fromsuch experiments it can be ascertained whether a proposedcompound is a precursor of an alkaloid. It is important tonote that the results are valid only for the specific plantproduct of a specific plant because different pathways forthe same product are known.

As an example, the biogenesis of the alkaloid papaver-ine from Papaver somniferum is given in Scheme 4. Ty-rosine is the precursor of papaverine in P. somniferumsince introducing the labeled tyrosine in position 2 by14C (=[2-14C]tyrosine) into the plant resulted in incorpo-ration into papaverine at positions C-1 and C-3. There-fore, two molecules of tyrosine are transformed into onemolecule of papaverine. Several chemical reactions areinvolved in the transformation of tyrosine to papaverine:for example, in tyrosine the aromatic nucleus is substi-tuted only by one oxygen atom, tyrosine contains a carbo-cyclic acid residue that is not present in the product, andin papaverine only one nitrogen atom is present, com-pared with the two in the precursor. Detailed analysisof the pathway, again using labeled precursors, showedthat tyrosine is first hydroxylated in plant cells into 3,4-dihydroxyphenylalanine (Dopa). Dopa is decarboxylated(loss of CO2) to give dopamine. Feeding experimentswith dopamine have clearly demonstrated that only onemolecule of dopamine is incorporated into papaverine.

The conclusion of this experiment is that tyrosine plays adouble role. It is hydroxylated and decarboxylated to formdopamine, and it is transformed into the intermediate 3,4-dihydroxyphenylpyruvic acid. The condensation reactionbetween an amine and a ketone, as well as the cyclizationof the intermediate imine, is a well-known organic syn-thetic process. In case of the biosynthesis of papaverine,norlaudanosoline is the product of condensation, cycliza-tion, and decarboxylation. With the aid of labeling ex-periments, it has been verified that norlaudanosoline is in-deed the precursor of papaverine. Furthermore, it has beenshown that the O-methylation of (−)-norlaudanosolineproceeds stepwise and partly parallel with dehydrogena-tion to papaverine.

IX. CHEMOTAXONOMY

As mentioned above, the occurrence of alkaloids in plantsis strongly related to the plant families. For a longtime, it has been known that closely related plants con-tain the same structurally similar alkaloids. For exam-ple, the two tropan alkaloids atropine and hyoscyaminewere isolated from the nightshade plants (Solanaceae)Atropa beiladonna (deadly nightshade), Datura stra-monium (thorn apple), and Mandragora officinarium



Alkaloids 489

(mandrake). Therefore, a classification of plant speciesbased on the structures of their alkaloids and other chem-ical ingredients has been attempted. Such a system ofclassification of plant species by chemical compounds isknown as plant chemosystematics or chemotaxonomy.

To investigate a group of alkaloids in the chemosystem-atic sense it is necessary to have a large body of statisticalmaterial available; otherwise, the results will be of uncer-tain value.

The distribution of indole alkaloids in plant species hasbeen quite well investigated. The number of structurallyknown alkaloids is ∼1200. They have been isolated frommore than 400 different plant species. Some alkaloids wereisolated from only one species; others were from manydifferent species. A total of 1200 indole alkaloids havebeen isolated nearly 3500 times. These data underline theimportance of this class of alkaloids for chemosystematicstudies. To establish a botanical classification system onthe structural basis of the indole alkaloids plants contain,it is important to investigate their biogenetic relationships.

The above-mentioned 1200 indole alkaloids taken intoconsideration are built up from tryptamine or tryptophan(the basic components) and the monoterpene derivativesecologanin. It was shown that tryptamine and secolo-ganin are directly transformed to the alkaloid strictosi-dine (Scheme 5). Most of the 1200 indole alkaloids arederivatives of strictosidine. They are distinguished fromstrictosidine by variations in the functional groups (e.g.,—H → —OH; NH → N—CH3) and/or loss of carbon

SCHEME 5 The indole alkaloid strictosidine is built up from thebase trytamine and the monoterpene secologanin.

atoms (e.g., —COOCH3 → —H) and/or opening or form-ing of chemical bonds. In Scheme 6 examples of struc-tural changes of strictosidine leading to the indole alka-loids ajmalicine, strychnine, and vincamine are given. Thestructural changes occur only in the secologanin part ofstrictosidine: in all cases the tryptamine part remains un-rearranged. Therefore, to emphasize the changes in thesecologanin part, the carbon skeleton of the latter is pre-sented in heavy lines in the scheme. These changes canoccur by rotations around or rearrangements of variouscarbon–carbon bonds. Changes of oxygen functionalityand double bonds and so on are not considered. The car-bon skeleton of ajmalicine can be derived from that ofstrictosidine by rotation of one carbon–carbon bond. Twoadditional rotations are necessary to explain the forma-tion of the carbon skeleton of strychnine. In addition, theCOOCH3 group should be eliminated and two additionalcarbon atoms (as an active acetate unit) have to be intro-duced into the molecule to reach strychnine. To explainthe formation of the carbon skeleton of vincamine fromthat of ajmalicine, a rearrangement of a three-carbon chainis necessary. Strictosidine, ajmalicine, and strychnine areexamples of indole alkaloids with a so-called nonrear-ranged carbon skeleton. All of the alkaloids with a nonre-arranged carbon skeleton have the same absolute configu-ration at C-15 (marked in Scheme 6) as in secologanin atC-7 (Scheme 5). In the rearranged types, the chirality ofC-15 disappears.

Nearly all (more than 99%) of the 1200 indole alka-loids mentioned above were isolated from plant speciesbelonging to only three plant families: Apocynaceae, Lo-ganiaceae, and Rubiaceae. It is remarkable that alkaloidsof the strictosidine and ajmalicine types were isolated fromthe plants of all three families. Alkaloids of the strychninegroup are known to occur only in the Loganiaceae. Al-kaloids with a rearranged secologanin skeleton (e.g., vin-camine) occur only in the Apocynaceae. Characteristic al-kaloids of the Rubiaceae are ajmalicine-type compoundsin various oxidation states.

The chemotaxonomic conclusion from these results areas follows. The enzyme system generating strictosidine-type alkaloids from tryptamine and secologanin must bethe same in all three plant families (strictosidine syn-thetase). The most primitive plant family seems to be theLoganiaceae because the species of this family produceonly the unrearranged skeleton types. Their special abil-ity consists in the incorporation of an additional C2 unit. Torearrange a carbon skeleton, more advanced enzyme sys-tems are necessary. Therefore, it is deduced that the Apoc-ynaceae are more developed than the Loganiaceae. Forother reasons not mentioned, Apocynaceae and Rubiaceaeare thought to be equivalent. Besides special oxidation



490 Alkaloids

SCHEME 6 Structural relation of strictosidine to other types of indole alkaloids. Carbon skeleton of the secologaninunit is given in heavy lines. % denotes equivalent to: ∧∧∧, cleavage; R, rearrangement of carbon–carbon bonds; C,conformation changes by rotation around carbon–carbon bonds.

reactions, they can use tyramine (decarboxylated tyrosine;Scheme 3) instead of tryptamine in the alkaloid formation(e.g., emetine). A more detailed analysis of alkaloid struc-tures in relation to plant families, subfamilies, tribes, gen-era, and species led to a proper classification of some plantspecies that had been incorrectly placed in the botanicalclassification system.

X. ROLE

A. In Natural Sources

Alkaloids can be described as products of secondary plantmetabolism like other complex natural compounds such asflavonoids, terpenoids, and steroids. The role of alkaloidsin alkaloid-containing plants is still a matter of specula-tion. For a long time, it has generally been accepted thatalkaloids are end products of metabolism and are wasteproducts of the plant. However, more recent investiga-tions have shown that alkaloids are involved in a dynamic

process; for example, the amount of the alkaloid coniinein Conium maculatum varies during a single day as wellas in the course of the development of the plant. Therehave been similar observations of the atropine alkaloidsin A. belladonna and the opium alkaloids in P. somniferum.In isotopic labeling experiments it was demonstrated thatthe turnover rate (half-life) of alkaloids in plants is veryfast. This clearly shows that plants do metabolize alka-loids, and therefore the theory that alkaloids are metabolicend products is no longer tenable. However, whether al-kaloids are of any use to the plant is still debatable sincein only a few cases has it been shown that alkaloid-containing plants are protected against consumption by an-imals. On the other hand, both alkaloid-free and alkaloid-containing plants can be attacked by parasites, fungi, andbacteria.

In contrast to their role in plants, the function of alka-loids in insects is much more established; for example,coccinelline, the tricyclic alkaloid N -oxide, is producedby beetles of the Coccinellidae and is used as a defensiveallomone against predators.



Alkaloids 491

TABLE IV Pharmacological Properties of Some Alkaloids

Alkaloid Pharmacological action

Aconitine Antipyretic, antineuralgic

Ajmalicine Antihistaminic

Ajmaline Antiadrenergic, antiarrhythmic

Atropine Anticholinergic, mydriatic, antispasmodic

Camptothecine Antitumor activity

Cocaine Local anesthetic, vasoconstrictor

Codeine Antitussive

Emetine Emetic, amebicide

Ephedrine Sympathomimetic

Ergot alkaloids α-Sympatholytic, vasodilator,antihypertensive

Homoharringtonine Antitumor activity

Morphine Analgesic, narcotic

Papaverine Antispasmodic

Physostigmine Parasympathomimetic, miotic

Pilocarpine Parasympathomimetic, diaphoretic, miotic

Quinidine Antiarrhythmic

Quinine Antimalarial

Reserpine Antihypertensive

Scopolamine Mydriatic, parasympatholytic

Tubocurarine chloride Muscle relaxant

Vincamine Vasodilator, antihypertensive

Vincaleucoblastine Antineoplastic

Vincristine Antineoplastic

B. Alkaloid Structures in Drug Design

The pharmacological importance of alkaloids is enor-mous. Since the discovery of morphine by Serturner, manyplant alkaloids have been used directly or by molecularmodification for therapeutic purposes. Even today a num-ber of alkaloids (or their hydro salts) are used clinically.Examples are given in Table IV; their structures are shownin Scheme 8.

O

CCN

H2H2

C2H5

H5C2

O

H

N COOCH3

Cocaine

Procaine

O

O

H3C

NH2

Other useful therapeutic agents are derivatives of alka-loids with minor changes in the molecular structure, suchas 2,3-dehydroemetine (amebicide), strychnine N -oxide(analeptic), and N ,N ′-diallylnortoxiferine dichloride [po-tent muscle relaxant; it differs from toxiferine dichloridein the substituents of two nitrogen atoms: (N)—CH2—CH CH2 instead of (N)—CH3]. Finally, another groupof compounds used therapeutically differ substantially instructure from the natural alkaloids. In their analgesic ac-tion cocaine and procaine are similar. The analgesic ac-tivity of morphine greatly stimulated molecular modifi-cation of its structure, even before all the details wereknown. Cyclazocine is considered to be ∼40 times morepotent analgetically than morphine given subcutaneouslyor orally. Because of side effects it is unacceptable as ageneral analgesic. Etorphine, another modified morphine,is 100–80,000 times more potent than morphine, depend-ing on the test system used. It is interesting that compoundA given in Scheme 7 has enhanced antitussive activity,whereas its analgesic power is lower than that of mor-phine. The pharmacology and “modification” of alkaloidsis a fascinating and at the same time very important aspectof medicinal chemistry.

In many cases the source of clinically used alkaloids isstill the natural plant (e.g., morphine, quinine), but in somecases industrial synthetic products are preferred (e.g., vin-camine, 2,3-dehydroemetine).

SCHEME 7 Morphine and three of its modifications.



492 Alkaloids

SCHEME 8 Structures of alkaloids in Table IV.



Alkaloids 493


ATOMIC SPECTROMETRY • HETEROCYCLIC CHEMISTRY

• MICROSCOPY (CHEMISTRY) • ORGANIC CHEMISTRY,COMPOUND DETECTION • ORGANIC CHEMISTRY, SYN-THESIS • PHARMACEUTICALS • X-RAY ANALYSIS

BIBLIOGRAPHY

Atta-ur-Rahman and Basha (1983). “Biosynthesis of Indole Alkaloids,”Oxford Univ. Press (Clarendon), London and New York.

Cordell, G. A., ed. (1950–2001). “The Alkaloids: Chemistry and Phar-macology,” Vols. 1–55, Academic Press, San Diego.

Cordell, G. A. (1981). “Introduction to Alkaloids: A Biogenetic Ap-

proach,” Wiley-Interscience, New York.Dalton, D. R. (1979). “The Alkaloids: The Fundamental Chemistry,”

Dekker, New York.Harborne, J. B., and Turner, B. L. (1984). “Plant Chemosystematics,”

Academic Press, New York.Hesse, M. (1981). “Alkaloid Chemistry,” Wiley, New York.

Mothes, K., and Schutte, H. R. (1969). “Biosynthese der Alkaloide,”Deutscher Verlag Wissenschaften, Berlin.

Pelletier, S. W., ed. (1983–1999). “Alkaloids: Chemical and BiologicalPerspectives,” Vols. 1–14, Pergamon, Elmsford, NY.

Specialist Periodical Report (1971–1983). “Alkaloids,” Vols. 1–13, TheChemical Society, London.

Waller, G. R., and Nowacki, E. K. (1978). “Alkaloid Biology andMetabolism in Plants,” Plenum, New York.

Wexler, P., ed. (1998). “Encyclopedia of Toxicology,” Academic Press,San Diego.

P1: ZCK 2nd Revised Pages Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology EN002G-51 May 25, 2001 13:44

Bioconjugate ChemistryClaude F. MearesUniversity of California, Davis

I. Chemical PrinciplesII. Multiple Sites of Chemical Conjugation:

Mathematics of Random LabelingIII. Molecular Cloning: Modern Preparative

MethodologyIV. Analysis of BioconjugatesV. Nomenclature

VI. Examples

GLOSSARY

Antibody A large protein molecule that binds reversiblyand with high specificity to a small molecule or a spe-cific region (epitope) of a macromolecule.

Bioconjugate Generally, a synthetic substance in whicha functional biological molecule is conjugated (linked)to another substance with useful properties. The lattermay be another biological molecule, a synthetic poly-mer, a small molecule, or even a macroscopic particleor a surface.

Enzyme A protein (or nucleic acid) that catalyzes a chem-ical reaction, such as hydrolysis of a peptide bond.

Epitope A molecular group, part of a larger molecule,recognized by an antibody or other binding macro-molecule. Typically, a small number of amino acidresidues in a particular geometric arrangement on thesurface of a protein.

Hapten A small molecule recognized by an antibody orother binding macromolecule. Sometimes referred toas a ligand. A hapten may be conjugated to another

molecule. For example, biotin binds strongly to theproteins avidin and streptavidin. Biotin may be con-jugated to other molecules (see Fig.1), which canthen be trapped by binding to immobilized avidin orstreptavidin.

Liposome A particle composed of multiple phospholipidmolecules, arranged in a bilayer structure to enclose asmall aqueous space (see Fig. 2).

ODN An oligodeoxynucleotide, a short, synthetic frag-ment of deoxyribonucleic acid (DNA).

PEG Poly(ethylene glycol), also called PEO, poly(ethy-lene oxide) (see Fig. 3). The degree of polymerizationn is usually specified according to the average numberof CH2 CH2 units in a polymer molecule.

BIOCONJUGATE CHEMISTRY involves the joining oftwo molecular functions by chemical or biological means.This includes, among other topics, the conjugation ofantibodies, nucleic acids, lipids, carbohydrates, or otherbiologically active molecules with each other or with other

93

P1: ZCK 2nd Revised Pages


94 Bioconjugate Chemistry

FIGURE 1 D-Biotin conjugated to another molecule, representedby R.

substances that add useful properties (polymers, drugs, ra-dionuclides, toxins, fluorophores, photoprobes, inhibitors,enzymes, haptens, ligands, surfaces, metal particles, etc.).

I. CHEMICAL PRINCIPLES

A. Nucleophiles and Electrophiles

A basic feature of almost all applications of bioconjugatechemistry is the attachment of one chemical species toanother. Most often this is done by forming an ordinary co-valent chemical bond, taking advantage of naturally occur-ring reactive groups where possible. For example, proteinscommonly have an abundance of nucleophiles (electron-rich atoms or molecular groups) on their surfaces (seeFig. 4), which can react with electrophiles on syntheticreagents.

B. Chemistry of Functional Groupsand Reagents

1. Thiols

The thiol ( SH) group, which occurs on the side chain ofthe amino acid cysteine, is a particularly reactive nucle-ophile. When treated with an appropriate electrophile, it isnot unusual that a single thiol will serve as by far the most

FIGURE 2 Example of a liposome (phospholipid vesicle) conjugated to functional molecules R.

FIGURE 3 PEG. R and R′ refer to attached groups.

reactive site in a protein containing many other, nitrogen-or oxygen-based, nucleophiles. Because of their affinityfor metals, thiol groups are also used to attach molecularlinkers to metal particles or surfaces.

2. Amines

The amino ( NH2) group, which occurs on the side chainof the amino acid lysine, is also a reactive nucleophile.Because lysine is a common amino acid (a typical an-tibody molecule of type IgG has more than 80 lysineresidues), it is often the target of chemical modification byelectrophiles.

3. Disulfides

The disulfide bond ( S S ), formed by mild oxidationof two thiols, plays a significant role because it is eas-ily broken by mild reducing agents. The disulfide bond isnot particularly reactive to nucleophiles or electrophiles,providing a strategy to protect a thiol from reaction. Adisulfide bond can also be used to temporarily attach onecomponent of a bioconjugate to the other; reductive cleav-age can easily reverse the attachment.

4. Active Esters

a. Carboxylic. Of the formula R CO X, where Xis a leaving group that is easily displaced, these reactivegroups are commonly used to acylate amino groups to



Bioconjugate Chemistry 95

FIGURE 4 Some biological nucleophiles commonly used in bioconjugate chemistry: cysteine, lysine, and a 5′ phos-phorothioate group from a synthetic nucleic acid. Nucleophilic atoms are shown in boldface.

form amides R CO NHR′, attaching molecular groupsR to lysine residues on proteins R′.

b. Imidic. Of the formula R CNH+2 X, used

similarly.

5. Isothiocyanates

Of the formula R N C S, these reactive groups are com-monly used to modify amino groups to form thioureasR NH CS NHR′, attaching molecular groups R to ly-sine residues on proteins R′.

6. Alkylating Agents

a. Alkyl halides. Of the formula R CO CH2 X,commonly used to alkylate thiol groups to form thioethersR CO CH2 SR′, attaching molecular groups R to cys-teine residues on proteins R′.

b. Alkenes. Containing a C C double bond near anactivating group, such as carbonyl, used similarly.

7. Photoprobes

Containing photosensitive groups such as aryl azideR N3, diazo CN2 , or ketones R CO R′, whichupon irradiation with ultraviolet light produce energeticreagents for covalent attachment to targets.

II. MULTIPLE SITES OF CHEMICALCONJUGATION: MATHEMATICSOF RANDOM LABELING

Biological molecules such as proteins or nucleic acids canhave many sites with similar chemical reactivity, such aslysine residues or phosphate residues. Random copoly-mers can also contain multiple reactive sites with simi-lar properties. Chemical conjugation involving these sitesleads to complex mixtures of products.

Consider a macromolecule with three equally reactivesites (n = 3) that can be modified by a reagent. We willuse a simple 3-bit notation to label each possible productin the resulting mixture (see Fig. 5).

If modification of one site on the macromolecule hasno effect on modification of the others (the usual approx-imation), we can define a conversion ratio s as a ratio ofconcentrations

s = [100]

[000]= [010]

[000]= [001]

[000].

(Here we compare each singly modified species to unmod-ified.)

Using the same assumptions, we discover that doublymodified macromolecules are related to singly modifiedones by the same factor s, and to unmodified ones by s2

[110]

[100]= s

[110]

[000]= [110]

[100]• [100]

[000]= s • s = s2.

(Note there are three of these.)

FIGURE 5 Illustration of species produced by random labeling.




FIGURE 6 Species resulting from random labeling of a macromolecule having n = 60 equally reactive sites, suchthat the average molecule has n = 5 haptens attached, according to Eq. (2).

Similarly, for triply modified macromolecules

[111]

[110]= s

and[111]

[000]= s3.

Another important factor is that there is more than oneway to produce a macromolecule with one site modified(e.g., there are three distinct singly-modified species inthis example). In general, if a macromolecule has n sitesand exactly r of them are randomly modified, then therewill be n!/(n − r )! • r ! distinct isomers as products.

The mole fraction of any of the species is given by, e.g.,X001 = [001]/

∑[SSS], where

∑

[SSS] = [000]

(∑ [SSS]

[000]

)

= [000] (I + s + s + s + s2 + s2 + s2 + s3)

= [000]n∑

r=0

n!

(n − r )! • r !sr = [000]Z

Note that Z is easy to calculate if we know s and n.So the mole fraction of any species is, e.g.,

X001 = [001]/∑

[SSS] = [001]/{[000]Z} = s/Z = s/(1 +3s + 3s2 + s3). In general, the mole fraction fr of productwith a particular degree of modification r is (Eq. 1):

fr =n!

(n − r )! • r !sr

∑

kn!

(n − k)! • k! sk

The average value of r in a real mixture of modified macro-molecules is denoted ν, which is given by Eq. (2):

υ =∑

r

fr • r =∑

rn!

(n − r )! • r !r • sr

∑

kn!

(n − k)! • k! sk

Figure 6 plots the fraction fr of macromolecules havingexactly r haptens attached, up to the value r = 11. Note thateach column contains contributions from n!/(n − r )! • r !distinct isomers. Eq. (2) can be used to calculate fr forlarger values of r .

III. MOLECULAR CLONING: MODERNPREPARATIVE METHODOLOGY

Gene fusion, whereby the gene coding for one protein isconnected to the gene coding for another, is one methodby which site-specific attachment of two proteins can beachieved. Site-directed mutagenesis, in which one aminoacid residue of a protein is genetically altered to producea mutant protein containing a single cysteine residue, issuited for site-specific conjugation of other molecules orsite-specific immobilization of the protein onto a surface.

IV. ANALYSIS OF BIOCONJUGATES

A. Methods

Characterization of bioconjugates frequently relies onassays of the properties of the components (enzyme



Bioconjugate Chemistry 97

activity, antibody binding, radioactivity, fluorescence,etc.). In addition, gel electrophoresis, high-performanceliquid chromatography, and mass spectrometry are fre-quently employed to confirm composition, molecularweight, homogeneity, and even details regarding sites ofattachment. Bioconjugates involving larger componentssuch as particles or surfaces are characterized using mi-croscopy and other surface-science techniques.

B. Challenges

The precise specification of the chemical linkages formedwhen a protein or other biomolecule is chemically attachedto another biomolecule, the product often having molecu-lar weight >100,000 daltons and multiple possible pointsof attachment, is rarely attempted. In the past and still to-day, researchers often work with mixtures of molecularconjugates having different points of attachment but sim-ilar composition. In special cases, a particularly reactivenaturally occurring residue such as cysteine permits easypreparation of homogeneous products.

V. NOMENCLATURE

The complexity of typical bioconjugates has so fardefeated efforts to systematize their nomenclature. It iscustomary to name products by referring to starting ma-terials; thus antibody enzyme conjugates are named ac-cording to the individual molecules involved.

VI. EXAMPLES

A. Targeted Therapy

An important application of bioconjugates involves thetreatment of disease by specifically targeting disease sites.Combining the properties of an antibody, synthetic poly-mer, or liposome with those of a drug, nucleic acid, orradionuclide is currently an important research activity.Some therapeutic products are now commercially avail-able, and many clinical trials are under way.

The important properties of antibodies are their spe-cific binding to molecular targets on diseased cells andthe physiological consequences of their large size, whichinfluence where they localize and how long they remainthere. Artificial polymers such as PEG share some of thelatter properties, as do liposomes. Smaller molecules canbe used in place of these targeting moieties, includingsmaller fragments of antibodies that retain binding speci-ficity (Fab fragments, single-chain antigen binding pro-teins, targeting peptides, etc.) and non-protein moleculesbased on nucleic acids (aptamers) or other structures foundto bind to or accumulate in biological targets. A typical

use of the polymer PEG is to modify the biological prop-erties of proteins. By attaching PEG groups to lysine sidechains, it is possible to reduce the likelihood that the im-mune system will mount a response to the protein. Thispermits repeated use of the protein over several cyclesof treatment. Another application of PEG conjugation isto modify the rate at which a protein leaves the circula-tion, allowing therapeutic bioconjugates to circulate forlonger periods. The important properties of nucleic acidsand their analogs involve modification of gene expressionprocesses in target cells. This may occur by the insertionof new genes into cells, or by interference with the expres-sion of existing genes by “antisense” binding. Antisenseoligodeoxyribonucleotides (ODNs) can inhibit gene ex-pression by hybridization to complementary messengerRNA sequences inside cells.

B. Imaging

Bioconjugates of low molecular weight are used exten-sively in imaging sites of human disease by nuclear tech-niques such as 99mTc scintigraphy (making images fromgamma photons). Here a coordination complex formedbetween the metal technetium and an organic ligand withtargeting properties (e.g., a small peptide, a hormone ana-log, etc.) serves to localize the radiotracer in the tar-get tissue. Positron Emission Tomography (PET) with18F-labeled fluorodeoxyglucose or other small moleculesprovides higher resolution images by detecting two anni-hilation photons in coincidence.

Magnetic Resonance Imaging (MRI) sometimes makesuse of bioconjugates containing paramagnetic metals suchas gadolinium or manganese conjugated to polymers tomodify their relaxivity properties.

Optical imaging probes that fluoresce in the near-infrared wavelength region can be conjugated with tar-geting molecules and used for in vivo imaging.

1. Separation and Analysis in vitro

Immobilization of biomolecules by linking them to sur-faces, such as the wells of microtiter plates, magneticbeads, or chromatography supports for subsequent, spe-cific binding of target molecules is often the first step to invitro analysis. The nature of the linkage may be as simpleas adsorption of a protein to a hydrophobic surface, or itmay involve covalent attachment of the biomolecule to re-active groups on the surface using the chemical principlesdiscussed above. With modern techniques, it is usuallypossible to accomplish this surface attachment withoutsignificant loss of biological properties.

Conjugates containing antibodies that specifically bindto a molecule of interest are widely used in bioanalytical




chemistry. Conjugates of antibodies with enzymes areused for assays of all sorts of biological molecules byenzyme-linked immunosorbent assay (ELISA). For ex-ample, a protein in a biological fluid can be analyzed byfirst trapping it using an immobilized antibody that bindsspecifically and strongly to an epitope on that particu-lar protein. After washing away unbound substances, asecond antibody that binds another epitope on the tar-get protein is allowed to bind to the immobilized target.This second antibody is conjugated to an enzyme; at thispoint, the sample contains a number of immobilized en-zyme molecules, related to the number of target proteinmolecules in the original biological fluid. The immobi-lized enzyme molecules are then quantified by standardtechniques of enzymology, such as catalytic productionof a light-absorbing product that can be measured spec-trophotometrically. Western blotting is another applica-tion of antibody conjugates, wherein electrophoretic sep-aration of proteins is followed by blotting the separatedbands onto a special membrane. The bands containing aspecific antigen are then stained and visualized by pro-cedures not unlike those used for ELISA, except that thecolored product is precipitated where the bands are.

2. Biophysical Studies

Bioconjugate chemistry plays a prominent role in studyingconformational changes in macromolecules or the bind-ing of one macromolecule to another. Techniques such aschemical cross-linking, fluorescence energy transfer, flu-orescence polarization, electron paramagnetic resonance,or various other spectroscopy or microscopy experiments,generally involve preparation of a bioconjugate tagged insome way by a reporter molecule. Fluorescence energytransfer from an excited donor moiety to an acceptor moi-ety depends on the distance between donor and acceptor;it is useful for measuring distances in the 10 100 A range,appropriate for macromolecular complexes. Fluorescencepolarization is sensitive to difference in the rotational dif-fusion (tumbling) of molecules in solution, and can readily

reveal the binding of a small molecule to a macromolecule.Electron paramagnetic resonance provides related infor-mation about changes in tumbling behavior, and can alsobe used to study the distances between paramagnetic cen-ters. Electron microscopy can show the spatial relationsbetween metal clusters attached to macromolecules asprobes. Recently, an approach has emerged that involvescleaving the sugar-phosphate backbone of a nucleic acidor the polypeptide backbone of a target protein, using acutting protein that binds to the target. The cutting proteinis prepared by conjugation with a small chemical reagent,which acts as an artificial nuclease or protease. Using toolssuch as nucleic acid sequencing or Western blotting, thesites on the target macromolecule that are within reach ofthe cutting reagent can be identified. Chemical reagentsdesigned for use in all these biophysical studies are com-mercially available from specialty vendors.


BIOPOLYMERS • CARBOHYDRATES • ENZYME MECHA-NISMS • NUCLEIC ACID SYNTHESIS • PROTEIN

SYNTHESIS

BIBLIOGRAPHY

Bioconjugate Chemistry, a journal published by the American ChemicalSociety (http://pubs.acs.org/BC ).

Hermanson, G. T. (1996). “Bioconjugate Techniques,” Academic Press,San Diego, CA.

Lundblad, R. L. (1995). “Techniques in Protein Modification,” CRCPress, Boca Raton, FL.

Means, G. E., and Feeney, R. E. (1971). “Chemical Modification ofProteins,” Holden-Day, San Francisco, CA.

Aslam, M., and Dent, A. (1998). “Bioconjugation: Protein CouplingTechniques for the Biomedical Sciences,” Grove’s Dictionaries Inc.,New York.

“Pierce Catalog and Handbook” (1994, 1995). Life Science & AnalyticalResearch Products, Pierce Chemical Company, Rockford, IL.

Wong, S. S. (1991). “Chemistry of Protein Conjugation and Cross-Linking,” CRC Press, Inc., Boca Raton, FL.

http://pubs.acs.org/BC

P1: LLL/LLL P2: FJU Final Pages Qu: 00, 00, 00, 00


CarbohydratesHassan S. El KhademThe American University

I. IntroductionII. Monosaccharides

III. OligosaccharidesIV. PolysaccharidesV. Oligonucleotides and Polynucleotides

GLOSSARY

Aldoses Chiral polyhydroxyalkanals having three ormore carbon atoms (polyhydroxyalkanals having fewerthan three carbons are achiral). Aldoses having enoughcarbon atoms to form five- or six-membered oxy-genated rings readily undergo cyclization by in-tramolecular nucleophilic attack of suitably located hy-droxyl groups onto the aldehydic groups. Saccharidespossessing five-membered rings are called furanoses,and those possessing six-membered ones, pyranoses.

Anomeric configuration The cyclization of an aldose ora ulose to form a furanose or a pyranose ring occursby intramolecular nucleophilic attack of a suitably lo-cated hydroxyl group on a carbonyl function (presentin the same molecule). This cyclization results in theformation of a chiral hemiacetal group in place of anachiral carbonyl group, which increases the numberof chiral centers present in the molecule by one. Theconfiguration of the newly created center is designatedby the greek letter α or β, followed by the D or L no-tation of the chiral center farthest from the carbonyl.Thus, one speaks of α-D-aldopento-furanoses or β-L-hexopyranuloses.

Furanoses Aldoses having four carbon atoms or moreand uloses having five carbon atoms or more can formfive-membered rings. Such cyclic monosaccharides arecalled furanoses because they are related to tetrahy-drofuran. Aldopentoses having such five-memberedrings are designated aldopentofuranoses, and hexu-loses, hexofuranuloses. In solution, furanoses existmainly in envelope and twist conformations, alternat-ing in a wavelike motion between them.

Monosaccharides Monomeric saccharides, which aregrouped according to the number of carbon atomsin their skeleton into pentoses, hexoses, and so on.Members of these groups may exist in the form ofpolyhydroxy aldehydes, called aldoses, or polyhydroxyketones, called uloses. Accordingly, certain monosac-charides are designated aldopentoses, other hexuloses,and so on.

Mutarotation Designates changes in optical rotationsobserved over a period of time. The phenomenon oc-curs when chiral substances equilibrate in solutionwith one or more isomeric forms. Saccharides usu-ally crystallize in structurally pure form; for example,D-glucose usually separates in the α-D-glucopyranoseform. When this is dissolved in water, it undergoes

369

P1: LLL/LLL P2: FJU Final Pages


370 Carbohydrates

equilibration to yield a complex mixture of six isomers(two furanoses, two pyranoses, an acyclic carbonylform, and a hydrated carbonyl form). The rotation of theequilibrium mixture is naturally different from that ofthe original pure α-D-glucopyranose, which accountsfor the observed change in rotation.

Oligosaccharides Low molecular weight polymers madeup of 2 to 10 monosaccharide units linked togetherby acetal linkages. Oligosaccharides are classified ac-cording to degree of polymerization into disaccha-rides, trisaccharides, and so on. They are also groupedinto reducing and nonreducing oligosaccharides. If oneof the terminal monosaccharides possesses a hemiac-etal group, the oligomer is referred to as a reducingoligosaccharide (since it is susceptible to mild oxi-dants). If, on the other hand, it does not possess ahemiacetal group (such as when two anomeric hy-droxyl groups on adjacent monosaccharides are in-volved in acetal bond formation), the oligosaccharideis referred to as nonreducing. Reducing oligosaccha-rides mutarotate, because in solution one of their ter-minal monosaccharides is in equilibrium with otherforms, while nonreducing oligosaccharides do notmutarotate.

Polysaccharides Polymers composed of more than10 monosaccharide units linked by acetal bridges. Mostnaturally occurring polysaccharides have degrees ofpolymerization ranging between 102 and 105. Linearpolysaccharides such as cellulose are usually micro-crystalline, are insoluble in water, and form strong filmsand fibers, whereas highly branched polysaccharidessuch as gums form gets and produce brittle films.

Pyranoses Aldoses having five carbon atoms or more anduloses having six carbon atoms or more can form strain-less six-membered rings. Such cyclic forms of saccha-rides are called pyranoses, because they are related totetrahydropyran. Aldopentoses having six-memberedrings are called aldopentopyranoses, and hexulosespossessing such rings are designated hexopyranuloses.The most stable conformations of pyranoses are the twochair conformations 1C4 and 4C1, of which the latter isusually the more stable.

Saccharide (Sugarlike; synonymous with carbohydrate).It is often prefixed by “mono,” “oligo,” or “poly” todesignate degree of polymerization.

Sugars Sweet-tasting saccharides (monosaccharides anddisaccharides are sweet tasting, whereas most trisac-charides and all higher saccharides are devoid of taste).Because the term sugar encompasses compounds pos-sessing a physiological rather than a chemical property,and because the term has been used indiscriminately todesignate different saccharides (sucrose by the publicand D-glucose by the medical profession), it is slowly

being replaced in chemistry texts by less ambiguousnames such as monosaccharides and disaccharides.

Uloses (also known as ketoses) Chiral polyhydroxyalka-nones that possess at least four carbon atoms in theirskeleton (lower members are achiral). Uloses that pos-sess a sufficient number of carbon atoms to cyclize instrainless rings do so and yield furanuloses and pyran-uloses.

CARBOHYDRATES are monomeric, oligomeric, orpolymeric forms of polyhydroxyalkanals or polymericforms of polyhydroxyalkanals or polyhydroxyalkanones,which are collectively called monosaccharides. In thenineteenth century, when the empirical formulas of or-ganic compounds were determined, it was found thatthe sugars and polysaccharides known at the time hadthe formula Cn(H2O)y . They were accordingly given thename carbohydrates (i.e., hydrates of carbon). Today’susage of the word carbohydrate applies to a large num-ber of monomeric, oligomeric, and polymeric compounds,which do not necessarily have their hydrogen and oxygenatoms in the molecular ratio of 2:1, but which belong tothe group of compounds called monosaccharides or can bereadily converted to members of that group by hydrolysis.

I. INTRODUCTION

A. Historical Background

The origins of carbohydrate chemistry can be traced backto the civilizations of antiquity. For example, the manufac-ture of beer and wine by alcoholic fermentation of grainstarch and grape sugar is well documented on the walls ofancient Egyptian tombs. The isolation of cellulose fibersfrom cotton and flax was started by the civilizations of theFar and Near East and was introduced by the Greeks toEurope. Gums and resins were valued commodities at thebeginning of the Christian era.

The isolation of sucrose from the juice of sugarcanemarks an important milestone in sugar chemistry. In theFar East, where this plant grew, sugar was isolated as ayellowish syrupy concentrate that crystallized, on stand-ing, into a brown mass, and a number of Chinese recipesdating from the fourth century describe in detail how thesugarcane juice was concentrated (Fig. 1). It is interestingthat the Sanskrit word sugar means “sweet sand,” whichaptly describes the properties of crushed, raw sugar. Sig-nificant progress in the manufacture of sugar occurred af-ter the French Revolution, when Europe under blockadehad to rely on the sugar beet for the manufacture of thisimportant commodity. Improved methods of isolation and



Carbohydrates 371

FIGURE 1 Reproduction of a seventh-century Chinese drawingrepresenting the manufacture of sugar.

refining were developed, including treatment of the syrupwith lime to precipitate the calcium complex, regenerationof the sugar with sulfur dioxide, and decolorization withanimal charcoal.

B. Importance of Carbohydrates

In addition to the manufacture of sucrose and paper, whichare today important industries, carbohydrates play a majorrole in a number of other industries. These include (1) thefood industry, which uses huge amounts of starch of var-ious degrees of purity in the manufacture of baked goodsand pastas, of gums in food processing, and of mono-and oligosaccharides as sweeteners and employs carbo-hydrates in fermentation to make beer and wine; (2) thetextile industry, which, despite the advent of synthetics, isstill dependent to a large extent on cellulose; (3) the phar-maceutical industry, particularly in the areas of antibiotics,intravenous solutions, and vitamin C; and (4) the chemicalindustry, which produces and markets several pure sugarsand their derivatives.

Carbohydrates also play a key role in the process of life;the master molecule DNA is a polymer made up of re-peating units composed of four nucleotides of 2-deoxy-D-erythro-pento-furanose (2-deoxy-D-ribose), the sequenceof which constitutes the coded template responsible forreplication and transcription. Saccharide derivatives alsoform part of many vital enzyme systems, specifically ascoenzymes.

The best-known use of carbohydrates is undoubtedly innutrition, as members of a major food type that are metab-olized to produce energy. Although the average percent-age of carbohydrates consumed by humans in comparisonwith other food types differs from country to country, and

figures are often inaccurate or unavailable, the percent-age for the world as a whole has been estimated to bemore than 80%. The exothermic reactions that produceenergy in the cell are the outcome of a number of com-plex enzyme cycles that originate with hexoses and endup with one-, two-, or three-carbon units. The energy re-leased from these reactions is stored in the cell in theform of a key sugar intermediate, adenosine triphosphate(ATP), and released when needed by its conversion to thedi- or monophosphate. Finally, carbohydrates are the mostabundant organic components of plants (>50% of the dryweight) and therefore constitute the major part of our re-newable fuels and the starting material from which mostof our fossil fuels were made.

C. Classification

Carbohydrates are classified according to their degree ofpolymerization into monomeric carbohydrates, which in-clude monosaccharides and their derivatives, and poly-meric carbohydrates, which comprise oligosaccharides,polysaccharides, DNA, and RNA. The polymeric carbohy-drates differ in the type of bridge that links their monosac-charide units. Thus, oligosaccharides and polysaccha-rides are polyacetals, linked by acetal oxygen bridges,whereas DNA and RNA are poly(phosphoric) esters,linked by phosphate bridges. In addition to these well-defined groups of carbohydrates, there exist a number ofderivatives, for example, antibiotics, that are best studiedas a separate group, because some of their members maybe monomeric, whereas others are oligomeric.

1. Monosaccharides

Monosaccharides are chiral polyhydroxyalkanals or poly-hydroxyalkanones that often exist in cyclic hemiacetalforms. Monosaccharides are divided into two majorgroups according to whether their acyclic forms possessan aldehyde group or a keto group, that is, into aldoses orketoses (glyculoses). Each of these, in turn, is classified ac-cording to the number of carbon atoms in the monosaccha-ride chain (usually 3–10) into trioses, tetroses, pentoses,hexoses, and so on. By prefixing “aldo” to these names,one can define more closely a group of aldoses, for exam-ple, aldopentoses, whereas for ketoses, it is customary touse the ending “ulose,” as in hexulose. Finally monosac-charides can be grouped according to the size of theirrings into five-membered furanoses and six-memberedpyranoses. It should be noted that, in order to form afuranose ring, four carbon atoms and one oxygen atomare needed, so that only aldotetroses and higher aldoses,and 2-pentuloses and higher ketoses, can cyclize in thisring form. Similarly, in order to form a pyranose ring five



372 Carbohydrates

TABLE I Monosaccharides: Aldoses

No. of chiralMonosaccharide Aldosea carbons Aldofuranose Aldopyranose

Triose Aldotriose 1 — —

Tetrose Aldotetrose 2 Tetrofuranose; aldotetraofuranose —

Pentose Aldopentose 3 Pentofuranose; aldopentofuranose Pentopyranose: aldopentopyranose

Hexose Aldohexose 4 Hexofuranose; aldohexofuranose Hexopyranose; aldohexopyranose

Heptose Aldoheptose 5 Heptofuranose; aldoheptofuranose Heptopyranose; aldoheptopyranose

Octose Aldooctose 6 Octofuranose; aldooctofuranose Octopyranose; aldooctopyranose

Nonose Aldononose 7 Nonofuranose; aldononofuranose Nonopyranose; aldononopyranose

Decose Aldodecose 8 Decofuranose: aldodecofuranose Decopyranose; aldodecopyranose

a Although an achiiral aldobiose (glycolic aldehyde) exists, it is not considered to be a saccharide because, by definition, a saccharidemust contain at least one asymmetric carbon atom.

carbon atoms and one oxygen atom are required, so thatonly aldopentoses and 2-hexuloses, as well as their higheranalogs, can cyclize in this form. Ultimately, by combiningthe ring type to the names used above (e.g., aldopentose orhexulose), such combination names as aldopentofuranoseand hexopyranulose can be formed, which define withoutambiguity the group to which a monosaccharide belongs(see Table I and II).

As their name denotes, monosaccharides are mono-meric in nature and, unlike the oligosaccharides andpolysaccharides, which will be discussed later (SectionsIII and IV), they cannot be depolymerized by hydrolysisto simpler sugars. Monosaccharides and oligosaccharidesare soluble in water; their solutions in water are often sweettasting, and this is why they are referred to as sugars.

2. Oligosaccharides

Oligosaccharides and polysaccharides are polyacetals thatrespectively have, as their names denote (oligo = “few”;poly = “many,” in Greek), a low (2–10) or a high (>10)degree of polymerization (DP). They are composed ofa number of monosaccharides linked together by acetal

TABLE II Monosaccharides: Ketoses (Glyculoses)

No. of chiralMonosaccharide Ketosea carbons Ketofuranose Ketopyranose

Tetrose Tetrulose 1 — —

Pentose Pentulose 2 Pentulofuranose —

Hexose Hexulose 3 Hexulofuranose Hexulopyranose

Heptose Heptulose 4 Heptulofuranose Heptulopyranose

Octose Octulose 5 Octulofuranose Octulopyranose

Nonose Nonulose 6 Nonulofuranose Nonulopyranose

Decose Deculose 7 Deculofuranose Deculopyranose

a Although an achiiral triulose (1,3-dihydroxyacetone) exists, it is not considered to bea saccharide because it lacks an asymmetric carbon atom.

oxygen bridges and yield on depolymerization (hydro-lysis) one or more types of monosaccharide. Oligosaccha-rides are further grouped into (1) simple (true) oligosac-charides which are oligomers of monosaccharides thatyield on complete hydrolysis only monosaccharides; and(2) conjugate oligosaccharides, which are oligomers ofmonosaccharides linked to a nonsaccharide, such as a lipidaglycon, usually a long chain fatty acid, alcohol or amine,or a carbocyclic steroid or terpenoid. The carbohydrateportion of these compounds is resposible for their speci-ficity and is involved in cell recognition. Two main typesof compounds are recognized: glycolipids, which can beof animal, plant, or microbial origin, and gangliosides,which contain sialic acid and are found in large amountsin cerebral tissues of patient defficient in the enzyme N-acetylhexosaminidase. There are two ways to classify sim-ple oligosaccharides further. The first is, according to DP,into disaccharides, trisaccharides, tetrasaccharides, and soon, and the second, according to whether the oligomerchain has at one end a hemiacetal or a hemiketal function(a latent aldehyde or keto group). Such terminal groups,if present, are readily converted to carboxylic groupsby mild oxidants, and accordingly, oligosaccharides



Carbohydrates 373

TABLE III Oligosaccharidesa

Reducing Nonreducing

Homooligosaccharides Heterooligosaccharides Homooligosaccharides Heterooligosaccharides

Simple oligosaccharides

Disaccharides

Maltose Lactose Trehalose Sucrose

4-α-D-Glcp-D-Glc 4-β-D-Galp-α-D-Glc α-D-Glcp-α-D-Glcp β-D-Fru f -α-D-Glcp

Cellobiose Lactulose Isotrehalose Isosucrose

4-β-D-Glcp-D-Glc 4-β-D-Galp-D-Fru β-D-Glcp-β-D-Glcp α-D-Fru f -β-D-Glcp

Isomaltose Melibiose

6-α-D-Glcp-D-Glc 6-α-D-Galp-D-Glc

Gentiobiose Turanose

6-β-D-Glcp-D-Glc 3-α-D-Glcp-D-Fru

Trisaccharides

Maltotriose Manninotriose Raffinose

4-α-D-Glcp-Maltose 6-α-D-Galp-Melibiose 6-α-D-Galp-SucroseTetrasaccharides

Maltotetraose Stachyose

4-α-D-Glcp-Maltotriose 6-α-D-Galp-Raffinose

Pentasaccharides

Hexasaccharides

Heptasaccharides

Octasaccharides

Nonasaccharides

Decasaccharides

Conjugate oligosaccharides Glycolipids

Gangliosides

a Glc, Glucose; Gal, galactose; Fru, fructose; f , furanose; p, pyranose.

possessing these groups are referred to as reducing, incontradistinction to those that resist such oxidation andare designated nonreducing (see Table III). Thus. whereasall monosaccharides are reducing, there are reducing andnonreducing disaccharides, trisaccharides, and so on. Be-cause monosaccharides and simple oligosaccharides aresoluble in water and sweet tasting, they are called sugars.It should be noted, however, that the sweetness of oligosac-charides decreases with increasing DP and disappears atDP 4–5.

3. Polysaccharides

Polysaccharides and oligosaccharides are polymeric innature and are structurally similar (both are polyac-etals having oxygen bridges linking the monosaccharidemonomers), but they may differ markedly in DP; thepolysaccharides may reach a DP of 105, whereas, by def-inition, the maximum DP for oligosaccharides is 10. Al-though, by convention, compounds having a DP of 11or more are designated polysaccharides, a difference be-tween the properties of a lower polysaccharide with a DP

of 11 and those of a higher (DP = 10) oligosaccharide canhardly be detected. However, most polysaccharides havea much higher DP than oligosaccharides, which rendersquite significant the sum of the gradual changes that occurin their physical properties with increasing DP. For exam-ple, because, with a rise in DP, the solubility decreases andthe viscosity increases, some higher polysaccharides, forexample, cellulose, are completely insoluble in water (alloligosaccharides are soluble), while the increase in vis-cosity may cause the solutions of other polysaccharides toset and gel.

There are several ways to classify polysaccharides; acommon one is to group them according to their sources,that is, into plant and animal polysaccharides, and thensubdivide the former into skeletal polysaccharides (cel-lulose, etc.), reserve polysaccharides (starch, etc.), gumsand mucilages, algal polysaccharides, bacterial polysac-charides, and so on. The disadvantage of this classificationis that it tells us very little about the chemistry of thesepolymers.

The classification used in chemistry texts usually dis-tinguishes between (1) simple (true) polysaccharides that



374 Carbohydrates

TABLE IV Polysaccharides

Homopolysaccharides Heteropolysaccharides

Linear Branched Linear Branched

Simple polysaccharides

Amylose (α-D-glucan) Amylopectin Mannans Gums

Cellulose (β-D-glucan) Glycogen Xylans Mucilages

Chitin (D-glucosaminan) Pectins

Algin

Agar

Bacterial polysaccharides

Conjugate polysaccharides Peptidoglycans

Glycoproteins

Lectins

afford on depolymerization only mono- and oligosac-charides or their derivatives (esters or ethers) and(2) conjugate polymers made up of a polysaccharidelinked to another polymer, such as a peptide or a protein(to form a glycopeptide or a glycoprotein). True polysac-charides are, in turn, grouped into two major classes: (1)homopolysaccharides, which are polymers having as a re-peating unit (monomer) one type of monosaccharide; and(2) heteropolysaccharides, which are made up of morethan one type of monosaccharide. Because the shape ofpolymers significantly influences their physical proper-ties, each of these types of polymer is further divided intolinear and branched polysaccharides (Table IV).

4. DNA, RNA, Nucleotides, and Nucleosides

Unlike oligo- and polysaccharides, which are polyacetalslinked by oxygen bridges, DNA and RNA are polyesterslinked by phosphate bridges. DNA is the largest knownpolymer; its DP exceeds 1012 in human genes and de-creases as the evolution ladder is descended. This giantmolecule plays a key role in replication and in transcrip-tion. It achieves the latter by doubling one of its strandswith a smaller polymer, mRNA, which, in turn, bindswith a string of oligomers, tRNA, to form the peptidechain. The monomers of both DNA and RNA are madeup of phosphorylated 2-deoxy-D-erythro-pentofuranosyl-and D-ribofuranosyl-purine and -pyrimidine bases, des-ignated nucleotides. The latter can undergo hydrolysis oftheir phosphoric ester groups to afford simpler monomers,the nucleosides. Thus it is apparent that this group can alsobe divided according to DP into monomers (nucleosidesand nucleotides), oligomers (tRNA), and polymers (DNAand mRNA).

Some carbohydrate derivatives may belong to morethan one of the above-mentioned groups. For example,

carbohydrate-containing antibiotics may be monosaccha-ride or oligosaccharide in nature.

II. MONOSACCHARIDES

Monosaccharides (monomeric sugars) exist as chiral poly-hydroxyalkanals, called aldoses, or chiral polyhydrox-yalkanones, called ketoses, which are further classified(see Tables I and II) according to the number of carbonatoms in their chains into trioses, tetroses, pentoses, hex-oses, and so on, and according to the type of ring they forminto furanoses and pyranoses (five- and six-memberedrings).

A. Structure, Configuration, and Conformationof Monosaccharides

1. Structure

The presence of an aldehyde group in aldoses such asD-glucose was established by the fact that (1) aldoses reactwith carbonyl-group reagents, giving oximes and hydra-zones; and (2) aldoses are oxidized to acids that possessthe same number of carbon atoms. Furthermore, by es-timating the number of acetyl groups in fully acetylatedaldoses, it is possible to determine the number of hydroxylgroups present in the starting aldoses (e.g., five OH groupsin aldohexoses) and to determine their structural formulas.

2. Configuration

The system commonly used to represent three-dimensional linear molecules such as monosaccharidestwo dimensionally is the Fischer projection formula. Thisaffords an unambiguous way of depicting monosaccha-rides, as follows. (1) The carbon chain is drawn vertically,



Carbohydrates 375

FIGURE 2 Models and formula of D-Erythrose. The two modelson the left are rotamers of the compound on the right.

with the carbonyl group at (or nearest to) the top and thelast carbon atom in the chain (i.e., the one farthest fromthe carbonyl group) at the bottom. (2) Each carbon atomis rotated around its vertical axis until all of the vertical(C C) bonds in the chain lie below an imaginary curvedplane such as that of a rolled piece of paper, and all ofthe horizontal bonds (parallel to the x axis) lie above theplane of the paper. The curved plane is then flattened, andthe projection of the molecule is represented as viewed(see Fig. 2).

a. Relative and absolute configuration. The rela-tive configuration of D-glucose was established by EmilFischer in 1891 and constituted at the time a monumentalachievement, for which he earned a Nobel prize. Nowa-days, the determination of the absolute configuration ofa monosaccharide offers no difficulty, because the con-figurations of a large number of related compounds areavailable. The unknown is simply converted to a com-pound of known configuration by means of reactions thatdo not affect the configuration at the chiral center(s).

Since D-glucose is an aldohexose, it must possess fourchiral carbon atoms and can exist in 24 = 16 stereoiso-mers (see Table I). In order to determine which of theseisomers is actually D-glucose, a reference compound ofknown absolute configuration is needed, which can beprepared from, or converted to, D-glucose. Because thefirst determination of the absolute configuration of an or-ganic compound had to await the advent of X-ray crystal-lographic techniques developed in the middle of the twen-tieth century, such a reference compound did not exist inFischer’s time. This is why he was able only to proposea relative configuration for the monosaccharides knownin his time. He chose, as the reference compound, thedextrorotatory form of glyceraldehyde, now designated D-(+)-glyceraldehyde, and arbitarily assumed that the OHattached to C-2 in this compound is to the right whenrepresented by a Fischer projection formula (see Fig. 3).

FIGURE 3 Fischer projection formula of D-glycer-aldehyde.

It was fortunate that the arbitary assignment was laterconfirmed by X-ray crystallography. Otherwise, all ofthe configurations assigned to carbohydrates during the60 years that followed Fischer’s assignment would havebeen in error, causing confusion and chaos in the chemicalliterature.

Once the absolute configuration at C-2 of the D-(+) iso-mer of glyceraldehyde had been established, it could berelated to the configuration of the corresponding center inany aldose that is obtained from this isomer by ascendingthe series or that yields this isomer by repeated descendingdegradations (Fig. 4). It should be stated at this point that,because C-2 in glyceraldehyde corresponds to the chiralcenter farthest from the carbonyl group in any aldose or ke-tose, it is possible to group monosaccharides into two fam-ilies: one related to D-(+)-glyceraldehyde and designatedby the prefix D, and one related to L-(−)-glyceraldehydeand designated by the prefix L. Monosaccharides of the D

family all have the (R) configuration at the chiral centerfarthest from the carbonyl (the OH attached to this chiralcarbon is to the right in a Fischer projection). Conversely,monosaccharides of the L family have the (S) configura-tion at this center (the OH group is to left).

b. Configuration of the acyclic form of aldoses.Fischer used an ingenious method to determine whetherin an aldose molecule, the configurations of the chiral cen-ters that are equidistant from the center (e.g., C-2 and C-4in a pentose) had the same sign. He determined whetherconversion of the two groups situated at the “top” andthe “bottom” of an aldose molecule (CHO and CH2OH)into the same type of group, for example, a carboxylicgroup, rendered the product achiral (afforded a meso com-pound). If so, he could conclude that a plane of symmetrywas created during the conversion (oxidation) and that theconfiguration of the chiral centers situated at equal dis-tances from this plane of symmetry was identical. He alsoreasoned that, if the dicarboxyclic acid possessed an axisof symmetry, it could be obtained from only one aldose,whereas if it lacked an axis, it could be obtained from twodifferent aldoses, an aldose having the aldehyde group atthe top and the primary hydroxyl group at the bottom, andanother aldose having the aldehyde group at the bottomand the primary hydroxyl group at the top. Using this rea-soning. Fischer was able to determine the configuration ofall the aldoses known in his time.

c. Cyclic structures and anomeric configuration.Soon after the configuration of the acyclic form of glu-cose and the other aldoses had been established, it be-came apparent that these structures could not representthe major components of the equilibrium mixture. Thus,the IR spectrum of D-glucose does not exhibit a strongcarbonyl band at 1700 cm−1, and its 1H NMR spectrum



376 Carbohydrates

FIGURE 4 Aldoses of the D family.

lacks an aldehydic proton at δ 10, also characteristic ofsuch a group.

Two forms of D-glucose were isolated and designated,α and β (a third one was also isolated, but it proved tobe an equilibrium mixture). Each of the two forms has, insolution, a characteristic optical rotation that changes withtime until it reaches a constant value (that of the equilib-rium mixture). The change in optical rotation with timeis called mutarotation, and it is indicative of molecularrearrangements occurring in solution:

α-D-Glucose ⇀↽ mixture ⇀↽ β-D-glucose[α]D + 112.2◦ + 52.7◦ +18.7◦

It is now known that D-glucose and the other aldoses thathave the necessary number of carbon atoms exist mainlyin the form of five- or six-membered cyclic hemiacetalsand rarely in seven-membered ones. These forms are theoutcome of an intramolecular nucleophilic attack by thehydroxyl oxygen atom attached to C-4 or C-5 on the car-bonyl group. The five-membered ring produced in the first

case is related to tetrahydrofuran and is therefore desig-nated furanose. The six-membered ring produced in thesecond case is related to tetrahydropyran and is calledpyranose. Because cyclization converts an achiral alde-hyde carbon atom to a chiral hemiacetal carbon atom, twoisomers are produced, which have been designated α andβ. Accordingly, a solution of an aldose at equilibrium con-tains a mixture of at least two (α and β) furanoses, two(α and β) pyranoses, as well as traces of the acyclic formand its hydrate (see Scheme 1).

Although usually there is a preponderance of the twopyranose forms, the composition of the equilibrium mix-ture and the contribution of each of the various forms differfrom one sugar to another, depending on the instability fac-tors dictated by the configuration. For example, becausethe interaction between cis OH groups on adjacent carbonatoms renders this arrangement considerably less desir-able in a furanose ring than a pyranose ring, it was foundthat the all-trans-β-D-galactofuranose contributes more tothe equilibrium (3%) than β-D-glucofuranose (0.1%).



Carbohydrates 377

SCHEME 1 Cyclic and acyclic forms of D-glucose existing insolution.

The anomeric configuration refers to the chirality atC-1 of a cyclic aldose or C-2 of a cyclic 2-ketose. Thesecenters are achiral in the acyclic forms, and their chiralityis the result of their conversion to cyclic hemiacetals. Twoanomeric furanoses and two anomeric pyranoses can beproduced from an acyclic monosaccharide, provided thatthe latter possesses the requisite number of atoms (fourcarbon atoms for a furanose ring and five for a pyranosering). The configuration of the anomeric center, designatedby the Greek letters α and β, is related to the last chiralcenter in the following way. In the D series, if the OH onthe anomeric carbon is to the right in a Fischer projectionor down in a Haworth formula (see below), the isomer isα, and if the OH is to the left or up, it is β. Conversely, inthe L series, the α anomer has the C-1 OH to the left in aFischer projection or up in a Haworth formula, and the βanomer has this OH to the right or down (see Fig. 5).

The common method of depicting monosaccharides intheir cyclic forms, without assigning a specific conforma-tion to the ring, is to use the Haworth formula. A pentagon

having 108◦ angles or a hexagon with 120◦ angles is rep-resented as seen by an observer situated at an angle of∼60◦ above the plane of ring, and as a precaution againstoptical illusions, regarding the side closer to the viewer,the bonds nearest to the observer are thickened, to give asense of perspective (see Fig. 5).

Although rings may be turned around their centers, it iscustomary to orient them in such a manner that C-1 is tothe right, and the ring oxygen is farthest from the viewer.

3. Conformation

Pyranose rings can exist in a number of inter-convertibleconformers, of which the chair forms are the most stable.The number of recognized forms of the pyranose ring aretwo chair (C), six boat (B), four half-chair (H), six skew(S), and six sofa forms. (See Fig. 6 for the two chair andtwo of the boat forms.) To designate each of these forms,the number of the ring atom(s) lying above the plane of thepyranose ring is superscripted before the letter designatingthe form (C, B, H, S, etc.), and the number of the ringatom(s) lying below the plane is subscripted after the letter,e.g., 4C1 (see Fig. 7).

It is possible to determine the conformation of a saccha-ride or glycoside either experimentally or by determiningon purely theoretical grounds which conformer of a givenring (pyranose or furanose) will be the most stable.

If the conformation in the solid state is desired, X-raycrystallography or neutron diffraction is prescribed. It willprovide the exact location of each atom in the crystallattice and, by so doing, show the anomeric configura-tion and the conformation of the ring. X-ray crystallogra-phy and neutron diffraction afford unambiguous diagramssuch as Fig. 8, which shows the conformation of methylα-D-glucopyranoside as deduced from neutron diffractionanalysis using a computer.

If, on the other hand, it is desired to know the conforma-tion of a saccharide in a given solvent, NMR spectroscopyis used. Use is made of the fact that the anomeric protonof cyclic sugars and glycosides is the only one attachedto two oxygen atoms, which deshield it. Accordingly, an1H NMR spectrum of such compounds will reveal at lowfield a well-defined doublet (coupled by the H-2 proton),whose coupling constant J1.2 will give the dihedral anglebetween protons 1 and 2. Then by using proton decouplingtechniques, it is possible to identify H-2 and determine thedihedral angle between it and H-3, and so on. In this way itis possible to determine the relative orientation of the pro-tons on successive pairs of carbon atoms, and by summingup the data, one can obtain the conformation of the wholemolecule, as well as the anomeric configuration. Figure 9shows the 1H-NMR spectrum of α-D-glucopyranose.



378 Carbohydrates

FIGURE 5 α and β configuration of aldoses and ketoses.

FIGURE 6 Two chair and two boat conformations.

FIGURE 7 The two chair conformations of β-D-glucose.

Conformation analysis of the possible pyranose formsshows that the two chair conformations are so much morestable than the others that the analysis should be restrictedto the choice between the 4C1 and the 1C4 conformers.Models of the two forms are checked for the presence ofany of the instability factors listed in Table V and appro-priate numerical value(s) assigned to each. These are thensummed, and the form having the smallest numerical totalis the more stable conformer.

The most important instability factor in the pyranosering (with a value of 2.5 units) is a situation calleddelta 2. This is when the OH on C-2 bisects the anglebetween the ring oxygen atom and the oxygen atom ofthe anomeric hydroxyl group, forming an isosceles trian-gle having one oxygen atom at each corner. An instabilityfactor of 2.0 units is given to an axial CH2OH. An ax-ial hydroxyl group on the ring has an instability factor of1.0 unit. The 1,3 interaction between an axial CH2OH andan axial OH group on the same side of the ring producesan unfavorable situation, of an additional 2.5 units.



Carbohydrates 379

FIGURE 8 Neutron diffraction of methyl α-D-glucopyranoside.

The anomeric effect destabilizes the anomer hav-ing an equatorial hydroxyl group on C-1 (e.g., β-D-glucopyranose) due to the dipole–dipole interactions be-tween its unbonded electrons and the unbonded electronsof the ring oxygen atom. The anomeric effect depends onthe polarity of the solvent, and this is why one does notgive it a high numerical value when computing the insta-bility factors of free sugars dissolved in water. It does,

FIGURE 9 1H NMR spectra of a solution of α-D-glucose in DMSO-d6 at 400 MHz: (a) OH protons coupled and(b) OH protons decoupled.

TABLE V Free Energy of UnfavorableInteractions

Type of interaction Free energy (kJ)

Gauche-1,2

O-1-e-O-e 2.30

O-1-e-O-a 4.19

O-e-O-e or -a 1.47

C-e-O-e or -a 1.88

Axial–axial-1.3

O–O 6.28

O–C 10.47

O–H 1.88

C–H 3.77

Anomeric effect 1.26

however, play an important role during glycoside forma-tion, favoring the conformer having an axial OR group onC-1, such as an α-D-glucopyranoside. Empirical calcu-lations of the conformational free energy by summationof the unfavorable interactions have been used success-fully in conformational analysis to determine the morestable conformer of a pyranose in a chair form. Two typesof interaction are used: those between two gauche-1,2substituents [either equatorial–equatorial (ee) or axial–equatorial (ae)], and those between two syn-diaxial-1,3groups.



380 Carbohydrates

TABLE VI Stable conformation of D-Aldopyra-noses in Aqueous Solutions

Aldose Conformation

Aldohexoses

α-D-Allose 4C1

β-D-Allose 4C1

α-D-Altrose 4C1, 4C4

β-D-Altrose 4C1

α-D-Galactose 4C1

β-D-Galactose 4C1

α-D-Glucose 4C1

β-D-Glucose 4C1

α-D-Gulose 4C1

β-D-Gulose 4C1

α-D-Idose 4C1, 1C4

β-D-Idose 4C1

α-D-Mannose 4C1

β-D-Mannose 4C1

α-D-Talose 4C1

β-D-Talose 4C1

Aldopentoses

α-D-Arabinose 1C4

β-D-Arabinose 4C1, 1C4

α-D-Lyxose 4C1, 1C4

β-D-Lyxose 4C1

α-D-Ribose 4C1, 1C4

β-D-Ribose 4C1, 1C4

α-D-Xylose 4C1

β-D-Xylose 4C1

Using these values, it can be calculated that the 4C1

form of β-D-glucopyranose has a conformational energyof 8.38, considerably lower than that of the 1C4 form witha value of 33.5, leaving no doubt that the first is the morestable conformer. Table VI shows the conformation ofD-aldopyranoses in aqueous solutions.

The principal conformers of the furanose ring are theenvelope (E) and the twist (T) forms, of which the latterare the most stable. To designate a particular conforma-tion, the method is the same as that used for pyranoses;namely, the letter used to designate the form (E or T) issuperseded by the number of the ring atom situated abovethe plane of the ring and is followed by the number of thering atom below the plane of the ring.

For furanose rings, the most stable conformers are theenvelope and twist forms; these can exist in 10 arrange-ments each. Due to the low energy barriers between thetwist and envelope conformers, a sugar in the furanoseform is believed to undergo rapid interconversions be-tween them. The slightly more favored twist conformerwould rapidly pass through an envelope conformation

(which is less favored because it possesses two eclipsedcarbon atoms) to go to the next twist form. Because inter-action between two eclipsed carbon atoms is greater thanbetween a carbon atom and an oxygen atom, the ring oxy-gen atom tends to occupy a position along the plane of thering and to leave the puckering to carbon atoms.

B. Reactions of Monosaccharides

In reactions involving monosaccharides, it is important toremember that the functional groups found in the vari-ous cyclic and acyclic forms will be present side by sidein the reaction mixture. Thus, in a reaction involving analdohexose, a carbonyl group and two types of hydroxylgroups will be provided by the acyclic form. These arethe primary hydroxyl group attached to the terminal posi-tion, and four (less reactive) secondary hydroxyl groups.In addition, each of the cyclic forms will contribute threetypes of hydroxyl groups: a hemiacetal hydroxyl functionat C-1, a terminal primary hydroxyl group, and three (lessreactive) secondary hydroxyl groups.

It is also important to remember that the course of areaction does not depend totally on the relative amountof a given species. Thus, free saccharides readily undergonucleophilic additions, characteristic of carbonyl groups,even though carbonyl groups are present only in theiracyclic form, which contributes little to the equilibrium

SCHEME 2 Representation of the anomeric configurationD-glucose in a Fischer (top right) and a Haworth (bottom left)formula.



Carbohydrates 381

mixtures. For example, solutions of hexoses contain lessthan 1% of the acyclic form, and yet they readily affordcarbonyl group derivatives in high yields. This is becausethe addition of a nucleophile to the carbonyl group of anacyclic form will immediately shift the equilibrium in fa-vor of this form, allowing more of it to be formed and toreact with the nucleophile.

Concerning the hydroxyl groups, it should be recog-nized that the hydroxyl groups can play a dual role. Suchgroups can act as leaving groups when stronger nucle-ophiles attack the carbon atom to which they are attached.These nucleophilic substitution reactions may be of theSN1 or SN2 type. Alternatively, the oxygen of the hydroxylgroup acts as a nucleophile, adding to carbonyl groups andcarbonium ions or displacing good leaving groups to af-ford esters, acetals, ketals and ethers, and other species.The hydroxyl group of the hemiacetal function at C-1 of analdose or C-2 of a ketose is the most reactive of all the hy-droxyl groups found in a monosaccharide. The next mostreactive hydroxyl group is the primary hydroxyl group atthe terminal position. This is followed in reactivity by thesecondary hydroxyl groups. The oxidation of carbonyl andhydroxyl groups has important applications in industrialprocesses.

1. Reactions of the Carbonyl Group

The reactions of the carbonyl group of a monosaccharideinclude the nucleophilic addition of a carbon, nitrogen,oxygen, or sulfur atom. It should be noted that, althoughthese additions afford acyclic products, the latter may cy-clize, so that the product may ultimately be acyclic orcyclic. On the other hand, intramolecular nucleophilic ad-dition, by a hydroxyl group attached to the sugar chain onthe carbonyl carbon atom, can afford only cyclic products.Note also that, in both types of additions, the reaction mayoccur with or without subsequent loss of water.

The addition of carbon nucleophiles to the carbonylgroup of aldoses has been widely used to extend the carbonchains of saccharides, that is, to ascend the series. Similaradditions to the keto group of glyculoses have been usedto prepare branched sugars. Of particular value in form-ing C C bonds are the nucleophiles −CN, −CH2NO2,and −CH2N2, as well as the ylides and organometallicnucleophiles involved in the Wittig and Grignard reac-tions (Scheme 3).

The nitrogen nucleophiles commonly used withmonosaccharides contain a primary amino group attachedto a carbon, a nitrogen, or an oxygen atom. The productsobtained from the first type of nucleophile are rarely ofthe acyclic, Schiff base type, because they readily cyclize.The other two types of nucleophile afford carbonyl-groupderivatives that usually exist in acyclic forms. The major

SCHEME 3 Addition of carbon nucleophiles to carbonyl groups.

pathway for the reaction between a nitrogen nucleophileand a free sugar is via the acyclic form of the monosac-charide, which usually affords an acyclic addition product.The latter may subsequently lose water to yield an acycliccondensation product, or it may cyclize. The same reac-tion product may be formed by nucleophilic substitution



382 Carbohydrates

SCHEME 4 Amadori rearrangement of aldimines via eneaminols.

on the hemiacetal function of a cyclic sugar. Nucleophlicsubstitution is much slower than nucleophilic addition,but in this case its contribution is enhanced by the largeproportions of cyclic forms of the sugar present in theequilibrium mixture.

Aldoses and ketoses react with ammonia and aminesto give 1-deoxy-1-imino- and 2-deoxy-2-imino deriva-tives, respectively, both of which exist mainly in cyclicforms, referred to as glycosylamines. During the reac-tion of aldoses with amines, the 1-amino-1-deoxyuloses,that is, glycosylamines that are first formed, often rear-range to give 1-amino-1-deoxyketoses, called Amadoricompounds. The reaction leading to their formation is alsonamed after its discoerer and referred to as the Amadorirearrangement. Unlike glycosylamines, which exist pre-ponderantly in cyclic forms, Amadori compounds may becyclic or acyclic. The mechanism of the reaction has beenextensively studied by Weygand, who proposed the mech-anism shown in Scheme 4, which involves a sigmatropicrearrangement of an aldimine to a 1,2-enaminol, whichlater ketonizes.

Sigmatropic rearrangements leading to the formation ofenaminols are by no means restricted to imines, as theyoccur during the conversion of hydrazones into osazones,and are involved in the cyclization of the latter compounds.

The analogous reaction between monosaccharides andamino acids or peptides is of great importance to the foodindustry. This is because its ultimate outcome is the forma-tion of the dark polymeric products known as malanoidins,which give baked goods their characteristic color. The ini-tial stages of this complicated reaction, collectively knownas the Maillard reaction (Scheme 5), are well documented.They involve the formation of N -glycosylamino acids,

which undergo Amadori rearrangement to ketose aminoacid derivatives and then dimerize to diketose amino acids.The latter then undergo a series of double-bond migra-tions, via 1,2-enolizations and 2,3-enolizations, to affordmono- and dideoxyhexosuloses. The glycosuloses are thenattacked by amino acids and, after undergoing polymeriza-tions and decarboxylations, yield polymeric melanoidins.

The hydrazine derivatives of sugars are more reactivethan the sugars from which they are prepared. This isbecause, when a hydrazone residue is introduced intoan aldose molecule, the number of nucleophilic groupsthat are capable of entering a reaction is increased byunity (the second nitrogen atom of the hydrazone actsas a strong nucleophile), whereas the number of groupscapable of undergoing nucleophilic attack remains con-stant because nucleophiles will add to the C N groupat position 1 in the same way that they do to the C Ogroup of the parent sugar. In the case of osazones andother bis(hydrazones), the reactivity is further enhancedbecause a second C N group is introduced (in place of aless reactive H C OH group). As a result, the capacityof osazones and bis(hydrazones) to undergo addition andcyclization reactions is greater than that of monohydra-zones, which in turn is greater than that of the free sugarsthat yielded them.

Saccharide hydrazones have been prepared from un-substituted, monosubstituted, and N,N-disubstituted hy-drazines. The substituents attached to the hydrazine in-clude alkyl, aryl, or heteroaryl groups, as well as acyl,aroyl, thioacyl, thioaroyl, and sulfonyl groups.

Sugar hydrazones can exist as equilibrium mixturesof various tautomeric forms. This is apparent from thecomplex mutarotation they exhibit in solution. The most



Carbohydrates 383

SCHEME 5 Millard reaction scheme.

important of these structures are the acyclic Schiff basesand the two pairs of anomeric cyclic forms (the two five-and two six-membered rings). However, when crystal-lized, one form of the hydrazone is usually isolated. Forexample, the crystalline form of D-galactose phenylhydra-zone was shown by the following sequence of reactions tobe the acyclic form (see Scheme 6).

Osazones and bis(hydrazones) are hydrazine deriva-tives of aldosuloses and diuloses, respectively. If the twohydrazone residues are attached to C-1 and C-2 of a sac-charide, the derivative is referred to as an osazone andis named by suffixing osazone to the name of the ketosethat possesses the same carbon chain and the same con-figuration (irrespective of the sugar used in its prepara-tion); for example, D-arabino-2-hexulose phenylosazoneis the name given to the osazone obtained from D-glucose,D-mannose, and D-fructose. The stability of osazones is at-tributed to their chelated rings (see Scheme 7).

Reducing sugars react with hydroxylamine to giveoximes, which have been prepared for use in the Wohl

SCHEME 6 Formation of the same galactose hydrazone pen-taacetate from a pyranose, a furanose, and an acyclic sugaracetate.



384 Carbohydrates

SCHEME 7 Chelated ring structures of D-arabino-hexulose phenylosazone, represented by a Fischer formula onthe left.

degradation for descending the series. This reaction makesuse of the fact that, on refluxing with acetic anhydride,the oxime readily loses water (or acetic acid, if it is firstN-acetylated) and is converted to the nitrile. Then usingthe fact that the addition of HCN to aldehydes is reversible,the nitrile formed is hydrolyzed to an aldose having onecarbon atom less than the starting oxime (see Scheme 8).If carried out with ammonia, the hydrolysis affords anundesired intermediate bis(acetamido) derivative, whichmust by hydrolyzed with acid before the desired aldosecan be liberated.

The addition of l mol of RO− and RS− nucleophiles tothe carbonyl group of a saccharide affords the hemiacetalor thiohemiacetal, which in solution remains in equilib-rium with the free saccharide. On the other hand, the ad-dition of 2 mol of the nucleophile affords stable acetals ordithioacetals (see Fig. 11).

If one of the OH groups of the acyclic form of the sugaris the nucleophile, the product is one of the four possiblecyclic structures of a sugar (α- and β- furanose and α- andβ-pyranose). If, instead, the nucleophile originates fromwithout the sugar molecule, hemiacetals or thiohemiac-etals are obtained, depending on the nucleophile. Becausethese compounds are constantly in equilibrium with theacyclic form, they are not isolated and have little impor-tance in preparative chemistry. Three types of acetal arepossible, depending on the source of the nucleophiles(s)involved in the reaction: (1) If both nucleophile moleculesdo not arise from the sugar molecule, the product is anacyclic acetal or dithioacetal; (2) if one nucleophile formspart of the sugar molecule and one is not a part of it, thereis obtained a cyclic acetal or thioacetal, which is referred

SCHEME 8 The Wohl degradation is used to reduce the number of C atoms in sugar.

to as a glycoside or thioglycoside; (3) if both nucleophilemolecules are part of the same sugar molecule, the productis a bicyclic acetal. Acetals of types 2 and 3 are, in real-ity, derivatives of the cyclic forms of sugars, and, as such,they do not belong among the reactions of the carbonylgroup.

In general, the best route for preparing the acyclic formsof sugars and acyclic sugar derivatives is that from thedithioacetals; another way is to prepare an acyclic hy-drazone of the sugar, acetylate it, and then remove thehydrazone residue from the acetate with benzaldehyde.Because of the importance of dithioacetals in synthesis,their preparation, as well as their conversion to acetalsand other useful derivatives, will be examined. Dithioac-etals are readily obtained in acidic media by the treatmentof monosaccharides with the desired thio alcohol, usu-ally ethanethiol, in the presence of concentrated HCl (seeScheme 9).

The demercaptalation (removal of the thioacetalresidues) is usually conducted in the presence of a mix-ture of yellow mercuric oxide and cadmium carbonate ormercury chloride. If hydrolysis is needed in order to pre-pare the aldehydo sugar, water is added to the mixture,care being taken that the reaction is performed on the per-acetylated dithioacetal (to prevent formation of the cyclicforms).

2. Nucleophilicity Quotient of Monosaccharides

The nucleophilicity quotient (NQ) is a measure of the sus-ceptibility of polyfunctional molecules to cyclization byintramolecular, nucleophilic addition. Sugars and sugar



Carbohydrates 385

FIGURE 10 Envelope and twist conformations of furanoses.

derivatives possess more nucleophilic species than nucle-ophile acceptors (e.g., there are five nucleophilic oxygenatoms vs. one carbonyl group in the acyclic form of an al-dohexose); accordingly, the number of nucleophile accep-tors exerts a greater, and sometimes controlling, influenceon the capacity of the molecule to cyclize, whereas thenucleophilicity of the attacking groups plays a lesser role.For this reason, integers are used to designate the numberof nucleophile acceptors (such groups as C O, C NR,C C C O, and C C C NR, which can undergo ad-dition reactions), and fractions are used to designate thenucleophilicity (n) of attacking species (e.g., OH or NHgroups) that are suitably situated to react with a nucle-ophile acceptor. The value of n, which is a measure of theaffinity of the nucleophile to a particular acceptor, can beobtained from tables or can be determined experimentallyfrom kinetic measurements using a Hammett-like equa-tion, namely, log(k/k0) = ns, where s is the sensitivity ofthe nucleophile acceptor to the attacking nucleophile. Freealdoses have an NQ value of 1.4, which signifies that (1)

FIGURE 11 Acetals hemiacetals, and their sulfer (thio)derivatives.

SCHEME 9 Formation of acyclic aldoses from diothioacetals.

these molecules possess one nucleophile acceptor (C-1 ofthe acyclic form; the carbon atoms in positions 2, 3, 4,5, and 6 are not counted, because their susceptibility tonucleophiles by SN reactions is much smaller than thatof C-1, which undergoes carbonyl addition reactions) and(2) the nucleophilicity of the attacking oxygen atom (asdetermined from tables) is 4. Saccharide hydrazones andoximes have similar NQ values, even though they possessadditional nucleophiles (NH or OH groups) whose nucle-ophilicity is enhanced by adjacent nitrogen atoms. This isbecause these additional nucleophiles are not suitably sit-uated to attack the nucleophile acceptor (the C N group).Saccharide phenylosazones, on the other hand, are muchmore reactive, and this is reflected by the high NQ value of2.6. They possess two nucleophile acceptors, namely, C-1and C-2, which form part of C NR systems, and contain anitrogen atom, which is suitably situated to attack a nucle-ophile acceptor (the C C C N group present in one ofthe tautomeric forms in equilibrium). The nucleophilicityof this atom is enhanced by the α effect of the adjacentnitrogen atom, giving it an n value of 6. Finally, if it is de-sired to obtain compounds that are even more reactive, an-other nucleophile acceptor could be added to the osazonemolecule, for example, a carbonyl group in the saccharidemoiety or in the hydrazone residue, which would bringtheir NQ values to 3.6. Examples of such highly reactivemolecules are L-ascorbic acid bis(phenylhydrazones) andglycosulose bis(benzoylhydrazones).

3. Nucleophilic Substitution Reactions of theAnomeric Carbon Atom

The anomeric carbon atom is the most reactive carbonatom in furanoses, pyranoses, and their derivatives. Al-though it is less susceptible to nucleophilic attack thanthe carbonyl group of acyclic sugars (because it under-goes nucleophilic substitution instead of addition), it issignificantly more reactive than the remaining hydroxyl-bearing carbon atoms. The affinity of the anomeric carbonatom to nucleophiles is attributed to its capacity to formresonance-stabilized carbonium ions that can undergo di-rect reactions via SN1 mechanisms or synchronous reac-tions by SN2 mechanisms. The rates of these reactions



386 Carbohydrates

depend on the nucleophilicity of the attacking group andthe nature of the leaving group, that is, how good a leavinggroup it is.

The nucleophilic substitution reactions that will be con-sidered in this section are (1) the displacement of the hemi-acetal hydroxyl group ( OH) or one of its esters, for ex-ample, an anomeric acetoxyl group ( O CO R), by anOR group from an alcohol to afford glycosides, or by theX group of hydrogen halides to afford glycosyl halides;and (2) the displacement of the OR groups of glycosidesby OH groups (hydrolysis), by OR groups (anomerizationand transglycosidation), or by a halogen.

a. Displacement of OH groups by OR groups(glycosidation). The anomeric hydroxyl groups ofcyclic sugars and their esters can be exchanged in acidmedia by the OR group of alcohols and phenols. The dis-placement of the hemiacetal hydroxyl group by an ORgroup is known as the Fischer method of glycoside for-mation, and the similar exchange of an anomeric estergroup is referred to as the Helferich method.

i. The Fischer glycosidation method. An example ofthe Fischer glycosidation method is the conversion ofD-glucose to two methyl D-glucofuranosides and twomethyl D-glucopyranosides by treatment with methano-lic HCl. The first products isolated are two furanosides,methyl α-D-glucofuranoside and methyl β-D-glucofura-noside. The furanosides are kinetically favored becauseclosure of five-membered rings is faster than that of six-membered rings. If the reaction time is prolonged or ifreflux temperatures are used, the thermodynamically fa-vored pyranosides will predominate, affording methyl α-D-glucopyranoside and methyl β-D-glucopyranoside inhigh yields (see Scheme 10). A study of the reaction

SCHEME 10 Formation of methyl α and β-D-glucofuranosidesand methyl α and β-D-glucopyranosides.

products of the Fischer reaction revealed the followingpoints. (1) At equilibrium, there is usually a preponder-ance of pyranose forms, since the six-membered rings arethe thermodynamically favored forms of glycosides. (2)As a rule, aldoses having adjacent bulky groups (OH orOMe groups) in the cis orientation will afford a smallerproportion furanosides on equilibration. This is becausethese substituents are closer together (and repel one an-other more) in furanosides than in pyranosides. For thesame reason, a trans relationship of adjacent bulky groupsis more favorable in furanosides than in pyranosides. Thisis why furanosides having C-1 and C-2 trans (e.g., theα anomers of D-arabinose and D-lyxose and the β anomersof D-ribose and D-xylose) are more stable and exceed theconcentration of their anomers. It was also found that theall-trans-methyl β-D-arabinofuranoside is present in highconcentration in the equilibrium mixture. In the case of themethylated derivatives of D-arabinose (which have bulkierOMe groups), the all-trans orientation is so favorable in thefive-membered ring that, at equilibrium, the concentrationof the furanosides exceeds that of the pyranosides. (3) Be-cause of the anomeric effect, the anomer having an axialmethoxyl group always predominates among the two pyra-nosides (in furanosides no substituent exists in a truly axialposition). (4) During the formation of methyl glycosidesby the Fischer method, a small proportion of the acyclicacetal is always produced. This by-product is formed bynucleophilic addition to the carbonyl group of the acyclicsugar, as well as by nucleophilic substitution on the fura-nose form of the sugar, which can cause the ring to open.

ii. The Helferich glycosidation method. TheHelferich method of glycoside formation involvesallowing a peracetylated sugar (obtained by treating afree sugar with acetic anhydride in pyridine) to reactwith a phenol or an alcohol in the presence of an acidcatalyst, a Lewis acid such as ZnCl2 or, a protic acidas p-toluenesulfonic acid. Because peracetylated sugarsusually exist in the pyranoid form, this method constitutesa convenient way in which to prepare glycopyranosides(see Scheme 11). However, if a peracetylated furanose isthe starting compound, the product will, evidently, be aglycofuranoside. If mild conditions are used (e.g., a shortheating time with p-toluenesulfonic acid), the conversioncan be achieved with retention of configuration. However,if drastic conditions are used, anomerization occurs,and that anomer having a bulky group axial on C-1 willpredominate. As in the Fischer method, the intermediatehere is a resonance-stabilized carbonium ion. Becausehalogens are good leaving groups, the glycosyl halidesare important intermediates in carbohydrate synthesis.The most important derivatives are the glycosyl chloridesand bromides; the iodides are usually too reactive andtherefore not stable over any length of time; and the



Carbohydrates 387

SCHEME 11 Formation of a phenylglucoside from a glycosylchloride.

fluorides are too inert, requiring drastic conditionsfor reaction. The glycosyl halides can be prepared bybubbling HCl or HBr into a solution of a peracetylatedor perbenzoylated monosaccharide. Alternatively, aswill be seen later, they can be obtained in the same wayfrom glycosides. Although many halides are isolated incrystalline form, they are often caused to react directlywith a nucleophile after removal of the dissolved acid.

b. Nucleophilic displacement of OR groups. Nu-cleophilic substitution reactions on the anomeric carbonatom include displacement of the anomeric OH groups,acetoxyl groups and halogens but may involve displace-ment of the OR group of glycosides by a nucleophile.The nucleophiles discussed in this section are the OH andOR groups, as well as halides. Acid catalysts are alwaysneeded in the anomeric displacement reactions in order toproduce the reactive species, a resonance-stabilized car-bonium ion. The latter is usually formed when the oxygenatom of the OR group attached to the anomeric carbonatom becomes protonated and the aglycon is eliminatedas ROH. If the glycoside reacts with the same nucleophilefound in its aglycon (e.g., if a methyl glycoside is treatedwith methanol in the presence of an acid catalyst), anomer-ization will result, that is, the two anomers of the sameglycoside will be produced, and the anomer having theOR group axially attached will predominate (the anomericeffect). If, on the other hand, the glycoside reacts with adifferent nucleophile, for example, if a methyl glycoside istreated with water (in a hydrolysis) or with another alcohol(in a transglycosidation), the OR group will be replacedby an OH group in the first case and by an OR′ group inthe second. In such reactions, if the OH group attached toC-4 of glycopyranosides and C-5 glycofuranosides is notblocked, the foregoing displacement reactions will affordmixtures of α- and β-furanoses (or furanosides) and α-and β-pyranoses (or pyranosides), irrespective of whetherthe starting glycoside is a furanoside or a pyranoside.

Another important nucleophile that can react with gly-cosides is X−, to produce glycosyl halides, which are syn-thetically valuable intermediates. In general, glycosidesare hydrolyzed by acids and are stable toward bases. Ex-ceptions to this rule are glycosides having, as the aglycon,a group derived from a phenol, an enol, or an alcohol hav-ing an electronegative group in the β position, as these arelabile toward bases.

Aldofuranosides are hydrolyzed much faster (50–200 times the rate) than the aldopyranosides, and, amongthe furanosides, the less stable isomers (those having ad-jacent bulky groups in the cis orientation) are hydrolyzedthe fastest. Similarly, methyl β-D-glycopyranosides arehydrolyzed faster than their (more stable) α-D anomers,because of the anomeric effect. (The situation is reversedwhen bulky aglycons are used, because of the considerableaxial–axial interactions.) Ketofuranosides and ketopyra-nosides are also hydrolyzed faster than the correspondingaldosides. Finally, aldopyranoses and aldofuranoses hav-ing a 2- or 3-deoxy or a 2,3-dideoxy functionality arehydrolyzed considerably faster than their hydroxylatedcounterparts; this is due to a decrease in steric interac-tion during the conversion of the chair conformer of theglycopyranoside to the half-chair conformer of the carbo-nium ion and during the bimolecular displacement of thealcohol by water in the glycofuranosides.

The accepted mechanism for the acid hydrolysis of gly-cosides starts, as usual, with protonation. Although pro-tonation of either the glycosidic oxygen atom (that of the

OR group attached to the anomeric carbon atom) or thering oxygen atom is possible, there is evidence that it isusually the former (the glycosidic oxygen atom) that isprotonated. The next step in the reaction, namely, the for-mation of a carbonium ion by shifting of the electronsof the C O R bonds, can be achieved in two ways: ei-ther by breaking the C O bond and shifting the positivecharge to the anomeric carbon atom or by breaking theO R bond and shifting the charge to the R group of theaglycon. The first route is favored, because the saccha-ride carbonium ion is stabilized by resonance with theform having the charge on the ring oxygen atom, andmost hydrolyses follow this route. In rare cases, for ex-ample, during the acid hydrolysis of tert-butyl glycosides,the bond between the oxygen atom and the R group ispreferentially broken, because of the remarkable stabilityof the tert-butyl carbocation. The foregoing mechanismshave been confirmed by conducting hydrolyses of glyco-sides in 18O-labeled water and locating the labeled oxy-gen atom in the products (the free sugar or the alcohol).With most aglycons the label was found to be attachedto the free sugar liberated, whereas with tert-butyl glyco-sides the label was found to be on the alcohol formed (seeScheme 12).



388 Carbohydrates

SCHEME 12 Mechanism of hydrolysis of glycosides.

Anomerization is the acid-catalyzed isomerization ofa group (OR) attached to the anomeric carbon atom ofa cyclic saccharide derivative. The reaction is initiatedby protonation of the OR group of the glycoside and isfollowed by elimination of the alcohol group (ROH) toafford a carbocation. This then undergoes nucleophilicattack by an OR group (identical to the OR originallypresent in the glycoside). Of the two anomers possible, thathaving the bulkier axial group predominates (the anomericeffect).

An important reaction of anomeric OR groups is theirdisplacement by X groups to form glycosyl halides. Thelatter are used in the Koenigs–Knorr method of glycosi-dation, which involves the reaction of a glycosyl halidewith an alcohol in the presence of a heavy-metal catalyst.For example, tetra-O-acetyl-α-D-glucopyranosyl bromidereacts with methanol in the presence of silver carbonate(which acts as an acid acceptor) to afford, in high yield,methyl tetra-O-acetyl-β-D-glucopyranoside. The reactionmechanism is presumed to involve an intermediate carbo-nium ion (see Scheme 13).

It should be noted that the direct replacement of a hemi-acetal hydroxyl group by a halogen atom (without pro-

SCHEME 13 Mechanism of glycoside formation.

tection of the hydroxyl groups) is not practical from thesynthetic point of view, because it is usually accompaniedby extensive isomerization and aromatization (includingthe formation of furan derivatives).

4. Nucleophilic Attacks by Hydroxyl Groups

There are two types of hydroxyl groups in the acyclicforms of monosaccharides: a primary hydroxyl group,which is, by definition, always terminal, and a number ofsecondary hydroxyl groups. In the cyclic forms of sugars,there exists a third type, namely, the glycosidic hydroxylgroup attached to the anomeric carbon atom that formspart of the hemiacetal group of aldoses, or the hemiac-etal group of ketoses. All three types of hydroxyl groupare strong nucleophiles that can add to the carbonyl groupof acid anhydrides, to afford esters, or replace the leavinggroups of alkylating agents, to afford ethers. They can alsoinduce substitution reactions at the anomeric center of thesame or a different sugar, to afford anhydro derivatives ordisaccharides, respectively. Finally, they can add in pairsto the carbonyl group of aldehydes and ketones, to givescyclic acetals.

The hemiacetal hydroxyl groups of cyclic sugars are themost reactive of the three types of hydroxyl group. Theyare followed in nucleophilicity by the reactivity of the ter-minal primary hydroxyl groups. Because acyclic sugars donot possess hemiacetal hydroxyl groups, their primary hy-droxyl groups are the most reactive hydroxyl groups in themolecule. The least reactive hydroxyl groups in both cyclicand acyclic sugars are the secondary hydroxyl groups.These may differ in reactivity, depending on whether theyare axially or equatorially oriented (eliminations are fa-vored by an antiperiplanar orientation) and according towhether the substituents are situated in a crowded environ-ment. Thus, it is often possible to block a primary hydroxylgroup and leave the secondary hydroxyl groups free bymaking use of the fact that, because secondary hydroxylgroups are in a more crowded environment, they mightnot react to any appreciable extent if mild reaction condi-tions and insufficient reagents are used. Stereo-chemicalconsiderations also play important, and sometimes deci-sive, roles in determining the course of competing reac-tions involving more than one hydroxyl group. For ex-ample, when methyl β-D-ribofuranoside, which possessesone primary and two secondary hydroxyl groups, reactswith acetone, the 2,3-O-isopropylidene derivative is pref-erentially formed by attack of the two cis-oriented sec-ondary hydroxyl groups, despite the fact that a primaryhydroxyl group (which is trans-oriented) is available (seeScheme 14). This is because, had the latter group reacted,it would have produced a highly strained, six-membered3,5-isopropylidene ring.



Carbohydrates 389

SCHEME 14 Formation of 2,3-O-ispropylidene rings.

The foregoing generalizations are useful when oneis planning multistage syntheses of complex sugarmolecules, especially when it is advantageous to blockcertain hydroxyl groups selectively and leave the otherones free for subsequent reaction. They are also useful inpredicting which group will preferentially react when alimited amount of reagent is available.

In the following section, four types of hydroxyl groupderivatives will be discussed: esters, ethers, anhydro sug-ars and disaccharides, and cyclic acetals.

a. Formation of esters. Esters are generally used toblock hydroxyl groups, that is, to deactivate their oxy-gen atoms and, by so doing, prevent them from attackingnucleophile acceptors. The esters most commonly usedfor this purpose are the acetates and benzoates. Occasion-ally, substituents are introduced in the para position of thephenyl ring of the latter esters in order to increase theircrystallizing properties. The O-p-nitrobenzoyl and the O-p-toluoyl derivatives have been found to be useful in thisrespect.

Peracetylation (full acetylation) can be achieved atroom temperature by treatment of the saccharide in pyri-dine with acetic anhydride or, at a higher temperature, byheating of the saccharide in a mixture of acetic acid andacetic anhydride. In both cases, the thermodynamicallyfavored pyranose derivative is obtained. If the furanosederivative is desired, the methyl furanoside is acetylated,and the product is subjected to acetolysis (hydrolysis andacetylation), to replace the OMe group by OAc. The lastreaction is conducted at low temperature with a mixture ofacetic acid, acetic anhydride, and a few drops of sulfuricacid (see Scheme 15).

To prepare benzoates, p-substituted benzoates, and sul-fonates, the necessary acid chloride is allowed to react inpyridine with the saccharides or saccharide derivatives.

In general, ester groups are more stable in acidic thanin basic media, and acetates are more readily hydrolyzedthan benzoates. To carry out a deacetylation, a solutionof sodium methoxide is added in catalytic amounts to

SCHEME 15 Formation of penta-O-acetyl-D-glucoyranose andfuranose.

the sugar acetate at 0◦C. Most esters are also saponifi-able with NaOH in acetone at low temperature. Acetyl,benzoyl, and p-nitrobenzoyl groups attached to the oxy-gen atom of the anomeric centers of cyclic sugars can bedisplaced by nucleophiles, especially in the presence ofLewis acid catalysts. This reaction is useful in the prepara-tion of glycosides and nucleosides. Ester groups attachedto the other carbon atoms of a saccharide, that is, linkedto nonanomeric carbon atoms, are considerably more in-ert toward nucleophiles and will not undergo substitutionreactions under mild reaction conditions. To achieve a nu-cleophilic substitution of esters attached to nonanomericcenters, the latter esters must themselves be good leav-ing groups, for example, tosylates, mesylates, and triflates(see Scheme 16).

b. Formation of ethers. True saccharide ethers havethe hydroxyl groups that are attached to the nonanomericcarbon atoms replaced by alkoxyl groups. These deriva-tives should be distinguished from the structurally similar

SCHEME 16 Replacement of an OH group with hydrogen.



390 Carbohydrates

alkyl glycosides, discussed earlier, which possess acetal-type OR groups. Methyl ethers have been used exten-sively in the structure elucidation of saccharides. Theywere first used by Haworth to determine the ring size ofmonosaccharides and the ring size and position of linkageof oligosaccharides. More recently, they have been usedto volatilize monosaccharides before they are subjected togas-chromotographic analysis. Because silyl derivativesare easier to prepare than methyl ethers, the former havenow replaced methyl ethers in the gas-chromatographicanalysis of monosaccharides. On the other hand, becausemethyl ethers are stable toward acid- and base-catalyzedhydrolysis, they continue to be used as a means of la-beling free hydroxyl groups in saccharides, a procedurefrequently used in the structure elucidation of oligosac-charides and polysaccharides. It is for this reason that theproblem of permethylating saccharides (etherifying all oftheir hydroxyl groups) continues to attract the attentionof carbohydrate chemists. The original methods of Purdie(MeI and AgOH) and of Haworth (Me2SO4 in alkali) havebeen much improved by the use of such aprotic solvents asHCONMe2 (DMF) or Me2SO (DMSO). Today, the mostwidely used methylating procedure is that of Hakomori,who used sodium hydride and Me2SO to generate the baseMeSOCH−

2 needed for this type of methylation.Benzyl ethers offer unique advantages in syntheses re-

quiring the selective blocking and deblocking of hydroxylgroups. They can be introduced under mild conditionsby the action of benzyl chloride in pyridine and can beremoved in neutral media by catalytic hydrogenolysis,which does not affect esters or cyclic acetals. The for-mer class of compounds is base labile, and the latter, acidlabile.

Other synthetically useful ethers are the triphenyl-methyl ethers, or as they are often called, trityl deriva-tives (Ph3C ). Their bulky phenyl groups render theirformation from secondary and tertiary hydroxyl groupsso difficult that they are normally obtained from primaryhydroxyl groups only. These ethers are therefore used toblock the primary hydroxyl groups selectively, so as toleave the secondary hydroxyl groups free for subsequentreactions. Removal of the trityl ethers is very facile and canbe achieved by mild acid hydrolysis. Such organic acids asacetic acid, which do not affect esters or cyclic acetals, se-lectively deblock the oxygen atom bearing the trityl groupand regenerate the primary hydroxyl group, with libera-tion of triphenylmethanol. Alternatively, the trityl ethergroup can be removed by catalytic hydrogenolysis, whichthen affords triphenylmethane and the free primary hy-droxyl group. The ease of hydrolysis of trityl ethers maysometimes be responsible for the occurrence of undesir-able hydrolysis during the course of a subsequent reaction,particularly when vigorous conditions are used.

c. Anhydrides and disaccharides. An intra or in-termolecular nucleophilic attack initiated by the oxygenatom of a suitably placed hydroxyl group on the anomericcarbon atom of the same or a different sugar will affordan anhydride or a disaccharide, respectively. The reactionis best performed by introducing a good leaving group,usually a halogen atom, on the anomeric carbon atomand blocking all of the hydroxyl groups except the one tobe involved in the subsequent reaction (see Scheme 17).In this way, the formation of undesired products can beavoided.

Disaccharides are often obtained enzymatically, bypassing the substrate or substrates (the saccharides in-volved in the dimerization) through a column filled witha polymer that had been prebound chemically to the spe-cific enzyme needed to perform the transformation. Thismethod has, of late, acquired attention in industrial op-erations, because it allows the enzymes to remain activeindefinitely on the column and prevents their being washedaway during the work-up.

d. Cyclic acetals. If oriented properly, any two adja-cent (but not necessarily contiguous) hydroxyl groups willreact with an appropriate aldehyde or ketone to yield an un-strained five- or six-membered cyclic acetal, respectively.The most commonly used carbonyl compounds are ben-zaldehyde, which affords mostly six-membered benzyli-dene acetals, and acetone, which yields five-memberedisopropylidene acetals. The latter derivatives have the ad-vantage of existing in one isomeric form, unlike benzyli-dene derivatives, which exist in two isomeric forms. Thisis because, on reacting with chiral glycols, all aldehydes,except formaldehyde, and all mixed ketones yield chiralacetals.

The formation of cyclic acetals is catalyzed by acidsand proceeds by two successive nucleophilic attacks; atfirst one hydroxyl group attacks the protonated carbonylderivative to form the hemiacetal. The latter, in turn, be-comes protonated and is attacked by the second hydroxylgroup.

The formation of isopropylidene acetals can beachieved by treatment of a saccharide derivative with ace-tone or its dimethyl acetal (2,2-dimethoxypropane) in thepresence of an acid catalyst such as HCl or H2SO4 anda dehydrating agent. The reaction is often conducted atroom temperature and can be monitored by 1H NMR spec-troscopy. The signals of two methyl groups are usuallywell resolved because of their chiral environment.

A useful starting material in many synthesesis 1,2 : 5,6-di-O-isopropylidene-α-D-glucofuranose, ob-tained by treating D-glucose with acetone. This derivativeaffords an easy means of access to furanoses having a freehydroxyl group on C-3.



Carbohydrates 391

SCHEME 17 Formation of disaccharides.

Benzylidene groups are also useful in the selectiveblocking of saccharide derivatives. They usually involvethe primary hydroxyl group of a pyranoside and the4-hydroxyl group if it is suitably positioned. Benzyli-dene groups are also useful for introducing a halogenatom instead of the primary hydroxyl group by theHanessian method using N -bromosuccinimide as shownin Scheme 18.

5. Oxidation

The oxidation of a sugar involves the breaking of C H,C C, O H bonds and the transfer of electrons to the ox-idant. Two types of oxidation are possible. The first in-volves the transfer of two electrons from one atom to theoxidant and is referred to as heterotytic oxidation. Theother, known as homolytic oxidation, occurs in two steps,each involving the transfer of one electron.

a. Heterolytic oxidations. Heterolytic oxidations,that is, those involving two-electron transfers, are exem-

plified by the oxidation of a secondary hydroxyl group toa keto group. This involves the elimination of two hydro-gen atoms and the formation of a double bond between thecarbon atom and the oxygen atom of the alcohol function.The first step takes place by breakage of a C H bond andelimination of a hydride ion. Because the hydride anionis a very poor leaving group, this bond breakage is gener-ally the slowest step of the oxidation (the rate-determiningstep). What makes the reaction possible, despite this, is thefact that the elimination is aided by the oxidant, which cap-tures the pair of electrons and converts the hydride ion to aproton. Another proton is concurrently liberated by disso-ciation of the O H bond of the alcohol, and both protonsreact with a base (usually a hydroxyl group to form wa-ter). Such an oxidation usually occurs by a concerted orE2 type of mechanism, in which the oxidant is not bondedto the sugar, except possibly in the transition state.

Oxidants, for example, chromic acid, have been shownto form intermediate esters, which subsequently decom-pose by bimolecular eliminations. The difference betweenthis and the previous oxidation is that the leaving group is



392 Carbohydrates

SCHEME 18 Replacement of three OH groups with hydrogen.

the reduced form of the oxidant, and the capture of elec-trons by the oxidant is the driving force of the reaction.Furthermore, the breaking of the C H bond, which occurssimultaneously, is the rate-determining step.

Of particular interest is the oxidation of primary hy-droxyl groups with pyridinium chlorochromate (PCC)to afford aldehydes. The oxidant may be generated insitu by dissolving chromium trioxide in HCl to formchlorochromic acid and adding pyridine to form thedesired salt. The oxidation is usually carried out indichloromethane.

Glycols are more acidic than monohydric alcohols, andthe C-1 group on an aldopyranose is even more acidicthan either, because of the inductive effect of the ring oxy-gen atom. It is for this reason that sugars and glycosidesare more readily oxidized than ordinary alcohols. Oxida-tion of free sugars at higher pH is often accompanied bycompeting processes of epimerization and degradation. Ingeneral, the β anomers of D sugars and their glycosidesare more rapidly oxidized than the α anomers. Similarly,the 2-hydroxyl group of methyl β-D-glucopyranoside ismore acidic than the corresponding group in the α-D

anomer.Halogens (usually bromine) and hypohalites (particu-

larly sodium hypoiodite) have been used to oxidize aldosesto aldonic acids and their lactones. The aldehydic groupof acyclic sugars is converted to a carboxyl group, andthe anomeric hydroxyl group is converted to lactones. Al-

though the terminal (primary) hydroxyl groups of sugarscan also undergo oxidation by these reagents, the reactionoccurs at a somewhat lower rate, so that the oxidation canbe stopped at the monocarboxylic acid (or lactone) stage.By the use of moderately vigorous conditions, oxidationsof the primary hydroxyl groups have been used to convertglycosides to glycosiduronic acids. The oxidation of sec-ondary hydroxyl groups to keto groups requires drasticconditions, and these may be accompanied by cleavageof C C bonds. A study of the above-mentioned oxida-tions has revealed that changes in the reaction conditions(temperature, acidity, and concentration) are often accom-panied by changes in the nature of the oxidizing species.For example, the concentration of hypohalous acid is verylow, but when sufficient alkali is added to the system, theconcentration of hypohalite ion increases dramatically (atpH 1, 82% of the total chlorine exists as free chlorineand 18% as hypochlorous acid; at pH 4, only 0.4% of thechlorine is free and 99.6% exists as hypochlorous acid;and at pH 8, 21% exists as hypochlorous acid and 79%as hypochlorite). The order of effectiveness of halogensas oxidants is Br2 > Cl2 > I2, the last being quite ineffec-tive. This order corresponds to the rate of hydrolysis of thefree halogen by water and to the solubility in water, whichexplains the ineffectiveness of free iodine as an oxidant.The mechanism of the oxidation of aldoses with chlorineand bromine in acid media was studied by Isbell, whofound that the active oxidants are the halogens and not thehypohalous acids.

The accumulation of hydrogen bromide during oxida-tions by bromine profoundly lowers the rate of further ox-idation. This effect is not due only to the simple increasein acidity, because, although other strong acids also in-hibit the rate, the effect is largest for hydrogen bromideand chloride.

To minimize this inhibiting influence, the reaction canbe conducted in the presence of a solid buffer, such asbarium carbonate. In general, the presence of a buffer in-creases the yield of aldonic acid and precludes the hy-drolysis of disaccharides. Yields of 96% of D-gluconicacid have been reported. When the oxidation period is ex-tended for unbuffered solutions, keto acids may be formedin small yields. Thus, hexoses afford hex-5-ulosonic acids.Under more drastic conditions, C C bonds are cleaved,yielding chain-shortened acids. The commercial produc-tion of aldonic acids has been achieved by electroysis ofdilute solutions of sugars in the presence of a bromideand a solid buffer, such as calcium carbonate. Presum-ably, there is formed, at the anode, free bromine, whichthen oxidizes the aldose to the aldonic acid and is itselfreduced to bromide. If the electrolytic method is not wellcontrolled, aldaric acids and glyc-2 and -5-ulosonic acidsmay be produced.



Carbohydrates 393

Ketoses are generally resistant to bromine oxidation,and the latter may be useful in the removal of aldose con-taminants from ketoses. However, by extending the pe-riod of oxidation, D-fructose affords D-lyxo-5-hexulosonicacid.

b. Homolytic oxidations. Most homolytic oxida-tions occur in two successive steps, each of which involvesa one-electron transfer. The most difficult (and slowest)step in the homolytic oxidation of a sugar is the first,which involves the abstraction of a hydrogen atom from aC H group, followed by the transfer of an electron fromthis hydrogen atom to the oxidant, to afford a proton. Inthis process, the sugar that lost a hydrogen atom is con-verted to a radical, which is stabilized by resonance (theunshared pair of electrons on the adjacent oxygen atomcan migrate to the carbon atom from which hydrogen hadbeen abstracted). This explains why the initial abstractionof hydrogen usually takes place by the removal of a hy-drogen atom attached to carbon (H C O H), ratherthan from the one attached to oxygen (H C O H). Itis also in agreement with the fact that compounds havingequatorial hydrogen atoms linked to carbon atoms (i.e.,carbon-linked hydrogen atoms, which are more accessibleto the oxidant than the axially oriented hydroxyl groupsattached to the same carbon atom) undergo more facile cat-alytic oxidation than axially oriented C H groups (whichare linked to equatorial hydroxyl groups). The final stepof the oxidation is a fast reaction involving the homolysisof the O H bond of the sugar radical, to afford a carbonylgroup plus a second hydrogen radical which, like the first,is converted by the oxidant to a proton.

Degradation of a carbohydrate having a carbonyl groupor potential carbonyl group begins with nucleophilic ad-dition of hydroperoxide anion to the group, forming a hy-droperoxide hydrin called the hydroperoxide adduct.

The adducts of substances having a hydroxyl groupattached to the carbon atom adjacent to the carbonylgroup usually decompose by a process that is calledthe β-hydroxy hydroperoxide cleavage reaction. With al-doses, this affords formic acid and the next lower aldose.

α,β-Diketones react rapidly with alkaline hydrogenperoxide with rupture of the C C bond and produc-tion of the two carboxyl groups. Degradation of an α,β-unsaturated ketone takes place by addition of hydroper-oxide anion to the double bond. The adduct decomposes,producing an epoxide which may react further, as in theoxidation of the enolic form of diketomyoinositol.

The degradation of a reducing disaccharide proceedsrelatively rapidly by the β-hydroxy hydroperoxide cleav-age reaction until the process is interrupted by the gly-cosidic linkage. At this point, the hydroperoxide adduct,lacking an α-hydroxyl group, decomposes largely by

elimination of the C-1 hydrogen atom, cleavage of theO H bond, and formation of the corresponding 2-O-glycosylaldonic acid or by a minor reaction involvingcleavage of the C-1 C-2 bond with formation of theO-formyl acetal of the next lower sugar.

6. Reduction of Carbohydrates

Numerous reagents have been employed to reduce carbo-hydrates and their derivatives. These include such inor-ganic hydrides as those of aluminum and boron; reactivemetals, for example, sodium in the presence of protondonors; and metal catalysts, such as palladium, platinum,or nickel in the presence of molecular hydrogen.

a. Lithium aluminum hydride. Reductions of car-bohydrate derivatives with lithium aluminum hydride areusually performed in nonaqueous solutions (such as ben-zene, ether, or tetrahydrofuran), because the reagent re-acts violently with water. The boric acid formed duringthe borohydride reduction is usually removed as volatilemethyl borate. Lithium aluminum hydride converts car-bonyl groups to hydroxyl groups and ether lactones toalditols. For example, 2,3,5,6-tetra-O-acetyl-D-glucono-1,4-lactone yields D-glucitol. (The acetyl groups are re-ductively cleaved by the reagent.)

If less reactive hydrides are desired, to avoid reducingthe aldehydes first formed to hydroxyls, then lithium tri-ter-butoxyaluminum hydride or diisobutylaluminum hy-dride (DIBAL-H) may be used at low temperature. Thefirst will reduce an aldonic acid chloride into an aldose andthe second an ester or a nitrile to an aldehyde (Scheme 19).

b. Borohydrides and diborane. In contrast tolithium aluminum hydride, sodium borohydride andlithium borohydride are relatively stable in water and canbe used to reduce lactones to sugars in acid media or tothe alditol in basic media. Thus, at pH 5, D-glucono-1,5-lactone affords D-glucose and, at pH 9, D-glucitol.

Sodium borohydride (or sodium borodeuteride) is alsoused to convert aldoses or oligosacchrides to alditol deriva-tives for examination by gas–liquid chromatography–mass spectrometry. Unlike hydride ions, which areelectron-rich nucleophiles, diborane is the dimer of anelectron-deficient electrophilic species (BH3). It is formedby reaction of a borohydride ion with strong acids andis used in 2.2′-dimethoxydiethyl ether (diglyme) as thesolvent. Diborane reacts faster with carboxylic acids, re-ducing them to primary alcohols, than with esters, be-cause the carbonyl character of the adduct of the first isstabilized by resonance, and that of the second is desta-bilized by resonance. Diisoamylborane, which does notcleave ester groups (as does diborane), reduces acetylated



394 Carbohydrates

aldono-1,4-lactones to the corresponding acetylatedaldofuranoses.

c. Sodium amalgam. Sodium amalgam in waterand sodium in liquid ammonia are commonly used as re-ducing agents. The metal alloy has an electron-rich surfacethat reacts with protons of a proton-donating solvent toform hydrogen atoms, which may combine to form molec-ular hydrogen or react with the substrate to give a radicalion. The same radical-ion can be directly obtained fromthe metal, which, in both cases, is oxidized to a cation,giving up electrons to the substrate. The hydrogen radi-cals and the substrate radical-ions, which are bound byadsorption to the electron-rich surface of the metal, com-bine to afford the alkoxide ion, which then accepts a protonfrom the solution. In the early literature, the reduction oflactones to aldoses (a key step in ascending the series)had been carried out with sodium amalgam in acid solu-tion (pH 3), using such buffers as sodium hydrogenox-alate, sulfuric acid, and, later, ion-exchange resins. Thesebuffers are needed because the reduction does not proceedin alkaline solution (formation of the sodium salt of thealdonic acid impedes reaction with the nucleophilic re-ductant). The free acids are not reduced, but esters can be.Thus, methyl D-arabinonate is converted to D-arabinose bysodium amalgam. The susceptibility of lactones to reduc-tion and the inertness of the free acids have been exploitedin the preparation of glycuronic acids by reduction of themonolactones of aldaric acids.

d. Catalytic hydrogenation. In the course of hydro-genation with palladium or platinum catalysts, surface re-actions occur between such electron-deficient species asthe double bonds of carbonyl groups (which are attractedto the surface of the catalyst) and the hydrogen atomsgenerated by splitting hydrogen molecules. Catalytic hy-drogenation is used to convert keto groups to secondaryalcohol groups. Thus, Raney nickel reacts with D-xylo-5-hexulosonic acid to give both D-gluconic and L-idonicacid. Usually, the axial alcohol is the preponderant prod-uct, because the reductant approaches the substrate fromthe less hindered side of the carbonyl group, to form anequatorial C H bond. Reduction of carbonyl groups canbe carried beyond the alcohol stage, to the deoxy stage.For example, methyl β-D-ribo-hexopyranosid-3-ulose isconverted to the 3-deoxy derivative by hydrogen in thepresence of platinum, and 1-deoxyalditols can be obtainedfrom aldoses by the reduction of their dialkyl dithioacetalswith Raney nickel.

The stereo specific syn hydrogenation of unsaturatedsugars (glycals and acetylenes) may be carried out het-erogeneously, using Lindlar catalyst (Pd/CaCO3) condi-tioned in quinoline, or homogeneously using Wilkinson’s

SCHEME 19 Action of reducing agents.

catalyst. Rh[(Ph)3P]3Cl. Alternatively, trans additionis achieved by reduction with Li in EtNH2S at lowtemperature.

III. OLIGOSACCHARIDES

Oligosaccharides are polymeric saccharides that have, astheir name denotes, a low degree of polymerization, DP(oligo means “few” in Greek). They are composed of 2 to10 glycosidically linked monosaccharides, which can beliberated by depolymerization (e.g., by acid hydrolysis).Oligosaccharides having DPs of 2 to 3 are sweet tastingand are included among sugars, whereas higher membersare devoid of taste and are not referred to as such.

Sucrose is the world’s most widely used sweetener andthe one produced in largest quantities. Its industrial prepa-ration from sugar cane and from beets is now well es-tablished. The process starts with pressing and concen-tration of the juice. This is followed by precipitation ofsucrose as a calcium complex with lime and its regenera-tion with SO2. Refining is achieved by charcoaling and re-crystallization. Recent innovations in the process includesshifting to continuous flow operations, which have con-siderably increased the efficiency of production. Up untilrecently the only competitor of sucrose was glucose syrupproduced by acid hydrolysis, or enzymatic hydrolysis of



Carbohydrates 395

corn starch. Now three additional products produced fromstarch have entered the market and lowered the consump-tion of sucrose. They are fructose syrup, very high fruc-tose syrup, and crystalline fructose. As a result, sucrosecan account for only half the sweeteners consumed in theUnited States. Oligosaccharides are grouped into simple(or true) oligosaccharides, which yield on deploymeriza-tion monosaccharides only; and conjugate oligosaccha-rides, which are linked to nonsaccharides such as lipidsand afford on deploymerization monosaccharides andaglycons. The simple oligosaccharides are further classi-fied (1) according to DP, into disaccharides, trisaccharides,tetrasaccharides, and so on; (2) according to whether theyare composed of one or more types of monosaccharides,into homo- and heterooligosaccharides; and (3) accordingto whether they do or do not possess a hemiacetal func-tion at one terminus of the molecule, into reducing andnonreducing oligosaccharides.

Related homooligosaccharides can form homologousseries; a homologous series of oligosaccharides is a groupof similarly linked oligosaccharides that are composed ofthe same monomer and whose DP increases in the se-ries one unit at a time. When homopolysaccharides arepartially hydrolyzed, they often afford homologous seriesof oligosaccharides. For example, the maltooligosaccha-rides obtained by partial hydrolysis of starch comprisedimers, trimers, tetramers, and so on composed of α-D-glucopyranose units linked by 1 → 4 acetal bonds.

A. Structure

The complete structure of oligosaccharides is establishedwhen the following points are determined:

1. The DP, that is, the number of monosaccharide unitspresent in the oligomer molecule

2. The nature of the monosaccharide monomer(s)3. In the case of heterooligosaccharides, the monosac-

charide sequence4. The ring size (pyranose or furanose) and the position

of linkage of the different monosaccharides (1 → ?)5. The anomeric configuration (α or β)6. The conformation of the monosaccharide rings

1. Determination of the Degree of Polymerization

The following procedures are recommended for determin-ing the DP of oligosaccharides:

a. Mass spectrometry. If the molecular weight ofthe oligosaccharide is less than 1000, mass spectrome-try can be used to determine its molecular weight andhence its DP. It must be remembered, however, that mass

spectra produced by electron impact seldom show themolecular ions of saccharides and that other ionizationmethods (e.g., chemical ionization) must be used to re-veal these ions. The silyl ethers of oligosaccharides or oftheir alditols (obtained by reduction and silylation) are of-ten used for mass spectrometric measurements, since theyare more volatile than free oligosaccharides. Alternatively,peracetylated disaccharides or esters of disaccharides orof their glycosides can be used in the chemical ionizationmode. Figure 12 shows three mass spectra of a disac-charide glycoside obtained by negative chemical ioniza-tion (a), by electron impact (b), and by positive chemicalionization (c). Only the chemical ionization mass spectra(negative and positive) revealed molecular ions.

b. Chromatography. It is often possible to deter-mine the DP of members of a homologous series of poly-meric saccharides from their position on chromatogramsrelative to a known member of the series. For example,partial hydrolysis of starch affords D-glucose and a ho-mologous series of maltooligosaccharides, consisting ofdi-, tri-, tetra-, pentasaccharides, and so on. This mix-ture separates on chromatographs according to DP. On pa-per chromatographs and on high-performance liquid chro-matographs, the mobility of the saccharides is inverselyproportional to their DP (i.e., the largest oligomers moveslowest), and on gel filtration chromatography, it is di-rectly proportional to DP (i.e., the largest oligomers movefastest). The fact that the oligomers are eluted in the orderof increasing or decreasing DP, makes it possible to deter-mine their DP by determining their order of elution relativeto a known member. Thus, by recognizing maltose (andglucose) in chromatograms of partially hydrolyzed starch(Fig. 13), it is possible to identify the subsequent bands asmaltotriose, maltotetrose, maltopentose, and so on.

c. Reducing power. The DP of reducing oligosac-charides, particularly those with low molecular weight,can be determined from their reducing power. For exam-ple, the reducing power of a maltooligosaccharide relativeto glucose (taken as 100%) is matched with the calculatedvalues for oligomers having different DPs. Thus, a disac-charide would be expected to have about half the reducingpower of glucose (actually 53%), a trisaccharide, a third(actually 35%), and a tetrasaccharide, a fourth (26%) thatof glucose, and so on. Accordingly, if a maltooligosac-charide exhibits a reducing power equal to 35.8% thatof D-glucose, it can be safely assumed that it is a trisac-charide. The differences in the reducing powers of suc-cessive members of a homologous series of oligosaccha-rides decrease with increasing DP, so that beyond a DPof 5, the differences become too small for reliable DPmeasurements.



FIG

UR

E12

Mas

ssp

ectr

aof

adi

sacc

harid

e.(a

)N

egat

ive-

ion

chem

ical

ioni

zatio

nin

met

hane

,(b

)el

ectr

onim

pact

ioni

zatio

n,an

d(c

)po

sitiv

e-io

nch

emic

alio

niza

tion

inm

etha

ne.



Carbohydrates 397

FIGURE 13 Liquid chromatogram of partially hydrolyzed starchshowing glucose and the maltooligosaccharides.

2. The Monosaccharide Components

The nature of the monomers in an oligosaccharide is es-tablished by identifying the monosaccharides liberated byacid-catalyzed hydrolysis. Homooligosaccharides, whichare composed of one type of monosaccharides, afford onlyone monosaccharide, which can be isolated from the hy-drolysate by conversion to crystalline derivatives. On theother hand, heterooligosaccharides afford, on hydrolysis, amixture of monosaccharides, which must be separated bychromatography. Paper, thin-layer, or high-performanceliquid chromatography (HPLC) can be used without pre-treating the monosaccharides, but a gas-chromatographicseparation requires prior silylation of the saccharides, tovolatilize them. It should be noted that silylated monosac-charides appear as double peaks (one peak for the α

anomer and one for the β), which tends to crowd the chro-matographs. To avoid this complication, the monosaccha-rides can be reduced before silylation (silylated alditolsappear as single peaks). It should also be noted that allthe above-mentioned chromatographic techniques cannotdifferentiate between D and L isomers, which mandatesthat polarimetric measurements (optical rotation, opticalrotatory dispersion or circular dichroism) of the monosac-charide component(s) or of their derivatives be includedin the structure elucidation of new oligosaccharides. In thecase of heterooligosaccharides, this necessitates prepara-tive chromatography of the hydrolysate before the opticalmeasurements are taken.

3. The Monosaccharide Sequence

Homooligosaccharides (whether reducing or nonreduc-ing) do not require a monosaccharide sequence determina-

tion, because they possess identical monomers all throughtheir chains. Heterooligosaccharides, except nonreducingdisaccharides, must have the sequence of their monomersdetermined. In the case of reducing heterodisaccharides,it must be determined because the monomer at the nonre-ducing end of the molecule is different from the one atthe reducing terminus, and the position of both monomersmust be determined. It was stated above that nonreducingdisaccharides do not require a monosaccharide sequencedetermination. This is because the two monomers are lo-cated at two similar (nonreducing) ends of a chain thathas no beginning (no reducing terminus). On the otherhand, nonreducing heterooligosaccharides with DP > 2are made up of a monosaccharide linked glycosidicallyto the hemiacetal function of a reducing oligosaccharide,the sequence of which must be determined.

The methods available for determining the monomersequence are presented here, in order of increasing DP ofthe oligosaccharides, starting with disaccharides.

a. Monosaccharide sequence in disaccharides.The sequence of monomers in disaccharides is deter-mined when the monomer located at the reducing ter-minus of the molecule is identified. This automaticallydetermines the position of the other monomer (it must belocated at the nonreducing end). To identify the monosac-charide located at the reducing terminus, use is made ofthe fact that this saccharide moiety exists in equilibriumwith an acyclic form and is therefore much more sus-ceptible to oxidants and reducing agents than the othermoiety. Thus, mild oxidation converts reducing disac-charides to aldobionic acids, which on hydrolysis affordaldonic acids (from the reducing termini) and reducingmonosaccharides (from the nonreducing ends). Reduc-tion affords aldobiitols, which on hydrolysis yield alditols,from the reducing moieties and monosaccharides fromthe nonreducing moieties. The use of this method ofstructure determination is exemplified by the oxidationand reduction of melibiose [6-(α-D-galctopyranosyl)-D-glucopyranose] to give, in the first case, melibionic acid[6-(α-D-galactopyranosyl)-D-gluconic acid] and, in thesecond, melibiitol [6-(α-D-galactopyranosyl)-D-glucitol](Scheme 20). Hydrolysis of these affords D-galactose fromthe nonreducing end of the molecule and D-gluconic acidin the first case and D-glucitol in the second (both formedfrom the reducing half of the molecule). This experimentindicates that D-galactose is the monomer located at thenonreducing end of the dimer and that D-glucose is theone at the reducing terminus.

b. Monosaccharide sequence in trisaccharides.Because it is easier to determine the structure of disaccha-rides than it is to investigate the structure of trisaccharides,



398 Carbohydrates

SCHEME 20 Oxidation and reduction of melibiose.

it is often advantageous to deduce the structure of the lat-ter by identifying (or determining the structure of) the twodisaccharides resulting from the partial hydrolysis (eitherchemical or enzymatic) of the trisaccharide under inves-tigation. If known, the disaccharides would be comparedwith authentic samples, and if not, the new disaccharideswould be investigated as shown above. In piecing togetherthe structures of a trisaccharide from two disaccharides,the following rules are used:

1. If the trimer is composed of three (different) typesof monomers (e.g., A–B–C), only one monomer, namely,B, will be found in both dimers (A–B and B–C). Thismonomer must be located in the center of the trimer, andthe remaining monomers (A and C) are made to occupythe same terminal positions in the trimer as they do in the

dimer. [The monomer present in the nonreducing end ofa disaccharide (A) will be at the nonreducing end of thetrisaccharide, and the monomer found at the reducing endof a disaccharide (C) will be put at the reducing end of thetrisaccharide.]

2. If the trimer is composed of only two types ofmonomers (e.g., (A–A–B or A–B–B or A–B–A orB–A–B), it is possible to have one or two commonmonomers in the resulting dimers. If one monomer onlyis common to the two dimers, for example, A in A–A andA–B (obtained from A–A–B) or B in A–B and B–B (ob-tained from A–B–B), the same rules mentioned above areused. However, if both monomers (A and B) are commonto the two dimers (e.g., A–B and B–A obtained from eitherA–B–A or B–A–B), an estimation (by gas chromatogra-phy or HPLC) of the two monomers in the trisaccharide



Carbohydrates 399

hydrolysate will reveal that one monomer (A in A–B–Aand B in B–A–B) occurs in double the amount of theother; this monomer must be terminally located in thetrisaccharide molecule. To illustrate this method, a nonre-ducing trisaccharide, raffinose, will be used as an exam-ple of a trisaccharide having three different monomers(A–B–C) (Scheme 21). On partial hydrolysis with acids,this trisaccharide affords a reducing disaccharide [6-(α-D-galactopyranosyl)-D-glucopyranose], and on enzymatichydrolysis it yields a nonreducing disaccharide sucrose[α-D-glucopyranosyl-β-D-fructofuranose]. The commonmonomer in both disaccharides, D-glucopyranose, musttherefore be located at the center of the trimer. The othertwo monomers (D-galactopyranose and D-fructofuranose)are then assigned positions at both ends of the nonreducingtrimer. Oxidation (or reduction) of the first disaccharide,followed by hydrolysis, will reveal that D-glucose is atthe reducing terminus and D-galactose is at the nonreduc-ing ends of the dimer molecule and confirm the locationof the latter at one of the nonreducing ends of the trimermolecule. Since the second disaccharide is nonreducing,the terminal D-fructofuranose must be attached throughC-2 to C-1 of D-glucose. This establishes the structureof raffinose as O-α-D-galactopyranosyl-(1 → 6)-O-α-D-glucopyranosyl-(1 → 2) O-β-D-fructofuranoside.

c. Monosaccharide sequence in tetrasaccharides.The sequence in tetrasaccharides can be deduced by deter-mining the sequence of monosaccharides in the oligosac-charides that result from their partial hydrolysis (twotrisaccharides and three disaccharides). If the tetrasaccha-ride is composed of four different monomers (A–B–C–D),the monosaccharide sequence can be deduced by deter-mining the monomer sequence in the disaccharides (A–B,B–C, and C–D) produced by partial hydrolysis. Use ismade of the fact that the monomers common to two dimers(B, found in A–B and B–C; and C, in B–C and C–D) mustbe linked together and must be located in the center of thetetramer. (Thus, B must be attached to A and C, and Cattached to B and D.) The same rule applies if only onemonomer is repeated and the two analogous monomersare contiguous (as determined from the fact that one ofthe dimers is a homodisaccharide). Examples of such ho-modisaccharides are A–A obtained from A–A–B–C; B–Bobtained from A–B–B–C; and C–C produced from A–B–C–C. If, on the other hand, the repeating monomers arenot contiguous but are separated by one monomer (as inA–B–A–C and A–B–C–B) or by two monomers (as inA–B–C–A), a study of the dimers alone may not sufficeto elucidate the structure of the tetramer, making it neces-sary to study also the structure of the two trisaccharidesresulting from its partial hydrolysis.

SCHEME 21 Formation of melibiose and sucrose from raffinose.

In general, the structure elucidation of polymers be-comes more difficult as the DP increases. However, be-cause higher oligosaccharides are composed of repeatingunits seldom larger than tetrasaccharide fragments (and of-ten composed of mono-, di- or trisaccharides), the numberof possible oligosaccharides liberated on partial hydrol-ysis remains relatively small. Thus, whereas a tetrasac-charide (A–B–C–D) will afford on partial hydrolysis twotrisaccharides (A–B–C and B–C–D) and three disaccha-rides (A–B, B–C, C–D), an octasaccharide (or for thatmatter a polysaccharide) composed of the same tetrasac-charide repeating unit will afford only two additional



400 Carbohydrates

trisaccharides (D–A–B and C–D–A) and one additionaldisaccharide (D–A).

4. Ring Size and Position of Linkage

The next stage in the structure elucidation of oligosaccha-rides involves the determination of the ring size and theposition of linkage of the monosaccharide constituents.For nonreducing disaccharides, the position of linkage isby necessity between anomeric carbons; otherwise the dis-accharide would be reducing. If the nonreducing disaccha-ride is composed of two aldoses, these two monomers mustbe linked by an acetal oxygen bridge joining C-1 of onealdoses to C-1 of the other, which is signified by (1 → 1);if the disaccharide is composed of two uloses (ketoses),the oxygen bridge must link C-2 of one ulose to C-2 ofthe other (2 → 2); finally, if the disaccharide is composedof one aldose and one ulose, the linkage is (1 → 2) (withthe oxygen bridge linking C-1 of the aldose to C-2 of theulose). Accordingly, when the monosaccharide compo-nents of a nonreducing disaccharide are identified, thereis no uncertainty about the positions of linkage, but onlyabout the size of the rings (and the anomeric configuration,which will be discussed later). Reducing disaccharides, onthe other hand, must have their position of linkage and ringsize determined. The position of linkage and/or ring size ofthe monosaccharide components of reducing and nonre-ducing oligosaccharides can be determined by labeling orby partial hydrolysis, as follows:

a. Labeling of free hydroxyl groups. The free hy-droxyl groups in oligosaccharides are attached to carbonatoms that are not involved in ring formation or in glyco-sidic bonds. These carbon atoms can be recognized if theyare marked with a suitable label, for example, by attach-ing to their oxygen atoms permanent blocking groups thatwill not be removed during the hydrolysis of the oligosac-charide. Methylation is often used, since many of the par-tially methylated monosaccharides that result from suchhydrolysis have been characterized by gas chromatogra-phy. Methylation and hydrolysis of a reducing disaccha-ride afford two methylated monosaccharides. The first hastwo free hydroxyl groups (one at position 1 and one wherethe ring was attached), and the second possesses three po-sitions free (one at position 1, one where the ring was at-tached, and one where the glycosidic bond was attached).The vacant positions in the latter might create ambiguity,since it is not known which of the unblocked positionswere due to the ring and which to the glycosidic bond. Toavoid this confusion, it is necessary to carry out anothermethylation on an acyclic disaccharide derivative, for ex-ample, the aldobionic acid obtained by oxidation of the

disaccharide or the aldobitol obtained by its reduction.These compounds have one acyclic moiety (the aldonicacid and the alditol, respectively), so that after hydrolysisof the alkylated dimer, the unmethylated position in theacyclic moiety will be the position of the glycosidic link-age. If the disaccharide is nonreducing, the position of theglycosidic linkage is known (C-1 for an aldose and C-2 foran ulose), and only one methylation experiment is needed(the unlabeled position is where the ring was attached).

The use of labeling in structure elucidation is exem-plified by the methylation and hydrolysis of melibioseto yield 2,3,4,6-tetra-O-methyl-D-galactose and 2,3,4-tri-O-methyl-D-glucose and by the oxidation of melibi-itose to melibionic acid or its reduction to melibiitol,then methylation of these, followed by hydrolysis of themethylated aldobionic acid or aldobiitol to yield 2,3,4,6-tetra-O-methyl-D-galactose and 2,3,4-5-tetra-O-methyl-D-gluconic acid in the first case and 1,2,3,4,5-penta-O-methyl-D-glucitol in the second (see Scheme 22).

b. Partial hydrolysis. It is possible to use the struc-ture of known disaccharides to determine the structure ofhigher oligosaccharides. Thus, the fact that the trisaccha-ride raffinose (discussed above) affords on partial hydrol-ysis melibiose and sucrose, whose structures are knownto be 6-(α-D-galactopyranosyl)-D-glucopyranose and α-D-glucopyranosyl β-D-fructofuranoside, respectively, es-tablishes that the trisaccharide molecule is composed of anα-D-galactopyranose ring attached through an α-glyco-sidic bond to position 6 of an α-D-glucopyranose ring,which in turn is attached glycosidically to the anomericposition of a β-D-fructofuranose ring. In other words,the trisaccharide must be 6-(α-D-galactopyranosyl)-α-D-glucopyranosyl β-D-fructofuranoside.

5. Anomeric Configuration

a. By partial hydrolysis. The hydrolysis of raffinoseto melibiose and sucrose, discussed earlier, suggests thatthe linkage between the galactose and the glucose units inthe trisaccharide is α-D, and the linkage between glucoseand fructose is α-D for glucose and β-D for fructose.

b. By enzymatic hydrolysis. The anomeric config-uration of the glycosidic bond of disaccharides can be de-termined by enzymatic hydrolysis. For example, emulsin,an enzyme obtained from bitter almonds, is known tohydrolyze β-D-glucosidic linkages and not α-D linkages.Accordingly, if a glucose-containing disaccharide is hy-drolyzed by emulsin, it can be concluded that it possessesa β-D linkage. It is essential for structural work involvingenzymes that the enzyme preparations be of the highest



Carbohydrates 401

SCHEME 22 Elucidation of the structure of melibiose.

purity in order to prevent misleading results caused bycontaminants.

6. Conformation

a. By NMR spectroscopy. The most commonmethod for determining both the anomeric configurationand the ring conformation of oligosaccharides is NMRspectroscopy. The coupling constant of the anomeric pro-ton (J1,2) is measured and used to determine the dihedral

angle between H-1 and H-2 in the disaccharide. Since thisangle depends not only on the anomeric configuration,but also on the ring conformation, both are determinedconcurrently. The sample is irradiated at the frequencyof the H-1, so that the signal of H-2 is identified (sinceit is decoupled partly), and J2,3 determined. The proce-dure is repeated by irradiating the sample at the frequencyof the H-2 absorption (to identify the signal due to H-3and measure J3,4) and then at the frequency of H-3 toidentify the signal due to H-4 and so on. When all the



402 Carbohydrates

FIGURE 14 NMR spectrum of α,α-trehalose octaacetate.

coupling constants of one ring are measured, the processis repeated for the other ring. Finally, the Karplus equationis used to determine the dihedral angle between the dif-ferent protons, which establishes the conformation of thering and the anomeric configuration. To identify all thesignals in an oligosaccharide spectrum, high-resolutionNMR instruments are needed (preferably ones with two-dimensional mapping capabilities). In the absence of suchequipment, it is still possible to determine the anomericconfiguration and ring conformation by measuring thecoupling constants of H-1 and H-4 (for pyranose rings).Figure 14 shows the NMR spectrum of octa-O-acetyl-α-D-glucopyranosyl-α-D-glucopyranoside (α,α-trehaloseoctaacetate), which clearly shows that this molecule pos-sesses two identical α-D-glucopyranosyl rings in the 1C4

conformation. This is apparent from the coupling of theanomeric proton and the fact that the two rings pro-duce identical signals, as well as from the couplingof H-4 (split by the two trans-diaxial protons at H-3and H-5).

b. By crystallography. Another way of determiningthe anomeric configuration and ring conformation is bycrystallography (using either X-ray or neutron diffrac-tion). Both techniques will afford the complete structure(in the solid state) of a crystalline oligosaccharide, includ-ing the orientation of the two rings vis-a-vis one another(angles φ and ψ). Figure 15 shows a diagram, deducedfrom X-ray defraction, of a dihydrate of α,α-trehalose.The anomeric configuration and the 4C1 conformation ofthe two rings are clearly revealed. Note also the remark-able symmetry of the molecule, which agrees with theNMR data discussed earlier.

B. Oligosaccharide Synthesis

Three types of oligosaccharide synthesis will be discussedin this section; to choose between them one should con-sider the complexity of the oligosaccharide, the length ofthe proposed scheme, and the availability of its startingmaterials. The synthetic methods are

1. The standard chemical method, which usuallyinvolves a nucleophilic substitution of a leavinggroup, previously introduced on a mono- oroligosaccharide, by an appropriate saccharide adduct.

2. The same chemical reactions described in (1), butcarried on a solid support. The operation resemblesthe automated synthesis of peptides andpolynucleotides carried out fixed beds of polymer orresin. In the present case, the nonreducing end of themonosaccharide is linked to a resin, via a linker. Themonosaccharide is then reacted, successively and inthe desired order, with the different monosaccharideadducts. When the oligomer is formed, it is separatedfrom the linker.

3. Biochemical methods which use one or moreenzymes to carry out the oligosaccharide synthesis,either in solution or on a solid support.

1. Chemical Methods of OligosaccharideSynthesis

The chemical synthesis of oligosaccharides usually in-volves reactions between saccharide derivatives that pos-sess good leaving groups at the anomeric position, suchas a halogen atom or an ester group, and an adduct, whichmay be a monosaccharide or an oligosaccharide. In the first

FIGURE 15 Conformation of α,α-trehalose dihydrate determinedby X-ray crystallography.



Carbohydrates 403

case, the glycosyl halide is subjected to a Koenigs–Knorrtype of reaction, with a halogen acceptor (e.g., AgCO3)as catalyst, whereas when an ester is used, a Helfrichtype of reaction is carried out with a Lewis acid cata-lyst, such as SbCl6, BF3, SnCl4, TiF4, etc. The glyco-syl halides are best prepared from peracetylated or per-benzoylated mono- or disaccharides by treatment withdichloromethyl methylether (DCMME). Other good leav-ing groups are thioglycosides, which requires a promotersuch as dimethyl(methylthio)sulfonium triflate (DMTST)and trichloroacetimidate which requires no activator andis therefore often used in oligosaccharide syntheses onfixed bed polymers, discussed below. The anomeric con-figuration of the formed glycosidic bond is controlled bythe nature of the protecting group adjacent to the leavinggroup. Participating groups such the acetyl group lead to1,2 trans glycosidic linkages and nonparticipating groupssuch as the benzyl group form 1,2-cis glycosidic linkages.In both cases the leaving group and the protecting groupmust be removed under mild conditions that do not af-fect the glycosidic linkages (bases in the first type andhydrogenation in the second).

Whichever glycosidation method is used in oligo-saccharide synthesis, two questions must be carefullyaddressed:

1. How can one ensure that only the oxygen atomin the desired position forms the glycosidic bond? Thisis achieved by ensuring either that the desired hydroxylgroup is the most reactive hydroxyl group in the adductmolecule or, better, that it is the only one available for re-action. Because hemiacetal hydroxyl groups are the mostreactive hydroxyl groups in cyclic saccharides, it is possi-ble to prepare nonreducing disaccharides by reacting gly-cosyl halides with unprotected monosaccharides. On theother hand, to form reducing disaccharides, it is necessaryto protect some or all of the nonreacting hydroxyl groupsin the adduct. Thus, if the desired disaccharide is linkedthrough a primary hydroxyl group [e.g., in the case of a(1→ 6)-linked disaccharide], it is necessary to block (byglycosidation) the anomeric position of the adduct (be-cause the hemiacetal hydroxyl is more reactive than theprimary hydroxyl group). Finally, if more than one pri-mary hydroxyl group is present in the adduct (such as inthe case of an ulose), or if one of the secondary hydroxylgroups is to form a glycosidic bond, it is advisable to blockall but this hydroxyl group to ensure that only the desiredglycosidic linkage is formed.

2. How can one ensure that the glycosidic linkageformed is of the desired anomeric configuration? The α

and β configuration of a newly formed glycosidic bondis determined, to a large extent, by the blocking groupspresent in the sugar moiety undergoing nucleophilic at-tack. Thus, if a participating group (e.g. an acetyl or a

benzoyl group) is attached to O-2 of the molecule un-dergoing nucleophilic attack, then a Koenigs–Knorr orHelfrich type of reaction will afford a glycoside hav-ing a trans-1.2 configuration. This isomer is favored be-cause the intermediate cyclic carbonium ion is attackedfrom the side opposing the ring. This is why β iso-mers are obtained from D-glucopyranosyl halides (and D-galactopyranosyl halides) and α isomers are formed fromD-mannopyranosyl halides.

To obtain a cis-1,2 configuration (e.g., a glycoside hav-ing an α-D-glucopyranosyl, an α-D-galactopyranosyl, ora β-D-mannopyranosyl configuration), the OH-2 of thesugar moiety undergoing nucleophillic attack must be pro-tected by a nonparticipating group, such as a benzyl group.In this case, a mixture of α and β isomers is obtained,which can be separated by chromatography. The compo-sition of this mixture is influenced by anomeric effects,which favor α-D anomers, and by temperature and dura-tion of reaction time, which favor the thermodynamicallymore stable product (i.e., the one with equatorial substi-tutents).

The following are examples of oligosaccharide synthe-ses that illustrate the ideas discussed above. Lactose [4-O-(β-D-galactopyranosyl)-D-glucose] possesses, as its namedenotes, a β-D-glycosidic bond linking C-1 of galactoseto position 4 of glucose. It could therefore be synthesizedby reacting a galactopyranosyl halide having participatingprotecting groups on O-2 with a glucose derivative havingall the hydroxyl groups blocked except for OH-4. Actuallythis synthesis was performed by reacting 2,3,4,6-tetra-O-acetyl-α-D-galactopyranosyl bromide with 2,3:5,6-di-O-isopropylidene-D-glucose diethylacetal under Koenigs–Knorr conditions, then hydrolyzing the acetal groups withacid and the ester groups with base (see Scheme 23).

SCHEME 23 Synthesis of lactose.



404 Carbohydrates

Consider now the synthesis of raffinose, which is O-α-D-galactopyranosyl-(1 → 6)-O-α-D-glucopyranosyl-(1 → 2) β-D-fructofuranoside. It is evident that thegalactopyranosyl halide needed for this synthesis must beprotected by a nonparticipating group in order to afford thedesired α-D-galactopyranosyl linkage (cis-1,2 configura-tion). Furthermore, since the adduct (sucrose) is a nonre-ducing disaccharide that possesses three primary hydroxylgroups (all of which are available for attack on the carbo-nium ion), it is necessary to block at least two of them toensure that the desired primary hydroxyl group, namely,the one attached to C-6 of the glucopyranose moiety, isthe one that reacts with the galactopyranosyl halide. Ex-perimentally, raffinose was synthesized by reacting, un-der Koenigs–Knorr conditions, a benzyl-protected galac-topyranosyl halide (tetra-O-benzyl-α-D-galactopyranosylchloride) with a sucrose derivative (2,3,4,1′,3′,4′,6′-hepta-O-acetylsucrose) having ester groups replacing all hy-droxyl groups except OH-6 of the glucopyranose moiety(see Scheme 24).

SCHEME 24 Synthesis of raffinose.

Although it is possible to achieve the synthesis of higheroligosaccharides by the addition of one monosaccharideat a time, it is advantageous to add the oligosaccharidecomponents of a large oligomer in blocks of two or moremonosaccharides (block synthesis). Thus, the trisaccha-ride and tetrasaccharide can be prepared by reacting adisaccharide glycosyl halide with a suitably protectedmono- or disaccharide. It is evident that a second roundof reactions would afford penta- and hexasaccharides.

Another interesting approach to the chemical and enzy-matic synthesis of oligosaccharide is the use of a stationarypolymer support to retain substrates (or enzymes) inside areactor (usually a column), while successive reagents arepassed through the polymer and are then washed out of thereactor. The advantage of this method is that it affords anefficient way of carrying out successive reactions withoutloss of the material attached to the column. For industrialenzymatic reactions involving costly enzymes, retainingthese inside the reactor would seem quite attractive.

2. Synthesis of Oligosaccharides on Fixed Bedsof Polymers

The importance of glycopeptides and glycolipids inmedicine and the small amount of oligosaccharides re-leased from natural sources necessitated the developmentof improved methods to prepare them including novel ap-proaches in their synthesis. As in the case of automaticnucleotide- and peptide-synthesizers, a serious search hasbeen made to develop suitable polymers for use as fixedbeds and as linkers to attach saccharides and to facilitatethe removal of the oligosaccharide formed at the end ofthe synthesis from the polymer support. The successivemonomers are added one at a time to the linker until thedesired oligomer is formed, then it is separated from thelinker. Many oligosaccharide syntheses on polymer sup-port use the commercially available polymer, polyethyleneω-monomethyl ether (MPEG) attached to the linker, α,α-dioxyxylyl diether (DOX). The product, (MPEG-DOX),needed for the oligosaccharide synthesis on solid supportis prepared as follows:

Me-O-(CH2-CH2)n-CH2-CH2-OH + Cl-CH2-(C6H4)-CH2-CL →(MPEG)

Me-O-(CH2-CH2)n-CH2-CH2-O-CH2-(C6H4)-CH2-Cl →Me-O-(CH2-CH2)n-CH2-CH2-O-CH2-(C6H4)-CH2-OH

(MPEG-DOX)

Scheme 25 illustrates the use of MPEG-DOX in thesynthesis of a pentasaccharide. The first monosaccha-ride added (A) is a benzyl blocked trichloroacetimidate



Carbohydrates 405

SCHEME 25 Synthesis of a pentasaccharide using a fixed bed of polymer (MPEG) attached to a linker (DOX). Thepentasaccharide is separated at the end of the synthesis by hydrogenation, which also rermoves the Bn blockinggroups.



406 Carbohydrates

SCHEME 26 Biochemical synthesis of a trisaccharide using two enzymes, β-(1 → 4)-galactosyltransferase to formthe disaccharide and an α-(2 → 3)-sialytransferase to form the desired product.

having an acetyl group where the next monomer is tobe attached. Its reactive leaving group, the trichloroace-timidate, is replaced by the OH of MPEG-DOX to givea blocked monomer-MPEG-DOX (B), which is treatedwith base (DBU) to remove the acetyl (Ac) group andthen with A to form a blocked dimer (C). Repetition ofthese operations will form the deacetylated blocked trimer,tetramer, and pentamer. When the desired number ofmonomers are linked together, the product is catalyticallyhydrogenated to remove the benzyl blocking groups andseparate the oligosaccharide from the linker and polymer(MPEG-DOX). The pentasaccharide is separated at theend of the synthesis by hydrogenation, which also removesthe Bn blocking groups (see Scheme 25).

3. Biochemical Synthesis of Oligosaccharides

This method is particularly useful for the synthesis of com-plex oligosaccharides that exist in small amounts difficultto characterize and study. Examples of these are the cellsurface oligosaccharides responsible for cell recognition.The most commonly used enzymes are glycosidases andglycosyltransferases, or a combination of both. The gly-cosidases are hydrolytic enzymes that can used for syn-thesis by carrying out the reaction in the reverse direction.Glycosyltransferases are enzymes that catalyze the syn-thesis of glycosides and oligosaccharides by the transferof monosaccharides from nucleotide donors to specific

substrate acceptors. The above scheme illustrates the en-zymatic synthesis of a trisaccharide by using a β-(1 → 4)-galactosyltransferase to form a disaccharide and anα-(2 → 3)-sialytransferase to form the target trisaccharide(see Scheme 26).

4. Reactions

Chemical modifications of oligosaccharides usually in-volve the reactive primary hydroxyl group, because se-lective blocking and deblocking of secondary hydroxylgroups is considerably more difficult. As the DP of anoligosaccharide increases, directing a substituent in a par-ticular moiety becomes more difficult. In a disaccharide,it is possible to direct a substituent toward the terminal(nonreducing) saccharide moiety by blocking the primaryhydroxyl group of the reducing moiety with a 1,6-anhydroring. The latter is introduced by treatment of the disaccha-ride glycoside (usually a phenyl glycoside) with base. Twoexamples of selective reactions involving the terminal ringof a disaccharide are depicted in Scheme 27.

1. The introduction of a carboxylic group by oxidationof the primary hydroxyl group of β-benzylcellobioside with oxygen in the presence of palladiumon charcoal occurs preferentially at the terminal ring.However, greater selectivity can be reached if the1,6-anhydro derivative is first formed.



Carbohydrates 407

SCHEME 27 Selective reactions of the nonreducing ring of cellobiose.

2. The introduction of an azido group in the 6′ positionof lactose can also be achieved by tosylating the1,6-anhydro derivative of cellobiose, tritylating, andfully acetylating the product (Scheme 28). When thetrityl group is removed by acid, the acetyl groupattached to position 4 migrates, yielding a4-hydroxy-6-O-acetyl derivative. Mesylating atposition 4 and displacing the leaving group with azideaffords the 4-azido derivative of lactose. This azidogroup can be reduced to an amino group or can bephotolyzed to afford an aldehydic group.

Of considerable importance as surfactants are the fattyacid esters of disaccharides. An example of these is su-crose monooleate, which possesses a hydophilic–lipo-philic balance of ∼15 and is ideal for use as a mild sham-poo. Such esters are obtained by reacting the disaccharidein DMF with the methyl ester of the fatty acid.

IV. POLYSACCHARIDES

Polysaccharides are polymeric saccharides linked byacetal (glycosidic) linkages. The repeating units aremonosaccharides or oligosaccharides, seldom larger thanpentasaccharides. Although some polysaccharides aremade up of fewer than a hundred sugar residues, the major-ity are much larger, reaching several thousand monomerunits in length.

Due to their high molecular weights, most polysaccha-rides exhibit the characteristic properties of high polymers(low solubility and high viscosity, etc.). This is why the

DP of polysaccharides can be determined by the samemethods used for the molecular weight determination ofsynthetic polymers.

Polysaccharides are usually given names that reflecttheir origin, for example, cellulose, the principal compo-nent of plant cell walls. Otherwise, a systematic nomencla-ture can be used that suffixes the ending “an” to the nameof the monomer. For example, glycan and mannan arepolymers of glucose and mannose, respectively, whereasgalactomannans are copolymers of galactose and man-nose. Many early polysaccharide names that have otherendings, such as pectin and chitin, are still in use.

Polysaccharides may exist as homologous series thathave average molecular weights, rather than possess dis-crete molecular weights as proteins do.

A. Structure

The types of monosaccharides found in polysaccha-rides can easily be determined after depolymerization(by acid-catalyzed hydrolysis), by chromatographic anal-ysis. Liquid chromatography may be used directly onthe hydrolysate, whereas gas chromatography is usu-ally carried out after sililation in order to volatilize themonosaccharides.

Because polysaccharides posses a certain naturally im-posed simplification, their structure elucidation is muchsimpler than would appear at first glance. A further sim-plification in polysaccharide structure is that only a veryfew monosaccharides are found in natural polysaccha-rides. There are only four commonly occurring hex-oses (glucose, mannose, fructose, and galactose), two



408 Carbohydrates

SCHEME 28 Conversion of a cellobiose derivative to a 4-amino-lactose.

pentoses (xylose, arabinose), two 6-deoxyhexoses (rham-nose, fucose), and two amino sugars (glucosamine andgalactosamine). Four uronic acids (glucuronic, galactur-onic, mannuronic, and iduronic acids) are found in sev-eral polysaccharides. Of these, D-glucuronic acid and D-galacturonic acid are the most common. The other two,D-mannuronic acid and L-iduronic acid, are found in al-ginic acids and in heparin, respectively.

Polysaccharides that afford on hydrolysis only onemonosaccharide type are termed homoglycans, whereasthose that afford two or more monosaccharide types are

termed heteroglycans. Each of these can be either linearor branched.

1. Linear Polysaccharides

Linear polysaccharides are formed when the hemiacetalhydroxyl group (on C-1 of an aldose) is replaced (substi-tuted) by a hydroxyl group from an adjoining monosac-charide unit. When this is repeated, a linear chain is formedthat has at one end (called the reducing end) a reducingmonosaccharide (one that has a free C-1 hydroxyl group)and at the other end (referred to as the nonreducing end)a glycosidically linked saccharide.

The glycosidic linkage in polysaccharides is usually re-peated in a regular manner, with little or no randomness.However, even such seemingly regular molecules as cel-lulose and amylose have irregularities in their structures.Rare as these may be (sometimes the frequency is onlyone in a thousand), these irregularities are real. They areattributable to the fact that within the enzyme–substratesystem there do occur during the course of time changes,either through abnormal action on the part of the princi-pal chainsynthesizing enzyme or through interference ofa second enzyme. The most abundant linear polysaccha-ride is cellulose, a glucan linked by β-D-(1 → 4) bonds,which is present in greater quantity than all other polysac-charides combined. Another important linear polysaccha-ride is amylose, which is a glucan linked by α-D-(1 → 4)linkages.

2. Branched Polysaccharides

When the hemiacetal hydroxyl groups of two monosac-charides are replaced by two hydroxyl groups belongingto a third sugar residue, a branching point is produced inthe molecule, which then becomes a branched polysac-charide. The molecule may contain a single or numerousbranch points. Examples of branched polysaccharides areamylopectin (the branched component of starch) and thegums and mucillages.

B. Polysaccharides of Industrial Importance

1. Starch

Most plants produce starch, but only a few concentrate it inamounts suitable for industrial production. These plantsinclude cereals, such as corn, which may contain up to80% starch, grains such as rice and tubers such as potatoes.Starch is found in the form of microscopic granules whichpossess characteristic shapes used to identify their source.



Carbohydrates 409

FIGURE 16 Partial structures of the two forms of starch; the linearamylose and the branched amylopectin.

The industrial preparation of starch starts by loosening thestarch granules away from their matrices with cold waterand letting them settle in tanks before filtration and drying.Purification of starch is achieved by repeated cycles ofdispersion in cold water, sedimentation, and filtration. Ifdesired, such treatment may be followed by fractionationto separate the linear polymer of starch, amylose from thebranched polymer, amylopectin (see Fig. 16).

Amylose: Amylose is the linear form of starch; it is pre-cipitated from a cold aqueous dispersion of starch granuleswith alcohols such as thymol or 1-butanol and filtered. Theaccompanying amylopectin is left in the supernatent solu-tion and if needed can be precipitated. Usually the ratio ofamylose to total starch ranges between 15 and 35%, how-ever, certain plant hybrids have been developed which arecapable of producing a much larger excess of one poly-mer over the other. Thus, amylomaize contains more than80% of amylose; whereas waxy maize starch containsonly 2% of this linear polymer. Amylose is composedof (1 → 4)-linked α-D-glucopyranosyl residues attachedby acetal bonds. Due to lengthy manipulation during endgroup assay, the DP of amylose measured chemically is in-variably lower than the DP measured by physical methods,such as light scattering or ultracentrifugation. For exam-ple, the DP calculated from the amount of glucose presentat the nonreducing end of the chain and estimated fromthe amount of 2,3,4,6-tetra-O-methyl-D-glucose producedupon methylation and hydrolysis of amylose ranges be-tween 200 and 300. This is much lower than the DPobtained by light scattering or ultracentrifugation (about6000). The difference has been attributed to degradationduring the chemical treatment. In neutral solution amy-lose exists as a random coil, but other conformers can beproduced by retrogradation, i.e., separating the differentinsoluble fractions of amylose that deposit on standing.In the presence of complexing agents amylose assumesa more organized conformation; namely that of a right-handed helix made up of 150 to 1000 turns and containingsix α-D-glucopyranose rings per turn. The dimensions ofthe amylose helix are such that it can accommodate aniodine atom, or a lipid molecule of appropriate size. Thelight absorption of iodine-amylose complexes is very sim-

ilar to that of iodine in nonoxygenated solvents, such ashydrocarbons. This observation has been attributed to thelack of interaction between the equatorial oxygen atomsof the six α-D-glucopyranose rings forming a turn of thehelix and an atom at its center. The latter is in a lipophilicsurrounding (of the axial C-H groups), very similar to thatof iodine in a hydrocarbon solution. The lipophilic cen-ter of the helix would also make it difficult to remove alipid molecule located in its center, which explains whythe 13C-NMR spectrum of pure amylose invariably showslipid absorptions.

Amylopectin: Amylopectin is a highly branchedα-D-glucose polymer that possesses a much largermolecular weight than amylose. It contains a numberof linear chains of disubstituted (1 → 4)-linked α-D-glucopyranosyl residues similar to those found in amylose,in addition it contains some (1 → 6) linked branches orig-inating from 1,4,6-tri-O-substituted α-D-glucopyranosylresidues. Light scattering data gives amylopectin a DPvalue in the order of 105 or higher if the starch fromwhich it was obtained had matured for a long period inthe plant. The accepted shape of the amylopectin macro-molecules is that of a three-dimensional tree or bush(shown two dimensionally in Fig. 17). This conformationis based on data obtained from methylation experimentsand enzymatic methodologies. For example, methylationof amylopectin yields a smaller amount amount of 2,3,6-tri-O-methyl-D-glucose than amylose. Because 2,3,6-tri-O-methyl-D-glucose is produced from linear portions of(1 → 4) linked α-D-glucopyranose polymers and becauseit is formed in a smaller ratio from amylopectin than fromamylose, one can conclude that the chains of amylopectinare considerably shorter than those of amylose. The highdegree of branching of amylopectin can also be deducedfrom the relatively large amount (4%) of 2,3,4,6-tetra-O-methyl-glucose, produced from the nonreducing ends ofthe amylopectin molecule, as compared to the amount pro-duced from amylose (<0.4%). In addition, amylopectinyields a unique product not usually found when amylose ismethylated and hydrolyzed, namely of 2,3-di-O-methyl-D-glucose, produced from the points of branching.

FIGURE 17 A two-dimensional depiction of the bush model ofamylopectin. The reducing end of the molecule is represented bya solid black circle; the other circles represent the nonreducingends.



410 Carbohydrates

a. Starch-based industries. Dry milling is used toproduce flours such as corn flour and meals such as oat-meal, whereas wet milling, is used as feedstocks in theproduction of sugar syrups, used in the manufacture offoods, soft drinks, and beer.

Sweeteners From Starch: The most widely used naturalsweetener is the disaccharide sucrose, which is availablein crystalline form as well as in the form of syrups un-der various names such as molasses and invert sugar. Upuntil recently the only serious competitor of sucrose wasthe cheaper glucose syrup produced since the early nine-teenth century by the acid hydrolysis of starch and morerecently by its enzymatic hydrolysis with α-amylases andglucoamylases. Because glucose syrup is significantly lesssweet than sucrose, greater quantities of this syrup mustbe used to yield a product of comparable sweetness to su-crose. This is a serious disadvantage, duly noted by thecalorie-conscious public and by the food industry, whichtook steps to remedy the situation. The glucose syrup waspartly converted to fructose by the enzyme glucose iso-merase obtained from Bacillus coagulans. The resultingproduct, designated fructose syrup, contains 42% fruc-tose and possesses nearly the same sweetness as sucrose.Recently, this sweetener was further improved upon; itwas converted into two products that are sweeter thansucrose. They are very high fructose syrups that con-tain about 90% fructose and crystalline fructose. Both areobtained from fructose syrup by purification and separa-tion on ion exchange columns. Fructose syrup and highfructose syrup have now replaced glucose syrup in manysoft drinks, jams, confectionery, and conserves and crys-talline fructose is slowly replacing sucrose. It is estimatedthat half the sweeteners consumed in the United Statesare now produced by the hydrolysis of starch in contin-uous flow biocatalyst reactors that efficiently reuse theenzymes.

Starch Fermentation Industries: Beer is manufacturedfrom barely whereas sake is obtained from rice. Ger-minated barely is used in breweries to produce the en-zyme β-amylase needed for the breakdown of starch,whereas a mold, Asparagillus oryzea, is used to producethe α-amylase needed to hydrolyze rice starch for the man-ufacture of sake. The wort in both cases is subjected tofermentation with the yeast, Saccharomyces serevisea. Asin many modern food-related industries, continuous flowreactors and immobilized enzymes have been used in themanufacture of beer, and more recently of wine, and cham-pagne. This has resulted in a considerable reduction in costand residence time.

Starch and starch-rich wastes from the food industryhave been proposed as feedstocks for the fuel industry toproduce gasohol, and the food industry to produce single-cell proteins for animal consumption.

b. Chemically modified starch. The properties ofstarch are often ameliorated by chemical modification.This is why today the majority of the starch producedfor industry is chemically modified. Examples of suchmodified starches shown below.

Oxidized Starch: Starch is oxidized to remove some ofits numerous OH groups that are responsible for hydrogenbonding and for gel forming. The resulting starch solutionsafter oxidation form weaker gels and are less susceptibleto autodecomposition or retrogradation. The oxidationsare usually carried out in basic media using oxidants suchas sodium hypochlorite or better sodium periodate. Oxi-dation with the latter is usually followed by electrolysis toregenerate periodic acid from the iodic acid formed duringthe process. The product formed by periodate oxidation isa 2,3-dialdehyde formed by cleavage of the 2,3- bonds ofstarch (see Scheme 29). This can be made to react with

SCHEME 29 Oxidation of starch with periodic acid gives a dialde-hyde, which can be grafted to natural polymers such as cellulose.Oxidation of starch with ceric ions gives a free radical which canbe grafted to acrilonitrile to give an absorbant graft polymer.



Carbohydrates 411

natural or synthetic polymers having OH groups to formcross-linked polymers. Cross linking with cellulose af-fords products possessing enhanced wet strength that areused in the paper industry to make paper towels. Crosslinkage with cotton affords shrink- and crease-resistantfabrics for the textile industry. It is also possible to re-act the free radicals produced by oxidation of starch withceric ions with monomers to form graft polymers (seeScheme 29). For example, copolymerization of oxidizedstarch with polyacrylonitrile yields excellent absorbentsused in the manufacture of disposable diapers capable ofabsorbing 5000 times their weight in water.

Cross-Linked Starches: Cross-linked starches are ob-tained by forming bridges between the hydroxyl groupsof the starch molecule and the hydroxyl groups of an-other molecule. The resulting compounds are extensivelyused by the food industry as thickeners or to impart addedmechanical stability to prevent the food from loosing itsshape during cooking. In theory any molecule possessingtwo groups capable of reacting with hydroxyl groups canbe used as a bridge between starch molecules. An exam-ple of such bridge-forming molecules is epichlorhydrin,which forms distarch glycerol depicted below.

Starch-OH + CH2-CH-CH2-Cl❙ �O

Epicholorohydrin

−→ Starch-O-CH2-CH-CH2-O-Starch

OH

Distarch glycerol

Starch Ethers: Starch ethers are produced by reactingstarch with ethylene or propylene oxide to the desired DS.For example, starch hydroxypropyl alcohol is producedby reacting starch with propylene oxide. The nucleophilicsubstitution is carried out in basic media and the result-ing starch ether, depicted below, is widely used in paper,textile, and food industries.

Starch-OH + CH2-CH-CH3 −→ Starch-O-CH2-CH-CH3❙ �O OH

Propylene oxide Starch ether

2. Glycogen

Glycogen is a homopolysaccharide found in the liver andmuscles of animals, where it is used to store energy. Chem-ically, glycogen is related to starch and closely resemblesamylopectin. It is composed of linear chains of disubsti-tuted (1→4)-linked α-D-glucopyranosyl residues attachedto (1→6) linked branches originating from 1,4,6-tri-O-substituted α-D-glucopyranosyl residues. The difference

FIGURE 18 Bush structure of glycogen. Note that the moleculeis more symmetric than that of amylopectin.

between glycogen and starch is in the source (starch isfound in plants) and the shape of its aggregates and theirsize. Instead of starch granules, glycogen is found in theform of spheres called β-partcles, having a DP of 105

and aggregates of spheres called α-particles having a DPof 107. The shape of the glycogen macromolecules resem-bles a bush or a tree (see Fig. 18). However, its agregatesare much more symmetric than those of amylopectin (seeFig. 17). This is why the aggregates of starch are calledgranules and those of glycogen, spheres.

3. Dextrans and Glucans

Dextrans are α-D-glucopuranose polymers linked for themajor part through (1 → 6) acetal bonds as well as bysome (1 → 3) and (1 → 4) linkages. Dextrans are highlybranched polymers having a DP in excess of 106. Al-though dextrans are obtained from a multitude of bac-teria, most of the industrially produced products are ob-tained from Leuconostoc mesenteroides and Leuconostocdextranicum.

D-glucans are obtained from fungi and may either beα-D-plucopyranose or β-D-glucopyranose polymers; theycontain mainly (1 → 3) and (1 → 4) glycosidic linkagesand rarely (1 → 6) bonds.

4. Cellulose

Cellulose is the most abundant organic compound onearth; its amount exceeds by far that of any other or-ganic polymer or monomer. Cellulose is found in a nearlypure form in cotton and flax and in much larger amounts,but less pure form in wood, from which it can be pro-duced by dissolving its contaminants, namely hemicel-lulose and lignin, with alkali. Cellulose-based industriesinclude the pulp and rayon industries where it is pro-duced and purified, respectively, and the paper and thetextile industries, where it is processed. Like amylose,cellulose is a linear polymer made up of (1 → 4)-linkedD-glucopyranosyl residues, which differ from those of



412 Carbohydrates

FIGURE 19 Structure of cellulose.

amylose in their anomeric configuration; β for celluloseand α for amylose. The DP of cellose, determined byphysical methods is around 104, much higher than thevalue deduced chemically by end group analysis, whichis only 200. The difference is attributed to its degradationduring chemical manipulation. A similar difference wasobserved with amylose. Upon methylation and hydrol-ysis, cellulose affords 2,3,6-tri-O-methyl glucopyranosein high yield (>90%). The presence of (1 → 4)-linkagescan be confirmed by partial hydrolysis to the disaccha-ride, cellobiose, whose β-D-configuration was confirmedby enzymatic hydrolysis (for a partial structure of cellulosesee Fig. 19). The actual shape of the cellulose aggregateswas studied by X-ray diffraction, which revealed that thereexist two types of cellulose: cellulose I and cellulose II.Cellulose I is formed naturally and is composed of highlyordered microfibrils that are held together by hydrogenbonds and are further associated to form fibers. Bundlesof cellulose I all have their reducing ends in the same sideof the macromolecules and the nonreducing end in theopposite side. A different type of bundle occurs in cel-lulose II, found in regenerated cellulose such as viscoserayon. The aggregates of cellulose II have their reducingends arranged in an antiparallel manner, i.e., their reduc-ing ends occur at both ends of the bundle. Cellulose IItypes can be further divided into IIII and IIIII and IVI andIVII, according to the conformation of their unit cells; theI family forms bent chains, whereas the II families adoptbent-twisted conformations.

a. Chemically modified cellulose. The most im-portant derivatives of cellulose are the esters, such ascellulose acetate and nitrate, the ethers, such as car-boxymethyl, hydroxyethyl, and hydroxypropyl cellulose,and the rayons, of which the most important are the acetaterayon, which is discussed next and viscose rayon obtainedfrom cellulose xanthate.

Cellulose Acetate: Cellulose acetate, also known as ac-etate rayon is obtained from wood pulp in three steps:(1) pretreatment (soaking with acetic acid and sulfuricacid); (2) acetylation with a mixture of acetic anhydrideand acetic acid, to form the 2,3,6-tri-O-acetyl derivative;and (3) partial hydrolysis to afford a product having a DSof about 2.6, which forms the best fibers and films.

Cellulose Nitrate: Cellulose can be esterified with amixture of nitric acid and some sulfuric acid added ascatalyst. The amount of water controls the degree of sub-

stitution, which in turn determines the appropriate usagefor the product; higher DS products are used as explosivesand lower ones as cements and lacquers.

Carboxymethyl Cellulose: This ether of cellulose is ob-tained by a modified Williamson synthesis whereby cel-lulose is first treated with sodium hydroxide and the alco-holate formed is reacted with the halide, namely sodiummonochloroacetate. Although a DS of 3 is attainable (po-sitions 3,4, and 6 can all be etherified), a DS of only 0.4 isall that is needed to achieve the desired viscosity and sol-ubility needed in the commercial products. Carboxmethylcellulose is widely used in the form of alkali metal saltsas surfactant in detergents, in oil drilling, mining, and inthe food, paper, and textile industries.

Hydroxyethyl Cellulose: Ether alcohols such ashydroxyethyl cellulose are used industrially to solubi-lize cellulose and to facilitate its handling. Hydroxy-ethyl cellulose is prepared by swelling the cellulose withsodium hydroxide and reacting it with ethylene glycol.As in the previous example, a low DS of 0.3 is sufficientto solubilize the product in alkali.

Cell-OH + CH2-CH2 −→ Cell-O-CH2-CH2-OH❙ �O

Cellulose Hydroxyethyl Cellulose

Hydroxypropyl cellulose is prepared in the same mannerexcept that propylene glycol is used in place of ethyleneglycol.

Cellulose Xanthate: Cellulose xanthate or viscose is anintermediate in the production of viscose rayon, used inthe manufacture of textile fibers and cellophane wrapping.The manufacture of viscose starts by swelling the cellulosewith sodium hydroxide. The mixture is then treated withcarbon disulfide to form the sodium xanthate of cellulose,which is known as viscose. It was given this name be-cause its viscosity increases with time. When the desiredconsistency is reached, the mass is extruded into an acidbath to form fibers or sheets depending on the extrusionorifice shape (circular holes or slits). The acid bath decom-poses the xanthate groups and regenerates cellulose nowcalled viscose rayon to distinguish it from other rayons.

Cell-OH −→ Cell-ONa + CS2 −→ Cell-OCSSNa

Cellulose Viscose

−→ Cell-OH

Viscose rayon

5. Mannans and Galactomannans

Mannans are highly insoluble polymers of β-D-man-nopyranose, found in the kernels of many palms, for ex-ample, ivory nut. This hard material is composed of nearly



Carbohydrates 413

pure mannan, and was, until the advent of plastics, usedfor making buttons. Many palm kernels contain less purepolymers of mannose, often containing galactose and re-ferred to as galactomannans.

6. Chitin and Chitosans

Chitin and Chitosans are polymers of β-D-glucosamine;the main difference between them is in DP, the first poly-mer has a DP around 105 whereas the second has only104. As a result chitosan is much more soluble than chitinand is therefore easier to handle and to shape; it is used inphotographic films, for immobilization of enzymes, andits phosphate esters as fire retardants.

7. Hemicelluloses

Hemicellulose is the most abundant naturally occurringorganic compound after cellulose. It differs from cellu-lose in that it dissolves in alkali to form a dark brownsolution known in the paper industry as “black liquor.”Some hemicelluloses are linear while others are highlybranched. For example, xylans, the most common compo-nent of hemicellulose, are mostly linear polymers madeup of (1 → 4) linked β-D-xylopyranosyl residues, whichmay occasionally branch at the 3-position and arabino-D-galactans which are highly branched polymers.

8. Pectin

Pectin is produced in large quantities as by-products of theapple and orange juice productions. The pulp of apples andthe peels of orange left over after pressing contain about

FIGURE 20 The four 2-deoxyribonucleotides found in DNA and the four ribonucleotides found in RNA.

15 and 30% pectin, respectively. The pectin extracted fromthese agricultural wastes is used in the manufacture of jamsand jellies. It is also used as a stabilizer of fruit drinks andpharmaceutical suspensions because of its great stabilityat low pH.

V. OLIGONUCLEOTIDES ANDPOLYNUCLEOTIDES

Oligonucleotides are DNA or RNA segments of lowmolecular weight; they are composed of nucleotidemonomers (see Fig. 20) linked by phosphoric ester bridgesspanning C-5′ of one unit and C-3′ of the other. Syntheticoligonucleotides have been widely used in the study ofDNA, in protein biosynthesis, and in induced mutagenesis.They were instrumental in deciphering the genetic code,and they have since been used to introduce mismatches inspecific sites of DNA strands to induce specific changesin enzymes.

More recently, antisense oligonucleotides have beenused in the fight against AIDS and cancer. Anitisenseoligonucleotides were synthesized that complement por-tions of mRNA which translate to the protein cover of theHIV virus or to a cancer specific oncogene. In the cell,the oligonucleotides complex with the targeted RNA andarrest the synthesis vital proteins (see Fig. 21).

A. Synthesis of Oligonucleotides

Both oligodeoxyribonucleotides (DNA-type oligomers)and oligoribonucleotides (RNA-type oligomers) havebeen synthesized, but the first have attracted most of the



414 Carbohydrates

FIGURE 21 Inhibition of gene translation with antisenese oligonucleotides. An oligonucleotide designed to complexeswith a segment of mRNA is used to arrest the synthesis of a vital protein.

attention of chemists because of their biological impor-tance and their relative ease of preparation. These char-acteristics become evident when the structures of the twotypes of nucleotide monomers are examined. Thus, theDNA monomers have only one free hydroxyl group (onC-3′), which, when esterified by the phosphate group ofanother monomer, will form a phosphate bridge in the de-sired position (linking C-3′ of one nucleotide to C-5′ ofthe other), whereas ribonucleotides have two free hydroxyl

SCHEME 30 Some reactions of ribonucleotides.

groups (one on C-2′ and one on C-3′), which necessitatesthat the 2′-OH group be protected to prevent the formationof the undesired 2′ → 5′ phosphate bridge.

B. Block Synthesis of Oligonucleotides

An oligonucleotide chain may be terminally phosphated(esterified) on C-5′ or C-3′. Oligomers of the first type areprepared from natural nucleotides by treating an acetylated



Carbohydrates 415

SC

HE

ME

31F

orm

atio

nof

olig

onuc

leot

ides

inau

tom

atic

synt

hesi

zers

usin

gfix

edbe

ds.



416 Carbohydrates

nucleotide (one having a phosphate group on C-5′ and anacetyl group on O-3′) with a nucleotide or an oligonu-cleotide having a blocked phosphate group on C-5′ and afree hydroxyl group in position 3′. The second type, thoseterminally phosphated at C-3′, are usually prepared fromnucleosides phosphated at C-3′ (which are not availablecommercially). The monomers are acetylated on O-5′ andtreated with other monomers having free hydroxyl groupson C-5′ and a blocked phosphate group on C-3′ to obtainthe desired oligonucleotide.

In both cases, it is necessary to block the phosphategroups to prevent the formation of anhydrides (pyrophos-phates). This is done by treating the nucleotide with3-hydroxypropanonitrile to obtain an ester removable withalkali (β elimination). Other phosphate-blocking groupsinclude PhCH2 O (CH2)2 NH2, which forms an amidehydrolyzable by acids.

The coupling of the two nucleotide moieties involvesformation of a phosphoric diester from a monoester (or atriester from a diester). The reaction is catalyzed by suchcondensing agents as dicyclohexylcarbodiimide (DCC), areagent extensively used in peptide synthesis, or arylsul-fonyl chlorides (for example, 2,4,6-trimethylphenyl- or2,4,6-triisopropylphenylsulfonyl chloride).

All nucleotides possessing primary amino groups (threeof the four bases present in DNA and RNA nucleotides),namely adenylic, guanylic, and cytidylic acid, must havetheir amino groups protected to prevent the formation ofamide ester bridges (instead of diester phosphate bridges).The primary amino group of adenylic acid is usually pro-tected by a benzoyl (or p-methoxybenzoyl) group. This isachieved by treating the nucleotide with an excess of ben-zoyl chloride and removing the undesired benzoyl groupswith alkali. Guanylic acid is usually protected by reac-tion with N ,N -dimethylformamide diethyl acetal to givea readily removable Schiff base. Finally, cytidylic acid isprotected by formation of a carbamate (see Scheme 30).

C. Formation on Automated OligonucleotideSynthesizers

The phosphite triester approach is commonly used withautomatic synthesizers. The phosphite ester formed is ox-idized at the end of each sequence to a phosphate ester.Typically, a deoxynucleoside 3′-phosphoramidite is addedto a growing DNA segment that is linked to a silica support.The silica is first reacted with a 3-aminopropyl triethoxysi-lane linker and then with an appropriate tritylated nu-

cleoside, such as a 5′-dimethoxytrityl-2′-deoxynucleosidehaving a 3′-p-nitrophenylsuccinoyl group. The additionof successive deoxynucleotides to this support-bounddeoxynucleoside is carried out by (a) removal of thetrityl (DMT) protecting group with trichloroacetic acid;(b) condensing with a tritylated deoxynucleoside 3′-phosphoramidite, to yield a deoxydinucleoside phosphitetriester; (c) oxidizing of the phosphite triester to a phos-phate triester with iodine (see Scheme 31). Repetitions ofthis sequence have been used to synthesize DNA segmentscontaining up to 200 deoxynucleotides. Once a synthesisis complete, the oligodeoxynucleotide is removed fromthe column, freed of protecting groups, and isolated bygel electrophoresis and purified by HPLC.


BIOPOLYMERS • PHYSICAL ORGANIC CHEMISTRY • PULP

AND PAPER

BIBLIOGRAPHY

Aspinall, G. O., ed. (1982). “The Polysaccharides,” Academic Press,New York.

Boons G. J., ed. (1998). “Carbohydrate Chemistry,” Academic and Pro-fessional.

Ginsberg, V., and Robbins, P. W. (1991). “Biology of Carbohydrates,”Jai Press, Tokyo.

Robert M. Giuliano, R. M., ed. (1992). “Cycloaddition Reactions inCarbohydrate Chemistry,” ACS Symposium Series, Washington D.C.

Gyorgydeak, Z., and Pelyvas, Z. F. (1998). “Monosaccharide Sugars,”Academic Press, New York.

Hanessian, S. (1983). “Total Synthesis of Natural Products,” Pergamon,Oxford,England.

Horton, D., ed. (1995). “Advances in Carbohydrate Chemistry andBiochemistry,” Vol. 50, Academic Press, New York.





Kennedy, J. F., ed. (1988). “Carbohydrate Chemistry,” Clarendon Press,Oxford, England.

Ogura, H., Hasegawa, A., and Suami, T. (1992). “Carbohydrate SyntheticMethods,” Kondasha, Tokyo.

Postema, M. H. D. (1995). “C–Glycosides Syntheses,” CRC Press,Cleveland.

P1: FPP Revised Pages Qu: 00, 00, 00, 00


Catalysis, HomogeneousPiet W.N.M. van LeeuwenUniversity of Amsterdam

I. Definition and Scope of Homogeneous CatalysisII. Elementary Steps for Homogeneous Catalysis

III. Description and Mechanisms of HomogeneousCatalytic Processes

GLOSSARY

Coordination compound Molecule or ion comprising ametal atom or ion surrounded by ions or moleculescalled ligands.

Coordination number Number of ligand atoms bondedto the metal atom or ion in a coordination compound.

Electrophile Refers to the reactivity of an atom, ion, ormolecule; an electron-deficient species ready to acceptan electron pair from another species thus forming achemical bond.

Enantiospecificity Preference for the formation of a sin-gle enantiomer of a compound.

Metal complex See Coordination compound.Migratory insertion Insertion of an unsaturated mole-

cule such as ethene or CO (Y) into a metal–X bond,whereby X migrates to one of the atoms of the unsatu-rated molecule coordinated to the metal giving M–Y–X.

Nucleophile Refers to the reactivity of an atom, ion, ormolecule; an electron-rich species ready to donate anelectron pair to another species thus forming a chemicalbond.

Oligomerization Coupling of several molecules of thesame kind forming a larger unit (oligomer).

Oxidative addition Oxidation of a metal compound ac-companied by an increase of the coordination number.

Polymerization Coupling of many molecules (mono-mers) forming a polymer (macromolecule).

Reductive elimination (Reverse of oxidative addition.)Loss of one or two ligands from a metal complex byreduction of the metal.

THE PHENOMENON CATALYSIS has become wellknown owing to the use of catalysts in automobile ex-hausts, which ensure the complete conversion of the lasttraces of fuel and decompose the toxic nitrogen oxides andcarbon monoxide to nitrogen and carbon dioxide. Cata-lysts are substances that enhance chemical reactions thatare not themselves consumed in the reactions. Exhaust cat-alysts are heterogeneous, as the catalyst is a solid materialwhile the reactants are in the gaseous phase. In this articlewe are concerned with homogeneous catalysts, that is tosay that catalyst and reactant are molecularly dispersed inthe same phase, often the liquid phase. Homogeneous cat-alysts for the gas phase are also known and a well-knownexample is the chlorine atom catalyzing the decomposi-tion of ozone in our stratosphere. Here we will focus onmetal catalysts in solution that have found widespread uti-lization in chemical industry for the manufacture of bulkchemicals, fine chemicals, and pharmaceuticals and in or-ganic synthesis in the laboratory. The catalysts presented

457

P1: FPP Revised Pages


458 Catalysis, Homogeneous

are important for a variety of reactions including hy-drogenation, isomerization of alkenes, oligomerization,polymerization, carbonylation, hydroformylation, hydro-cyanation, metathesis, polyester formation, etc. Manycatalysts display high rates and selectivities for conver-sions that would have taken many steps in the absence ofthe catalyst or would not have been possible at all. Manyof the catalytic reactions show a high atom economy, i.e.,a high percentage of the atoms of the starting materialsending up in the final product as by-product formation isminimized.

I. DEFINITION AND SCOPE OFHOMOGENEOUS CATALYSIS

A. Definition

Berzelius coined the term catalysis in 1836 when he hadnoticed changes in substances when they were brought incontact with small amounts of certain species called “fer-ments.” Many years later in 1895 Ostwald came up withthe definition that: A catalyst is a substance that changesthe rate of a chemical reaction without itself appearinginto the products. This means that according to Ostwald acatalyst can also slow down a reaction. The definition usedtoday reads as follows: A catalyst is a substance that in-creases the rate at which a chemical reaction approachesequilibrium without becoming itself permanently involved.The “catalyst” may be added to the reactants in a differentform, the catalyst precursor, which has to be brought intoan active form (“activated”). During the catalytic cyclethe catalyst may be present in several intermediate formswhen we look more closely at the molecular level. An ac-tive catalyst will pass a number of times through this cycleof states; in this sense the catalyst remains unaltered. Thenumber of times that a catalyst goes through this cycle isthe turnover number. The turnover number (t.o.n.) is thetotal number of substrate molecules that a catalyst convertsinto product molecules. The turnover frequency (t.o.f.) isthe turnover number in a certain period of time. Substratesare present in larger amounts than the catalyst; when wereport on catalytic reactions the ratio of substrate to cat-alyst is an important figure. An inhibitor is a substancethat retards a reaction. An inhibitor is also present in “cat-alytic” or substoichiometric amounts. In a radical chainreaction an inhibitor may be a radical scavenger that inter-rupts the chain. In a metal-catalyzed reaction an inhibitorcould be a substance that adsorbs onto the metal making itless active or blocking the site for substrate coordination.We also talk about a poison, a substance that stops thecatalytic reaction. We will often see the word co-catalyst,a substance that forms part of the catalyst or that playsanother role somewhere in the catalytic cycle.

Kinetics is an important part of catalysis; after all, catal-ysis is concerned with accelerating reactions. Comparisonof catalysts and comparison of catalyzed and noncatalyzedreactions is not a straightforward task. Suppose we havea bimolecular reaction of species A and B with a rate ofproduct formation:

d[P]/dt = k1[A][B]

We don’t know what the rate equation for the catalyzedreaction might look like, but it is reasonable that at leastthe catalyst concentration will occur in it, e.g.:

d[P]/dt = k2[Cat][A][B]

Hence the dimension of the reaction is different, even inthe simplest case, and hence a comparison of the two rateconstants has little meaning. If both reactions occur si-multaneously in a system we may determine what part ofthe product is made via the catalytic route and what partisn’t. In enzyme catalysis and enzyme mimics one oftencompares the k1 of the uncatalyzed reaction with k2 of thecatalyzed reaction; if the mechanisms of the two reactionsare the same this may be a useful comparison. In prac-tice the rate equation may take a much more complicatedform than the ones shown above. The rate equation tellsus something about the mechanism of the reaction.

Before we turn to “mechanisms” let us repeat how acatalyst works. We can reflux carboxylic acids and alco-hols and nothing happens until we add traces of mineralacid that catalyzes esterification. We can store ethene incylinders for ages (until the cylinders have rusted away)without the formation of polyethylene, although the for-mation of the latter is exothermic by more than 80 kJ/mol.We can heat methanol and carbon monoxide at 250◦C and600 bar without acetic acid being formed. After we haveadded the catalyst the desired products are obtained at ahigh rate.

A catalyst lowers the barrier of activation of a reaction,i.e., it lowers the activation energy. When protons or Lewisacids are the catalysts this description seems fairly accu-rate. Take for instance a Diels-Alder reaction catalyzed bya Lewis acid (Fig. 1):

The catalyst makes the dienophile more electrophilic. Itlowers the energy level of the LUMO and the interactionbetween the LUMO of the dienophile and the HOMO

FIGURE 1 Lewis acid catalyzed Diels-Alder reaction.



Catalysis, Homogeneous 459

of the diene increases. As a result the reaction becomesfaster than the uncatalyzed one. Accidentally it becomesalso more regioselective. In this instance the catalyzedreaction is very similar to the uncatalyzed reaction. Oftenthis is not the case. In particular when the reagents arevery unreactive toward one another (ethene versus ethene,methanol versus CO, see above) the simple picture of acatalyst that lowers the reaction barrier is beyond the truth.

The following description is more general. A cata-lyst provides a more attractive reaction pathway for thereagents in order to arrive at the products. This new path-way may involve many steps and may be rather compli-cated. Summarizing, a catalyst provides a new reactionpathway with a low barrier of activation, which may in-volve many intermediates and many steps. The sequenceof steps is called the mechanism of the reaction. Mech-anism also refers to the more detailed description of areaction at the molecular bonding level. During the pro-cess, the catalytic cycle, the catalyst participates in many“complexes” all of which might be named “the” catalyst.It cycles continuously from one species to another. In thissense the catalyst itself remains unchanged during the cat-alytic conversion (Ostwald, first paragraph). The catalystmay spend most of its time in the form of a certain species,which is often the species undergoing the slowest reactionof all in the catalytic cycle. This species may be observedby in situ spectroscopy and it is referred to as the restingstate of the catalyst.

B. Scope of Homogeneous Catalysis

Homogeneous catalysis, by definition, refers to a catalyticsystem in which the substrates for a reaction and the cata-lyst components are brought together in one phase, mostoften the liquid phase. More recently a narrower defini-tion has become fashionable according to which homoge-neous catalysis involves organometallic complexes as thecatalysts (strictly speaking an organometallic compoundshould contain a bond between a carbon atom and themetal, but this is not true for all catalysts to be discussed, aswe will include coordination complexes in our definition).As a shorthand notation it will also be used for the presentcontribution, but it should be borne in mind that thereare many interesting and important reactions employinghomogeneous catalysts that are not organometallic or co-ordination complexes. Examples of such systems or cat-alysts are general acid and base catalysis (ester hydroly-sis), organic catalysts (thiazolium ions in Cannizzarro re-actions), enzymatic processes, nitrogen oxides, chlorideatoms, etc.

Organometallic catalysts consist of a central metal sur-rounded by organic (and inorganic) ligands. Both the metaland the large variety of ligands determine the properties

FIGURE 2 Selectivity in nickel catalyzed butadiene reactions.

of the catalyst. The success of organometallic catalystslies in the relative ease of catalyst modification by chang-ing the ligand environment. Crucial properties to be in-fluenced are the acceleration of the reaction and the se-lectivity to certain products. One metal can give a varietyof products from one single substrate simply by changingthe ligands around the metal center. Figure 2 shows theproducts that can be obtained from butadiene with nickelcatalysts.

Several types of selectivity can be distinguished in achemical reaction: chemoselectivity; regioselectivity; di-astereoselectivity; and enantioselectivity. See Fig. 3 for anexplanation. Note the difference between diastereoselec-tivity and enantioselectivity; the former starts with a chiralsubstrate and an achiral catalyst to produce diastereomers,while the latter starts with an achiral (often called prochi-ral) substrate and the optical activity is induced by a chiralcatalyst.

II. ELEMENTARY STEPS FORHOMOGENEOUS CATALYSIS

A. Creation of a “Vacant Site” andCoordination of the Substrate

To bring the reactants together a metal center must have avacant site. Metal catalysis begins, we could say, with thecreation of a vacant site. In the condensed phase solventmolecules will always be coordinated to the metal ion and“vacant site” is an inaccurate description. Substrates arepresent in excess and so may be the ligands. Therefore, acompetition in complex formation exists between the de-sired substrate and other potential ligands present in the so-lution. Often a negative order in one of the concentrationsof the “ligands” present can be found in the expression for




FIGURE 3 Selectivities involved in chemical conversions.

the rate of product formation. When the substrate coordi-nates strongly to the metal center this may give rise to azeroth order in the concentration of the substrate, i.e., sat-uration kinetics (c.f. Michaelis-Menten kinetics as knownfrom enzyme kinetics). Also, strong coordination of theproduct of the reaction may slow down or inhibit the cat-alytic process. These phenomena are similar to desorptionand adsorption in heterogeneous catalysis. Two mecha-nisms are distinguished for ligand/substrate displacement,an associative and a dissociative one; see Fig. 4.

B. Insertion and Migration Reactions

Insertion and migration refer to the process in which an un-saturated molecule inserts to a metal-anion bond. Figure 5

FIGURE 4 Dissociative and associative processes for ligandexchange.

shows how a new carbon-carbon is made when a carbonmonoxide molecule inserts into a platinum methyl bond,or more accurately, when the anionic methyl group “mi-grates” to the unsaturated CO molecule. The empty siteleft by the methyl group will be occupied by another lig-and, a solvent molecule in this figure.

A second important migration reaction involves alkenesinstead of carbon monoxide. Figure 6 gives a schematicrepresentation of a hydride that migrates to a coordinatedethene molecule cis to the hydride. The figure shows thehydride migration resulting in an empty space in the coor-dination sphere of the metal. This coordinative unsatura-tion can be lifted in two ways: first an agostic interactionwith the β-hydrogens may occur and secondly an incom-ing ligand may occupy the vacant site.

The migration reaction of hydrides to alkenes can bedescribed as a 2 + 2 addition reaction. The reaction takesplace in a syn fashion with respect to the alkene; the twoatoms M and H add to the same face of the alkene (Fig. 7).

C. β-Elimination and De-Insertion

The reverse reaction of the migration of η1-bonded anionicgroups to coordinated alkenes is named β-elimination

FIGURE 5 Migration reaction.




FIGURE 6 Hydride migration to ethene.

(Fig. 8). The migration reaction diminishes the total elec-tron count of the complex by two, and creates formallya vacant site at the metal; β-elimination does the oppo-site. β-Elimination requires a vacant site at the complex(neglecting solvent coordination) and during the processthe electron count of the complex increases by two elec-trons. The reaction resembles the β-elimination occurringin many organic reactions, but the difference lies in theintramolecular nature of the present process, as the elim-inated alkene may be retained in the complex. In organicchemistry the reaction may well be a two-step process,e.g., proton elimination with a base followed by the leavingof the anion. In transition metal chemistry the availabilityof d-orbitals facilitates a concerted cis β-elimination.

Instead of β-elimination one will also find the terms de-insertion and extrusion, especially for CO. The process iscompletely analogous because

1. a vacant site is needed for the reaction to occur;2. the electron count of the metal increases with two

during the de-insertion (provided that the carbonmonoxide remains coordinated to the metal andsolvent coordination is neglected).

D. Oxidative Addition

In an oxidative addition reaction a compound XY addsto a metal complex during which the XY bond is brokenand two new bonds are formed, MX and MY. X and Y arereduced and will at least formally have a minus one chargeand hence the formal valency of the metal is raised by two.The coordination number of the metal increases by two.The electron count around the metal complex increasesby two, but the d-electron count of the metal decreases bytwo. The 16-electron square planar complex is converted

FIGURE 7 Syn additin of metal hydride to alkene.

FIGURE 8 β-hydride elimination.

into an octahedral 18-electron complex. In Fig. 9 we havedepicted the oxidative addition of methyl iodide to Vaska’scomplex (L = phosphine).

The oxidative addition of acids to zerovalent metalssuch as nickel is also an instructive example. It resem-bles the reactions with alkyl halides and may result in an“amphoteric” hydride:

L4Ni + H+ → L4HNi+

The starting material is an 18-electron nickel zero complexwhich is protonated forming a divalent nickel hydride.This can react further with alkenes to give alkyl groups,but it also reacts as an acid with hard bases to regeneratethe nickel zero complex.

The oxidative addition of alkyl halides can proceed indifferent ways, although the result is usually a trans addi-tion independent of the mechanism. In certain cases thereaction proceeds as an SN2 reaction as in organic chem-istry. That is to say that the electron-rich metal nucleophileattacks the carbon atom of the alkyl halide, the halide be-ing the leaving group. This process leads to inversion ofthe stereochemistry of the carbon atom (only when thecarbon atom is asymmetric this can be observed). Thereare also examples in which racemization occurs. This hasbeen explained on the basis of a radical chain mechanism.The reaction sequence for the radical chain process readsas follows:

LnM + R• → LnRM•LnRM• + RX → LnRMX + R•

Oxidative additions involving C-H bond breaking is a topicof interest, usually referred to as C-H activation; the ideais that the M-H and M–hydrocarbyl bonds formed willbe much more prone to functionalization than the unre-active C-H bond. Intramolecular oxidative additions ofC-H bonds have been known for quite some time, seeFig. 10. This process is named orthometalation. It occursfrequently in metal complexes, and is not restricted to “or-tho” protons.

Oxidative addition reactions involving carbon-to-oxygen bonds are of relevance to the catalysis with

FIGURE 9 Oxidative addition of methyl iodide to Ir(I).




FIGURE 10 Oxidative addition of a carbon-hydrogen bond.

palladium complexes. The most reactive carbon-oxygenbond is that between allylic fragments and carboxylates.The reaction starts with a palladium zero complex and theproduct is a π -allylic palladium(II) carboxylate (Fig. 11).

The point of interest is the “amphoteric” character of theallyl anion in this complex. On the one hand it may reactas an anion, but on the other hand it is susceptible to nu-cleophilic attack by, for example, carbon centered anions.This has found widespread use in organic synthesis. Thereaction with the anion releases a zerovalent palladiumcomplex and in this manner palladium can be employedas a catalyst.

E. Reductive Elimination

Reductive elimination is simply the reverse reaction ofoxidative addition; the formal valence state of the metalis reduced by two (or one, in a bimetallic reaction) andthe total electron count of the complex is reduced by two.While oxidative addition can also be observed for maingroup elements, this reaction is more typical of the transi-tion elements in particular the electronegative, noble met-als. In a catalytic cycle the two reactions occur pairwise.At one stage the oxidative addition occurs, followed by,e.g., insertion reactions, and then the cycle is completedby a reductive elimination of the product.

Reductive elimination of molecules with carbon-carbonbonds has no counterpart in oxidative addition reactions.First, because the metal-carbon bonds energies may notalways be large enough to compensate for the energyof the carbon-carbon bond, and secondly the carbon-carbon bond is much less reactive than a carbon-hydrogenbond or a dihydrogen bond due to repulsive interactions.In organic synthesis the palladium- or nickel-catalyzedcross-coupling presents a very common example involv-ing oxidative addition/reductive elimination. A simplifiedscheme is shown in Fig. 12. The third step shows the re-ductive elimination.

FIGURE 11 Oxidative addition of allyl acetate to Pd(0).

FIGURE 12 Palladium-catalyzed cross coupling.

F. α-Elimination Reactions

α-Elimination reactions have been the subject of muchstudy since the mid-1970s mainly due to the pioneeringwork of Schrock. The Early Transition Metals are mostprone to α-elimination, but the number of examples ofthe later elements is growing. A classic example is shownin Fig. 13. The sequence of elementary steps, the natureof the reagents, and the reaction conditions are pertinentto the success of such reactions, but these details do notconcern us here. Dimethyl complexes of many metals leadto formation of methane via an α-elimination process, butoften the putative metal-alkylidene species is too reactiveto be isolated.

Metal alkylidene complexes find application in themetathesis of alkenes, the cyclopropanation of alkenes,Wittig-type reactions, and the McMurry reaction. In suit-able complexes α-elimination can occur twice yieldingalkylidyne complexes. Fig. 14 shows a schematic exam-ple for tungsten. Alkylidyne complexes can be used ascatalysts for the metathesis of alkynes.

G. Cyclometallation

Cyclometallation refers to a process of unsaturated moi-eties forming a metallacyclic compound. Examples ofthe process are presented in Fig. 15. Metal complexesrevealing these reactions comprise: M = L2Ni for reac-tion a, M = Cp2Ti for reactions b and c, M = Ta for d,and M = (RO)3W for e. The latter examples involvingmetal-to-carbon multiple bonds have only been observedfor early transition metal complexes, the same ones men-tioned under α-elimination, Section F.

The reverse reaction of a cyclometallation is of impor-tance for the construction of catalytic cycles. The retro-cyclometallations of reactions a and b are not productive,unless the structures were obtained via another route. Forc–e the following retro reactions can be envisaged leadingto new products; see Figs. 16 and 17.

FIGURE 13 α-elimination leading to a metal alkylidne complex.




FIGURE 14 α-elimination leading to a metal alkylidyne complex.

H. Activation of a Substrate TowardNucleophilic Attack

1. Alkenes

Coordination of an alkene to an electronegative metal ac-tivates the metal toward attack of nucleophiles. After thenucleophilic attack the alkene complex has been convertedinto a σ -bonded alkyl complex with the nucleophile at theβ-position. With respect to the alkene (in the “organic”terminology) the alkene has undergone anti-addition ofM and the nucleophile Nu; see Fig. 18.

As indicated under section B the overall result is thesame as that of an insertion reaction, the difference beingthat insertion gives rise to a syn-addition and nucleophilicattack to an anti-addition. In Fig. 18 the depicted nucle-ophile is anionic, but Nu may also be a neutral nucleophilesuch as an amine. The addition reaction of this type is thekeystep in the Wacker-type processes catalyzed by palla-dium.

2. Carbon Monoxide

Coordinated carbon monoxide is subject to activation to-ward nucleophilic attack. Through σ -donation and π -backdonation into the antibonding CO π∗ orbitals the carbonatom has obtained a positive character. This makes the car-bon atom not only more susceptible for a migrating anionat the metal center, but also for a nucleophile attackingfrom outside the coordination sphere. In this instance it ismore difficult to differentiate between the two pathways.The nucleophilic attack by alkoxides, amines, and wateris of great interest to homogeneous catalysis. A dominantreaction in syn-gas systems is the conversion of carbonyls

FIGURE 15 Schematic overview of cyclometallation reactions.

FIGURE 16 Retrocyclometallation reactions.

with water to metal hydrides and carbon dioxide (“shiftreaction”); see Fig. 19.

I. σ-Bond Metathesis

A reaction that is relatively new and not mentioned inolder textbooks is the so-called σ -bond metathesis. It isa concerted 2 + 2 reaction immediately followed by itsretrograde reaction giving metathesis. Both late and earlytransition metal alkyls are prone to this reaction, but itsoccurrence had to be particularly invoked in the case ofthe early transition metals; the latter often lack d-electronsneeded for other mechanisms such as oxidative addition.This alternative does not exist for d0 complexes such asSc(III), Ti(IV), Ta(V), W(VI), etc., and in such casesσ -bond metathesis is the most plausible mechanism. InFig. 20 the reaction has been depicted.

J. Dihydrogen Activation

In the above reactions we have not explicitly touched uponthe reactions of dihydrogen and transition metal com-plexes. Here the reactions that involve the activation ofdihydrogen will be summarized, because they are verycommon in homogeneous catalysis and because a compar-ison of the various mechanisms involved may be useful.Four reactions are usually distinguished for hydrogen:

1. oxidative addition (see Fig. 9),2. bimetallic oxidative addition,3. heterolytic cleavage, and4. σ -bond metathesis (see Fig. 20).

1. Oxidative Addition

Oxidative addition of dihydrogen commonly involvestransformation of a d8 square planar metal complex intoa d6 octahedral metal complex, or similar transformations

FIGURE 17 Cyclization through reductive elimination.




FIGURE 18 Nucleophilic anti-attack.

involving d2 → d0, d10 → d8, etc. The oxidative additionof dihydrogen to low-valent metal complexes is a com-mon reaction in many catalytic cycles. In spite of the highstrength of the dihydrogen bond the reaction proceedssmoothly to afford cis dihydrido complexes. The bond en-ergy of a metal hydrogen bond is in the order of 240 +/−40 kJmol−1 which is sufficient to compensate for the lossof the H-H bond (436 kJmol−1). The hydride is formallycharged with a minus one charge and this electron countgives dihydrogen the role of an oxidizing agent. The clas-sic example of oxidative addition to a d8 metal complexis the reaction discovered by Vaska and Diluzio:

trans-IrCl(CO)(PPh3)2 + H2 → IrH2Cl(CO)(PPh3)2

In rhodium complexes the reaction has found widespreadapplication in hydrogenation. In model compounds thereaction reads:

Rh(diphosphine)+2 + H2 → RhH2(diphosphine)+2The second reaction, bimetallic oxidative addition—

also referred to as homolytic splitting—involves a reac-tion of a dimeric complex with H2 in which two metalcenters participate. For instance, a dimer of a d7 metalcomplex reacts with dihydrogen to give two d6 species.In this process dihydrogen also turns formally into twohydride anions. A well-known example is the conversionof dicobaltoctacarbonyl into hydridocobalttetracarbonyl:

Co2(CO)8 + H2 → 2 HCo(CO)4

Another example implies two molecules of a Co(II), a d7

complex, which does suggest that radical reactions may be

FIGURE 19 Nucleophilic attack at coordinated CO followed by a“shift” reaction.

FIGURE 20 σ -Bond metathesis.

occurring, such as in Co(CN)−35 . Indeed in the subsequent

reactions the proton is transferred as a hydrogen radicalto the activated alkenes (e.g., acrylates, styrene) and thereference to the process as a homolytic dissociation is cer-tainly applicable to the reactions with alkenes. Formally,dihydrogen is transformed into two hydride anions and thereaction is again an oxidative addition.

2 Co(CN)−35 + H2 → 2 HCo(CN)−3

5

2. Heterolytic Splitting

Heterolytic splitting of dihydrogen refers to the splittingof dihydrogen into a proton and a metal-bonded hydride. Itis a common reaction, probably for many transition metalcations, but there are only a few cases for which there isclear proof for its occurrence. In the ideal case the het-erolytic splitting is catalyzed by the metal ion and a basewhich assists in the abstraction of the proton. In this re-action there is no formal change in the oxidation state ofthe metal. The mechanism has been put forward for Ru(II)complexes which can react with dihydrogen according to:

RuCl2(PPh3)3 + H2 → RuHCl(PPh3)3 + HCl

Ruthenium has a sufficient number of d-electrons to al-low for oxidative addition of dihydrogen, which couldthen be quickly followed by reductive elimination ofHCl. Observations on dihydrogen complexes of rutheniumhave thrown light on the mechanism heterolytic splitting.CpRu(L)((η2-H2)+ reacts rapidly with NEt3 as can be de-duced from the dynamic 1H-NMR spectra which indicatea rapid exchange of the dihydrogen complex with its con-jugate base, CpRu(L)(H) (Fig. 21).

K. Activation by Lewis Acids

In Section H we have discussed the activation of carboncontaining fragments toward nucleophilic attack by coor-dination of the fragment to a transition metal. Here we will

FIGURE 21 η2-H2 complex and heterolytic splitting.




FIGURE 22 Lewis acid catalyzed Diels-Alder reaction.

describe a few examples of activation of reagents whencomplexed to Lewis acids. In organic textbooks one willfind a variety of reactions catalyzed by Lewis acids.

1. Diels-Alder Additions

The Diels-Alder reaction is the reaction of a diene with amono-ene to form a cyclohexene derivative, an importantreaction for the construction of organic intermediates. Oneof its attractions is the atom efficiency of 100% with noby-products being formed. The mono-ene, or dienophilewhich may also be an alkyne, has a LUMO of low energywhile the diene is usually electron rich with a high-lyingHOMO. The interaction of these two orbitals starts thereaction between the two molecules (Fig. 22).

The reaction can be accelerated by complexation of thedienophile to a Lewis acid, which further lowers the levelof the interacting LUMO: two regioisomers are formedin the example shown, we shall not consider the stereo-chemistry. Typically in these systems governed by frontierorbitals, the reactions not only become much faster witha Lewis acid catalyst, but also more regioselective, that isto say the pathways to different isomers may experiencedifferent accelerations.

2. Epoxidation

Alkenes can be transformed into epoxides by hydroperox-ides and a catalyst, which often is a high-valent titanium ormolybdenum complex acting as a Lewis acid. The mech-anism is not clear in great detail; in Fig. 23 a suggestedmechanism is given. The key factor is the action of themetal on the peroxo group making one oxygen atom elec-trophilic.

FIGURE 23 Epoxidation catalyzed by Lewis acids.

FIGURE 24 Wilkinson’s hydrogenation cycle.

3. Ester Condensation

An application of industrial importance of Lewis acidicmetal salts is the condensation of carboxylic diacids anddiols to give polyesters. This is an acid catalyzed reactionthat in the laboratory is usually catalyzed by protic acids.For this industrial application salts of manganese, nickel,or cobalt and the like are used.

III. DESCRIPTION AND MECHANISMSOF HOMOGENEOUS CATALYTICPROCESSES

A. Hydrogenation

1. Wilkinson’s Catalyst

Undoubtedly the most popular homogeneous catalyst forhydrogenation is Wilkinson’s catalyst, RhCl(PPh3)3, dis-covered in the 1960s. The reaction mechanism, its depen-dence on many parameters, and its scope have been studiedin considerable detail.

The catalytic cycle is shown in Fig. 24. Note that thereactions have been drawn as irreversible reactions whilemost of them are actually equilibria. In this scheme Lstands for triarylphosphines and S for solvent (ethanol,toluene). The alkene is simply ethene. For simplicity we




FIGURE 25 Re- and si-faces on an alkene substituted at oneatom.

have chosen Wilkinson’s complex as the starting point inour cycle. As in many other cases to follow the numberof valence electrons switches during the cycle betweentwo numbers differing by two electrons; in this instancethe valence electron counts switch between 16 and 18. A14-electron count for the unsaturated species occurring atthe beginning of the cycle has also been discussed. Thefirst step in this sequence is the dissociation of one ligandL that is replaced by a solvent molecule.

After ligand dissociation an oxidative addition reac-tion of dihydrogen takes place. As usual this occurs incis fashion and can be promoted by the substitution ofmore electron-rich phosphines on the rhodium complex.The next step is the migration of hydride forming the ethylgroup. Reductive elimination of ethane completes the cy-cle. Obviously, employing electron-withdrawing ligandscan increase the rate of this step. Within a narrow windowof aryl phosphines small changes in rates have been ob-served which could indeed be explained along the linessketched above. Strong donor ligands, however, stabilizethe trivalent rhodium(III) chloro dihydride to such an ex-tent that the complexes are no longer active.

2. Asymmetric Hydrogenation

a. Introduction “prochirality.” Planar moleculespossessing a double bond such as alkenes, imines, andketones, which do not contain a chiral carbon in one ofthe side chains, are not chiral. When these moleculescoordinate to a metal a chiral complex is formed, unlessthe alkene possesses C2V symmetry. A simple silvercation Ag+ suffices. In other words, even a simple alkenesuch as propene will form a chiral complex when it

FIGURE 26 Synthesis of L-dopa (oAn = ortho-anisyl = 2-methoxyphenyl).

coordinates to a transition metal, so will trans-2-butene,but cis-2-butene won’t. If a bare metal atom coordinatesto cis-2-butene the complex has a mirror plane, andhence the complex is not chiral. In summary, with fewexceptions, complexation of a metal to the one face of analkene gives rise to a certain enantiomer and complexationto the other face gives rise to the other enantiomer.

For complexation of planar molecules to metals ruleshave been introduced to allow us to denote the faces ofthe planar molecule; they are called the re-face and thesi-face. Usually this simple annotation takes into accountthat only one carbon atom is used. It may be more compli-cated when the two carbon atoms of the alkene give riseto two stereocenters. In Fig. 25 we have drawn how wecan distinguish the two faces of a simple alkene, or ratherthe side of attack of a specific atom of the alkene.

When a metal complex is chiral, either because it con-tains a chiral ligand or a chiral metal center, and then formsa complex to a “prochiral alkene,” the resulting complexis a diastereoisomer. Thus, a mixture of diastereomers canform when the chiral complex coordinates to each faceof the alkene. As usual, these diastereomers have differ-ent properties and can be separated. Or, more interestinglyfor catalysis, the two diastereomers are formed in differentamounts.

b. Enantioselective hydrogenation. The asymmet-ric hydrogenation of cinnamic acid derivatives has beendeveloped by Knowles at Monsanto. The synthesis of L-dopa (Fig. 26), a drug for the treatment of Parkinson’sdisease, has been developed and applied on an industrialscale. The reaction is carried out with a cationic rhodiumcomplex and an asymmetric diphosphine as the ligandthat induces the enantioselectivity. Surprisingly, the reac-tion is not very sensitive to the type of diphosphine used,although it must be added that most ligands tested arebis(diphenylphosphino) derivatives. On the other hand,the reaction is very sensitive to the type of substrateand the polar substituents are prerequisites for a success-ful asymmetric hydrogenation. Fig. 26 shows the overallreaction.




FIGURE 27 Diastereomeric alkene complexes.

The hydrogenation reaction is carried out with a substi-tuted cinnamic acid. The acetamido group is of particularimportance because it functions as a secondary complexa-tion function in addition to the alkene functionality. In thefirst step the alkene coordinates to the cationic rhodiumspecies for which there are two possibilities, binding to there-face and the si-face (Fig. 27). This new chiral center,combined with the ligand chirality, leads to the formationof two diastereomers which have different free energiesand reactivities. Thus, a preference for just one enantiospe-cific pathway may result. For a few systems in situ char-acterizations have been carried out and both complexeshave been identified. The final product is not necessarilyderived from the most abundant isomer; all that mattersis through which pathway does the fastest reaction takeplace that produces the most product.

The formation of the re and si adducts is reversible andthus in this step no chirality is determined yet. As soonas an irreversible step occurs—after chirality has beenintroduced—the chirality of the final product has beendetermined. In the present reaction this step may be eitherthe oxidative of addition of H2 or the migratory insertion(Fig. 28).

A large series of asymmetric ligands have been devel-oped most of which have the asymmetric “center” in the

FIGURE 28 Asymmetric hydrogenation.

FIGURE 29 C2-Chiral diphosphines.

bridge rather than at the phosphorus atoms as is the casein DIPAMP, the ligand shown in Fig. 28 for the L-dopasynthesis. Their synthesis is similar to the one developedby Kagan for DIOP (see Fig. 29) starting from asymmetricacids which can be commercially obtained (tartaric acidin the case of DIOP). Crucial to the success of this familyof ligands is the C2 “propeller”-type symmetry which di-vides the space around rhodium (or any metal) into fourquadrants, two relatively empty and two filled ones (seeFig. 29).

Two phenyl groups at one phosphorus of the chelat-ing ring adopt an axial and an equatorial position. Thisexplains the similarity of this large family of ligands inthe enantioselective reactions. Coordination of the dehy-droalanine derivative (or enamides in general) will nowtake place in such a way that the auxiliary donor atomcoordinates to one site, and the phenyl-substituted alkenewill coordinate to the other site with the face that gives theleast interaction of its substituents with the phenyl groupof the ligand pointing into the same quadrant. This givesthe predominant metal alkene adduct.

Three other ways of achieving chirality in phosphinesshould be mentioned. The first one also concerns C2 chiralligands of the type DuPHOS (Fig. 30) developed by Burk.




FIGURE 30 Chiral diphosphine ligands.

The chirality resides neither on the backbone nor on thephosphorus atom, but on the cyclic substituents. Modifi-cation of the R-groups leads to highly efficient catalysts.C1-chiral ligands can also be very selective as has beenshown by Togni who developed a range of ligands, hereexemplified by Josiphos. The ligand contains “two” chiralcenters, one at the carbon atom and the other involves fa-cial chirality of the cyclopentadienyl plane. Derivatives ofthis type are used for the commercial production of chi-ral pharmaceuticals and agrochemicals. Another way toachieve chirality is by hindered rotation around molecularaxes (“axial” chirality) as will be shown in the next sectionwhere BINAP will be introduced. Application DIOP?

B. Isomerization

1. Insertion and β-Elimination

A catalytic cycle which involves only one type of ele-mentary reaction must be a very facile process. Isomer-ization is such a process since only migratory insertionand its counterpart β-hydride elimination are required.Hence the metal complex can be optimized to do exactlythis reaction as fast as possible. The actual situation isslightly more complex due to the necessity of vacant sites,which have to be created for alkene complexation and forβ-elimination. As expected, many unsaturated transitionmetal hydride complexes catalyze isomerization. Exam-ples include monohydrides of Rh(I), Pd(II), Ni(II), Pt(II),and Zr(IV). The general scheme for alkene isomerizationis very simple; for instance it may read as shown in Fig. 31.

FIGURE 31 Simplified isomerization scheme involving β-hydrideelimination.

FIGURE 32 Simplified allylic isomerization scheme.

2. Allylic Mechanism

A second mechanism that has been brought forward in-volves the formation of allylic intermediates. The pres-ence of a hydride on the metal complex is not required inthis mechanism which can best be described as an oxida-tive addition of an “activated” C-H bond (i.e. an allylichydrogen) to the metal. The allyl group can recollect itshydrogen at the other end of the allyl group and the resultis also a 1,3 shift of hydrogen (Fig. 32).

3. Asymmetric Isomerization

An important application of an isomerization is found inthe Takasago process for the commercial production of(−)menthol from myrcene. The catalyst used is a rhodiumcomplex of BINAP. The BINAP complex is an asym-metric ligand based on the atropisomerism of substituteddinaphthyl (Fig. 33). It was first introduced by Noyori.Atropisomers of diphenyl and the like are formed whenortho substituents do not allow rotation around the cen-tral carbon-carbon bond. As a result two enantiomers areformed.

For asymmetric hydrogenation, transfer hydrogenation,and isomerization of double bonds using both rutheniumand rhodium complexes BINAP has been extensivelyused. The synthesis of menthol is given in the reactionscheme, Fig. 34. The key reaction is the enantioselectiveisomerization of the allylamine to the asymmetric enam-ine. It is proposed that this reaction proceeds via an allylicintermediate.

This is the only step that needs to be steered to the cor-rect enantiomer, since the other two are produced in the

FIGURE 33 The two enantiomers of BINAP.




FIGURE 34 The Takasago process for (−)menthol.

desired stereochemistry with the route depicted. Of theeight possible isomers only this one (1R,3R,4S) is impor-tant. After the enantioselective isomerization the enamineis hydrolyzed. A Lewis acid catalyzed ring closure givesthe menthol skeleton. In a subsequent step the isopropenylgroup is hydrogenated over a heterogeneous Raney nickelcatalyst. Asymmetric catalysis involving metal-catalyzedhydrogenations and isomerization is becoming increas-ingly important in the production of pharmaceuticals,agrochemicals, and flavors and fragrances.

C. Carbonylation of Methanol

1. Monsanto Acetic Acid Process

The carbonylation of methanol has been developed byMonsanto in the late 1960s. It is a large-scale operationemploying a rhodium/iodide catalyst converting methanoland carbon monoxide into acetic acid. Since the 1990s aniridium/iodide-based catalyst has been used by BP.

CH3OH + CO → CH3COOH

�G, standard conditions, −17.8 kcal/mol

The same carbonylation reaction can be carried out witha cobalt catalyst. The rhodium and iridium catalysts haveseveral distinct advantages over the cobalt catalyst; theyare much faster and far more selective. The higher rate isin process terms translated into much lower pressures (thecobalt catalyst is operated at pressures of 700 bar).

For years the Monsanto process has been the most at-tractive route for the preparation of acetic acid. The twocomponents are rhodium and iodide, which can be addedin many forms. A large excess of iodide may be present,

and both on a weight basis as well as on a molecularbasis, it would be fair to say that the catalyst is iodideand the promoter is a trace of rhodium. Methanol andthe iodide component form under the reaction conditionsmethyl iodide. Rhodium is present as the anionic speciesRhI2(CO)−2 . The first step of the catalytic cycle (seeFig. 35) is the oxidative addition of methyl iodide to thisrhodium complex. Migration of the methyl group gives anacetyl rhodium complex. CO complexation and reductiveelimination of acetyl iodide complete the cycle. Acetyl io-dide hydrolyzes to give acetic acid and hydrogen iodide.The latter regenerates methyl iodide in a reaction withmethanol.

It is important that in the two “organic” equilibriainvolving iodide the one with methanol involves com-plete conversion to methyl iodide, whereas acetyl iodideis completely converted into acetic acid and hydrogeniodide:

CH3OH + HI −→ H2O + CH3I

CH3COI + H2O −→ CH3CO2H + HI

The rate-determining step in this process is the oxidativeaddition of methyl iodide. The iodide enables the forma-tion of a methyl rhodium complex; methanol is not suf-ficiently electrophilic to carry out this reaction. Withinthe operation window of the process the reaction rate isindependent of the carbon monoxide pressure. Further-more, the methyl iodide formation from methanol is al-most complete, which makes the reaction rate also prac-tically independent of the methanol concentration. As aresult of the kinetics and the equilibria mentioned above,all iodide in the system occurs as methyl iodide. Hence,up to high conversions the rate does not depend on the




FIGURE 35 Monsanto carbonylation of methanol.

concentrations of the two reactants. The process schemeis presented in Fig. 36.

2. Acetic Anhydride

In many applications acetic acid is used as the anhydrideand the synthesis of the latter is therefore equally impor-tant. In the 1970s Halcon (now Eastman) and Hoechst(now Celanese) have developed a process for the con-version of methyl acetate and carbon monoxide to aceticanhydride. The process has been on stream since 1983and several 100,000 are being produced annually togetherwith some 10–20% acetic acid. The reaction is carried outunder similar conditions as the Monsanto process, andalso uses methyl iodide as the “activator” for the methylgroup. The reaction scheme follows that of the Monsantoprocess except for the “organic” cycle, in which acetic

FIGURE 36 Simplified flow-scheme of the Monsanto process foracetic acid.

acid replaces water, and methyl acetate replaces methanol(Fig. 35):

CH3COOCH3 + HI ⇀↽ CH3COOH + CH3I

CH3COI + CH3COOH ⇀↽ CH3COOOCH3 + HI

The first reaction generates methyl iodide for the oxidativeaddition, and the second reaction converts the reductiveelimination product acetyl iodide into the product and itregenerates hydrogen iodide. There are, however, a fewdistinct differences between the two processes. The ther-modynamics of the acetic anhydride formation are lessfavorable and the process is operated much closer to equi-librium. Under standard conditions the �G values are ap-proximately:

MeOAc + CO → Ac2O −2.5 kcal/mol

Two more differences are

1. in the Eastman process 5% of H2 is added to thecarbon monoxide to accomplish reduction of trivalentrhodium, and

2. the addition of cations such as Li+ or Na+ isnecessary.

D. Hydroformylation

1. Introduction

Functionalization of hydrocarbons from petroleumsources is mainly concerned with the introduction ofoxygen into the molecule. Roughly speaking, two waysare open: oxidation and carbonylation. Oxidation is thepreferred route inter alia for aromatic acids, acrolein,maleic anhydride, ethene oxide, propene oxide, andacetaldehyde. Hydroformylation (older literature and




technical literature say “oxo” reaction) is employed forthe large-scale preparation of butanal and butanol, 2-ethylhexanol, and detergent alcohols. Butanal and butanolare used in many applications as a solvent, in esters,in polymers, etc. The main use of 2-ethylhexanol is inphthalate esters, which are softeners (plasticizers) inPVC. The catalysts applied are based, again, on cobaltand rhodium.

2. Cobalt-Based Oxo-Process

Roelen accidentally discovered the hydroformylation ofalkenes in the late 1930s while he was studying the con-version of synthesis gas to liquid fuels (Fischer-Tropschreaction) using a heterogeneous cobalt catalyst. It tookmore than a decade before the reaction was taken further,but now it was the conversion of petrochemical hydrocar-bons into oxygenates that was the driving force. It wasdiscovered that the reaction was not catalyzed by the sup-ported cobalt but, in fact, by HCo(CO)4 formed in theliquid state.

A key issue in the hydroformylation reaction is the ra-tio of “normal” (linear) and “iso” (branched) product be-ing produced. Figure 37 explains this colloquial expres-sion. The linear (“normal”) product is the desired product;the value of butanal is higher because this is the productwhich can be converted to 2-ethyl hexanol via a base-catalyzed aldol condensation and a hydrogenation. The de-tergent alcohols should be preferably linear because theirbiodegradability was reported to be better than that of thebranched product. The linearity obtained in the cobalt-catalyzed process is 60–80%. The reaction mechanismfor cobalt is similar to that of rhodium, which will be dis-cussed in the next section.

Thermodynamics of hydroformylation and hydrogena-tion at standard conditions are as follows:

H2 + CH3CH=CH2 + CO → CH3CH2CH2C(O)H

�G 15 −33 −28(l) = −10 kcal/mol

�H 5 −26 −57 = −36 kcal/mol

H2+ CH3CH=CH2 → CH3CH2CH3

�G 15 −6 = −21 kcal/mol

�H 5 −25 = −30 kcal/mol

FIGURE 37 The hydroformylation reaction.

Thus the reaction is highly exothermic and favored bythermodynamics at temperatures roughly below 200◦C.Hydrogenation of the alkene to alkane is thermodynami-cally even more attractive. Often this reaction is observedas a side reaction.

In industrial practice the older cobalt catalyst is stillused today for the conversion of higher alkenes to deter-gent aldehydes or alcohols (>C12). The cobalt processrequires high pressures (70–100 bar) and temperatures(140–170◦C). Aldol condensation and hydrogenation ofthe alkene to alkane (∼10%) are undesirable side reac-tions for the detergent alcohols. Interestingly, the cheaperinternal alkenes can be used for this process and yet theoutcome is mainly a terminally hydroformylated, linearaldehyde. Often the separation of catalyst, product, by-product, and starting material is tedious. For propenethe most economic processes are rhodium-based catalystscommercialized in the 1970s.

3. Rhodium-Based Hydroformylation

Fundamental work by Nobel laureate Wilkinson demon-strated that rhodium triphenylphosphine catalysts allowedthe operation of the hydroformylation reaction at muchlower pressure (1 bar was reported by Wilkinson) andtemperature than the cobalt process. The selectivity wasalso reported to be considerably higher, virtually no hy-drogenation was observed and the linearity was in somecases as high as 95%. The rhodium catalysts were re-ported to be three orders of magnitude faster in rate. Theresulting milder reaction conditions would give much lesscondensation products. In 1971 Union Carbide Corpo-ration, Johnson and Matthey, and Davy Powergas (nowKvearner) joined forces to develop a process based onthis new finding. As yet it is only applied for propene.Hydroformylation of propene is of prime importance andworldwide probably more than 7 million tons of butanalare produced this way annually.

A convenient catalyst precursor is RhH(CO)(PPh3)3.Under ambient conditions it will slowly convert 1-alkenesinto the expected aldehydes. Internal alkenes exhibithardly any reaction. At higher temperatures (100–120◦C)pressures of 10–30 bar are required. Unless a large excessof ligand is present the catalyst will also show some iso-merization activity, but the internal alkenes thus formed




will not be hydroformylated. The 2-alkene concentrationwill increase and the 1-alkene concentration will decrease;this will slow down the rate of the hydroformylation re-action. This makes the rhodium catalyst less suited forthe conversion of alkenes other than propene for whichisomerization is irrelevant. To date, hydroformylation ofhigher alkenes is industrially still carried out with cobaltcatalysts.

Propene hydroformylation can be done yielding a lin-earity ranging from 60 to 95% dependent on the phosphineconcentration. At very high phosphine concentration therate is low, but the linearity achieves its maximum value.The commercial process operates presumably around 30bar of syn-gas, at 120◦C, at high phosphine concentra-tions, and linearities around 92%. The estimated turnoverfrequency of moles of product per mole of rhodium com-plex per hour is in the order of 300. Low ligand concentra-tions, with concomitant low linearities, will give turnoverfrequencies in the order of 10,000 at 10 bar and 90◦C.In the presence of carbon monoxide this rhodium cata-lyst has no activity for hydrogenation and the selectivitybased on starting material is virtually 100%. The n-butanalproduced contains no alcohol and can be converted bothto butanol, to 2-ethyl-hexanol-1, and to other products asdesired.

The most likely mechanism for the reaction is givenin Fig. 38. Note the trigonal bipyramidal structure (tbp)of the rhodium catalysts. The σ -bonded hydrido groupor the alkyl group are bound in an apical position of thetbp structure. This leaves three equatorial and one apical

FIGURE 38 Rhodium-catalyzed hydroformylation.

position for the remaining four ligands, carbon monox-ide, phosphorus ligands, and the alkene substrate. Asusual, the ligand replacement equilibria have been sim-plified. After an insertion step a 16-electron species isformed which restores its 18-electron count with a newCO molecule. There is consensus in the literature that thecatalytic species with two phosphine ligands plays a ma-jor role, certainly when L = PPh3. The high preference forlinear aldehyde formation is ascribed to steric factors—congestion at the rhodium center favors the formation oflinear alkyl and acyl species. Especially when two phos-phine ligands coordinate bis-equatorially, there is a strongpreference for the formation of n-alkyl groups. Later wewill see how one can make use of this concept to obtainhighly selective catalysts.

For rhodium-phosphine catalysts, the kinetics showsthat one of the first steps, complexation of alkene or migra-tory insertion of the alkene, is the rate-determining step. Afirst order in alkene is observed, while there is a negativedependence in the CO and/or phosphine concentration,which also stems from the replacement reaction. The rateof reaction is independent of the hydrogen concentration.

Another limiting case is the hydroformylation us-ing electron-poor rhodium catalysts (e.g., HRh(CO)4 orHRh(CO)3L, L being an electron-poor phosphite). In thisinstance the oxidative addition of dihydrogen is the slow-est step of the cycle and now the reaction shows a negativeorder in CO pressure, a first-order dependency in H2, andthe reaction rate is independent of alkene concentration,i.e., saturation kinetics with respect to alkene. Often theexpression is more complicated and does not reveal thepresence of a single slow step in the process.

4. Ligand Effects

Phosphine-based ligands have found widespread applica-tion not only in organometallic chemistry but also in indus-trial applications of homogeneous catalysis. Their stericand electronic properties have been widely studied to es-tablish structure-activity relationships. The σ -basicity andπ -acidity of phosphorus ligands can be compared by look-ing at the stretching frequencies of the coordinated carbonmonoxide ligands in complexes such as NiL(CO)3 whereL is the phosphorus ligand. Strong σ–donor ligands give ahigh electron density on the metal and hence a substantialback-donation to the CO ligands and lowered IR frequen-cies. Strong π–acceptor ligands will compete with CO forthe electron back-donation and the CO stretch frequencieswill remain high. The IR frequencies represent a reliableyardstick of the electronic properties of a series of phos-phine ligands toward a particular metal.

Thus, the electronic parameter for ligands with the samedonor atom can be fairly well measured and applied in a




FIGURE 39 Tolman’s cone angle.

qualitative sense. Many attempts have been undertaken todefine a reliable steric parameter complementary to theelectronic parameter. Most often used is Tolman’s param-eter θ (theta). Tolman proposed to measure the steric bulkof a phosphine ligand from CPK models in the followingway. From the metal center, located at a distance of 2.28A from the phosphorus atom in the appropriate direction,a cone is constructed which embraces all the atoms of thesubstituents on the phosphorus atom (see Fig. 39).

The cone angle is measured, and these cone angles θ(simply in degrees) are the desired steric parameters. Crys-tal structure determinations have shown that in practicethe angles realized in the structures are smaller than theθ -values would suggest. In reality, intermingling of theR-substituents leads to smaller effective cone angles.

Some typical values are:

Ligand PR3, R=: θ -value:H 87CH3O 107n-Bu 132PhO 128Ph 145i-Pr 160C6H11 170t-Bu 182

5. Ligand Effects in Hydroformylation

Ligand effects on the selectivity of the rhodium-catalyzedhydroformylation has been extensively studied. The ratemay vary several orders of magnitude and also the se-lectivity to either the branched or the linear product mayvary dramatically with minor changes in the electronic andsteric properties of the ligand.

6. Bidentates with Constrained Bite Angles

A few classes of new ligands have been introduced thatgive very high selectivity to linear products. All use theconcept that a bis-equatorial coordination of the phospho-rus ligands is needed to raise the selectivity to linear prod-uct. As mentioned above for triphenylphosphine, it was

FIGURE 40 Preferred coordination modes of Xantphos (left) anddppe.

thought that also in this instance this coordination modegave rise to the desired selectivity. The two structures areshown in Fig. 40.

Bidentate ligands have been made and used exten-sively in many studies on coordination compounds andorganometallic compounds. The majority of bidentate lig-ands, however, lead to “bite” angle of the bidentate on themetal between 75 and 95◦. If a bidentate is to coordinatein a bis-equatorial fashion, it should have a bite angle of∼120◦. Only very few ligands are available. Three of themare shown in Fig. 41. Molecular modeling and crystallo-graphic studies by Casey and van Leeuwen have learnedthat indeed the “natural” bite angle of these ligands is110–120◦. They give linear to branched product ratios upto 100.

7. Two-Phase Hydroformylation: Water-SolubleCatalysts

In the 1980s a new process has come on stream employinga two-phase system with rhodium in a water phase and thesubstrate and the product in an organic phase. The catalystused is a rhodium complex trisulfonated triphenylphos-phine (tppts) (Fig. 42) which is highly water-soluble (inthe order of 1 kg of the ligand “dissolves” in 1 kg of wa-ter). The ligand forms complexes with rhodium that arevery similar to the ordinary triphenylphosphine complexes(i.e., RhH(CO)(PPh3)3). The rhodium complex resides inthe aqueous phase. The substrates, propene/CO/H2, arevery slightly water-soluble. Upon reaction a second phaseforms, the organic phase, consisting of the product bu-tanal and dissolved propene and syn-gas. The two phasesare intensely stirred to ensure a fast transport of the gasesfrom the gas phase to the organic and water phase.

The two phases are removed from the reactor and sepa-rated in a settler tank after the heat and the gases have beenremoved. Often in organic synthesis the separation of twolayers is not such an easy process step, but in this particularinstance a very clean separation of the two layers occurs.Most importantly, the rhodium/ligand components remaincompletely in the aqueous phase. This type of separationis very attractive in a large-scale, continuous process. Forthe flow-scheme of the process, see Fig. 43.

Ruhrchemie has commercialized the process, after theinitial work had been done by workers at Rhone-Poulenc,




FIGURE 41 Ligands leading to high l/b ratios in rhodium-catalyzed hydroformylation.

for the production of butanal from propene. The linearityof the product amounts to 92%. The process is not appli-cable to the hydroformylation of higher alkenes because

� the isomerization cannot be completely suppressedwith aryl phosphines, thus leading to the formation of2-alkenes which cannot be separated from the1-alkenes in the recycle,

� and the solubility of higher alkenes in water is verylow.

E. Alkene Oligomerization

1. Introduction

1-Alkenes, or linear α-olefins as they are called in indus-try, are desirable starting materials for a variety of prod-ucts. Polymers and detergents are the largest end uses. Wemention a few applications:

C4 PolybutyleneC6−8 Comonomers in HDPE, LLDPE,

synthetic estersC6−10 Alcohols (hydroformylation) as

phthalates for PVC plasticizersC8−10 As trimers in synthetic lub-oilsC10−14 After hydroformylation, detergentsC14−16 Sulfates and sulfonates in detergents

Industrially, alkenes are obtained from several reactions,one being ethene oligomerization. Three processes are

FIGURE 42 Sulfonated triphenylphosphine (tppts).

available, two based on aluminum alkyl compounds orcatalysts and one on nickel catalysts.

2. Shell-Higher-Olefins-Process

Many transition metal hydrides will polymerize ethene topolymeric material or, alternatively, dimerize it to butene.Fine-tuning of these catalysts to one that will give a mix-ture of, for example, predominantly C10 to C20 oligomersis not at all trivial. Nickel complexes have been extensivelystudied by Wilke and his coworkers for their activity asalkene oligomerization catalysts. In the late 1960s Keimand coworkers at Shell discovered a homogeneous nickelcatalyst which selectively oligomerizes ethene to higherhomologs (Fig. 44).

a. Oligomerization. The catalyst is prepared in aprereactor from nickel salts with boron hydrides as the re-ductor under a pressure of ethene and then ligand is added.Polar solvents such as alcohols are used for the dissolu-tion of the catalyst. The catalyst solution and ethene areled to the reactor, a stirred autoclave, which is maintainedat 80–120◦C and 100 bar of ethene (Fig. 45).

b. Separation. The product alkenes are insoluble inthe alcohol and phase separation takes place. After set-tling, the alcohol layer goes to a regeneration unit. Thealkene layer is washed and ethene is recycled to the reac-tor. The products are distilled and the desired fractions arecollected.

c. Purification and metathesis. The lower alkenesand the heavy alkenes must be “disproportionated” to givethe full range of alkenes. To this end the light and heavyalkenes are sent to an isomerization reactor after havingpassed a purification bed, a simple absorbent to removealcohol and ligand impurities. The isomerized mixture isthen passed over a commercial molybdenum metathesiscatalyst (CoMox), also a fixed bed reactor, to give a broadmixture of internal alkenes. After distillation the C11−14




FIGURE 43 Flow-scheme of Ruhrchemie/Rhone-Poulenc process.

fraction (±15%) is used as a feedstock for alcohol pro-duction via cobalt catalysts. The light and heavy productsare recycled. A bleed stream of the heavy ends must limitthe buildup of branched products and polymers.

The total production of higher olefins via this and sim-ilar routes is estimated to be 2 million tons annually. Alarge part of the alkenes are produced for captive use.

3. Aluminum Process

Two aluminum-based processes are being used. The one-step process is operated at high temperature using alu-minum alkyls as the catalyst. This process resemblesthe nickel process discussed; as a matter of fact Zieglerand coworkers were working with this aluminum cata-lyst when they accidentally found the influence of nickel,which under these drastic conditions gave the unwantedbutenes.

The second aluminum-based process is a two-step pro-cess that employs aluminum stoichiometrically by eachpass, the alkyls are grown at a lower temperature, decom-posed in the next reactor at higher temperature, and thealuminum complexes are recycled. This procedure leads

FIGURE 44 SHOP catalyst.

to “peaking” in the C10–C16 range. The oligomers arenow formed according to a Poisson distribution, which isvery narrow compared with the Schulz-Flory distribution.When each initiator makes only one oligomer the Mw/Mn

approaches 1 (Poisson) whereas Mw/Mn approaches 2when chain transfer occurs (Schulz-Flory). In poly-mer synthesis these distributions also play an importantrole.

In the late 1990s a new group of catalysts was dis-covered comprising iron and cobalt complexes containingpyridinediimine ligands. Extremely fast catalysts were re-ported (Fig. 46). Turnover frequencies as high as severalmillions per hour were recorded. This represented a totallyunexpected development showing that a new combinationof ligand and metal can lead to surprises. Catalyst activi-ties are very high and a separation of catalyst and productmay not be needed.

F. Zirconium-Catalyzed Polymerizationof Alkenes

1. Synthesis

Polymers can be made by condensation reactions and byaddition polymerizations. An example of the former is thecondensation of diols and diesters forming polyesters; allmolecules are involved in the steady growth of speciesof higher molecular weight. Addition polymerization in-volves the reaction of initiating species with monomers,thus building up this limited number of polymer molecules




FIGURE 45 Simplified flow-scheme of SHOP.

in an excess of monomers. Four types of reaction can bedistinguished for organic monomers:

1. radical polymerization;2. anionic polymerization;3. coordination polymerization (e.g., Ziegler-Natta

polymerization); and4. cationic polymerization.

Transition metal complexes play the key role in coordina-tion polymerization.

2. Introduction to Ziegler-Natta Polymerization

The titanium catalyst for the stereoselective polymeriza-tion of propene to isotactic polymer was discovered byZiegler and Natta in the mid-1950s. Soon after its discov-ery it was commercially exploited and even today isotacticpolypropylene is one of the most important commoditypolymers. Polymers with side-arms, polypropylene (PP)being the simplest one containing methyl groups, maypossess stereoregularity in the arrangement of these side-arms. For example, if all methyl groups in PP are ori-ented toward the same direction in the stretched polymerchain, we say this polypropene is isotactic (see Fig. 47).If the methyl groups are alternatingly facing to the frontand the back, we say the polymer has a syndiotactic mi-crostructure. Random organization of the methyl groups(not shown) is found in atactic polymers. The stereoreg-ularity of a polymer has an enormous influence on the

FIGURE 46 Example of new catalyst for ethene oligomerization.

properties of the polymer. Isotactic PP is the well-knownsemicrystalline material that we all know, while atacticPP is a rubber-like material. The degree of isotacticity de-termines the “melting point” of the polymer. Isotactic PPwith less than one mistake in any of its 200 insertionsmelts at 166◦C.

The catalyst and the concomitant technology have un-dergone drastic changes over the years. The titanium cat-alysts can be prepared by the interaction of TiCl4 andalkylaluminum compounds in a hydrocarbon solvent. Thisreaction can be carried out with numerous variations togive a broad range of catalysts. It is a heterogeneous high-surface TiCl3 material of which the active site is titaniumin an unknown valence state. It is quite likely that alkyltita-nium groups at the surface are responsible for the coordi-nation polymerization. Since the 1980s, titanium catalystssupported on magnesium salts have been used.

3. The Cossee-Arlman Mechanism

The mechanism proposed for the solid titanium chloridecatalysts is essentially the same for all catalysts and itis usually referred to as the Cossee-Arlman mechanism.Titanium is hexa-coordinated in the TiCl3 catalysts or sup-ported analogs by three bridging chlorides, one terminalchloride, and one terminal chloride that is replaced by analkyl group by the alkylating agent (Et2AlCl or Et3Al),and a vacancy that is available for propene coordination

FIGURE 47 Microstructures of polypropylene.




FIGURE 48 Cossee-Arlman site.

(see Fig. 48). This simple picture yields an asymmetrictitanium site, but by considering the lattice surface fourbridging chlorides may also lead to an asymmetric site.

In Cossee’s view, one way or another, the site has tocontrol the way a propene molecule inserts, by doing thisin a very controlled manner one can imagine that a stere-oregular polymer will form. This seems obvious, sincedifferent catalysts give different stereospecifities (realityis more complicated as we will see). The asymmetry ofthe site regulates the mode of coordination of the propenemolecule, in other words it steers the direction to whichthe methyl group will point.

4. Site Control Versus Chain-end Control

Over the years two mechanisms have been put forwardas being responsible for the stereo-control of the grow-ing polymer chain, first the site-control mechanism andsecondly the chain-end control mechanism. In the site-control mechanism the structure of the catalytic site de-termines the way the molecule of 1-alkene will insert(enantiomorphic-site control). As we have seen previ-ously, propene is prochiral and a catalyst may attack eitherthe re-face or the si-face. If the catalyst itself is chiral asthe one in Fig. 48, a diastereomeric complex forms andthere may be a preference for the formation of a particu-lar diastereomer. If the catalyst adds to the same face ofeach subsequent propene molecule, we say isotactic PP isformed (a definition proposed by Natta). Thus, we see thatstereoregular polymerization is concerned with asymmet-ric catalysis and indeed the way the problems are tackledthese days have much in common with asymmetric hy-drogenation and related processes.

When we look more closely at the intermediate polymerchain we see an alternative explanation emerging. Afterthe first insertion has taken place a stereogenic center is ob-tained at carbon 2; see Fig. 49. Coordination with the nextpropene may take place preferentially with either the re-

FIGURE 49 Enantiomorphic chain-end control.

face or the si-face, or as displayed in Fig. 49, with themethyl group pointing up or down.

In summary, in the chain-end control mechanism thelast monomer inserted determines how the next moleculeof 1-alkene will insert. Proof for this stems from catalystsnot containing a stereogenic center that give stereoregularpolymer. Secondly, whatever site-control we try to induce,the chain that we are making will always contain, by def-inition, an asymmetric center. These two points wouldstrongly support the chain-end control mechanism. As wehave mentioned above, the nature of the solid catalystshad an enormous influence on the product, and this un-derpins the Cossee site-control mechanism. Thus both areoperative and both are important. Occasionally, chain-endcontrol only suffices to ensure enantiospecifity. The anal-ysis of the products using high resolution 13C-NMR hasgreatly contributed to the mechanistic insight and distinc-tion between the various catalysts. NMR analysis gives adetailed picture of the relative orientation of the methylgroups in the chain, i.e., the regular ones, but more inparticular the mistakes that were made.

5. Homogeneous Catalysts

In the 1970s the first claims appeared concerning thehomogeneous stereospecific polymerization, but they re-ceived relatively little attention as during the same yearsthe first highly active heterogeneous titanium catalysts,immobilized on magnesium salts, were reported and theindustrial interest in homogeneous catalysts diminished.

The development of the new family of homoge-neous catalysts based on biscyclopentadienyl Group 4metal complexes for the stereoselective polymerization ofalkenes is mainly due to Kaminsky, Ewen, and Brintzinger.In 1980 Kaminsky and Sinn reported on an extremelyfast homogeneous catalyst for the polymerization ofethene formed from the interaction of Cp2Zr(CH3)2 and(CH3AlO)n. At 8 bar of ethene and 70◦C an average rateof insertion of ethene amounting 3 × 107 mole of etheneper mole of Zr per hour was reported. For propene thiscatalyst led to completely atactic polymer. Ewen was thefirst to report the synthesis of stereoregular polymers withsoluble Group 4 metal complexes and alumoxane as theco-catalyst. He found that Cp2TiPh2 with alumoxane andpropene gives isotactic polypropene. This catalyst does notcontain an asymmetric site that would be able to controlthe stereoregularity. A stereoblock polymer is obtained,see Fig. 50. Formation of this sequence of regular blocksis proof of the chain-end control mechanism.

Using an intrinsically chiral titanium compound (racethylene-bis-indenyl titanium dichloride), first describedby Brintzinger, Ewen obtained polypropene that was inpart isotactic. Kaminsky and Brintzinger have shown that




FIGURE 50 Stereo-block isotactic PP with an achiral catalyst.

highly isotactic polypropene can be obtained using theracemic zirconium analog of the ethylene-bis(indenyl)compound. Modification of the cyclopentadienyl ligandshas led to a very rich chemistry and a great variety of mi-crostructures and combination thereof has been obtainedincluding isotactic polymer with melting points above160◦C, syndiotactic polypropene, block polymers, hemi-isotactic polymers, etc.

6. Mechanistic Explanations

Fundamental studies have led to a detailed insight into themechanism of the polymerization and the control of themicrostructure through the substituents on the cyclopen-tadienyl ligands. The nature of the catalyst has been thetopic of many studies and is now generally accepted tobe a cationic Cp2ZrR+ species. The first requirement forobtaining an active catalyst is to ascertain the formationof cationic species. Secondly, the metallocene (or relatedligand) must have the correct steric properties.

The achiral Cp2TiPh2 catalyst activated with alumoxanecan produce PP with some isotacticity. Hence, the naturaltendency of the chain in this 1,2 insertion is to catalyzethe formation of isotactic PP. How do we explain the im-provements made with the bridged, chiral metallocenes?We need a catalysts with a specific preference for coordi-nation to either the re- or si-face of the incoming propene.The metallocene can be equipped with a chiral substituentand thus we can try if this is indeed effective. By trialand error and molecular modeling we will find the bestsubstituent at our Cp-ring. Two sites are needed in ourcatalyst: one for the migrating chain and the other for thepropene molecule. We know that insertion takes place ina migratory manner. The one asymmetric site that we cre-ated controlling the coordination of propene is insufficient,because after this migration the alkyl chain will reside onthe other site. The next propene molecule will coordinateto the former alkyl site. Thus, the two sites must be thesame and yet chiral. In asymmetric hydrogenation we haveseen how this problem can be solved, viz. with a C2 sym-metric site. The first compounds introduced having thisfeature are shown in Fig. 51.

Zirconium is tetrahedrally surrounded by two cyclopen-tadienyl ligands and two chlorides. The latter two ligands

are not essential as they are replaced before the polymer-ization can start by an alkyl group (e.g., CH3 from methy-lalumoxane) and by a solvent molecule, thus generatingthe required cationic species.

The two cyclopentadienyl anions are linked togetherby a bridge, here an ethane bridge, and extended with anorganic moiety which renders the molecule its chirality.In the example of Fig. 51 a 1,2-ethane bridged bis(1,1-tetrahydroindenyl) dianion has been drawn. Two com-pounds are obtained when this complex is synthesized,the rac-isomer mixture and the meso-isomer. Often theycan be separated by crystallization. The sterically less hin-dered rac-isomers may be formed in excess. The isomerscan only interchange by breaking of a cyclopentadienyl-metal bond and reattaching the metal at the other faceof the cyclopentadienyl ligand. This process is veryslow compared to the rate of insertion reactions withpropene.

The rac-isomers have a twofold axis and therefore C2-symmetry. The meso-isomer has a mirror plane as the sym-metry element and therefore Cs-symmetry. For polymer-ization reactions the racemic mixture can be used since thetwo chains produced by the two enantiomers are identicalwhen begin- and end-groups are not considered. The twosites (X, in Fig. 51) have the same absolute configuration,and coordination at both sites will occur with a preferencefor the same face of the propene molecule. A simple wayof looking at it would be that the cyclohexyl rings enforcea certain position of the methyl group of propene. When

FIGURE 51 Structure of 1,2-ethanediyl-tetrahydro-indenyl-ZrX2.




FIGURE 52 Bridged fluorenyl-cyclopenadienyl complex givingsyndiotactic PP.

we rotate the molecule around the twofold axis we seethat the site carrying propene is equivalent to the other.In Natta’s words, an isotactic polymer should form (ne-glecting mistakes, and thermodynamic versus kinetic fac-tors). The meso-isomer has no preference for either faceof propene and gives atactic polypropylene.

7. Site and Chain-End Control ReenforceIsotactic Specificity

When dealing with cationic titanium and zirconium met-allocene catalysts, the insertion is a 1,2 mode. As a result,the two control mechanisms reenforce one another. In-deed, the best results in terms of isotacticity are obtainedwith these catalysts.

8. Syndiotactic PP by Site Control

A catalyst that contains two binding sites with the sameabsolute configuration gives isotactic PP. When we canmake a catalyst that contains two identical sites but withthe opposite chirality, each site should coordinate to theopposite face of propene. Following Natta’s definition thisshould give syndiotactic polymer. In other words, a cat-alyst containing a mirror plane—in such a way that thetwo coordination sites are mirror images—gives syndio-tactic polymer, because one site should coordinate to oneface of propene, and vice versa. To this end, Ewen madeisopropyl(1-fluorenyl-cyclopentadienyl) ligands and theirmetallocenes; see Fig. 52. This complex has no chirality(i.e., the dichloride precursor). The catalyst indeed givessyndiotactic polymer.

FIGURE 54 Palladium(0)-catalyzed allylic substitution.

FIGURE 53 “Single site” catalyst for polyethylene.

The above initial findings were the beginning of a newgeneration of catalysts for making polyolefins. It has ini-tiated a lot of research both in industry and academia.Especially interesting is the molecular design of both cat-alysts and polymers. For process reasons, however, thenew homogeneous catalysts, for polypropylene at least,should be heterogenized on solid supports.

Only one commercialization of the new metallocenecatalysts will be mentioned here as an example, the so-called “Single Site Catalyst” developed by Dow and shownin Fig. 53. It is not a stereospecific catalyst as it simplypolymerizes ethene and adds higher alkenes. It gives anarrow MW and a high rate of (re-)insertion of higheralkenes. The control of branching and the extent of long-chain branching leads to a product that can be easilyprocessed.

G. Palladium Catalysis

Palladium catalysts are nowadays an important ingredientof many multistep syntheses both in industry and the lab-oratory. The versatility of palladium is enormous. Herewe will present a few typical examples. In several in-stances iso-electronic complexes of nickel and sometimesrhodium can be used or they lead to even better results.The basic principles are the same.

1. Allylic Alkylation

The palladium-catalyzed allylic alkylation has become avery important tool for organic chemists. It has been de-veloped by Trost, following an initial report by Tsuji. Thereagents used are all very mild and are compatible withmany functional groups. The method has been applied in




FIGURE 55 Regioselectivity in allylic substitution.

the synthesis of many complex organic molecules. Duringthe reaction a new carbon-carbon bond is formed and theresulting molecule still contains a double bond that mightbe used for further derivatization.

The reaction starts with an oxidative addition of an al-lylic compound to palladium zero. Allyl halides, carboxy-lates, etc., can be used. At first we will consider triph-enylphosphine as the ligand, but often ***large ligandinfluences have been detected. A π -allyl-palladium com-plex forms. Formally, the allyl group is an anion in thiscomplex, but owing to the high electrophilicity of palla-dium, the allyl group undergoes attack by nucleophilicreagents, especially soft nucleophiles. After this attack,palladium(0) “leaves” the allyl group and the product isobtained. Palladium zero can reenter the cycle and hencethe reaction is catalytic in palladium. As a side-producta salt is formed. For small-scale industrial applications itis not a problem to make salts in stoichiometric amounts.The mechanism and a few reagents are shown in Fig. 54.

Regioselectivity is high in this reaction and dependson the ligand used. Ligand effects can ensure substitutionat the allylic carbon carrying most carbon substituents,or just the reverse. When a diphosphine ligand is useda hexenyl group instead of the allyl group substitutionoccurs mainly at the terminal carbon, see Fig. 55.

An asymmetric application of this reaction has been de-veloped. The model substrate studied most is the symmet-

FIGURE 56 Asymmetric allylic substitution.

rically 1,3-diphenyl substituted allylpalladium complexshown in Fig. 56. The ligand used is a C2 chiral diphos-phine ligand, but also C1 asymmetric ligands have beensuccessfully applied. The orientation of the two phenylsubstituents at carbons 1 and 3 of the allyl fragment isdifferent under the influence of the chiral C2 ligand. With-out the chiral C2 ligand carbons 1 and 3 are mirror im-ages; palladium is attached to the “local” re- and si-face.The chiral C2 ligand makes carbons 1 and 3 diastero-topic, i.e., they are chemically different. As a result onespecific carbon atom (carbon 1) undergoes selective at-tack by the nucleophile. This way a chiral compound isobtained.

A great variety of ligands have been used. Trost has beenespecially successful using a chiral bidentate phosphinewith a very large bite angle (110◦); see Fig. 57.

This ligand “embraces” the metal and thus exerts its in-fluence on the allyl group. Even substrates carrying smallsubstituents can now be asymmetrically substituted.

2. Cross Coupling

The making of carbon-to-carbon bonds from carbo-cations and carbo-anions is a straightforward and sim-ple reaction. Easily accessible carbo-anions are Grignardreagents RMgBr and lithium reagents RLi. They can beconveniently obtained from the halides RBr or RCl and the




FIGURE 57 One of Trost’s asymmetric ligands for “AAA,” asym-metric allylic alkylation.

metals Mg and Li. They are both highly reactive materials,for instance, with respect to water.

The reactions of these carbon centered anions with po-lar compounds such as esters, ketones, and metal chlo-rides are indeed very specific and give high yields. Thereaction of Grignard reagents and the like with alkyl oraryl halogenides, however, is extremely slow giving manyside-products, if anything happens at all.

The “cross coupling” reaction has found wide applica-tion both in organic synthesis in the lab and in industrialenvironment. The transition metal catalysts are usuallynickel and palladium. In addition to organomagnesium andorganolithium a great variety of organometallic precursorscan be used. Also, many precursors can serve as startingmaterials for the carbocation. Last but not least, the ligandon the transition metal plays an important role in deter-mining the rate and selectivity of the reaction. Here wewill present only the main scheme and take palladium asthe catalyst example. Figure 58 gives the general schemeof the palladium-catalyzed cross-coupling reaction.

We can start with palladium(II) or palladium(0), but forthe present explanation the latter is more convenient. Ox-idative addition of an aryl halide (PhBr in Fig. 58) to palla-dium zero takes place and a square planar Pd(II) complexis formed. Subsequently the inorganic bromide reacts withthe Grignard or lithium alkyl reagent (here 2-BuMgBr)giving a diorganopalladium complex and magnesium di-

FIGURE 58 Palladium-catalyzed cross coupling.

bromide. The third reaction that occurs is the reductiveelimination giving the organic cross-coupling product andthe palladium(0) catalyst in its initial state.

Less reactive organometallics derived from tin andboron can also be used. These reagents do not react withwater, but they are still able to alkylate the palladiumbromide intermediate. As mentioned above, their forma-tion does involve one more step because they are alsomade via Grignard type reagents. The coupling using tinorganometallics is referred to as “Stille” couplings. Thesecond reagent based on boron was introduced by Suzuki.The boronic ester derivative is made from trimethyl borateand an aryl anion reagent followed by hydrolysis of thetwo remaining methyl ester groups. This phenylboronicacid is soluble in water and the coupling reaction can evenbe carried out in water. See Fig. 59 for a scheme of theSuzuki coupling.

The cross-coupling reaction is applied industrially inthe synthesis of alkyl-aryl compounds that are used in liq-uid crystals, aryl-aryl compounds in agrochemicals, phar-maceuticals, etc.

3. Heck Reaction

A reaction related to the cross coupling is the Heck reac-tion. The reaction involves the coupling of an aryl halide(or pseudo halide) with an alkene in the presence of a base.Owing to the latter, as in the cross-coupling reaction anequivalent of base is formed. In the base case, the Heckreaction will produce substituted styrenes. These productscan also be made in the cross-coupling reaction discussedabove. It is attractive that the Heck reaction does not in-volve Grignard-type reagents and thus it allows the pres-ence of groups reactive toward Grignard reagents such asesters, acids, ketones, etc.

The reaction starts again with the oxidative addition ofan aryl halide (Br or I) to palladium zero. The next stepis the insertion of an alkene into the palladium carbon




FIGURE 59 Suzuki cross-coupling reaction.

bond just formed. The third step is β-elimination givingthe organic product and a palladium hydrido halide. Thelatter reductively eliminates HX that reacts with base togive a salt. Figure 60 shows the reaction scheme.

In modern Heck chemistry the halide is replaced bynoncoordinating anions such as O3SCF3. The advantageis that the palladium center is now much less saturated andmore positively charged (and hence more reactive) duringthe alkene coordination and insertion steps.

4. Wacker-type Reactions

The Wacker process is the oxidation of ethene by diva-lent palladium to ethanal in the presence of water. TheWacker-Hoechst process has been studied in great detailand in all textbooks it occurs as the example of a homoge-neous catalyst system illustrating nucleophilic addition to

FIGURE 60 Heck reaction.

alkenes. Palladium chloride is the catalyst and palladiumactivates coordinating ethene toward a nucleophilic attackby water. The overall reaction reads:

C2H4 + 1/2 O2 → CH3CHO �H = −244 kJ/mol

Thus the reaction is highly exothermic as one might expectfor an oxidation reaction.

Figure 61 shows the reaction scheme. First coordina-tion of ethene to palladium has to take place. One chlorideion is replaced by water and one by ethene. Ethene isactivated toward nucleophilic attack by coordination tothe electrophilic palladium ion. Then the key step occurs,the attack of water (hydroxide) to the activated ethenemolecule. The nucleophilic attack of water or hydroxidetakes place in an anti-fashion; i.e., the oxygen attacksfrom outside the palladium complex and the reactionis not an insertion of ethene into the palladium oxygen




FIGURE 61 Reaction scheme of the Wacker ethanal production.

bond. The nucleophilic attack of a nucleophile on analkene coordinated to palladium is typical of “Wacker-”type reactions. The rate equation is in agreement with thereplacement reactions:

v = k[

PdCl2−4

]

[C2H4][

H3O+]−1[Cl−]−2

After a rearrangement of the hydroxyethyl group, ac-etaldehyde (ethanal) forms and palladium zero is the re-duced counterpart. Including the palladium component thereaction reads:

C2H4 + Pd2+ + H2O → Pd(0) + 2H+ + CH3CHO

This reaction had been known since the beginning of thecentury. Reoxidation of palladium directly by oxygen isextremely slow. The invention of Smidt (Wacker Chemie)involved copper as the intermediate in the recycle of di-valent palladium:

2 Cu2+ + Pd(0) → 2Cu+ + Pd2+

2 Cu+ + 1/2 O2+ 2H+ → 2 Cu2+ + H2O

All kinetics are in support of the formation of a β-hydroxyethylpalladium complex as the intermediate.

An important application of the Wacker chemistry isthe reaction of ethene, acetic acid, and dioxygen over aheterogeneous catalyst containing palladium and a reox-idation catalyst for the commercial production of vinylacetate.

FIGURE 62 Oxypalladation of 2-allylphenol.

5. Wacker-type Reactions

A recent development is the use of Wacker activation forintramolecular attack of nucleophiles to alkenes in thesynthesis of organic molecules. The nucleophilic attackis intramolecular; the rates of intermolecular reactions arevery low. The reaction has been applied in a large varietyof organic syntheses and is usually referred to as Wacker(type) activation of alkene (or alkynes). If oxygen is thenucleophile, we also say oxypalladation. An example isshown in Fig. 62. Under these conditions the palladiumcatalyst is often also a good isomerization catalyst, whichleads to the formation of several isomers.

As shown above, palladium can catalyze a variety of re-actions. In addition to the ones shown palladium can alsocatalyze insertions of carbon monoxide, followed by esterformation with alcohols or aldehyde formation with dihy-drogen. Combinations of these reactions in one system areknown. From a point of view of economy this can be veryattractive as several steps are combined in one reactor, us-ing just one catalyst. These “cascade-” or “tandem-” typereactions will become very important for fine chemicalsynthesis in the near future.

H. Asymmetric Epoxidation

1. Introduction

Epoxidation of alkenes (Fig. 63) is a powerful reactionfor the introduction of oxygen in hydrocarbons. Oxygencan be used for the oxidation of ethene over a silver cata-lyst, but for all other alkenes hydroperoxides and relatedreagents are the source of oxygen to avoid further oxi-dation. As catalysts one can use titanium alkoxides, andmolybdenum and tungsten compounds.

Epoxides are interesting starting materials for furtherderivatization. In most instances the reaction will lead tothe formation of mixtures of enantiomers (that is when the

FIGURE 63 Catalytic epoxidation of alkenes.




FIGURE 64 Disparlure.

alkenes are prochiral, or when the faces are enantiotopic).Three important reactions should be mentioned in thiscontext the:

1. Katsuki-Sharpless epoxidation of allylic alcohols2. Sharpless asymmetric hydroxylation of alkenes with

osmium tetroxide3. Jacobsen asymmetric epoxidation of alkenes

All three reactions find wide application in organic syn-thesis. Here we will only discuss the Sharpless epoxida-tion of allylic alcohols. This reaction finds industrial ap-plication in Arco’s synthesis of glycidol, the epoxidationproduct of allyl alcohol, and Upjohn’s synthesis of dispar-lure (Fig. 64), a sex pheromone for the gypsy moth. Thesynthesis of disparlure starts with a C10 allyl alcohol inwhich the alcohol is replaced by the other carbon chainafter asymmetric epoxidation. Perhaps today the Jacobsenmethod can be used directly on a suitable alkene, althoughthe steric differences between both ends of the moleculesare extremely small.

2. Katsuki-Sharpless Asymmetric Epoxidation

The reaction uses allylic alcohols and hydroperoxides.The catalyst is a chiral Ti(IV) catalyst. The chiralityis introduced in the catalyst by reacting titaniumtetra-isopropoxide with one mole of a simple tartrate diester.The reaction is shown in Fig. 65.

The enantioselectivity is not very sensitive to the natureof the allylic alcohol. By contrast, titanium and tartratesare essential to the success. Note the difference with theL-dopa asymmetric hydrogenation, which can be carried

FIGURE 65 Katsuki-Sharpless asymmetric epoxidation.

out with hundreds of C2 chiral diphosphines and specificsubstrates only.

I. Terephthalic Acid

1. Introduction

Terephthalic acid (1,4-benzenedicarboxylic acid) is usedfor the production of polyesters with aliphatic diols as thecomonomer. The polymer is a high-melting, crystallinematerial forming very strong fibers. It is the largest volumesynthetic fiber and the production of terephthalic acid isthe largest scale operated process based on a homogeneouscatalyst. More recently the packaging applications (PET,the recyclable copolymer with ethylene glycol) have alsogained importance. Terephthalic acid is produced fromp-xylene by oxidation with oxygen. The reaction is car-ried out in acetic acid and the catalyst used is cobalt (ormanganese) acetate and bromide. Phthalic anhydride ismade from naphthalene or o-xylene by air oxidation overa heterogeneous catalyst. The main application of phthalicanhydride is in the dialkylesters used as plasticizers (soft-eners) in PVC. The alcohols used are, for instance, 2-ethylhexanol obtained from butanal, a hydroformylationproduct.

2. The Chemistry

The overall reaction reads as follows (Fig. 66):The reaction produces two moles of water per mole

of terephthalic acid. The oxidation reaction is a radicalprocess. The key step of the reaction scheme involvesthe cobalt(III)-bromide-catalyzed H-abstraction to give abenzylic radical, hydrogen bromide, and a divalent cobaltspecies (Fig. 67). This is the initiation reaction of the rad-ical chain reaction. Without cobalt or manganese and bro-mide the reaction would be very slow.

When one methyl group has been oxidized 4-toluic acidis formed. This intermediate is soluble in the solvent used,acetic acid. In the next step it is oxidized to terephthalic




FIGURE 66 Formation of terephthalic acid.

acid, which is almost insoluble in acetic acid. The interme-diate product is 4-formylbenzoic acid. This cocrystallizeswith the final product. Since it is a mono-acid it is willcause a termination of the polymerizing chain.

High molecular weights are needed to obtain strongmaterials. Thus, the presence of mono-acids is detrimen-tal to the quality of terephthalic acid and they have to becarefully removed. One approach involves a second ox-idation carried out at higher temperatures and the othermethod involves reduction of the formylbenzoic acid totoluic acid. The latter in more soluble and stays behind inthe liquid. The reduction is done with a palladium catalyston carbon support. This method gives the highest qualityterephthalic acid.

The polymerization with ethylene glycol can be carriedout directly from the acid if this is very pure. The esterifi-cation/polymerization is conducted at high temperatures(200–280◦C) using metal carboxylates as homogeneouscatalysts.

3. Process Description

Most of the terephthalic acid is produced with a catalystsystem developed by Scientific Design. It was purchasedby Amoco and Mitsui and is referred to as the Amoco Ox-idation. The solvent is acetic acid. Compressed air is usedas the source for oxygen. The catalyst dissolved in aceticacid and the two reactants are continuously fed into thereactor. The temperature is around 200◦C and the pressureis approximately 25 bar. The reaction is very exothermic(1280 kJ/mol). This heat of reaction is removed by evap-

FIGURE 67 Structures involved in terephthalic acid synthesis.

oration/condensation of acetic acid. This heat can be usedin the solvent distillation/purification part of the plant; seeFig. 68.

The oxidation proceeds in two steps, p-toluic acid is theintermediate. This mono-acid has a high solubility in thismedium and undergoes the second oxidation. The conver-sion is high, approaching 100% and the yield based onp-xylene is ∼95%. The solubility of terephthalic acid inmost solvents is extremely low, even at this high temper-ature. The solubility in acetic acid at room temperature is0.1 g/l. At 200◦C it is still only 15 g/l. Thus, most of theproduct crystallizes while it is being formed and, as men-tioned above, the intermediate formylbenzoic acid cocrys-tallizes. The mixture of liquid and solid is led to a filter inwhich the crude product containing 0.5% of mono-acid iscollected. The liquid, containing traces of xylene, p-toluicacid, water, the catalyst, and some product is sent to a dis-tillation unit to remove water and to recover acetic acidand the catalyst. The catalyst is relatively cheap, but theheavy metals cannot be simply discarded and a recycle ofthese robust catalysts would seem to be in place.

As outlined above a purity of “only” 99.5% is a seriousproblem in making a polymer feedstock and the mono-acid intermediate has to be removed either by oxidationunder more forcing conditions (Mitsubishi) or hydrogena-tion to the more soluble 4-toluic acid (Amoco). The finalproduct contains amounts of 4-formylbenzoic acid as lowas 15 ppm. The proportion of toluic acid may be an orderof magnitude higher. Aldehydes usually cause coloring ofpolymers and that is an additional reason to remove themcarefully.




FIGURE 68 Simplified scheme for p-xylene oxidation.

J. Adiponitrile by Hydrocyanation

The transition metal-catalyzed addition of HCN to alkenesis potentially a very useful reaction in organic synthesis,and it certainly would have been widely applied if its at-traction was not largely offset by the toxicity of HCN.Industrially the difficulties can sometimes be minimizedto an acceptable level, and Du Pont has commercializedthe addition of HCN to butadiene for the production ofadiponitrile (NC(CH2)4CN), a precursor to 6,6-nylon.

The catalyst precursor is a nickel(0) tetrakis (phosphite)complex which is protonated to form a nickel(II) hydride.Alternatively, we could write an oxidative addition ofHCN to nickel zero. Subsequently coordination and inser-tion of the diene takes place followed by reductive elim-ination of a pentenenitrile with concurrent regenerationof the nickel zero complex. Two isomerization reactionsmust occur in order to achieve a high selectivity, and thenthe HCN addition cycle is performed for the second timeon the much less reactive 4-pentenenitrile. In Fig. 69 thehydrocyanation mechanism in a simplified form is givenfor ethene; the basic steps are the same as for butadienebut the complications due to isomer formation are lacking.

Hydrocyanation of butadiene is more complicated thanthat of ethene; it requires two hydrocyanation steps andseveral isomers can be observed. The isomers obtained inthe first step of the HCN addition to butadiene are shownin Fig. 70. The addition first leads to compounds 1 and 2,which equilibrate perhaps via the retroreaction. In orderto obtain adiponitrile, 2 should isomerize to 4, and not tothe thermodynamically more stable 3 (conjugation). Thethermodynamic ratio is 2:3:4 = 20:78:1.6. The isomer-ization of 2 to 4 happens to be favorably controlled by

the kinetics of the reactions; the reaction 2 to 4 reachesequilibrium, but the reaction 2 to 3 does not. Note that thenickel complex not only is responsible for the addition ofHCN but that it is also capable of catalyzing selectivelythe isomerization. The final step is the addition of HCNto 4 to give 5, adiponitrile.

K. Alkene Metathesis

Metathesis of alkenes has been known for quite some time.The initial catalysts were based on tungsten, molybdenum,and rhenium. Both homogeneous and heterogeneous cat-alysts find application. Well-known, older homogeneouscatalysts can be formed from WCl6 and an alkylatingreagent. After dialkylation with reagents such as EtAlCl2

FIGURE 69 Hydrocyanation of ethene.




FIGURE 70 Hydrocyanation of butadiene leading to adiponitrile.

or R4Sn, α-elimination occurs and a metal alkylidene isformed, Cl4W=CR2, see elementary steps. Often oxy-gen containing promoters including O2, EtOH, or PhOHare added. Turnover frequencies as high as 300,000 mol(product) · mol−1(cat) · h−1 can be achieved, even at roomtemperature. The development of well-defined catalystsbased on molybdenum and tungsten is mainly due toSchrock. Synthetically useful reactions, including acyclicolefin metathesis, ring-opening metathesis polymerization(ROMP), alkyne polymerization, acyclic diene metathe-sis polymerization (ADMET), and ring-closing metathe-sis (RCM) have been catalyzed by early-transition-metalalkylidenes. Porri introduced in situ catalysts based onruthenium or iridium in polar media. The synthesis ofwell-defined ruthenium alkylidene complexes of the typeRuCl2(=CHPh)(PR3)2 was developed by Grubbs; seeFig. 71.

The peculiar property of the ruthenium (and also irid-ium catalysts) is that they are active in the presence ofpolar substituents. The initial catalysts, generated fromhigh-valence metal chlorides with alkylating agents werenot resistant to oxygen-containing compounds. Schrockdeveloped sterically hindered alkylidene complexes, suchas the one shown in Fig. 72. The in situ prepared catalystscontaining bulky phenols have also been reported. By re-placing the two alkoxides by a chiral bulky bisnaphtholasymmetric metathesis has been achieved.

FIGURE 71 An example of Grubbs’ ruthenium catalysts.

Many applications of these new catalysts to organicchemistry are known as they can serve as a means to makelarge rings, for instance, after two allyl groups have beenintroduced to a nonlinear fragment; see Fig. 73. The re-sulting double bond can be further functionalized or hy-drogenated if needed.

L. Cyclopropanation

As in the previous section about metathesis, cyclopropa-nation is concerned with the transfer of carbene or alkyli-dene species from a metal to an organic molecule. Thereaction involves the metal-catalyzed decomposition of adiazo compound to give a reactive metal-alkylidene com-plex followed by the transfer of the “carbene” to an alkene.The product of this very convenient reaction is a cyclo-propane derivative in a single step (see Fig. 74). Catalystsused for this reaction are based on copper, rhodium, andruthenium. A common, relatively stable diazo compoundis N2CHCO2Et. The example shown in Fig. 74 is carriedout commercially using a copper complex containing achiral imine dialkoxide ligand. After conversion of the

FIGURE 72 Schrock’s catalyst for metathesis of functionalalkenes.




FIGURE 73 Ring-closing metathesis.

acid to a substituted amide Cilastatin, a pharmaceutical isobtained.

Cyclopropanation of dienes leads to an important classof compound used as insecticides, the pyrethroid-typecompounds related to chrysanthemic acid, a natural insec-ticide obtained from East African chrysanthemums. Cat-alysts used comprise rhodium (II) dimers of carboxylatesand related compounds. Chiral complexes can lead to highenantioselectivities. This chemistry has been successfullydeveloped by Doyle. In Fig. 75 the synthesis of the ba-sic skeleton has been depicted, which requires replace-ment of the ethyl group by other groups. Interestingly,the “carbene” fragment can also be inserted into N−Hor C−H bonds. The former reaction is used for makingβ-lactams.

The activity of ruthenium complexes for this reactionaffords a bridge between the metathesis reaction in theprevious section and cyclopropanation, an area that hasbeen explored by Noels. A common intermediate can beimagined for the two reactions leading either to metathesisor cyclopropanation, as shown schematically in Fig. 76.

M. Hydrosilylation

The catalytic addition of a silicon compound contain-ing an Si−H bond to a C=C double bond is the mostcommon way for the synthesis of silane or silicon com-pounds. The silicon compounds are used to make poly-mers (silicone rubbers) or a large variety of agents usedfor the modification of surfaces. For making polymers thereagent used is, for instance, methyldichlorosilane, whichis added to vinyl compounds using a catalyst (Fig. 77). Thedichlorosilane compounds are hydrolyzed to give siliconepolymers.

For industrial processes the oldest and preferred cata-lyst for the hydrosilylation reaction is Speier’s catalyst,H2PtCl6, which can be used in parts per million quanti-ties. The catalyst is reduced in situ and probably alcoholsand/or other oxygen-containing compounds are involved

FIGURE 74 Cyclopropanation using diazocompounds as “car-bene” precursor.

FIGURE 75 Cyclopropanation yielding pyrethroid insecticides.

in this reaction. Schematically the reaction mechanismcan be imagined as a sequence involving alkene complex-ation, oxidative addition of the silane breaking the Si−Hbond, and insertion and reductive elimination, as shown inFig. 78. It is referred to as the Chalk-Harrod mechanismand it follows exactly the elementary steps described in theintroduction. The precise nature of the platinum catalystfor this process, however, has not been delineated. The cat-alysts are extremely sensitive to the conditions applied andin practice one encounters incubation times, capricious ki-netics, alkene isomerization, formation of unsaturated sideproducts, etc.

A large variety of metal complexes catalyze the hy-drosilylation reaction. Other double bonds can serve asthe silane acceptor, such as alkynes, ketones, and imines.For the latter two substrates asymmetric catalysts also havebeen developed.

N. Polyesters

One of the largest applications of homogeneous catal-ysis involves the manufacture of polyesters. In a sim-ple condensation process diacids and diols are poly-merized forming the polymers. The simplest one ispolyethyleneterephthalate (PET), which is made fromethylene glycol and terephthalic acid as discussed above.

FIGURE 76 Relation between metathesis and cyclopropanationmechanism.




FIGURE 77 Hydrosilylation and silicone polymer formation.

FIGURE 78 The Chalk-Harrod mechanism for hydrosilylation.

Instead of terephthalic esters the dimethyl esters are usedand the condensation reactions are in part transesterifica-tions. The monomers have to be extremely pure, becauseany monomeric impurity will end the growing chain; seeFig. 79.

The catalysts used for the transesterification reaction aresimple salts, such as zinc or calcium acetate. For the di-rect esterification a mixture of Sb2O3 and titanium alkox-ides are used. Similar salts are used for the manufactureof polyamides, polyurethanes, and polycarbonates. Themechanism involves interaction of the oxygen atoms of

FIGURE 79 The two steps for polyester formation.

the monomers and the protons of the monomers withthe anions added. Apparently a subtle balance betweenthese interactions is needed in order to obtain effectivecatalysts. As for other reactions involving oxygenates ordioxygen, the knowledge about the various catalysts isempirical and a large variety of cations and anions havebeen tested. This is different from many reactions dis-cussed above that involve discrete formation of metal-to-carbon bonds, via elementary steps for which we some-times know how they can be influenced by ligand variation,such as the oxidative addition and reductive eliminationsteps.


CATALYSIS, INDUSTRIAL • CATALYST CHARACTERIZA-TION • ELECTRON TRANSFER REACTIONS, GENERAL • KI-NETICS (CHEMISTRY) • ORGANOMETALLIC CHEMISTRY

BIBLIOGRAPHY

Beller, M., and Bolm, C., eds. (1998). “Transition Metals for Or-ganic Synthesis; Building Blocks and Fine Chemicals,” Wiley-VCH,Weinheim, Germany.




Cornils, B., and Herrmann, W. A., eds. (1996). “Applied Homoge-neous Catalysis with Organometallic Compounds,” VCH, Weinheim,Germany.

Cornils, B., and Herrmann, W. A., eds. (1998). “Aqueous PhaseOrganometallic Catalysis–Concepts and Applications,” Wiley-VCH,Weinheim, Germany.

Diederich, F., and Stang, P. J., eds. (1998). “Metal-Catalyzed Cross-Coupling Reactions,” Wiley-VCH, Weinheim, Germany.

Fink, R., Mulhaupt, G., and Brintzinger, H. H., eds. (1995). “ZieglerCatalysts,” Springer, Berlin.

Noyori, R. (1995). “Asymmetric Catalysis in Organic Chemistry,” Wiley,New York.

Parshall, G. W., and Ittel, S. D. (1992). “Homogeneous Catalysis,” Wiley,New York.

van Leeuwen, P. W. N. M., and Claver, C., eds. (2000). “RhodiumCatalyzed Hydroformylation,” In “Catalysis by Metal Complexes,”Kluwer Academic Publishers, Dordrecht.

van Santen, R. A., van Leeuwen, P. W. N. M., Moulijn, J. A., andAverill, B. A., eds. (1999). “Catalysis, an Integrated Approach,” 3rded., Elsevier, Amsterdam.

P1: GTY/GWT P2: GLM Final Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology EN006K-933 June 29, 2001 12:14

Fuel ChemistrySarma V. PisupatiPennsylvania State University

I. FuelsII. Properties for UtilizationIII. Combustion of CoalIV. Fixed Bed CombustionV. Pulverized Coal Combustion

VI. Fluidized Bed CombustionVII. Environmental Issues in Coal Combustion

VIII. Coal GasificationIX. Coal LiquefactionX. Liquid FuelsXI. Properties for UtilizationXII. Major Combustion Methods

XIII. Gaseous Fuels

GLOSSARY

Atomization A process of breaking a liquid stream intofine droplets.

Combustion Rapid oxidation of fuels generating heat.Equivalence ratio Ratio of the amount of air required for

stoichiometric combustion to the actual amount of airsupplied per unit mass of fuel.

Flame A chemical interaction between fuel and oxidantaccompanied by the evolution of heat and light.

Fluidization A process of making a bed of particles be-have like a fluid by a jet of air.

Fuel Any substance that can be used to produce heat andor light by burning.

Gasification Conversion of solid fuel into a combustiblegas.

Liquefaction Conversion of solid fuel into liquid fuel.Selective catalytic reduction A method used to se-

lectively reduce nitrogen oxides to nitrogen using acatalyst.

ENERGY is the life-blood of a modern society. The qual-ity of life of any society is shown to be directly propor-tional to the energy consumption of the society. In thepast century (1900–2000), the population of the worldand the United States has increased by 263 and 258%, re-spectively. The energy demand increased by a startling 16

253

P1: GTY/GWT P2: GLM Final


254 Fuel Chemistry

TABLE I Total Primary Energy Consumptionand Their Sources

1999 (Quadrillion Btus)

Source World U.S.

Petroleum 152.20 37.71

Natural Gas 86.89 22.1

Coal 84.77 21.7

Nuclear 25.25 7.73

Hydro Electric 27.29 3

Renewable 2.83 4.37

Total 381.88 96.6

and 10 times, respectively. A vast majority of this energy(about 85%) comes from fossil fuels. These fuels—coal,oil, natural gas, oil shale, and tar sands were formed overmillions years by compression of organic material (plantand animal sources) prevented from decay and buried inthe ground. Table I shows the total primary energy useand the sources. Most of the fuels are used to generateheat and/or power. This chapter deals with the fuels ori-gin, properties, and utilization methods and chemistry ofthese processes.

I. FUELS

Fossil fuels are hydrocarbons comprising of primarily car-bon and hydrogen and are classified as solid, liquid, andgaseous based on their physical state. Solid fuels includenot only naturally occurring fuels such as wood, peat, lig-nite, bituminous coal, and anthracites, but also certainwaste products by human activities like petroleum cokeand municipal solid waste. Approximately 95% of the coalmined in the United States is combusted in boilers and fur-naces to produce heat and/or steam. The other 5% is usedto produce coke for metallurgical uses. Liquid fuels aremostly produced in a refinery by refining naturally occur-ring crude oil, which include gasoline, diesel, kerosene,light distillates, and residual fuels oils. Each of these hasdifferent boiling range and is obtained from a distillationprocess. Gaseous fuels include natural gas, blast furnacegas, coke oven gas, refinery gases, liquefied natural gas,producer gas, water gas and coal gas produced from vari-ous gasification processes. Except for natural gas, most ofthe gaseous fuels are manufactured.

A. Origin of Fossil Fuels

Earth is a closed system with respect to carbon, and there-fore carbon on this planet has to be used and reused. Atotal account of carbon in the world would explain fos-

sil fuel formation. Global carbon cycle illustrates the fatecarbon in the world. Carbon exists in the world in threemajor reservoirs: in the atmosphere as CO2, in the rocks asCO−−

3 , and in the oceans, which occupy two thirds of theplanet’s surface, as carbonate (CO−−

3 ) and bicarbonates(HCO−

3 ). The CO2 in the atmosphere has a vital role in theformation of fossil fuels. The CO2 in the atmosphere reactswith water vapor in the presence of sunlight to form the or-ganic matter and oxygen by photosynthesis reaction. Theorganic matter can be of microscopic plant (phytoplank-ton) or microscopic animal (zooplankton) or higher plants.The dead organic matter through decay reaction combineswith oxygen and forms CO2 and H2O. This decay reac-tion is exactly the reverse of the photosynthesis reaction.Fossil fuels have formed by minimization or preventionof the decay reaction by possibly inundating the organicmatter by water or covering by sediments. The organicmatter, microscopic plants and animals, and higher plants,is chemically comprised of protiens, lipids, carbohydrates,glycosides, resins, and lignin. Of these, lignin is predom-inantly present in higher plants. Other components arepredominantly present in zoo, phytoplankton, and algae(microscopic organic matter). Coal is a complex materialcomposed of microscopically distinguishable, physicallydistinctive, and chemically different organic substancescalled macerals and inorganic substances called minerals.

During the transformation of organic matter to coal,there is a significant loss in oxygen and moderate lossof hydrogen with an increase in carbon content by ini-tial aerobic reactions and subsequent anaerobic reactions.This process leads to the formation of kerogen. The or-ganic matter, which is rich in algae, forms alginitic kero-gen (type I kerogen); whereas the organic matter, whichis rich in fatty acids and long-chain hydrocarbons, leadsto the formation of liptinic or Type II kerogen. The or-ganic matter consisting of lignin structure forms Type IIIkerogen and leads to the formation of coal. Formation ofkerogen from the organic matter is by a process called“diagenesis,” and the transformation of kerogen to fossilfuels is by a process called “catagenesis.” The tempera-ture in the earth’s surface increases with depth at a typicalrate of 10–30◦C/km. As the temperature and pressure dueto overburden increase, peat is transformed into lignitein about 30–50 million years. The primary reactions thatare believed to occur during this transformation are de-hydration, decarboxylation, and condensation leading toa loss of oxygen, hydrogen, and some carbon. Progres-sive transformation of lignite to subbituminous coal tobituminous coal and then to anthracite occurs in a timeperiod ranging from 50 to 300 million years. The reac-tions responsible for these changes resulting in rapid lossof hydrogen are dealkylation, aromatization, and conden-sation. Formation of anthracites from bituminous coals



Fuel Chemistry 255

TABLE II Compositional Analysis of Various Solid Fuels

BituminousLignite, coal (hvAb), Anthracite,Darco Subbituminous Upper Clarion Primrose Petroleum

Fuel component Wooda Peata seam, TXb coalb seamb seamb Cokea

Moisture (as rec’d) 48.00 — 32.60 27.12 1.73 3.77 5.58

Volatile matter (d.b.)c 72.80 75.00 67.39 47.56 39.41 3.71 10.41

Fixed carbon (d.b.) 24.2 23.1 21.34 38.12 51.68 82.38 88.89

Ash (d.b) 3.00 2.70 11.27 14.32 8.91 13.91 0.71

Heating value (d.b) (Btu/lb) 9,030 8,650 11,375 10,842 13,390 12,562 15,033

Carbon 55.00 8,650 74.90 73.88 83.67 96.65 88.64

Hydrogen 5.77 5.60 4.58 6.28 5.33 1.25 3.56

Nitrogen 0.10 0.70 1.75 1.22 1.46 0.78 1.61

Sulfur 0.10 0.17 0.78 1.78 4.82 0.52 5.89

Oxygen 39.10 40.10 18.78 18.62 9.54 1.32 0.30

a C, H, N, S, and O are given on a dry ash free basis.b C, H, N, S, and O are given on a dry mineral matter free basis.c Dry basis.

requires higher pressures and temperatures in excess of200◦C. The higher temperatures are encountered whenthere is a magma nearby in the ground, and high lateralpressures are encountered where land masses collide lead-ing to the formation of mountains. Both of these scenarioslead to the formation of anthracites. The nature of the con-stituents in coal is related to the degree of coalification,the measurement of which is termed rank. Coals may beclassified according to (1) rank, based on the degree ofcoalification; (2) type, based on megascopic and micro-scopic observations (physical appearance) that recognizedifferences in the proportion and distribution of variousmacerals and minerals; and (3) grade, based on value fora specific use. Typical properties of these solid fuels areshown in Table II.

II. PROPERTIES FOR UTILIZATION

Coals vary widely from place and to place, and sometimeseven within a few feet in a particular seam because ofthe nature of the precursor material and the depositionalenvironment. Therefore, coals and other solid fuels areanalyzed for certain important properties for utilization(summarized in Table III).

Coal rank is usually determined from an empirical anal-ysis called the proximate analysis and calorific value or op-tical reflectance of vitrinite. The proximate analysis con-sists of determination of moisture, volatile matter, and ashcontents, and, by difference from 100%, the fixed car-bon content of a coal. The American Society for Testingand Materials (ASTM) method of classification of coals

uses the dry, mineral matter-free volatile matter to classifycoals above the rank of medium volatile bituminous. Forcoals with greater than 69% volatile matter, the methoduses the moist (containing natural bed moisture but notsurface moisture), mineral matter-free calorific value toclassify coals below the rank of high volatile bituminous.The moisture content is obtained by heating an air-driedcoal sample at 105–110◦C under specified conditions un-til a constant weight is obtained. The moisture content, ingeneral, increases with decreasing rank and ranges from 1to 40% for the various ranks of coal. The moisture contentis an important factor in both the storage and the utiliza-tion behavior of coals. The presence of moisture adds un-necessary weight during transportation, reduces the avail-able heat consuming latent heat of vaporization, and poses

TABLE III Important Properties for Utilization

Property Factors affecting

Compositional analysis Proximate analysis

Ultimate analysis

Heating value

Grindability Coal rank

Moisture

Ash

Combustibility Proximate analysis

Surface area

Porosity

Petrographic Analysis

Inorganic constituents Associated with the organic structure

Discrete inorganic minerals



256 Fuel Chemistry

some handling problems. Volatile matter is the materialdriven off when coal is heated to 950◦C in the absence ofair under specified conditions, and is determined practi-cally by measuring the loss of weight. It consists of a mix-ture of gases, low-boiling-point organic compounds thatcondense into oils upon cooling, and tars. Volatile matterincreases with decreasing rank. In general, coals with highvolatile matter are known to ignite easily and are highlyreactive in combustion applications. The calorific value isthe amount of chemical energy stored in the coal, whichis released as thermal energy upon combustion and is di-rectly related to the rank. The calorific value determinesin part the value of a coal as a fuel for combustion appli-cations. The ultimate analysis includes elemental analysisfor carbon, hydrogen, nitrogen, and sulfur and oxygen ona dry ash free basis. Oxygen content in the fuel is obtainedby subtracting the sum of the percentages of C, H, N, andS from 100.

The grindability of a coal is a measure of its resistanceto crushing. The ball-mill and Hardgrove grindability testsare the two commonly used methods for assessing grind-ability. The Hardgrove method is often preferred to theball-mill test because the former is faster to conduct. Thetest consists of grinding a specially prepared coal samplein a laboratory mill of standard design. The percentage byweight of the coal that passes through a 200-mesh sieve(screen with openings of three-thousandths of an inch or74 µm) is used to calculate the Hardgrove GrindabilityIndex (HGI). Two factors affecting the HGI are the mois-ture and ash contents of the coal. A correlation betweenthe HGI and rank indicates that, in general, lignites andanthracite coals are more resistant to grinding (low in-dices) than are bituminous coals. The index is used as aguideline for sizing grinding equipment in the coal pro-cessing industry.

Inorganic constituents in coal upon oxidation produceash. The inorganic constituents are present in the form ofdiscrete mineral particles or bonded to the organic coalstructure or sometimes water associated. The manner inwhich the inorganic constituents are present in raw coalsdetermines the intermediate and final form of ash. Lower-rank coals (lignites and subbituminous) tend to have moreorganically associated inorganics compared to high-rankcoals.

Combustibility is measured by proximate analysis andsome empirical laboratory tests. Volatile matter is an in-dicator of ease of ignition and the fraction of coal that isburnt in gas phase. The higher the volatile matter the less isleft as char that needs to be burnt. The combustion of chartakes a longer time because of the heterogeneous reactionbetween carbon and oxygen. Combustion of high volatilecoals therefore is easier to ignite and releases most of theheat closer to the burner. This requires an increase in the

velocity of the fuel and air mixture jet to keep the flameat a distance from the burner tip. However, coals withlower volatile matter have a delay in ignition and there-fore cause flame stability problems. These coals requireadditional design measures such as high swirl or a “bluffbody” to promote recirculation of hot gases to stabilizethe flame. Combustibility of coal is also characterized bybench-scale laboratory techniques such as drop tube reac-tor, entrained flow reactor, or thermogravimetric analyzerto determine the reactivity of fuels. One such method isused to determine a burning profile. This is widely usedto determine the combustion behavior of an unknown fuelrelative to a reference fuel. This method originally wasdeveloped by Babcock and Wilcox. A burning profile isa plot of the rate at which a solid fuel sample loses itsweight as a function of temperature, when heated at con-stant rate in air. The burning profile is used to determinethe temperatures at which the onset of devolatilization andchar combustion, peak reation rate, and the completion ofchar oxidation occur. The temperatures range over whichthese take place also indicates the heat release rates andzones.

For a better understanding of behavior of coal ina furnace, pilot-scale combustion tests are performedto characterize ignition and flame stability, combustionefficiency, gaseous and particulate emissions, slaggingand fouling, and the erosion and corrosion aspects of afuel.

III. COMBUSTION OF COAL

Combustion is rapid oxidation of fuels generating heat.Combustion of fuels is a complex process the understand-ing of which involves chemistry (structural features of thefuels), thermodynamics (feasibility and energetics of thereactions), mass transfer (diffusion of fuel and oxidantmolecules for a reaction to take place, and products awayfrom the surface), reaction kinetics (rates of reactions),and the fluid dynamics (bulk movement of the fuel andcombustion gases) of the process.

Combustion of solid fuels generally involves threesteps—drying or evaporation of moisture, thermal de-composition and devolatilization, and oxidation of solidresidue or char. Since solid fuels differ widely in termsof physical and chemical properties, a general qualitativediscussion on the combustion behavior of a coal parti-cle is given here. The duration and chemistry of each ofthese processes depend on the type of fuel burnt and thesize of the particles, heating rate, furnace temperature, andparticle density. For example, wood, peat, and lignite fu-els contain a large percentage of moisture and the dryingtime is quite long. Also the volatile matter is quite high



Fuel Chemistry 257

and therefore the time required for heterogeneous com-bustion of solid residue (char) is shorter. The drying timeof a small particle is the time required to heat the particleto the boiling point of the water and evaporate the wa-ter. This depends on the amount of water in the particle,particle size, and the heating rate of the particle. The rateof heat transfer to the particle depends on the tempera-ture difference between the boiler furnace and the particletemperature.

Devolatilization of a fuel particle involves thermal de-composition of the organic structure, thereby releasing theresulting fragments and products. The rate of devolatiliza-tion depends on the rate of heat transfer to the particle tobreak weak bonds resulting in the primary decompositionof products. The primary products of decomposition travelfrom the interior of the particle to the surface through thepores of the solid. While doing so, the primary products,depending on their nature, could react with each other orwith the char surface and result in secondary products ordeposits on the walls of the pores. The devolatilizationprocess begins at about 350◦C and is a strong functionof temperature. The products consist of water, hydrogen-rich gases and vapors (hydrocarbons), carbon oxides, tars,light oils, and ammonia. The product distribution for bitu-minous coal and lignite is shown in Fig. 1. These volatilesmay be released as jets and play an important role in ig-nition and char oxidation steps. A carbon-rich solid prod-uct called char/coke also is produced. The compositionof the products of decomposition depends on the typeof the fuel, peak temperature, and rate of heating of theparticle.

Combustion of char is a slower process compared tovolatile combustion and is critical in determining the totaltime for combustion of a coal particle. The steps that areinvolved in the oxidation of char are

1. Diffusion of oxygen from the bulk to the char surface2. Diffusion of oxygen from the surface to the interior of

pores of the char3. Chemisorption of reactant gas on the surface of the

char4. Reaction of oxygen and carbon5. Desorption of CO and CO2 from the char surface6. Diffusion of CO and CO2 to the surface through the

pores7. Diffusion of products from the exterior surface to the

bulk of fluid

The properties of the char are important for its oxida-tion. Pulverized coal particles, during devolatilization, areknown to swell by 5–15% depending of the coal type andheating rate. Various types of chars with varying physicalstructures are produced. The volatile matter when released

FIGURE 1 The product distribution for bituminous coal and lig-nite. From Lawn, C. W. (1987). “Principles of Combustion,” Aca-demic Press.

from a coal particle produces void space or porosity. Theporosity of the char gradually increases as the conver-sion progresses. Ultimately the char particle becomes soporous that the particle fragments, as shown in Fig. 2. Thecombustion char that is produced can take place as sur-face reaction a (constant density and shrinking diameter)or a volumetric reaction (constant diameter and changingdensity). The surface reaction generally occurs when thetemperatures are high and chemical reaction is fast anddiffusion is the rate-limiting step, whereas volumetric re-action dominates when the temperatures are low and theoxidant has enough time to diffuse into the interior of theparticle. The concentration profile of the reactant is shownin Fig. 2.

Ash formation occurs predominantly by two mecha-nisms. Volatile inorganic species are vaporized during charcombustion and subsequently condense when the temper-ature is low downstream. The ash particles formed by thismechanism tend to be submicron in size; whereas, min-eral inclusions come into contact with one another, andsince the temperature in the pulverized coal combustion



258 Fuel Chemistry

FIGURE 2 Simplified schematic of combustion of coal particles.

units is high enough to melt the molten ash, particles coa-lesce to form larger particles. Hence, the size distributionof pulerized coal ash tends to be bimodal.

A. Major Combustion Processes

Combustion methods of solid fuels are classified into fourcategories. These are fixed-bed combustion, fluidized-bedcombustion, pulverized or suspension firing, and cyclonefiring. These methods differ in the fuel size used and thedynamics of the fuel particles.

IV. FIXED BED COMBUSTION

Combustion of coal in fixed beds (e.g., in stokers) is theoldest method of coal use and it used to be the most com-mon. Large-size coal pieces, usually size graded between1/8 and 2 in., are supported on a grate and air is suppliedfrom the bottom through the grate. The coal particles arestationary, hence the term “fixed” bed. Fixed-bed combus-

tion systems can be further classified based on the coal feedsystem—overfeed, underfeed, spreader stoker, or travel-ing grate. Large size limits the rate of heating of the par-ticles to about 1◦C/sec and requires about 45 to 60 minfor combustion. A schematic of a traveling grate combus-tion furnace is shown in Fig. 3. Coal particles initiallydevolatilize and the combustible volatile gases are burntabove the bed by the overfire air. The overfire air is crucialin obtaining complete combustion of the volatiles. In mostof the stoker systems, coals which exhibit caking proper-ties can be a problem due to clinker formation. Clinkersprevent proper air distribution combustion, the air lead-ing to high unburnt carbon and higher carbon monoxideemissions.

When the fuel is drawn on to the grate from the coalhopper by the moving grate, the bed is approximately 10–15 cm. The coal particles initially require less air for dryingand combustion. As the coal bed ignites and the plane ofignition travels down from the top to the grate, combustionair demand increases to a maximum. As the coal bed burnsand reaches the other side, the demand for air decreases


Encyclopedia of Physical Science and Technology EN006K-933 July 12, 2001 15:6

Fuel Chemistry 259

FIGURE 3 Schematic diagram of a traveling grate stroker.

to a minimum. Ash particles produced are dropped into arefuse pit on the other side of the combustion chamber. Asshown in Fig. 4, the oxidant and the products of combus-tion have to travel through the layer of ash produced. Thiscombustion method tends to produce a high amount of un-burnt carbon and high(er) CO levels. Gaseous emissionsfrom stoker systems are hard to control.

Since the coal bed height is constant, an increase in thethroughput (larger size) can only be achieved by increas-ing the size of the grate. Due to this physical limitation ofthe grate size, stoker combustion units have an upper sizelimit of about 100 MMBtu/hr or 30 MW(thermal). There-fore, the application of this type of combustion method islimited to industrial and small on-site power generationunits.

V. PULVERIZED COAL COMBUSTION

In the pulverized coal combustion system, coal groundto a very fine size (70–80% passing through a 200-meshscreen or 75 µm) is blown into a furnace. Approximatelyabout 10% of the total air required for combustion is usedto transport the coal. This primary air, preheated in theairheater, is used to drive out the moisture in the coal andtransport the coal to the furnace. A majority of the com-bustion air is admitted into the furnace as secondary air,which is also preheated in the air preheater. The particlesare heated at 105–106◦C/sec, and it takes about 1–1.5 secfor complete combustion. This increases the throughput tofurnace and, hence, it is the most preferred form of com-

FIGURE 4 Combustion of a large coal particle.

FIGURE 5 A schematic of a pulverized coal firing system.

bustion method by electric utilities. Based on the burnerconfiguration, pulverized coal combustion systems can bedivided into wall-fired, tangentially fired, and down-firedsystems. In wall-fired units the burners are mounted on asingle wall or two opposite walls. Wall-fired systems withsingle wall burners are easy to design. However, due to un-even heat distribution, the flame stability and combustionefficiency are poor compared to the opposed wall-firedsystem. Because of the possibility of flame impingementon the opposite wall, this system may lead to ash slag-ging problems. An opposite wall-fired boiler avoids mostof these problems by providing even heat distribution andbetter flame stability. A diagram of the opposed firing sys-tem is shown in Fig. 5. However, the heat release rate canbe higher than single wall burner systems and could leadto slagging and fouling problems. A tangential firing sys-tem has burners installed in all four corners of the boiler(as shown in Fig. 6). The coal and primary air jets are is-sued at an angle, which is tangent to an imaginary circle atthe center of the furnace. The four jets from four cornerscreate a “fireball” at the center. A top view of the furnaceis provided in Fig. 6. The swirling flame then travels upto the furnace outlet. This method provides the highesteven heating flame stability. The temperature in the “fire-ball” can reach as high as 1700–1800◦C. The burner is akey component in the design of the combustion system.It also determines the flow and mixing patterns, ignitionof volatiles, flame stability, temperature, and generation ofpollutants. Burners can be classified into two types—swirland nonswirl. Primary air transports coal at velocities ex-ceeding 15 meters/sec (the flame velocity) of the fuel. Theflame velocity is a function of volatile matter. Secondaryair ignites the fuel, determines the mixing pattern, and



260 Fuel Chemistry

FIGURE 6 A plan view of a tangentially fired furnace.

also determines the NOx formation. High volatile bitumi-nous coals tend to have higher flame velocities comparedto low volatile coals. Swirling flows promote mixing andesatablish recirculation zones. In swirling flows, the axialflux of angular momentum G and the axial flux of axialmomentum Gx are conserved.

Gx =∫ r2

r1

2πr (p + ρu2) dr = constant

Gφ =∫ r2

r1

2πρuwr2 dr = constant,

where u and w are the axial and tangential components ofthe velocity at radius r, p is the static gage pressure, andr1 and r2 are radial limits of the burner. Swirl number is ameasure of swirl intensity

S = Gφ

Gxrb,

where rb is the radius of the burner. A strong swirl usuallyhas a value of S > 0.6. Stronger swirl is necessary whenburning is difficult to ignite coals.

VI. FLUIDIZED BED COMBUSTION

Fluidization is a process of making solids behave like a“fluid.” When a jet of air is passed though a bed of solidsfrom underneath, the force due to drag pushes the particles

up while the gravity pulls the particles down. When thevelolity of the air increases above a certain point, the dragforce exceeds the weight of the particles and the particleswill start to fluidize. Under this condition, the solid bedmaterial is made to exhibit properties such as static pres-sure per unit cross section, bed surface level and flow ofsolids through a drain etc., similar to that of a fluid in thebed. This fluidization behavior promotes rapid mixing ofthe bed material, thus promoting gas–solid and solid–solidheat transfer.

In a typical fluidized bed combustion process, solid, liq-uid, and or/gaseous fuels, together with noncombustiblebed material such as sand, ash, and/or sorbent are flu-idized. The primary advantage of this mode of combustionover pulverized and/or fixed-bed combustion is its operat-ing temperature (800–900◦C). At this temperature, sulfurdioxide formed by combustion of sulfur can be capturedby naturally occurring calcium-based sorbents (limestoneor dolostone). The reactions between the sorbent andSO2 are thermodynamically and kinetically balanced inthis range of temperatures. The operating temperaturerange of FBC boilers is low enough to minimize thermalNOx production (which is highly temperature dependent)compared to other modes of combustion and yet highenough to achieve good combustion efficiency. In additionto these advantages, FBC boilers can be adapted to burn awide range of fuels such as low-grade, low-calorific-valuecoal wastes, and high-ash and/or high-sulfur coals in anenvironmentally acceptable manner. Fluidized beds can befurther subdivided into bubbling or circulating fluidizedbeds based on the operating characteristics, and into atmo-spheric and pressurized fluid beds based on the operatingpressure.

Bubbling fluidized beds characteristically operate witha mean bed particle size between 1000 and 1200 µm (1.0to 1.2 mm). The operating velocity in a BFBC boiler isabove the minimum fluidizing velocity and less than athird of the terminal velocity (1–4 min/sec). Under theconditions of low operating velocity and large particlesize, the bed operates in the bubbling mode, with a definedbed surface separating the densely loaded bed and the lowsolids freeboard region. In bubbling fluidized beds the goalis to prevent solids from carrying over, or elutriating, fromthe bed into the convective pass(es). Therefore, particlessmaller than 500–1500 µm are not used in BFBCs so asto avoid excessive entrainment.

Circulating fluidized beds operate with a mean particlesize between 50 and 1000 µm (0.05 to 1 mm). The hydro-dynamics of the fast fluidized beds allow CFBC boilersto use sorbent particles as fine as 50 µm. CFBC boilersoperate typically at about twice the terminal velocity ofthe mean size particles, ranging from 3 to 10 min/sec.This high operating gas velocity, recirculation rate, solids



Fuel Chemistry 261

FIGURE 7 A schematic of a circulating fluidized bed boiler.

characteristics, volume of solids, and the geometry of thesystem create a hydrodynamic condition known as “fastbed” or “dilute phase refluxing.” As a result of this con-dition solid agglomerates are continuously formed, dis-persed, and reformed again. This motion also increasesthe slip velocity between the gas and solid phases, whichincreases the attrition or fragmentation of char and sorbentparticles. The air required for the combustion is usuallyadmitted in two streams: as primary and secondary air.In CFBC boilers, as shown in Fig. 7, the furnace can beconceptually divided into three sections: (1) the lower fur-nace zone, below the secondary air level where the solidsflow regime is similar to a bubbling or turbulent bed andthe particles range in size from approximately 1000 to500 µm; (2) the upper furnace zone, above the secondaryair level where fast fluidization occurs and the flow pat-tern approaches almost plug flow and particles from 500to 150 µm are distributed in the circulating zone; and the(3) cyclone, where the material coarser than the cut pointof the cyclone is recycled back into the combustor. Thelarger particles remain in the system for an extended pe-riod of time, ranging from hours for the dense phase inthe lower zone to several minutes for the circulating phasein the upper zone of the furnace. Some of the fines makeonly one pass through the combustor and exit the cyclone,so that their residence time is similar to that of the gas.

The heating rate of a coal particle in an FBC boiler isapproximately 103–104◦C/sec depending on the particlesize and the total average residence time of a particle inthe combustion chamber is estimated to be 20 min.

Some of the advantages of CFB boilers include fuelflexibility (any fuel such as coal, peat, wood chips, agri-cultural waste, Refuse Derived Fuel (RDF), and oil, withsufficient heating value to rise the fuel and the combus-tion air to above the ignition temperature of the fuel can besuccessfully burnt in a CFB boiler), high combustion effi-ciency, smaller furnace cross section: good load followingcapabilities, higher turn down, and more environmentallyfriendly.

A. Behavior of Ash in Boiler Furnaces

The mineral matter in coal probably is the factor most re-sponsible in considering coal a “dirty” fuel as comparedto oil or gas. Because of the dominant role played by coalmineral matter in boiler performance, the need for a betterunderstanding by R&D on the behavior of this impurity isnow more important than ever. The most commonly ob-served problems associated with mineral matter in coal areslagging and fouling. Slagging is the most serious problemwith pulverized-coal-fired steam generators, particularlywall slagging, the buildup of slag deposits on water walltubes. This has several undesirable aspects:

1. It decreases heat transfer appreciably.2. It can interfere with gas flow when massive deposits

accumulate near burners.3. It results in masses of slag that can cause physical

damage to the furnace bottom when they becomedislodged.

4. It plugs normal ash handling facilities withexcessively large masses of bottom ash.

5. It provides a cover to encourage tube wastage byexternal corrosion.

The standard methods that are used to predict the slaggingand fouling behavior depend on the ash analysis, whichis reported as oxides of various inorganic elements. Slag-ging characteristics of coal ash can be predicted to someextent using the silica percentage. Generally, ash with alow silica percentage will cause the most problems withslagging, and coals with a high silica percentage will bethe least troublesome. Thus coal benefeciation to controlsilica percentage provides one way of modifying slaggingbehavior, although the effectiveness of coal washing willdepend on the chemical composition of the ash in theproduct. However, there are still a lot of factors that causeslagging (boiler operating parameters) that are not wellunderstood.

Fouling of superheaters and reheaters causes problemssuch as plugging of gas passages, interference with heattransfer, and erosion of metal tubes. Of the factors lead-ing to such deposits, the physical ones involving particle



262 Fuel Chemistry

motion, molecular diffusion, and inertial impaction are, tosome extent, controllable by the design of the unit. Thechemical factors associated with the ash, particularly thealkalies in the flue gas which are condensed on the sur-face of fly ash particles and the sintering characteristicsof these particles depend upon the mineral species in thecoal. Two alkalies, sodium and potassium, are generallyheld responsible for superheater and reheater fouling, withsodium given the most attention.

A fouling index has been developed to correlate sinter-ing strength more closely with total ash composition:

Fouling Index = Base/Acid × Na2O,

where the base is computed as Fe2O3 + CaO + MgO +Na2O + K2O, and the acid as SiO2 + Al2O3 + TiO2 is ex-pressed as weight percentage of the coal ash.

The fuel ash properties which boiler manufacturers gen-erally consider important for designing and establishingthe size of coal-fired furnaces include:

� The ash fusibility temperatures� The ratio of basic to acidic ash constituents� The iron/calcium ratio� The fuel–ash content (mg/MJ)� The ash firability

In addition to the furnace design, the ash content and char-acteristics also affect the superheater, reheater, and con-vective pass heat transfer surface design features. Thesecharacteristics along with a few others translate into therelative furnace sizes. It is recognized that the ash deposi-tion is much better understood if the inorganic constituentsin the coal are known rather than the ash composition af-ter oxidation of the minerals. With the advancement in thecomputing power and image analysis in the late 1980sand early 1990s, Computer Controlled Scanning Elec-tron Microscopy (CCSEM) has been a very useful toolin predicting ash behavior in furnaces. The determinationof organically associated inorganic constituents is beingmade by chemical fractionation. These advanced charac-terization methods are now being more effectively used tohelp understand and solve the ash depositional problems infurnaces.

VII. ENVIRONMENTAL ISSUESIN COAL COMBUSTION

A. Sulfur Dioxide Emissions

Coal contains sulfur and nitrogen which vary typicallybetween 0.5–5% and 0.5–2%, respectively. Sulfur uponcombustion forms sulfur dioxide. About 65% of the total

sulfur dioxide emissions are from electric utilities burn-ing fossil fuels. Sulfur, 97–99%, in the fuel forms SO2

and a small fraction of it is oxidized to SO3. The equi-librium level of SO3 in fuel lean combustion products isdetermined by the overall reaction

SO2 + 1/2 O2 <=> SO3.

The equilibrium constant (K p = 1.53 × 10 − 5 e 11,760/Tatm −1/2) indicates that the SO3 formation is not favor-able at combustion temperatures. SO3 concentration in theflue gas is typically 1–5% of the SO2.

The SO2 control techniques can be classified intoprecombustion, during combustion, and postcombustionmethods.

1. Precombustion Control Methods

Precombustion control technologies can be fuel switchingor fuel desulfurization. Since sulfur dioxide emissions aredirectly proportional to the amount of sulfur in the fuel,switching to a low sulfur fuel is a choice when permitted.However, fuel switching may not be an alternative if theregulations require a certain percentage of reduction inSO2 emissions regardless of the fuel sulfur content. Fueldesulfurization involves reduction of sulfur from the fu-els prior to combustion. Sulfur in coals is present in threeforms: pyritic, organic, and sulfatic. Pyritic sulfur is thesulfur that is present as a mineral pyrite (FeS2). The den-sity of pyrite is higher than coal. Therefore, when coal iscrushed and washed in water, the particles that are richin the pyrite sink and the particles that are rich in carbon(organic) float. The sinks are rejected and the floats withless sulfur are used as clean coal. The degree of desulfu-rization depends on the distribution of sulfur in the coal.The finer the pyrite particles are, the more difficult thecoal is to clean. These characteristics are of a coal that canbe evaluated by washability curves based on a standardfloat and sink analysis. The organic sulfur is present incoal bonded to the organic structure of the coal. This formof sulfur is finely distributed throughout the coal matrix,and therefore is not possible to remove by physical coalcleaning.

2. During the Combustion Control Method

In an FBC system, limestones or dolostones are introducedinto the combustion chamber along with the fuel. In thecombustion chamber, limestones and dolostones undergothermal decomposition, a process commonly known ascalcination. The decomposition of calcium carbonate, theprincipal constituent of limestone, proceeds according tothe following equation:



Fuel Chemistry 263

CaCO3 CaO + CO2(1)

MgCa(CO3)2 MgO + CaO + 2CO2.

Calcination, an endothermic reaction, occurs at tempera-tures above 760◦C. Some degree of calcination is thoughtto be necessary before the limestone can react with gaseoussulfur dioxide. Calcined limestone is porous in nature dueto the voidage (pores) created by the expulsion of carbondioxide.

Capture of the gaseous sulfur dioxide is accomplishedvia the following reaction, which produces a solid product,calcium sulfate:

CaO + SO2 + 1/2 O2 CaSO4. (2)

The reaction of porous calcium oxide with sulfur diox-ide produces a continuous variation in the physical struc-ture of the reacting solid as the conversion proceeds. Be-cause of the relatively high molar volume of CaSO4 ofCaO, the pore network within the reactant can be pro-gressively blocked as conversion increases. For pure CaOprepared by the calcination of reagent grade CaCO3, thetheoretical maximum conversion of CaCO3 to CaSO4 hasbeen calculated to be 57%. In practice, the actual conver-sion obtained using natural limestones is much lower dueto the nature of the porosity formed upon calcination. Cal-cium utilizations as low as 15–20 mol% have been reportedin some cases, although utilizations of about 30–40 mol%are typical. MgO will not react with sulfur dioxide at tem-peratures above 760◦C; therefore, the sulfation reaction ofdolomite is basically the reaction of sulfur dioxide withcalcium oxide.

In pressurized fluidized bed combustion, however, thepartial pressure of CO2 is so high that calcination doesnot proceed because of thermodynamic restrictions. Forexample, at 850◦C, calcium carbonate does not calcine ifthe CO2 partial pressure exceeds 0.5 atmosphere. Underthese conditions, the sulfation reaction is

CaCO3 (s) + SO2 + 1/2 O2 CaSO4 + CO2.

Increasing the pressure from one to five atmospheressignificantly increases the sulfation rate and calciumutilization.

3. Flue Gas Desulfurization

Dry sorbent use in the case of pulverized coal combustionunits is not efficient because of the operating tempera-ture of the combustion chamber. At higher temperatures(>1100◦C), CaO is known to sinter with a loss in theporosity and, therefore, the conversion. Also, at high tem-peratures, CaSO4 is not stable and decomposes to CaOand SO2 limiting this technology to FBC units.

Flue gas desulfurization systems in principle use al-kaline reagents to neutralize the SO2 and are classifiedas throwaway or regenerative types. This classificationis based on the product fate. While, in a throwaway pro-cess the product produced by absorbing medium is thrownaway (discarded), in a regenerative process, the SO2 is re-generated from the product.

Most of the flue gas desulfurization systems operatingin the United States use limestone or lime slurry scrubbing.In this system, limestone is finely ground (90% passingthough a 325-mesh screen or 45 µm) and made into slurry.This slurry is finely sprayed in the absorption (scrubber)column. This slurry absorbs SO2 in water as shown here

SO2 (g) + H2O <=> SO2 ·H2O (dissolving gaseous SO2)

SO2 ·H2O <=> H+ + HSO3−(hydrolysis of SO2).

This slurry with absorbed SO2 is sent to a retention tankwhere the precipitation of CaSO3, CaSO4 and unreactedCaCO3 occurs. Calcium carbonate has low solubility inwater. Low pH promotes dissolution of CaCO3 but low pHalso lowers solubility of SO2 in the scrubber. Therefore,a careful balance of pH is needed for this system. Thefollowing reactions take place in the retention tank, where

H+ + CaCO3 <=> Ca2+ + HCO3−1

(dissolution of limestone)

Ca2+ + HSO3− + 2H2O <=> CaSO3 ·2H2O + H+

(precipitation of calcium sulfate)

H+ + HCO3−1 <=> CO2 ·H2O(acid-base neutralization)

CO2 ·H2O <=> CO2 (g) + H2O (CO2 stripping),

the overall reaction being

CaCO3 + SO2 + 2H2O CaSO3 ·2H2O + CO2.

B. Nitrogen Oxides

Nitrogen oxides, NO and NO2, collectively known asNOx, are formed during combustion in three ways. About85–90% of the NOx emitted from the combustion cham-ber is NO and 5 or 10% as NO2. The three types of NOx

that form are by thermal, fuel, and prompt mechanisms.Nitrogen oxide emissions from coal combustion can

occur from three sources. Thermal NOx primarily formsfrom the reaction of nitrogen and oxygen in the combus-tion air. The Fuel NOx is a component that forms mainlyfrom the conversion of nitrogen in the fuel to nitrogenoxides. Prompt NOx is formed when hydrocarbon radicalfragments in the flame zone react with nitrogen to formnitrogen atoms, which then form NO.



264 Fuel Chemistry

No definite rules exist to determine which nitrogen ox-ide formation mechanism dominates for a given stationarycombustor configuration because of the complex interac-tions between burner aerodynamics and both fuel oxida-tion and nitrogen species chemistry. But in general, fuelnitrogen has been shown to dominate pulverized coal firedboilers, although thermal NO is also important in the post-flame regions where over-fire air is used. Thermal NO con-tributions only become significant at temperatures above2500◦F in coal flames. Prompt NO formation is not typ-ically an important mechanism during coal combustion.In the absence of fuel nitrogen, fuel NO is not a problemfor natural gas flames. Prompt NO, however, is impor-tant in the vicinity of the inlet burners where reacting fuelfragments mix with the oxidizing air. Thus, in natural gasburners, both prompt and thermal NO contribute to theformation of nitrogen oxides.

C. Thermal NO

The principal reactions governing the formation of thermalNO are

N2 + O NO + N

N + O2 NO + O.

These two reactions are usually referred to as thethermal-NO formation mechanism or the Zeldovich mech-anism. In fuel-rich environments, it has been suggestedthat at least one additional step should be included in thismechanism:

N + OH NO + H

The reactions (1), (2), and (3) are usually referred to as theextended Zeldovich mechanism. Experiments have shownthat as the complexity of the reactors and fuels increase, itis difficult to evaluate the rate forms. The general conclu-sion is that although the nonequilibrium effects are impor-tant in describing the initial rate of thermal-NO formation,the accelerated rates are still sufficiently low that very littlethermal NO is formed in the combustion zone and that themajority is formed in the postflame region where the res-idence time is longer.

D. Prompt NO

Prompt NO occurs by the fast reaction of hydrocarbonswith molecular nitrogen in fuel-rich flames. This Fenni-more mechanism accounts for rates of NO formation in theprimary zone of the reactor, which are much greater thanthe expected rates of formation predicted by the thermalNO mechanism alone. NO measurements in both hydro-carbon and nonhydrocarbon flames have been interpreted

to show that prompt NO is formed only when hydrocarbonradicals are available. It has been shown that the amountof NO formed in the fuel-rich systems is proportional tothe concentration of N2 and also to the number of car-bons present in the gas phase. Hence, this mechanism ismuch more significant in fuel-rich hydrocarbon flames orin the reburning flames. Two reactions are believed to bethe most significant to this mechanism:

CH + N2 HCN + N (3)

C + N2 CN + N. (4)

Reaction (4) was originally suggested by Fenimore in1971. Reaction (5) is considered to be only a minor, butnonnegligible, contributor to prompt NO with its impor-tance increasing with increasing temperature.

E. Fuel NO

Fuel NO is by far the most significant source of nitric ox-ide formed during the combustion of nitrogen-containingfossil fuels. Fuel NO accounts for 75 to 95% of the totalNOx accumulation in coal flames and greater than 50% infuel oil combustors. The reason for fuel NO dominancein coal systems is because of the moderate temperatures(1500–2000 K) and the locally fuel-rich nature of mostcoal flames. Fuel NO is formed more readily than ther-mal NO because the N H and N C bonds common infuel-bound nitrogen are weaker than the triple bond inmolecular nitrogen which must be dissociated to producethermal NO.

The main step, at typical combustion temperatures, con-sists of conversion of fuel nitrogen into HCN, step whichaccording to some investigators (Fenimore, 1976; Reeset al., 1981) is independent of the chemical nature of theinitial fuel nitrogen. Once the fuel nitrogen has been con-verted to HCN, it rapidly decays to NHi (i = 1, 2, 3) whichreacts to form NO and N2.

F. NOx Control Technologies

In order to comply with the regulations for nitrogen ox-ides emissions, various abatement strategies have been de-veloped. The most common methods used for NOx con-trol are air staging, fuel staging, flue gas recirculation,selective noncatalytic reduction (SNCR), or selective cat-alytic reduction (SCR). A determination of the most ef-fective and least expensive abatement technique dependson specific boiler firing conditions and the emission stan-dards. The principle of air staging is mainly reduction thelevel of available oxygen in zones where it is critical forNOx formation. By doing so, the amount of fuel burntat the peak temperature is also reduced. Air staging is



Fuel Chemistry 265

accomplished in the furnace by splitting the air stream forcombustion. A part of the combustion air is introduceddownstream as over-fire air (OFA), intermediate air (IA),or over-burner air (OBA). The principle of fuel staging orreburning involves reduction of the NOx already formedin the flame zone by reducing it back to nitrogen duringcombustion. This is usually accomplished by injecting fuelinto a second substoichiometric combustion zone in orderto let the hydrocarbon radicals from the secondary fuel re-duce the NOx produced in the primary zone. To completethe combustion process of the reburn fuel, additional airis introduced downstream. However, the main drawbackof this system is the short residence time available for thereburn fuel for complete oxidation. Therefore, this methoduses mostly natural gas or any other highly reactive fuelso that combustion can be completed within the residenceavailable.

In SNCR, chemicals are injected into the boiler, whichthen react with NOx and reduce it to N2.

Most commonly used chemicals are ammonia or urea.Other chemicals used in research work include amines,amides and amine salts, and cyanuric acid. Good mixingis essential for the success of this process and the optimumtemperature window at which the reactions take place is900–1100◦C. This happens to be in a region where theheat transfer surfaces are present. Achieving good mixingis difficult.

When the limits of NOx cannot be met by combustioncontrol or by SNCR, SCR methods are used. NOx concen-tration in the flue gas is reduced by injection of ammoniain the presence of a catalyst. The use of a catalyst reducesthe optimum temperature window. The reaction productsare water and nitrogen and this reaction is accomplished atmuch lower temperatures than SNCR (between 300 and400◦C). At lower temperatures, the unreacted ammoniacan react with sulfur trioxide in the presence of waterto form ammonium bisulfate (NH4HSO4), a sticky com-pound, which can cause corrosion, fouling, and blockingof downstream equipment.

This temperature is usually high enough to prevent con-densation of (NH4HSO4). This system can reduce NOx

emissions by about 90–95%.Most of the electric utilities are installing SCR systems

to meet the current and future NOx emission regulations.

VIII. COAL GASIFICATION

Gaseous fuels are easy to handle and use. They produceless by-products. As naturally occurring gaseous and liq-uid fuel resources deplete, synthetic gases and liquids pro-duced from coal can act as suitable replacements. Coalgasification is a process of conversion of solid coal into

combustible gases such as synthetic gas mixture (CO andH2) or by methanation reaction conversion to syntheticnatural gas (SNG). This is usually done by thermal de-composition, partial combustion with air or oxygen, andreaction with steam or hydrogen or carbon dioxide. Coalcontains impurities such as sulfur, nitrogen, and minerals,which generate pollutants when burned. Most of these im-purities and pollutant gases can be removed during the pro-cess of converting the coal into synthetic fuels. Therefore,gasification of coal is desirable from an environmentalviewpoint and to overcome the transportation constraintsassociated with solid fuels. While the goal of combustionis to produce the maximum amount of heat possible byoxidizing all of the combustible material, the goal of gasi-fication is to convert most of the combustible solids intocombustible gases (such as CO, H2, and CH4) with thedesired composition and heating value.

A. Gasification Reactions

During gasification, coal initially undergoes devolatiliza-tion (thermal decomposition) and the residual char under-goes some or all of the reactions listed here.

solid–gas

2C + O2 = 2CO (partial combustion)

C + O2 = CO2 (combustion)

C + CO2 = 2CO (Boudward reaction)

C + H2O = CO + H2 (water gas)

C + 2H2 = CH4 (Hydrogasification)

gas–gas

CO + H2O = CO2 + H2 (Shift)

CO + 3H2 = CH4 + H2O

For the thermodynamic and kinetic considerations, coalchar is normally considered to be 100% carbon. The heatsof reactions of the combustion or partial combustion reac-tions are exothermic (and fast), whereas most other gasi-fication reactions (boudward, water gas, and hydrogasi-fication) are endothermic and relatively slower. Usually,the heat requirement for the endothermic gasification re-actions is met by partial combustion of some of the coal.It is also important to note that gasification reactions aresensitive to the temperature and pressure in the system.Equilibrium considerations for the reversible gasificationreactions show that high temperature and low pressure aresuitable for the formation of most of the gasification prod-ucts, except for methane. Methane formation is favoredat low temperature and high pressure. High temperaturestend to reduce methane, carbon dioxide, and water vapor



266 Fuel Chemistry

in the products. Coal char is considered to be a mixtureof pure carbon and some inorganic impurities and struc-tural defects. Certain impurities and structural defects areknown to be catalytic; the absolute reaction rate dependson the amount and nature of the impurities and structuraldefects and also on physical characteristics such as surfacearea and pore structure. The physical structure controls theaccessibility of gasification medium to the interior sur-face. These physical structural characteristics depend onthe feed coal and the devolatilization conditions (heatingrate and peak temperature).

There has been considerable interest in gasifying coalswith several catalysts in order to minimize secondary re-actions and produce the desired product gas distribution.However, the major problem with using catalysts is poi-soning by various hydrocarbons and other impurities incoal. A major research effort in this area has been di-rected toward the gasification of carbon using alkali metalscompounds. Potassium chloride and carbonate have beenfound to increase the amount of carbon gasified to CO andH2. The same catalysts are found to reduce methane yield.

B. Classification of Gasification Processes

Gasification processes are primarily classified accordingto the operating temperature, pressure, reactant gas, andthe mode of contact between the reactant gases and thecoal/char. The operating temperature of a gasifier, usuallydictates the nature of the ash removal system. Operat-ing temperatures below 1000◦C allow dry ash removal,whereas temperatures between 1000 and 1200◦C causethe ash to partially melt, become “sticky,” and form ag-glomerates. Temperatures above 1200◦C result in melt-ing of the ash and it is removed mostly in the form ofliquid slag. Gasifiers may operate at either atmosphericor elevated pressure. Both temperature and pressure af-fect the composition of the final product gases. Gasifica-tion processes use one or a combination of three reactantgases: oxygen, steam, and hydrogen. The heat requiredfor the endothermic reactions (heat absorbing) is suppliedby combustion reactions between the coal and oxygen.Table IV illustrates the effect of gasification medium onthe product species and the calorific value. Methods ofcontacting the solid feed and the gaseous reactants in agasifier are of four main types: fixed bed, fluidized bed,entrained flow, and molten bath. The operating principleof fixed bed, fluidized bed, and entrained flow systemsis similar to that discussed for combustion systems (seeprevious section). The molten bath approach is similar tothe fluidized bed concept in that reactions take place ina molten medium (either slag or salt) with high thermalinertia and the medium both disperses the coal and acts asa heat sink, with high heat transfer rates, for distributing

TABLE IV Effect of Gasification Medium on Products andCalorific Value

Gasification Calorific value of themedium Products products (MJ/m3)

Air and steam N2, CO, H2 and CO2 5.6–11.2

Oxygen and steam CO, H2 11.2–14.5

Hydrogen CH4 35–38

the heat of combustion. Table V summarizes some impor-tant gasification processes, conditions, and product gascompositions.

C. Major Gasification Processes

The most important fixed-bed gasifier available commer-cially is the Lurgi Gasifier. It is a dry-bottom, fixed-bedsystem usually operated between 30 and 35 atmospherespressure. Since it is a pressurized system, coarse-sizedcoal (25–37 mm) is fed into the gasifier through a lockhopper from the top. The steam–oxygen mixture (gasify-ing medium) is introduced through the grate located in thebottom of the gasifier. The coal charge and the gasifyingmedium move in opposite directions (counter-currently).The gasifier is operated at about 980◦C and the oxygen re-acts with coal to form carbon dioxide, thereby producingheat to sustain the endothermic steam–carbon and carbondioxide–carbon reactions. The raw product gas consist-ing mainly of carbon monoxide, hydrogen, and methaneleaves the gasifier for further cleanup. Besides participat-ing in the gasification reactions, steam prevents high tem-peratures at the bottom of the gasifier so as not to sinter ormelt the ash. Therefore, this gasification system is mostsuitable for highly reactive coals. Hot ash is periodicallyremoved through a lock hopper at the bottom. Large com-mercial gasifiers measuring about 4 m in diameter and6.3–8.0 m in height are capable of gasifying about 50 tonsof coal per hour. Improved versions of the Lurgi Gasifierhave been developed but not yet commercialized.

The Winkler gasifier is a fluidized-bed gasification sys-tem, which operates at atmospheric pressure. In this gasi-fier, crushed coal is fed using a screw feeder and isfluidized by the gasifying medium (steam–air or steam–oxygen mixture depending on the desired calorific valueof the product gas) entering through a grate at the bottom.The coal charge and the gasification medium move cocur-rently (in the same direction). In addition to the main gasi-fication reactions taking place in the bed, some may alsotake place in the freeboard above the bed. The temperatureof the bed is usually maintained at 980◦C (1800◦F) and theproduct gas consists primarily of carbon monoxide and hy-drogen. The low operating temperature and pressure limitthe throughput of the gasifier. Because of the low operating



Fuel Chemistry 267

TAB

LE

VS

um

mar

yo

fS

om

eG

asifi

cati

on

Pro

cess

es

Pro

duct

gas

com

posi

tion

b(d

ryba

sis)

Pro

duct

gas

Tem

pera

ture

Pre

ssur

eG

asifi

cati

onT

ype

ofca

lori

ficva

lueb

Pro

cess

(◦C

)(P

sia)

med

ium

coal

feed

CO

CO

2H

2C

H4

C2H

6O

ther

s(B

tu/s

cf)

Com

men

ts

Lur

gia

980

400

O2+

stea

mN

onca

king

68.5

29.5

40.4

9.4

—1.

230

2

Win

kler

a81

515

O2+

stea

mA

nyco

al33

.420

.541

.93.

1—

1.1

275

Synt

hane

980

1000

O2+

stea

mA

nyco

al16

.728

.927

.824

.50.

81.

340

5

Hyg

asst

eam

–oxy

gen

980

1000

O2+

stea

mA

nyco

al23

.824

.530

.218

.60.

62.

337

4

Kop

pers

–Tot

zeka

1510

20O

2+

stea

mA

nyco

al55

.86.

236

.6—

—1.

429

8

2+

stea

mA

nyco

al21

.529

.332

.115

.61.

536

7

Texa

co14

8060

0O

2A

nyco

alin

46.6

11.5

38.7

0.7

2.7

300

the

form

ofco

al–w

ater

slur

ry

Cog

as98

075

Stea

m,a

ndai

ras

Any

coal

7.4

9.3

26.2

34.0

8.3

14.8

726

heat

sour

ce

Prod

ucer

980

15A

ir+

stea

mA

nyco

al29

.04.

012

.03.

0—

52.0

130

Wat

erga

s98

015

Air

+st

eam

Any

coal

415.

049

.00.

5—

4.5

300

Wes

tingh

ouse

1930

20A

ir+

stea

mA

nyco

al19

.29.

414

.42.

854

.214

0O

xyge

nca

nal

sobe

used

U-g

as10

4035

0A

ir+

stea

mA

nyco

al9.

812

.010

.3—

67.9

75

Com

bust

ion

1930

20A

irA

nyco

al24

.44.

110

.7—

—12

5O

xyge

nca

nal

sobe

used

engi

neer

ing

60.6

CO

2ac

cept

or81

515

0A

ir+

stea

mL

owra

nk17

.06.

653

.820

.90.

444

0

1.0

Hyd

rane

815

>10

00H

2A

nyco

al3.

9—

22.0

73.2

——

826

Kel

log

salt

930

1200

O2+

stea

mA

nyco

al33

.513

.345

.07.

5—

0.7

348

Oxy

gen

can

also

beus

ed

aC

omm

erci

ally

avai

labl

epr

oces

ses.

bC

ompo

sitio

nva

ries

with

coal

used

.



268 Fuel Chemistry

temperatures, lignites and subbituminous coals that havehigh ash fusion temperatures are ideal feedstocks for thistype of gasifier. Units capable of gasifying 40–45 tons perhour are commercially available.

The Koppers–Totzek gasifier has been the most success-ful entrained-flow gasifier. This process uses pulverizedcoal (usually less than 74 µm) entrained (blown) into thegasifier by a mixture of steam and oxygen. The gasifier isoperated at atmospheric pressure and high temperaturesof about 1600–1900◦C. The coal dust and gasificationmedium flow cocurrent (in the same direction) in the gasi-fier and because of the small coal particle size, the particleresidence time is approximately 1 sec. Although this resi-dence time is relatively short, high temperatures enhancethe reaction rates, and therefore almost any coal can begasified in this system. Tars and oils are evolved at mod-erate temperatures and crack at higher temperatures, andtherefore, there is no condensible tarry material in theproducts and the ash melts and flows as slag in the K–Tgasifier. Gasifiers with two diametrically opposite nozzles(also called heads) are most commonly used. However,the use of four nozzles doubles the throughput (over thetwo, nozzle design) and such configurations are also inuse. The product gas is mainly synthesis gas (a mixtureof CO and H2) and is primarily used for ammonia man-ufacture. Since no heavy-duty moving parts are involvedin this system, maintenance is minimal and availability isexpected to be high.

D. Advanced Coal Gasification Systems

Many attempts have been made to improve the first-generation commercial gasifiers. British Gas Corporationhas converted the Lurgi gasifier into a slagging type byincreasing the operating temperature and thereby accom-modating higher-rank coals which require higher temper-atures for complete gasification. Another version of theLurgi gasifier is the Ruhr-100 process with operating pres-sures about three times higher than that of the basic Lurgiprocess. Developmental work on the Winkler gasificationprocess has lead to a pressurized version called the pressur-ized Winkler Process with an aim of increasing the yield ofmethane to produce Synthetic Natural Gas (SNG). Otherprocesses in the developmental stages are the U-gas, Hy-Gas, Cogas, Westinghouse, and Synthane processes. TheTexaco gasification system appears to be most promisingentrained bed gasification system that has been developed.In this system coal is fed into the gasifier in the form ofcoal–water slurry and the water in the slurry serves asboth a transport and a gasification medium. This systemoperates at 1500◦C and the ash is removed as molten slag.Experience on demonstration units has indicated that theprocess has a potential to be used with combined-cycle

plants for power generation. Another entrained bed gasifi-cation process under development is a pressurized versionof the K–T process called the Shell–Koppers system.

These developing processes are primarily aimed at ei-ther increasing the operating pressure to increase thethroughput and provide pressurized product gas for ad-vanced power systems or increasing the operating tem-peratures to accommodate a variety of fuels, or both.

IX. COAL LIQUEFACTION

Similar to coal gasification, coal liquefaction is the processof converting solid coal into liquid fuels. The main differ-ence between naturally occurring petroleum fuels and coalis the hydrogen to carbon molar ratio. Petroleum has ap-proximately a hydrogen to carbon ratio of 2 where as coalhas about 0.8. Coal also contains higher concentrationsof oxygen, nitrogen, and significantly higher amounts ofmineral matter than petroleum. Therefore, conversion ofcoal into liquid fuels involves hydrogenation (additionof hydrogen either directly or indirectly). Direct hydro-genation either from gaseous hydrogen or from a hydrogendonor solvent is termed direct liquefaction. If the hydro-gen is added indirectly through an intermediate series ofcompounds, the process is called indirect liquefaction. Indirect liquefaction processing, the macromolecular struc-ture of the coal is broken down ensuring that the yieldof the correct size of molecules is maximized and that theproduction of the very small molecules that constitute fuelgases is minimized. In contrast, indirect liquefaction meth-ods break down the coal structure all the way to a synthesisgas mixture (CO and H2), and these molecules are usedto rebuild the desired liquid hydrocarbon molecules. Thiscan be achieved by a variety of gasification techniques asdiscussed in the Coal Gasification section.

A. Liquefaction Reactions

Since coal in a complex substance it is often representedby an average composition and the reactions occurringduring direct and indirect liquefaction can be illustratedas follows.

1. Direct Liquefaction

CH0.8S0.2O0.1N0.01 + H2 −→ (RCHx)

+ CO2, H2S, NH3, H2O.

2. Indirect Liquefaction

CH0.8S0.2O0.1N0.01 −→ CO + H2

+ H2S, NH3, H2O (Gasification)



Fuel Chemistry 269

CO + 2H2 −→ CH3OH (Methanol Synthesis)

2CH3OH −→ CH3OCH3 + H2O

(Methanol-to-Gasoline)

nCO + 2nH2 −→ ( CH2 )n + H2O

(Fischer–Tropsch Synthesis)

nCO + (2n + 1)H2 −→ ( CH2 )n+1 + H2O.

Direct liquefaction of coal can be achieved with and with-out catalysts using pressures of 250 to 700 atmospheresand temperatures ranging between 425 and 480◦C. In theindirect liquefaction process, coal is gasified to producesynthesis gas and cleaned to remove impurity gases andsolids. The processes used to clean the gases depend onthe impurities.

The principal variables that affect the yield and distri-bution of products in direct liquefaction are the solventproperties such as stability and hydrogen transfer capa-bility, coal rank and maceral composition, reaction condi-tions, and the presence or absence of catalytic effects. Bi-tuminous coals are the most suitable feed-stock for directliquefaction as they produce the highest yields of desir-able liquids, although most coals (except anthracites) canbe converted into liquid products. Medium rank coals arethe most reactive (react fast) under liquefaction conditions.Among the various petrographic components, the sum ofthe vitrinite and liptinite maceral contents correlates wellwith the total yield of liquid products.

3. Liquefaction Processes

The Bergius process was the first commercially availableliquefaction process. It was developed during the FirstWorld War and involves dissolving coal in a recycled sol-vent oil and reacting with hydrogen under high pressuresranging from 200 to 700 atmospheres. An iron oxide cat-alyst is also employed. The temperatures in the reactorwere in the range of 425–480◦C. Light and heavy liquidfractions are separated from the ash to produce gasolineand recycle oil, respectively. In general, 1 ton of coal pro-duces about 150–170 liters of gasoline, 190 liters of dieselfuel, and 130 liters of fuel oil. Separation of ash and heavyliquids, and erosion due to cyclic pressurization, posed dif-ficulties which caused the process to be taken out of useafter the war.

In the first generation, in a commercially operated, in-direct liquefaction process called Fischer–Tropsch syn-thesis, coal is gasified first using the high-pressure Lurgigasifier and the synthesis gas is reacted over an iron-basedcatalyst either in a fixed-bed Arge reactor or a fluidized-bed Synthol reactor. Depending on the reaction conditions,the products obtained consist of a wide range of hydro-carbons. Although this process was developed and used

widely in Germany during the Second World War, due topoor economics it was discontinued after the war. Thisprocess is used in South Africa (Sasol) for reasons otherthan economics.

4. Developments in Liquefaction Processes

Lower operating temperatures are desirable in direct liq-uefaction processes since higher temperatures tend to pro-mote cracking of molecules producing more gaseous andsolid products at the expense of liquids. Similarly, lowerpressures are desirable from an ease and cost-of-operationpoint of view. Recent research efforts in the area of directliquefaction have concentrated on reducing the operatingpressure, improving the separation process by using a hy-drogen donor solvent (Consol Synthetic Fuels Process),operation without catalysts (Solvent Refined Coal (SRC)),and by using a solvent without catalysts (Exxon DonorSolvent) but using external catalytic rehydrogenation ofthe solvent. Catalytic effects in liquefaction are due to theinherent mineral matter in the coal and to added catalysts.Recent research efforts have focused on the area of multi-stage liquefaction to minimize hydrogen consumption andmaximize overall process yields.

Later versions of Sasol plant (Sasol 2 and 3 units whichalso use indirect liquefaction process) have used only syn-thol reactors to increase the yield of gasoline and havereacted excess methane with steam to produce more COand H2.

Recent developments include producing liquid fuelsfrom synthesis gas through an intermediate step of con-verting the synthesis gas into methanol at relatively lowoperating pressures (750–1500 psi) and temperatures(205–300◦C). Methanol is then converted into a range ofliquid hydrocarbons. The use of zeolite catalysts (as de-veloped by Mobil) has enabled the direct production ofgasoline from methanol with high efficiency.

X. LIQUID FUELS

Liquid fuels are obtained by refining naturally occurringcrude oil. Like coal, crude oil from different places candiffer in composition because of the precursor materi-als and the conditions for transformation organic matterto crude oil. Crude oil is a complex naturally occurringliquid containing mostly hydrocarbons and some com-pounds containing N, S, and O atoms. Crude oil consists ofparaffins (straight-chain and branched-chain compounds),naphthenes (cyclo paraffins), and aromatics (benzene andits derivatives).

The average composition of crude oils from variousparts of the world does not vary significantly. However,because of the variations in viscosity, density, sulfur, and



270 Fuel Chemistry

TABLE VI Typical Yield from a Barrel of Gasoline

Product Yield (gallons)

Gasoline 19.5

Distillate fuel oil 9.2

Kerosene 4.1

Residual fuel oil 2.3

Lubricating oil, asphalt, wax 2

Chemicals for use in manufacturing 2(petrochemicals)

boiling points, these are separated into different fractionsin a refinery. The most common refining operations aredistillation, cracking, reforming alkylation, and coking.The demand for various products changes with the seasonand the lifestyle of the society. Typical yield from a barrelof crude oil is shown in Table VI.

XI. PROPERTIES FOR UTILIZATION

The majority of products (first four) are burnt in various de-vices. Gasoline is a mixture of light distillate hydrocarbonswith a boiling range of 25–225◦C consisting of paraffins,olefins, naphthenes, aromatics, oxygenates, lead, sulfur,and water. The exact composion varies with the sea-son and geographic location. During summer months lowvolatile components are added to reduce the vapor pres-sure, whereas in winter months low boiling componentsare added to make it more volatile. Under the 1990 CleanAir Act Amendments (CAAA), the U.S. EnvironmentalProtection Agency (EPA) developed reformulated gaso-line (RFG) to significantly reduce vehicle emissions ofozone-forming and toxic air pollutants. RFG is requiredto be used in the nine major cities with the worst ozoneair pollution problems. Similar to normal gasoline, RFGwill contain oxygnates. Oxygenates increase the combus-tion efficieny and reduce emission of carbon monoxide.Table VII provides a comparison of properties for variousgasolines.

Diesel fuel is also a mixture of light distillates but withhigher boiling point components with a boiling point rangeof 185–345◦C consisting of lower volatile and more vis-cous compounds. The average molecular weight is ap-proximately 200.

Kerosene fuels are used in jet engines. Kerosene fu-els have a wide range of boiling points. The aromatics inkerosene are limited due to their tendency to form soot.

Residual fuel oils are classified into five categoroies.Some of the important properties are listed here for goodatomization.

1. Specific gravity—ratio between the weight of anyvolume of oil at 60◦F to the weight of equal volumeof water at 60◦F

2. Viscosity—a measure of resistance to motion of afluid

3. Flash point—The temperature at which the vaporsgenerated “flash” when ignited by external ignitionsource

4. Pour point—The temperature at which the oil ceasesto flow when cooled under prescribed conditions.

A. Combustion of Liquid Fuels

Fuel oil-fired furnaces, diesel engines, and distillate fuel-fired gas turbines utilize fine liquid sprays to increase therate of evaporation and combustion rate of the fuel. In gen-eral the combustion of a liquid fuel takes place in a seriesof stages: atomization, vaporization, mixing of the vaporwith air, ignition, and maintenance of combustion (flamestabilization). Recent advances have shown the atomiza-tion step to be one of the most important stages of liquidfuel combustion. The main purpose of atomization is toincrease the surface area to volume ratio of the mixture.For example, breaking up of a 3-mm droplet into 30-µmdrops results in 106 droplets. This increases the burningrate by 10,000 times. The finer the atomization spray thegreater the subsequent benefits are in terms of mixing,evaporation, and ignition. The function of an atomizer istwofold: atomizing the oil and matching the momentumof the issuing jet with the aerodynamic flow in the furnace.

The atomizers for larger boiler burners are usually ofthe swirl pressure jet or internally mixed two fluid types,producing hollow conical sprays. Less common are theexternally mixed two fluid types. The principal consider-ations in selecting an atomizer for a given application areturn-down performance and auxiliary costs.

There are differences in the structures of the sprays be-tween atomizer types which may affect the rate of mixingof fuel droplets with the combustion air and hence theinitial development of a flame.

For distillate fuels of moderate viscosity, (30 mm2

sec−1) at ordinary temperatures, a simple pressure atom-ization with some type of spray nozzle is most commonlyused. Operating typically with a fuel pressure of 700–1000 kPa (7–10 atm) such a nozzle produces a distri-bution of droplet diameters from 10 to 150 µm. Theyrange in design capacity of 0.5–10 or more, cm3 sec−1.A typical domestic oil burner nozzle uses about 0.8 cm3

sec−1 of No. 2 fuel oil at the design pressure. Althoughpressure-atomizing nozzles are usually equipped with fil-ters, the very small internal passages and orifices of thesmallest tend to be easily plugged, even with clean fuels.With decreasing fuel pressure the atomization becomes



Fuel Chemistry 271

TABLE VII Properties of Various Types of Gasoline

Fuel parameter values (national basis)

GasoholConventional gasoline (2.7 wt% oxygen) Oxyfuel Phase I RFG

Avga Rangeb Avg Avg Avg

RVP3 (psi) 8.7-S 6.9-15.1 9.7-S 8.7-S 7.2/8.1-S

11.5-Wc 11.5-W 11.5-W 11.5-W

T50 (◦F) 207 141–251 202 205 202

T90 (◦F) 332 286–369 316 318 316

Aromatics (vol%) 28.6 6.1–52.2 23.9 25.8 23.4

Olefins (vol%) 10.8 0.4–29.9 8.7 8.5 8.2

Benzene (vol%) 1.60 0.1–5.18 1.60 1.60 1.0

(1.3 max)

Sulfur (ppm) 338 10–1170 305 313 302

(500 max)

MTBE4d (vol%) — 0.1–13.8 — 15 11

(7.8–15)

EtOH4 (vol%) — 0.1–10.4 10 7.7 5.7

(4.3–10)

a As defined in the Clean Air Act.b 1990 MVMA survey.c Winter (W) higher than Summer (S) to maintain vehicle performance.d Oxygenate concentrations shown are for separate batches of fuel; combinations of both MTBE and

ethanol in the same blend can never be above 15 vol% total.

progressively less satisfactory. Much higher pressures of-ten are used, especially in engine applications, to produce ahigher velocity of liquid relative to the surrounding air andaccordingly smaller droplets and evaporation times. Othermechanical atomization techniques for production of moremonodisperse sprays or smaller average droplet size (spin-ning disk, ultrasonic atomizers, etc.) are sometimes usefulin burners for special purposes and may eventually havemore general application, especially for small flows.

Conventional spray nozzles are relatively ineffectivefor atomizing of fuels of high viscosity such as No. 6or residual oil (Bunker C) and other viscous dirty fuels.In order to transfer and pump No. 6 oil, it must usually beheated to about 373 K, at which its viscosity is typically40 mm2 sec−1. Relatively large nozzle passages and ori-fices are necessary for the possible suspended solids. Drysteam may also be used in a similar way, as is commonpractice in the furnaces of power plant boilers using resid-ual oil.

Combustion of fuel oil takes place through a series ofsteps, namely, vaporization, gasification, ignition, disso-ciation, and finally attaining the flame temperature. Va-porization or gasification of the fine spray of fuel dropletstakes place as a physicochemical process in the combus-tion chamber. The temperature of vaporization for fueloil is in the range of 100–500◦F, depending on the grade

of the fuel. Gasification takes place at about 800◦F. Thefinal flame temperature attained is between 2000 and3000◦F. The combustion of an oil droplet takes placein 2–20 msec depending on the size of the droplet. Atypical characteristic of an oil flame is its bright luminousnature, which is due to incandescent carbon particles inthe fuel-rich zone.

τ =d2

pp

β.

Figure 8 illustrates the combustion of a single liquiddroplet. Evaporation of liquid supplies the gaseous fuel

FIGURE 8 Evaporation and combustion of a liquid fuel droplet.



272 Fuel Chemistry

that burns in the gas phase. Evaporation is caused by heattransfer to the surface of the droplet. The time required forcomplete evaporation is given by

β = 8λ

ρl cpln (1 + BT ),

where, BT is transfer coefficient, lambda is thermalConductivity, ρl = liquid density, and cp is the heatcapacity.

XII. MAJOR COMBUSTION METHODS

Internal combustion engines are devices that producework using the hot combustion gases directly rather thansteam. Three major types of IC engines are commonlyused today using the top three products from a refinery.They are (1) spark engine (gasoline engine), (2) compres-sion engine (diesel engine), and (3) gas turbine (aircraftengines).

A. Spark or Gasoline Engines

The most common engine is a 4-stroke engine. During theintake stroke, the fuel and air mixture is drawn into thecylinder with the exhaust valve closed. Then the air andfuel mixture is compressed in a compression stroke. Atthe top of the stroke, the spark plug ignites the mixture.During the expansion or power stroke, the high-pressurecombustion gases expand moving the piston down anddelivering the power. The gases expand completely, theexhaust valve opens, and the gases are expelled out dur-ing the exhaust stroke. The fuel and air are atomized andpremixed in a carburetor. The higher the compression ratiothe higher the efficiency of the engine. However, highercompression ratios also require higher octane number fu-els. The octane number of a fuel is indicative of its an-tiknock properties. At equivalence ratios below 0.7 andabove 1.4, the mixtures are generally not combustible.The equivalence ratio changes as the power requirementchanges. For example, as the vehicle accelerates, hightorque and power are required for which a fuel-rich mix-ture is used, whereas when the vehicle is cruising at highspeeds the vehicle needs fuel lean mixtures. Therefore, inIC engines it is difficult to maintain the air to fuel ratioconstant. Combustion in IC engines takes place in bothoxygen-deficient and oxygen-rich environments, and theair and fuel mixtures are preheated by compression. Ev-ery time a fresh batch of fuel comes in flame is producedand quenched resulting in unsteady combustion. This re-sults in continuous changes in the pollutant generation andemission.

B. Diesel Engines

Diesel engines are also IC engines. However, in Dieselengines, there is no carburetor. Only air is compressed tomuch higher pressures and the fuel is injected into thecompressed air. As the fuel and air are mixed, the fuelevaporates and ignites (hence called compression igni-tion). The pressures used in the engines are almost twicethose of the gasoline engines. Rate of injection and mixingof fuel and air determine the rate of combustion. Dieselengines are classified based on fuel injection, direct in-jection (DI), and indirect injection (IDI). Fuel quality ismeasured by cetane number (CN).

C. Gas Turbines

Another class of internal combustion engine is the gas tur-bine. Air is compressed to high pressures (10–30 atm) ina centrifugal compressor. Fuel is sprayed into the primarycombustion zone where the fuel burns and increases thetemerpature of the gases. The gas volume increases withcombustion and the gases expand though a turbine. Thepower generated exceeds that required for the compres-sor. This drives the shaft to run an electric generator. Inthe aircraft applications, the gases are released at high ve-locity to provide the thrust. These systems are light weightcompared to land-based systems. Land-based systems useeither distillate oil or natural gas. Gas turbine-based powergeneration is used commonly to meet the peak power re-quirements rather than for base load operation.

D. Environmental Challengesfor Liquid Fuel Utilization

Carbon monoxide is present in any combustion gas fromany carbon containing fuel. The main factor that leads to itsformation is incomplete combustion and in the IC enginescontinuous change from fuel-lean to fuel-rich conditionsresults in large emissions of CO. More than 70% of theCO emitted in the United States is from the transportationsector. CO emissions are also a function of vehicle speed.At lower speeds the emissions are higher. Cold starts alsoto contribute to higher emissions of CO. Oxygenates inthe fuel aid complete combustion and result in a decreaseof CO emissions. Catalytic converters placed at the endof the exhaust pipe oxidize the CO catalytically at lowertemperatures.

Most of the hydrocarbon emissions are emitted throughthe exhaust. However, methane, ethane, acetylene, propy-lene, and aldehydes were found in the exhaust but werenot present in the fuel. It can be deduced that thesewere formed during combustion. A significant amount ofhydrocarbon emissions also come from the combustionchamber wall crevices and solid deposits. These hydro-carbon emissions reduce the NOx emissions. However,



Fuel Chemistry 273

with the increase in the air to fuel ratio, CO and hydrocar-bon emissions are reduced but NOx emissions increase.To reduce all the emissions, three-way catalysts are beingused. They not only oxidize CO and HCs but also reduceNOx to N2.

XIII. GASEOUS FUELS

A. Combustion of Gaseous Fuels

In any gas burner some mechanism or device (flame holderor pilot) must be provided to stabilize the flame againstthe flow of unburned mixture. This device should fix theposition of the flame at the burner port. Although gas burn-ers vary greatly in form and complexity, the distributionmechanism is fundamentally the same in most. By keep-ing the linear velocity of a small fraction of the mixtureflow equal or less than the burning velocity, a steady flameis formed. From this pilot flame, the main flame spreadsto consume the main flow at a much higher velocity. Thearea of the steady flame is related to the volume flow rateof the mixture:

Vmix = Af × Su

where,

Vmix = volumetric flow rateAf = area of the steady flameSu = burning velocity

The volume flow rate of the mixture is, in turn, propor-tional to the rate of heat input:

Q = Vmix × HHV

where,

Vmix = volumetric flow rateHHV = higher heating value of the fuel

Q = rate of heat input

In the simple Bunsen flame on a tube of circular crosssection, the stabilization depends on the velocity varia-tion in the flow emerging from the tube. For laminar flow(parabolic velocity profile) in a tube, the velocity at a ra-dius, r , is given by:

v = const(R2 − r2)

where,

v = laminar flow velocityR = tube radiusr = flame radiusconst = experimental constant

Most of the commercial gas–air premixed burners arebasically laminar-flow Bunsen burners and operate at at-mospheric pressure—i.e., the primary air is induced from

the atmosphere by the fuel flow with which it mixes in theburner passage leading to the burner ports where the mix-ture is ignited and the flame stabilized. The induced airflow is determined by the fuel flow-through momentumexchange and by the position of a shutter or throttle at theair inlet. Hence, the air flow is a function of fuel velocityas it issues from its orifice or nozzle, or of the fuel supplypressure at the orifice. With a fixed fuel flow, the equiv-alence ratio is adjusted by the shutter, and the resultinginduced air flow also determines the total mixture flow,since desired air–fuel volume ratio is usually 7 or more,depending on the stoichiometry. Burners of this generaltype with many multiple ports are common for domesticfurnaces, heaters, stoves, and for industrial use. The flamestabilizing ports in such burners are not always round andmaybe slots of various shapes conform to the heating task.

Atmospheric industrial burners are made for a heat re-lease capacity of up to 50 kJ/sec−1 and even despite theirvaried designs their principle of stabilization is basicallythe same as that for the Bunsen burner. In some the mix-ture is fed through a fairly thick-walled pipe or casting ofappropriate shape for the application and the desired dis-tribution of flame. The mixture issues from many smalland closely spaced drilled holes, typically 1–2 mm in di-ameter, and burns, as rows of small pilot flame, spark orheated wire, usually located near the first holes, to avoidaccumulation of the unburned mixture before ignition. Therate of total heat release for a given fuel–air mixture canbe scaled with the size and number of holes—e.g., for2-mm-diameter holes it would be 10–100 J/(hole) or ingeneral 0.3–3 kJ cm2 sec−1 of port area, depending onthe fuel. The ports may also be narrow slots, sometimespacked with corrugated metal strips, to improve the flowdistribution and lessen the tendency to flashback.

Gas burners that operate at high pressures are usuallyintended for much higher mixture velocity or heating in-tensity and the stabilization against blowoff must thereforebe enhanced. This can be achieved by a number of meth-ods such as (1) surrounding the main port with a numberof pilot ports and (2) using a porous diaphragm screen.

In order to achieve high local heat flow the port velocityof the mixture should be increased considerably. In burnersthat achieve stabilization by causing pilot ports, most ofthe mixture can be burned at a port velocity as high as100 Su to produce a long pencil-like flame, suitable foroperations requiring a high heat flux.


CARBON CYCLE • COMBUSTION • COAL PREPARATION •COAL STRUCTURE AND REACTIVITY • ENERGY FLOWS IN

ECOLOGY AND IN THE ECONOMY • ENERGY RESOURCES

AND RESERVES • FOSSIL FUEL POWER STATIONS—COAL



274 Fuel Chemistry

UTILIZATION • FUELS • INTEGRATED GASIFICATION

COMBINED-CYCLE POWER PLANTS • INTERNAL COMBUS-TION ENGINES • PETROLEUM REFINING • POLLUTION, AIR

BIBLIOGRAPHY

Elliot, M. A. (ed.) (1980). “Chemistry of Coal Utilization,” 2nd Supple-mentary Volume, Wiley, New York.

Borman, G. L., and Ragland, K. W. (1998). “Combustion Engineering,”WCB/McGraw-Hill, Boston, MA.

Lawn, C. J. (ed.), (1987). “Principles of Combustion Engineeringfor Boilers” Academic Press, Harcourt Brace Jovanovich, NewYork.

Turns, S. R. (2000). “An Introduction to Combustion: Concepts andApplications,” 2nd ed., Mc Graw-Hill Inc., New York.

Wen, C. Y., and Lee, E. S. (eds.) (1979). “Coal Conversion Technology,”Addison-Wesley, MA.

Heinsohn, R. J., and Kabel, R. L. (1999). “Sources and Control of AirPollution,” Prentice-Hall, Upper Saddle River, NJ.

Williams, A., Pourkashanian, M., Jones, J. M., and Skorupska, N. (2000).“Combustion and Gasification of Coal,” Applied Energy TechnologySeries, Taylor and Francis, New York.

P1: GOL Final Pages Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology EN007F-314 July 6, 2001 16:59

Heterocyclic ChemistryCharles M. MarsonUniversity College London

I. General Aspects of Heterocyclic SystemsII. Physical PropertiesIII. Chemical PropertiesIV. Synthesis of HeterocyclesV. Major Classes of Heterocyclic Compounds

GLOSSARY

Alkaloid Cyclic organic compound distributed in livingorganisms and containing nitrogen. Many alkaloids arebases of significant pharmacological activity.

Annelation Attachment, by synthesis, of a fragment to astructure of one or more rings to produce a bi- or poly-cyclic molecule. It can also refer to electronic changesthat may occur when a structure is annelated.

Betaine Zwitterion of net neutral charge containing non-adjacent anionic and cationic sites. A hydrogen atomis not bonded to the cationic site.

Cyclodehydration Ring-forming reaction that proceedswith loss of water.

Degradation Procedure by which a molecule is cleavedto fragments, which are often identified, thereby assist-ing structural elucidation of the parent molecule.

Delocalization energy Calculated additional bonding en-ergy that results when the assumption that electronsare constrained to isolated double bonds is removed.Corresponds to the resonance energy in valence bondtheory.

Electrophile Electron-pair acceptor, such as H+, AlCl3,or Br+.

Enantioselective A reaction that affords predominantlyone enantiomer. Enantiomers of a compound arerelated to each other as nonsuperimposable mirrorimages.

Pi-excessive Heterocycle that contains a greater densityof electronic charge than does a carbon atom of ben-zene. The converse holds for a pi-deficient heterocycle.

Mesoionic Five-membered heteroaromatic mesomericbetaine that cannot be satisfactorily represented by onepolar structure and that possesses a sextet of electronsassociated with the heterocyclic ring.

Nucleophile Electron-pair donor, such as RO− or R3N.Protonation Formal addition of a proton (H+).Quaternization Alkylation of a tertiary amine to give a

quaternary ammonium salt.Resonance structures Two or more structures of the

same compound that differ only in the pairing arrange-ment of the electrons. These structures have no differ-ing chemical or physical properties; their identity isdepicted by a double-headed arrow (↔) inserted be-tween them. Resonance structures are nearly synony-mous with canonical forms.

Tautomerism Structural isomers of different energiesthat are interconvertible by surmounting a low-energy

321

P1: GOL Final Pages


322 Heterocyclic Chemistry

barrier. Isomerization involves the migration of a groupor atom (especially H, referred to as proton tau-tomerism).

Ylid 1,2-Dipolar species, of net neutral charge, compris-ing a negatively charged carbon atom bonded to a pos-itively charged heteroatom.

Zwitterion Species of net neutral charge that containsseparate cationic and anionic sites.

HETEROCYCLIC COMPOUNDS possess a cyclicstructure with at least two different kinds of atoms inthe ring. By far the most common type, and that exclu-sively considered here, contains carbon together with oneor more heteroatoms. The most common heteroatoms arenitrogen, oxygen, and sulfur, but many other elements canact as heteroatoms, such as phosphorus, tin, and silicon.

Heterocyclic compounds are extensively distributed innature; many play crucial roles in the biochemistry of liv-ing organisms. For example, the genetic material deoxyri-bonucleic acid (DNA) contains a sequence of nitrogenheterocycles held together by hydrogen bonds across theheterocyclic rings. Many sugars exist in the form of five- orsix-membered oxygen heterocycles. Most members of thevitamin B group are nitrogen heterocycles. Many naturallyoccurring pigments, antibiotics, and alkaloids are hetero-cyclic compounds. Owing to great advances in organicsynthesis, many dyes, plastics, pharmaceuticals, pesti-cides, and herbicides contain heterocycles synthesized inthe laboratory and not necessarily found in nature.

I. GENERAL ASPECTS OFHETEROCYCLIC SYSTEMS

A. Comparison with Carbocyclic Compounds

Molecules of organic chemical compounds are built upfrom a framework of carbon (C) atoms and associatedhydrogen (H) atoms, to which oxygen (O) or other het-eroatoms may or may not be attached. Because each car-bon atom normally has four bonds to other atoms, carbon issaid to be tetravalent. Hydrocarbons, which consist solelyof carbon and hydrogen atoms, may be linear, branched,or cyclic. In the last case, such compounds are referred toas carbocyclic.

Cyclohexane (C6H12) is a typical carbocycle.

CH2

H2C

H2C

H2CCH2

CH2

cyclohexane

A convenient method of depicting organic structuresuses polygonal forms in which the carbon atoms reside atapices and are bonded to the appropriate number of hy-drogen atoms if no other symbols (e.g., O for oxygen) arepresent. Replacement of a CH2 group in cyclohexane by

CH2

H2C

H2C

H2CNH

CH2 NH

piperidine

NH (nitrogen is usually trivalent) gives a heterocycliccompound called piperidine. It is often useful to comparea heterocycle to its carbocyclic analog; conversely, anynumber of heterocycles can be drawn by replacing thecarbon atoms in carbocyclic compounds.

N N N

benzene

pyridine

Removal of hydrogen (H2) from a CH CH bond resultsin the formation of a C C bond, an unsaturated linkagereferred to as olefinic. An aromatic structure contains al-ternating single and double bonds in a ring, such as ben-zene. Formal replacement of a CH in benzene by N givesthe heterocycle known as pyridine. In general, the proper-ties of aromatic and heteroaromatic compounds are quitedifferent from those of their saturated (or partially satu-rated) counterparts. That is because a special stability isassociated with those pairs of electrons (the so-called pielectrons) in a ring. Consequently, aromatic compoundsare usually very stable and do not suffer addition of otheratoms across the ring, unlike most other unsaturated com-pounds. A circle placed within a ring is used to denote acyclic and delocalized array of six electrons in pi orbitals.

Saturated or partly unsaturated heterocycles usually re-semble the acyclic (noncyclic) analogs closely in physi-cal and chemical properties. Thus, piperidine has muchin common with the acyclic amine diethylamine. Sim-ilarly, ethyl vinyl ether has much in common with2,3-dihydrofuran.

NH

O O

diethylamine

2,3-dihydrofuran ethyl vinyl ether1

2

34

5

P1: GOL Final Pages


Heterocyclic Chemistry 323

B. Heteroaromaticity

A nitrogen atom in a ring can be neutral or can carry apositive or a negative charge. Oxygen and sulfur atomsin a ring either are in the neutral form or carry a positivecharge. For aromaticity to be observed in a heterocycle,electrons in an orbital perpendicular to the plane of thering must overlap with the pi orbitals of other atoms inthat ring. The lone pair of pyrrole is in such an orbital andthe bonding can be expressed in the language of valencebond theory.

Z Z Z NH+

_

pi orbitals of pyrrole

+_

valence bond representation of pyrrole Z=NH

Because substantial negative charge resides at all thering carbon atoms of pyrrole (see valence bond structures,or canonical forms), it is referred to as a pi-excessive het-erocycle. Owing to the excess negative charge at carbon,it is unsurprising that the chemistry of pi-excessive hete-rocycles is dominated by electrophilic attack.

Pyridine is more aromatic than pyrrole, but for quite dif-ferent reasons. Here, the three double bonds in the pyridinering are markedly delocalized, and this accounts for thehigh degree of aromaticity of pyridine (see Table I).

N N N N+

_

+

_+

_

valence bond representation of pyridine

However, the lone pair on nitrogen is in an orbital par-allel with the plane of the ring and hence cannot donatenegative charge into it. Because nitrogen is a stronglyelectronegative (electron-attracting) element, the net re-sult is to remove electronic charge from the ring, as justdepicted in the canonical forms of pyridine. For this rea-son, pyridine undergoes electrophilic attack only undervery forcing conditions and is referred to as a pi-deficientheterocycle.

Considerations similar to the above can be applied toall “aromatic” heterocycles, at least in principle. An in-teresting series is made up of the resonance energies of3,4-benzoheterocycles, which are substantially lower than

TABLE I Approximate Aromatic StabilizationEnergies for Benzene and Some HeteroaromaticRings (kcal/mole)

Benzene 36 Pyridine 31 Pyrazine 23

Pyrrole 20 Thiophene 25 Furan 18

Pyrazole 25 Imidazole 20 1,2,4-Triazole 20

those of the 2,3-analogs, which retain much of the aro-maticity associated with the benzene ring.

ZZ

3,4-benzoheterocycle 2,3-benzoheterocycle

The quantitative measurement and even the precise def-inition of aromaticity are problematic. One common mea-surement of aromaticity involves determining the differ-ences in energy, by combustion, between a heterocycleand its carbocyclic analog; examples of values are givenin Table I.

A second method involves measurement of magneticring currents induced in aromatic systems by externalmagnetic fields. A nuclear magnetic resonance spectrom-eter records the observed magnetic effect, which can berelated to the degree of aromaticity.

C. Nomenclature and Numberingof Heterocycles

Single heterocycles are extensively referred to by theirwell-known or “trivial” names, such as pyrrole, furan, andthiophene. Numbering of a monocyclic heterocycle, suchas thiazole,

NH

O S

pyrrole furan thiophene

starts at the heteroatom of highest priority (oxygen takingpriority over sulfur, which takes priority over nitrogen,etc.). Thus, morpholine and thiazine are numbered asshown.

O

NH

S

NH

5

morpholine 1,4-thiazine4

3

2

1

6

5

4

3

2

1

6

In all heterocyclic systems, the prefixes “oxa,” “thia,”and “aza” denote oxygen, sulfur, and nitrogen atoms, re-spectively. Also, two, three, or four identical heteroatomsbear the prefixes di, tri, and tetra, respectively, as in“dithia” and “triaza.” Partially saturated rings are indicatedby the prefix “dihydro” and “tetrahydro,” etc., and/or 1H -,2H -, etc., to identify saturated carbon(s) not taken intoaccount by the prefixes of hydrogenation. Examples aretetrahydrofuran and 2H -pyran.

O O1

2

34

2

2H-pyran1

5

3

4

5

6

tetrahydrofuran

P1: GOL Final Pages



TABLE II Names of Monocyclic Heterocyclesa

One nitrogen in ring One oxygen in ring

Ring Completely One double Completely One doublesize unsaturated bond Saturated unsaturated bond Saturated

3 Azirine Azirine Aziridine Oxirene Oxirene Oxirane

4 Azeteb Azetine Azetidine Oxetiumb Oxetene Oxetane

5 Pyrrole Pyrroline Pyrrolidine Furan Dihydrofuran Tetrahydrofuran

6 Pyridine Piperideine Piperidine Pyryliumb Dihydropyran Tetrahydropyran

7 Azepine Dihydroazepine Tetrahydroazepine Oxepine Dihydrooxepine Oxepane

a Parts of names italicized are endings common to heterocycles of the given ring size.b Cationic; other rings neutral.

Table II gives the common names of heterocyclic ringscontaining one nitrogen atom and, on the right-hand side,one oxygen atom in the ring. The complete name is that forthe nitrogen or oxygen heterocycle, whereas that part un-derlined is the suffix for any ring containing one nitrogenatom and, on the right-hand side, any ring without nitrogen(oxygen has been chosen here). All the compounds listedcontain trivalent nitrogen or divalent oxygen, except theoxetium and pyrylium rings, which contain cationic triva-lent oxygen.

Systems containing two or more fused heterocyclicrings do not simply follow the rule of highest priorityof each heteroatom for their numbering. Other things be-ing equal, numbering is chosen so as to minimize the sumof the numbers assigned to the heteroatoms. Fusion oftwo or more heterocyclic rings is designated by letter-ing, starting with “a” between the 1,2 bond of the hetero-cyclic component. Thus, cinnoline could be referred to asbenzo[c]pyridazine.

NN

NN

cinnoline

4

3

43

1

6

5

1

6

5

7

8a

b

cd

e

f

pyridazine naphthalene

4

22

3

2

1

6

54a

7

88a

The numbering of cinnoline itself is based on that ofthe corresponding hydrocarbon, in this case naphthalene.An atom at a ring junction adopts the number of the atomthat precedes it in a clockwise sense, such as 4, 4a, 4b.The use of heterocyclic components is the most commonmethod of naming heterocycles. Many factors may need tobe considered, in a designated system of priorities. A typ-ical example that involves specifying unsaturation using“H” is 4H -1,3-dioxolo[4,5-d]imidazole.

O

ON

NH

1

3

4

4H-1,3-dioxolo [4,5-d] imidazole

An important alternative to the “heterocyclic compo-nents” method is replacement nomenclature in which thename of a fused heterocycle is derived from that of theparent hydrocarbon. Thus, the following derivative ofnaphtho[1,2-b:8,7-c]dipyran can be referred to as an hy-drogenated dioxabenzo[c]phenanthrene:

OO

3,4,9,10-tetrahydro-2H,12H-1,11-dioxabenzo [c] phenanthrene

6

8

12

5

7

9

10 3

4

11

12

D. Occurrence and Isolation

Both the number of sources in nature and the numberof known, naturally occurring heterocycles are vast. Verysimple heterocycles occur in nature, such a furan, foundin certain low-boiling wood oils, and indole, present injasmine and orange blossoms. The leaves of “lily of thevalley” afford azetidine-2-carboxylic acid. Pyrrolidine oc-curs in carrot green.

NH

NH

COOH

azetidine-2-carboxylicacid

pyrrolidine

Tryptophan is an amino acid essential in human nutri-tion but not synthesized by the body; casein contains about1% of tryptophan.

NH

NH2

COOHH

O

O

N

O

tryptophan piperine

P1: GOL Final Pages



Piperine (L . piper = pepper) is a piperidine found inblack pepper. Aspergillic acid is an antibiotic found inAspergillus flavus. Folic acid is a B vitamin widelydistributed in living matter.

N

N

O

OH

N

N

HN

N

O

H2N

HN

HN

COOH

COOH

Ofolic acid

aspergillic acid

More complex molecules frequently involve an array ofchiral centers. Riboflavin (vitamin B2) is an isoalloxazineinvolved in nutrition; its richest natural source is yeast.

N

N

NH

N

OH

HO

OH

OH

O

Oriboflavin

Other examples are 1-hydroxymethylquinolizidine,found in yellow lupin seeds, and thromboxane A2, an im-portant component in blood platelet aggregations.

N

H

OH

1-hydroxymethylquinolizidine thromboxane A2

O

OCOOH

OH

Structural elucidation of many natural products pavesthe way for their synthesis; multistriatin was synthesizedand used to trap the male elm bark beetle, which carriesthe fungus causing Dutch elm disease. Many naturally oc-curring heterocycles, such as cocaine (isolated from cocoaleaves) and morphine (from the opium poppy), have im-portant pharmacophysiological properties.

MeN

OCOPh

COOMe

cocaine

Exceedingly complex structures have invariably provedfascinating to the organic chemist. Solanidine, found inpotato shoots, is a formal fusion of a steroid skele-

ton (rings A–D) with a perhydroindole nucleus (ringsE and F). Strychnine, the notorious poison, first isolatedfrom Strychnos ignatii in 1818, contains several fused het-erocyclic rings; it was the subject of a brilliant synthesisin the 1960s.

N

HO

H

N O

O

H

H

H

H

N

D

A

C

B

FE

solanidine strychnine

II. PHYSICAL PROPERTIES

Physical properties offer good criteria for the purity ofheterocycles, just as they do for other organic compounds.The melting or boiling points were the characteristics for-merly most frequently used, but optical spectra (based onabsorption of light) and nuclear magnetic resonance spec-tra have become increasingly used. As for very many or-ganic compounds, heterocycles display precise and readilyreproducible physical properties.

A. Melting and Boiling Points

Melting and boiling points of common heterocycles aregiven in Tables III and IV. The boiling points of manyof the heterocycles listed are substantially higher thanthose of the analogous cycloalkanes. Many simple hete-rocycles are mobile, pungent liquids at room temperature.Because molecular weight is related to vapor density,smaller molecules are, on the whole, more volatile thanlarger ones. An example can be seen in the compari-son of alkylfurans with alkylbenzenes, the latter analogsbeing heavier by C2H2 O =10 mass units; the furans boilat about 40◦C lower than the corresponding benzenes.

TABLE III Boiling Points of Saturated Heterocycles and Cor-responding Carbocycles (◦C at 1 atm)

Heteroatom typeNumber andRing orientation of Saturatedsize heteroatoms NH O S cycloalkane

3 One 56 11 55 −34

4 One 63 48 94 13

5 One 87 66 121 49

6 One 106 88 141 80

6 Two (1,2) 165a — — 80

6 Two (1,3) 150 106 — 80

6 Two (1,4) 145 101 200 80

7 One 138 120 174 119

a Corrected to atmospheric pressure by the method of Hass andNewton (1970). In “Handbook of Chemistry and Physics,” 51st ed.(Weast, R. C., ed.), CRC, Cleveland, OH.

P1: GOL Final Pages



TABLE IV Melting and Boiling Points of Heteroaromatic Compounds Compared with Benzene (◦C at 1 atm)

SubstituentRing system(with locationof substituent) H CH3 C2H5 CO2H CO2C2H5 CONH2 NH2 OH OCH3 Cl Br

Benzene 80 111 136 122a 211 130a 184 43a 37a 131 155

Pyridine-2 115 128 148 137a 243 107a 57a 107a 252 171 193

Pyridine-3 115 144 163 235a 223 129a 65a 125a 179a 150 173

Pyridine-4 115 145 171 306a 219 156a 157a 148a 93a 147 174

Pyrrole-1 130 114 129 95a,b 180 166a —c — — — —

Pyrrole-2 130 148 181 205a,b 39a 174a — — — — —

Pyrrole-3 130 158 179 148a 78a,b 152a — — — — —

Furan-2 31 64 92 133a 34a 142a 68a 80a 110 78 103

Furan-3 31 65 — 122a 179 168a — 58a — 80 103

Thiophene-2 84 113 133 129a 218 180a 214 217 154 128 150

Thiophene-3 84 115 135 138a 208 178a — — — 136 157

Pyrazole-1 70a 127 137 — 212a — — — — — —

Pyrazole-3 70 205 209 214a,b 160a — 285 164a — — —

Isoxazole-3 95 118 138 149a,b — 134a — — — — —

Isoxazole-5 95 122 — 149a 174a — — — — — —

Imidazole-1 90 199 226 — 218a — — — — — —

Imidazole-2 90 141a 80 164a,b — — — 250a,b — — 207a

Imidazole-4 90 56a — 275a,b 157a 215a — — — — 130

Pyrimidine-2 123 138 — 270a — — 127a 320a — 65a —

Pyrimidine-4 123 141 — 240a,b — — 151a 164a — — —

Pyrazine-2 57a 135 — 229a,b — 189a — 119a 187a 160 180

a Melts above 30◦C.b Melts with decomposition.c A dash indicates that the compound is unstable unknown, or that the data are not readily available.

However, nitrogen heterocycles form a less regular se-ries than do the furans because of significant hydrogenbonding and other intramolecular interactions.

Alkyl groups attached to heteroaromatic rings usuallyincrease the boiling point in a regular manner: about 20–30◦C for each methyl group and 50–60◦C for each ethylgroup. Where an NH group becomes substituted givingan NR group, a significant decrease in the boiling pointoccurs because of much diminished hydrogen bonding.The carboxylic acids and amides are all solids; a hydroxylor an amino group in a nitrogen heterocycle usually resultsin a relatively high-melting solid.

B. Ultraviolet, Infrared, Nuclear MagneticResonance, and Mass Spectra

Ultraviolet (electronic absorption) spectra (Table V) givecharacteristic absorption bands for aromatic and many un-saturated heterocycles and can yield important informa-tion on their electronic structure. Absorption in the visi-ble region (400–800 nm) occurs for extensive conjugatedsystems, especially if charge can be delocalized over sev-eral rings. This is the basis of many heterocyclic dyes.

Infrared spectra of heterocycles (and organic com-pounds in general) provide a molecular “fingerprint” ofeach compound (Table VI), in addition to allowing therecognition of many polar substituents, such as C O andN H groups.

Proton magnetic resonance spectra (Table VII) provideinformation on the number and nature of the environmentof hydrogen atoms present in a molecule. Carbon mag-netic resonance spectra (carbon isotope 13) allow eachcarbon atom in a molecule to be identified, together withthe number of hydrogen atoms bonded to it. It is not sur-prising, therefore, that nuclear magnetic resonance has hadan enormous impact on heterocyclic chemistry, allowingrapid structural identification, among many other impor-tant features. Mass spectrometers are capable of detectingmany molecular ions, and ions formed by fragmentation,of heterocycles, even when only trace amounts are present.

C. Tautomerism

Tautomerism is very important in heterocyclic chemistryand many different types are known. Heteroolefinic com-pounds show tautomerism similar to their acyclic analogs.

P1: GOL Final Pages



TABLE V Ultraviolet Spectral Maxima and Extinctions for Heteroaromatic Compounds

Neutral form Monoprotonated form Neutral form Monoprotonated form

Wave length Intensity Wave length Intensity Wave length Intensity Wave length IntensityCompound (nm) (log10 ε) (nm) (log10 ε) Compound (nm) (log10 ε) (nm) (log10 ε)

Pyrrole 210 4.20 241 3.90 Pyridine 257 3.42 256 3.70

Furan 208 3.90 — — Pyridazine 247 3.04 — —300 2.51 238 3.21

Thiophene 231 3.87 — — Pyrimidine 243 3.51 242 3.64

Pyrazole 210 3.53 217 3.67 Pyrazine 261 3.77 — —300 2.93 266 3.86

Isoxazole 211 3.60 — — Indole 270 3.77 280 3.68Quinoline 275 3.51 — —

Isothiazole 244 3.72 — — 299 3.46 313 3.79212 3.52 — —

Thiazole 233 3.57 — — Isoquinoline 306 3.38 270 3.30319 3.47 332 3.63

The azoles show annular tautomerism of the type exem-plified next.

Hydroxyamino, mercapto, and methyl substituents arecapable of tautomerism when attached to a heterocyclicring containing a pyridine-like nitrogen atom. This typeof tautomerism is found in the pyridines.

Pyridones predominate strongly over their tautomericforms, the 2- and 4-hydroxypyridines. For the nitrogenanalogs, the reverse is true, 2- and 4-aminopyridines beingfavored over the pyridonimine forms. This difference canbe rationalized by considering the mesomerism of the al-ternative forms. The charge-separated form of 4-pyridone,being aromatic, is greatly preferred over the nonaromatic,charge-separated form of 4-hydroxypyridine. In the caseof 4-aminopyridine, the charge-separated version of theimino form affords little stabilization, despite being aro-matic, because nitrogen accommodates a negative charge

N N N

O

H

N

O

H

OHOH

N N N

NH

H

N

NH

H

NH2NH2

_

4-hydroxypyridine

+

+_

_

+

+

4-aminopyridine

4-pyridone

4-pyridonimine

_

much less readily than does oxygen.Hydroxypyridine 1-oxides are also tautomeric; 4-

hydroxypyridine 1-oxide exists in approximately equal

amounts of the pyridinium and 4-pyridone forms.

N

OH

O

N

O

OH

tautomerism of 4-hydroxypyridine 1-oxide

+

_

N -Hydroxypyrroles have, in principle, three tau-tomeric forms, but for 1-hydroxy-2-cyanopyrrole, only theN -hydroxy form has been observed.

N

OH

CN N

O

CN

H N

O

CN

HH

tautomerism of 1-hydroxy-2-cyanopyrrole

_

+

_

+

For 2,3-dihydroxy-furan, -pyrrole, and -thiophene, theketoenol forms are favored over the corresponding diketoforms.

XO

O

XO

OH

tautomerism of 2,3-dihydroxyheterocycles (X = O, NH, S)

TABLE VI Some Characteristic Infrared Ab-sorption Maxima for Ring-Stretching Modes ofCommon Heteroaromatic Compounds

Compound Absorption frequencies (cm−1)

Pyrroles 1560–1530 1510–1480 1410–1390

Furans 1600–1560 1520–1470 1410–1370

Thiophenes 1540–1505 1440–1405 1370–1340

Pyridines 1610–1590 1580–1570 1485–1465

P1: GOL Final Pages



TABLE VII Chemical Shifts in Proton Reso-nance Spectra of Various Heterocycles

Chemical shift at given positiona

Compound 2 3 4

Aziridine 1.48 — —

Oxirane 2.54 — —

Thiirane 2.27 — —

Azetidine 3.58 2.32 —

Pyrrole 6.62 6.05 —

Furan 7.40 6.30 —

Thiophene 7.19 7.04 —

Pyridine 8.50 7.06 7.46

Pyridazine — 9.17 7.68

a Parts per million on delta scale.

A somewhat similar system is 2,3-dihydroxyindole;here, the 2-hydroxy form is the sole tautomer at 20◦C, butthe equilibrium shifts progressively toward the 3-hydroxyform, which is favored by 95:5 at 100◦C.

NH

O

OH

H NH

O

OHH

23

2

3

1

tautomerism of 2,3-dihydroxyindole

1

Whereas 2-aminothiophene exists in the amino formrather than the imino form, benzannelation can completelyalter the tautomeric preference; thus, 1-aminobenzo[c]

S

NH2

S

NH

tautomerism of 2-aminothiophene

thiophene exists preferentially in the imine form, theamino form possessing appreciably less resonance energy.

S

NH2

S

NH

tautomerism of 1-aminobenzo [c] thiophene

Many other tautomeric equilibria of heterocycles havebeen studied; theoretical calculations on the positions ofthese equilibria often agree with experimental values.

D. Conformation

Whereas many fully unsaturated rings are very nearlyplanar, their saturated or partially saturated analogscontaining tetrahedrally coordinated atoms are necessarilythree-dimensional. This frequently gives rise to a preferredorientation of the atoms in a molecule, differing from other

possible orientations (conformers) by rotation about sin-gle bonds. Thus, cyclohexane exists predominantly in thechair form, although boat and twist forms of cyclohexaneare also known.

chair boat twist

conformations of cyclohexane

With saturated heterocycles, conformational analy-sis is subject to additional considerations, such asthe inversion of a lone pair on a heteroatom as forN -substituted piperidines, the effects of hydrogen bondingto a heteroatom, and dipole–dipole interactions involving aheteroatom (anomeric effects).

N RN

R

inversion at nitrogen of N-substituted piperidines

The geometries of the four-membered heterocyclicrings oxetane and thietane differ substantially from thatof cyclobutane, the carbocyclic equivalent. Eclipsing

H

H

HH

S

H H

12

1

2

cyclobutane thietane

1,2-interactions are relieved by the puckered structure ofcyclobutane. Similar eclipsing interactions are not signif-icant between the sulfur and adjacent hydrogen atoms ofthietane; consequently, this molecule is effectively planar.

Like piperidine, tetrahydropyran adopts a chair confor-mation. Hydroxylated and alkoxylated tetrahydrofuransare important because such ring systems occur in vari-ous sugars and RNA. 2-Alkoxytetrahydropyrans, unlikealkoxycyclohexanes, exist preferentially with the alkoxysubstituent in the axial position. In methoxylcyclohexane,for example, unfavorable 1,3-diaxial interactions make theequatorial position for the alkoxy substituent favored.

O

OMe

O OMe

conformations of 2-methoxytetrahydropyran

However, for the corresponding tetrahydropyrans, theaxial lone pair on the ring oxygen atom can donate elec-tron density into the C OMe antibonding (σ ∗) orbitalif the 2-alkoxy substituent is in the axial position. Ac-cordingly, since such donation of electron density low-ers the molecular energy, 2-alkoxy substituents exhibit astrong preference for the axial position (referred to as the

P1: GOL Final Pages



anomeric effect). The anomeric effect is a guiding princi-ple in predicting the conformation of numerous carbohy-drates.

O

O

HO

HO

HO

O

OH

OH

HO

OHO

sucrose

H

Sucrose (“sugar”) is an important example of thesestereoelectronic effects. There is little doubt that confor-mation is crucial to the function of many heterocyclesfound in nature.

III. CHEMICAL PROPERTIES

A. Chemical Reactivity

All chemical reactions involve the making and breakingof a bond or bonds. For each bond broken, one new one ismade. A bond can be formed or broken in a reaction stepthat is either (1) ionic, (2) free radical, or (3) electrocyclic.Ionic mechanisms involve the transfer of two electrons(depicted by a curved arrow) from either a lone pair or amultiple bond to another atom:

A B

A BD A D

_A

+B

B_+

In a free-radical step, each atom contributes one electron(depicted by one single-headed arrow) to form a new bond:

A BA B

In electrocyclic reactions, several bonds are formed andbroken as part of a ring:

CHOCHO

In heterocyclic chemistry, three patterns of reactivityare especially important:

1. The ability of a multiply bonded heteroatom to acceptelectrons at the site alpha to the heteroatom (Z):

C Z C Z+Nu Nuor

The attack is easier when Z carries a positive charge.2. The ability of a heteroatom attached to multiply

bonded carbon to donate a pair of electrons:

C CZ

C CZ

or_

Again, this nucleophilic behavior is easier when Z carriesa negative charge.

3. The tendency of heterocycles to revert to their initialdegree of unsaturation if disturbed. Usually this meansrearomatization.

B. Nucleophilic Attack

Electron displacement toward the nitrogen atom often al-lows nucleophilic reagents to attack alpha to a nitrogenheterocycle, such as in the Chichibabin reaction, in whichalkali-metal amides induce amination.

N N Nu

HN Nu

HN Nu

+

_

X

_Nuα

HX

nucleophilic addition-elimination at Cα of pyridine

_

This type of reactivity is observed in benzenoid chem-istry only when electron-withdrawing substituents arepresent. However, a powerful nucleophile is needed toform appreciable quantities of the anionic nitrogen spe-cies, which is of high energy, having been formed bydearomatization of the initial heterocycle. In the case ofheterocyclic cations, nucleophilic attack is much easier,chiefly because of the resulting neutralization of oppositecharges.

N N

H Nu

N

CH=CH-CH=CHNu

nucleophilic attack resulting in ring opening

+

_Nu

In many cases subsequent ring-closure occurs, to givea new heterocyclic ring. In certain cases, such as theDimroth rearrangement of N -alkylated (or arylated) het-erocycles to the corresponding alkylamine (or arylamine)heterocycles, the nature of the heterocyclic ring is not al-tered.

N

N NHMe

N

N NH2

Me

NaOH

I

+

_

Dimroth rearrangement

C. Electrophilic Attack

Electrophilic attack of pi-deficient heterocycles such aspyridine usually requires forcing conditions. Thus, nitra-tion of pyridine at 300◦C yields 3-nitropyridine in onlypoor yield.

N N

NO2300 °C+ H2SO4/KNO3/SO3

As with nucleophilic attack, the site of electrophilicattack can be rationalized, and is often predicted, by con-sidering the relative stability of the structure formed prior

P1: GOL Final Pages



to the products. In terms of valence bond theory, the firstdrawn canonical form, formed by ortho attack on pyridineby nitronium ion, is of high energy because it carries a pos-itive charge on the electronegative nitrogen atom. Hencethis form contributes little to the overall stability of thecation formed by attack at the ortho position. The samearguments apply to attack at the para position. However, inthe meta orientation all three forms contribute to loweringthe energy of the cation; in practice the nitrated productis found to be almost exclusively m-nitropyridine. Muchof the electrophilic and nucleophilic chemistry of het-erocycles has been successfully rationalized along theselines.

N N

HNO2

N

HNO2

N

HNO2

N N NO2

HN NO2

HN NO2

H

m-attack

NO2

+

NO2

+

+

attack of nitronium ion on pyridine

+

o-attack

+++

+

The canonical forms of pyridine N -oxide predict thatthe ring should be attacked by both nucleophiles and elec-trophiles. Both are indeed observed, such as the forma-tion of 2- and 4-chloro- and 4-nitropyridine N -oxides,respectively.

N

O

N

O

N

O

N

O

N

O

N

O

N

O

___

+

+ +

_

+

canonical forms of pyridine N-oxide

_

_+_+ +

D. Radical Reactions

Halogenation of heterocycles at high temperature givesproducts in which the substitution pattern suggests thatfree radicals (rather than ionic halogen) are involved intheir production.

N N Br

+450 °C

Br2

bromination of quinoline

Although neutral heteroaromatics often undergo alky-lation by radical species in poor yield, if protonatedheterocycles are used, good yields are frequently ob-tained. Under suitable conditions, pyridines, quinolines,

and isoquinolines can be alkylated in the 2-, 2-, and 1-positions, respectively. Acylation of protonated pyridines,

N N R

acid

alkylation of pyridine

+ ROOH

quinolines, pyrazines, and quinoxalines via free radicalsall proceed in fair yield. Thus, treatment of pyridine withfree radicals generated from formamide gives isonico-tinamide and the 2-isomer.

N N N CONH2

CONH2

isonicotinamide

HCONH2+ +ButOOHFe2+

+

In marked contrast, homolytic arylation of heterocy-cles often gives low yields and is unselective. Phenylradicals attack pyridine to give a mixture of 2-, 3-, and4-phenylpyridines.

E. Electrocyclic Reactions

Electrocyclic reactions are very important both for thesynthesis of heterocycles and in their transformation tonew cyclic structures. Occasionally, intramolecular Diels–Alder reactions are used to form large heterocyclic rings.

O

O∆

More often, two fragments undergo [4 + 2]cycloadditionto give a heterocyclic ring, as exemplified for the synthesisof pyridines and furans from oxazoles. Similar transfor-mations include the conversion of pyrroles into azabicyclo

O

N

EtO COOMe

COOMe N

HO

COOMe

COOMe

N

EtO

COOMe

COOMe

OHCl

+∆

EtOH

O

N

Ph NPh

COOH

O

COOH

O

COOH

synthesis of pyridines and furans via cycloaddition reactions

- PhCN∆+

compounds and pyridazines into pyridines.

N

COOMe

COOMe

COOEtN

COOMe

COOMeCOOEt

N

Me

NEt2

NN

NEt2COOEt

COOMe

N

MeOOC Me

NEt2

- HCN

synthesis of azaheterocycles via cycloaddition reactions

∆+

∆+

P1: GOL Final Pages



1,3-Dipolar cycloaddition reactions are similar toDiels–Alder reactions, both proceeding through a cyclictransition state of six pi electrons, but differ in thatthree atoms supply the four-electron unit (in contrastto four atoms in a Diels–Alder reaction). Consequently,a five-membered ring results. Many cycloadditions,including 1,3-dipolar cycloaddition reactions, proceedwith high stereocontrol. Sometimes the heterocyclic ringcan be cleaved to give acyclic compounds of definedstereochemistry.

O N

C6H5 C6H5

OH NH2N

O

C6H5heat LiAlH4

_+

Four-membered heterocycles are often synthesized by[2 + 2]cycloadditions. An important example is the reac-tion of ketenes with imines to give beta-lactams.

R1

N

R3R2

C

CRR

ONR1

O

R3R

RR2

[2 + 2] cycloaddition route to β-lactams

+

3-Substituted furans can be prepared by means of a pho-tolytic cycloaddition. Many other [2 + 2]cycloadditionsgiving heterocycles are known.

O O

COORH

O

COOR

HO

O

O

COOR

hν

synthesis of 3-substituted furans

+p-TsOH

F. Use of Heterocycles in Synthesis

The following examples show how heterocycles can beused to prepare molecules that often cannot be readilysynthesized by other routes. Thus, alkyl halides (R′X) areconverted into alpha-substituted aldehydes via dihydro-1,3-oxazine intermediates.

O

N

R1

O

N

R1

R2

H

O

R1

R2R2X 1. NaBH4

2. H2O_

synthesis of aldehydes using dihydro-1,3-oxazines

Certain furans readily undergo Diels–Alder reactions toform adducts that can eliminate water to give substitutedbenzenes, and in certain cases benzene gives naphthalenes.

COOH

O

COOMe

1. ∆

2. acid+

synthesis of naphthalene-1-carboxylic acid

Alpha-pyrone, on photolysis, yields the highly reactivecyclobutadiene.

O O

hν+ CO2

preparation of cyclobutadiene

Some alpha-pyrones can act as dienes, affording poly-substituted benzenes.

COOMe

COOMeCOOMe

COOMe

MeOOC

O O

MeOOC

180 °C

cycloaddition of an α-pyrone

+

Certain cyclic sulfones extrude sulfur dioxide whentreated with N ,N -dimethylaminofulvene, affording inky-blue azulenes.

R

RNMe2

S R

R

O

O+

synthesis of an azulene

Heterocyclic rings can be used to protect functionality.Thus, the formation of cyclic ketals can be used to protecteither ketones or aldehydes, or 1,2-diols.

O

O

R1

R2

R3HO

HO

R1

R2

R3

Op-TsOH

H3O++

protection of functionality as a cyclic acetal

Deprotection occurs under mildly acidic aqueous con-ditions. The Gabriel phthalimide synthesis enables alkylhalides to be converted to primary amines. The action ofammonia itself on alkyl halides produces a mixture ofamines and quaternary salts.

N

O

O

N

O

O

R

COOH

COOH

RNH2 +KH2O+_ RI

use of phthalimides in the Gabriel synthesis

P1: GOL Final Pages



A general means of deoxygenating ROH into RH in-volves conversion of the hydroxy group into a thiocar-bonylimidazolide followed by reductive cleavage usingtributyltin hydride in the presence of the radical initiatorazobisisobutyronitrile (AIBN).

O

OREtOOCN

O

NHCOOEt

O

OREtOOCN

NHCOOEt

NN

O

H H

Bu3SnH

AIBN

Benzotriazole ia an inexpensive and stable heterocyclethat is soluble in a variety of common solvents. It is easilyintroduced into a molecule and can act as a satisfactoryleaving group. Such properties enable its use in a widevariety of synthetic transformations, including displace-ments to form unsymmetrical secondary aliphatic amines.

NNHR2

R1

NN

+ NHR2

R1

R3R3MgX

When borane becomes bonded to certain chiral oxaz-aborolidines, hydride transfer to ketones takes place, giv-ing mainly one enantiomer of the alcohol. The larger groupof the ketone points away from the heterocyclic ring, min-imizing steric hindrance.

N B

O

HPh Ph

CH3

N B

O

HPh Ph

Ph

OBH

HH

MePh Cl

O

Cl

Ph Cl

H OH

+ -- +

BH3.THF

The ring opening of certain modified carbohydrates bymercury(II) acetate proceeds with subsequent cyclization(Ferrier rearrangement) to give a substituted cyclohex-anone that can be reduced to an inositol derivative. Inos-itol phosphates are key messenger molecules in cellularsignaling.

O

OMe

RO

OR

ROOCOCH3

OH

RO

OR

RO O

OCOCH3i Hg(OCOCH3)2

ii NaCl

Sulfur heterocycles readily form sulfur ylides that canundergo electrocyclic ring expansion to give a medium-size ring. The sulfur can then be used to link two carbonatoms with extrusion of sulfur dioxide to form a cyclicalkene (Ramberg–Backlund reaction).

S

COOEt

S

SOO

COOEt

COOEt

SOO

COOEt

COOEt

+ _

m-CPBA NaH

C2Cl6

- SO2

IV. SYNTHESIS OF HETEROCYCLES

The synthesis of heterocyclic rings has much in commonwith carbocyclic rings. Ring formation can be via cycload-dition, ring closure, ring expansion or contraction, ring-atom interchange (ANRORC reactions), or insertion ofreactive intermediates into double bonds.

c

d

a

b c

da

b

ba

cd

e

fa

b

cd

e

f

e

cb

a

de

cb

a

d

hν

[4 + 2] cycloaddition

1,3-dipolar cycloaddition

+

+

[2 + 2] cycloaddition

_

+

The Woodward–Hoffmann rules predict the regio- andstereochemistry of substituents in rings formed by cy-cloaddition reactions. In practice, a number of cycload-dition reactions are rarely or never observed, and theseare forbidden by these important rules. The extremelyuseful Diels–Alder reaction is a [4 + 2]cycloaddition andwas originally applied to carbocycles (atoms a–f = C), butit has been adapted to the synthesis of heterocycles.

O

NPh

O

NPh

∆

synthesis of a dihydro-1,2-oxazine

+

Ring-closure reactions follow well-established pat-terns, which are summarized in Baldwin’s rules. For ex-ample, the ring closure of the terminal radical below leadsto the so-called 5-Exo-Trig product rather than the 6-Endo-Trig product.

P1: GOL Final Pages



XX X

5-exo-trig product 6-endo-trig product

Ring expansion of aziridine to a tetrahydrooxazole isaccelerated by the release of angle strain in the former.

O

H

n-C6H13

HN

O

NH

n-C6H13+

a ring expansion of aziridine

V. MAJOR CLASSES OFHETEROCYCLIC COMPOUNDS

A. Three- and Four-Membered Rings

1. Synthesis

The saturated three-membered rings containing a nitro-gen, oxygen, or sulfur atom are respectively referred to asaziridine, oxirane, and thiirane. They are usually readilyformed by a ring closure that is entropically favored. Theanalogous formulation of four-membered rings (n = 2) isless favored and, unsurprisingly, less common.

CH2BrHZ Z CH2

SBr Br

(CH2)n (CH2)n

Na2Sthietane

cyclizations that give small heterocyclic rings

Epoxides (oxiranes) are extremely useful synthetic in-termediates (see Section V.A.2 below). Most double bondscan be epoxidized by reaction with m-chloroperoxy-benzoic acid (m-CPBA). Electron-deficient double bondsare an exception, but can be epoxidized using alkalinehydrogen peroxide.

Some alkenes undergo enantioselective epoxidationupon treatment with hypochlorite in the presence of a chi-ral manganese catalyst (Jacobsen reaction). The epoxida-tion of allylic alcohols by means of an alkylhydroperoxideattached to a titanium alkoxide–tartrate system is knownas the Sharpless asymmetric epoxidation and is one ofthe most general and widely used asymmetric processesknown; high enantioselectivity is usually observed.

Carbonyl compounds react with diazoalkanes with lossof nitrogen to give epoxides. Indeed, addition of a car-bene (R2C:) or a nitrene (RN:) (or a carbene or nitreneequivalent) to a double bond is another general route tothree-membered heterocycles.

OH OH

O

R1

R2O

R1

R2 O

O

Ph Ph

O

ButOOH, Ti(OiPr)4

(-)-diethyl tartrate

synthesis of epoxides

+ CH2N2

m-CPBA

dichloromethane

NaOCl

manganese salen

complex

The conversion of oxirane into thiirane proceeds via asSN ANRORC mechanism.

O O S

C

N

SO

N

NCO

S

S

_

oxirane

_

_

preparation of thiirane

KSCN

_

+ NCO

The formation of aziridine is illustrated by the followingsequence.

HN

H3N

SO3

formation of aziridine

NaOH+

_ H2O

Important routes to four-membered rings containingone (or more) heteroatoms are [2 + 2]cycloadditions, aspreviously discussed. Ring expansion of cyclopropanone

O

O

C C ONPh

O

Ph

HCHO PhCH=NPh

an azetidinone

to azetidinones is also practical.

O HO NHCH2COOH HO NHCH2COOR

Cl

NCH2COOR

OtBuOCl Ag+

azetidinones via ring expansion

P1: GOL Final Pages



2. Properties and Importance

The reactivity of three- and four-membered heterocyclesis dominated by ring strain, which facilitates ring opening.Four-membered rings often cleave to two two-memberedfragments. Cationic three- and four-membered rings un-dergo nucleophilic attack when undergoing ring opening.Propiolactone undergoes ring opening, with the productbeing dependent on nucleophile and acidity of the reac-tion medium.

O

O

HO

O

Nu Nu

O

OHA B

_Nu

A

B

ring opening reactions

Anions of small heterocycles are uncommon. Extrusionof small gaseous fragments (e.g., N2, CO2) is very com-mon for small, neutral heterocycles. Cis–trans isomeriza-tions via 1,3-diradicals are often observed when three-membered heterocycles are heated or irradiated.

HN

R RHH

HN

R RHH

∆

HN

R HRH

HN

R HRH

cis-trans isomerization of aziridines

Ring valence isomerizations of small rings are known.Ring expansion of vinyl heterocycles is well known. In

MeN

NMe

valence isomerization of an aziridine

many reactions of this form, especially when a three-membered heterocycle reacts inter molecularly with anolefin, the heterocycle acts as a 1,3-dipole, by a formal ringopening. Other ring expansions include the formation of

NNH

∆

formation of pyrrole by ring expansion

indole from 3-phenyl-1-azirine. Such ring expansions aretypical of three- and four-membered heterocycles, chieflybecause of relief of ring strain.

NH

N

indole

∆

3-phenyl-1-azirine

Epoxides (oxiranes) are frequently used as syntheticintermediates and are usually prepared from alkenes bytreatment with peracids or metal–hydroperoxide systems(see Section V.A.1 above). Epichlorohydrin has been usedto prepare glycerol commercially. The strain of the smallring leads to ready attack of epoxides by a wide variety ofnucleophiles.

OHHO

OH

ClO

NaOH

glycerolepichlorohydrin

RHO

O RMgX

O

H

HO

NHRH

RNH2

An useful procedure for inverting of alkene geometryinvolves stereospecific epoxidation and ring opening.

OH

H

O SiMe3

HH

_

trans-3-hexene

K

RCO3H

+Me3SiK

cis-3-hexene

N

O

O

H2N

OCONH2

NH

OCH3

O

P

N

N

Cl

Cl

OH

cyclophosphamide mitomycin C

Azetidine, the four-membered analog of aziridine, isa colorless liquid, stable at room temperature and lessreactive than aziridine. Interest in the cyclic amide aze-tidinone rose rapidly in 1943, when it was correctly sug-gested that penicillins contained this heterocyclic ring.The fused azetidinone ring is the key to the reactivity ofthe two related series of antibioties, the penicillins and thecephalosporins.

N

S

OCOOH

HNHH

O

Ph

benzylpenicillin

P1: GOL Final Pages



Several of the early chemotherapeutic agents used in awide variety of cancer treatments depend upon an three-membered ring undergoing nucleophilic attack by an oxy-gen or nitrogen atom in DNA of the cancer cells. Examplesinclude cyclophosphamide (which can eject chloride withthe formation of an aziridinium ring, a powerful elec-trophile) and mitomycin. In contrast, epoxides derived bymetabolism of many polybenzenoid hydrocarbons arehighly toxic because the epoxides form covalent linkageswith normal cell DNA.

NO

HNH

O

S

COOH

OAc

H

O

HO

NH2H

cephalosporin C

B. Five-Membered Heterocycleswith One Heteroatom

1. Synthesis

Pyrrole, furan, and thiophene are the most important par-ent heterocycles of this class, and their annelated deriva-tives constitute a vast array of both naturally occurring andsynthetic compounds.

The ring closure of a four-carbon chain with the incor-poration of one heteroatom is very easy.

SBr Br

sulfide

tetrahydrothiophene

NH

OCOOH

H2N

heat

pyrrolidone

A special case of this is the Hofmann–Loeffler–Freytagreaction, in which pyrrolidines (or piperidines) are formedby the decomposition of protonated N -haloamines.

N N

R

O

R

Cl N

R

HH

H+

conversion of N-chloramines into pyrrolidines

+

This particular free-radical mechanism confers high se-lectivity in the formation of products owing to a preferredsix-membered transition state. Another common route tofive-membered heterocycles is via 1,3-dipolar cycload-dition.

NH

MeOOC

Ph COOMe

NPh COOMeMeOOC AgNO3, Et3N

MeCN+

formation of pyrrolidines via cycloaddition

Fully unsaturated five-membered heterocyclic rings areoften synthesized by the ring closure of 1,4-dicarbonylcompounds to give, with ammonia or primary amines,pyrroles (Paal–Knorr synthesis); with dehydrating agents,furan; and with phosphorus pentasulfide, thiophenes.

NR1R2

R3

R1

O

R2

O

R3NH2+

Paal-Knorr synthesis of pyrroles

∆

Another important synthesis of pyrroles is the Knorrpyrrole synthesis, in which alpha-aminoketones or theirequivalent are condensed with carbonyl compounds con-taining active alphamethylene groups.

NH

COOH

COOEtCOOEt

COOMeONH3

O

+KOH+

Cl

Knorr pyrrole synthesis

_

Indoles (benzo[b]pyrroles) are frequently prepared bythe Fischer indole synthesis, in which aryl hydrazonesundergo rearrangement on heating, usually in the presenceof acid. The important drug indomethacin, used to combatrheumatoid arthritis, has been synthesized by this method.

MeO

COOH

ONH

NH2

MeO

NH

N

COOH

+

NH

COOH

MeO

N

COOH

MeO

O

Cl

synthesis of indomethacin

2. Naturally Occurring Five-MemberedHeterocycles

Pyrrole (Gk·pyr = fire) was first isolated from bone py-rolyzate; it imparts a bright red color to pinewood moist-ened with concentrated hydrochloric acid.

Condensation of suitable pyrroles in the presence ofacid leads to dipyrromethene cations, which are precursorsto porphyrins, a group of compounds found in all livingmatter and forming the basis of respiratory pigments.

HNNH

O

NH HN

H

+

Br

a dipyrromethene cation

HBr +

_

P1: GOL Final Pages



The oxidative condensation of certain pyrroles can leaddirectly to the porphyrin nucleus, as in the case of oc-taethylporphyrin.

NNH

N HNNH

COOEt

NEt2

octaethylporphyrin

1. KOH

2. AcOH, ∆, air

Heme, the coloring portion of hemoglobin, and thechlorophylls, the green pigments of plants, are bothporphyrins.

NN

N N

COOHHOOC

Fe

NN

N N

C20H39OOC

Mg

OMeOOC

chlorophyll aheme

NN

N N

CONH2

Co

CN

H2NOC

H2NOCCONH2

CONH2

H2NOC

H

N

N

O

OHO

H

HO

ONH

PO

O

O

cyanocobalamin

+

_

Vitamin B12 (cyanocobalamin) is used to combat per-nicious anemia; it was synthesized by Woodward and Es-chenmoser in brilliant and unparalleled sequences of over30 reactions. The skeleton of vitamin B12 is the corrinmacrocycle, differing from a porphyrin ring by having oneless carbon atom in the linking of the four pyrrole rings.Here, the nitrogen atoms of the pyrrole rings bind (act asligands to) a central carbon atom. Moreover, a nitrogenatom of a benzimidazole (heterocyclic) ring also acts as a

ligand. In this biologically active and important molecule,it is unsurprising to find a further type of heterocycle: aribose ring, one of the sugars based on a hydrogenatedfuran ring.

Other examples of naturally occurring furans are vita-min C (ascorbic acid), first isolated from orange juice in1928, and furfural, the artificial oil of ants, from which thesolvent tetrahydrofuran is made.

O

HO OH

OHO

HOO

CHO

ascorbic acid furfural

Many important indoles are found in nature: tryptamine,the parent of the essential amino acid tryptophan; thepowerful vasoconstrictor serotonin; and the powerful hal-lucinogen lysergic acid diethylamide (LSD) are a fewexamples.

NH

NH2

NH

NH2HO

serotonintryptamine

NH

NMe

O

Et2N

H

lysergic acid diethylamide

Indoles have been used since ancient times; woad, thecoloring matter worn by Boudicca’s warriors, was an im-pure form of indigotin (indigo) extracted from the plantIndigofera tinctoria. Tyrian purple, the highly prized dyeworn on the garments of emperors, is 6,6′-dibromoindigo,and was extracted from the mollusc murex.

NHR

OHN R

OR = H indigoR = Br Tyrian purple

Thiophene itself is found in coal tar. Biotin, the humangrowth factor vitamin H, contains a tetrahydrothiophenering alpha-Lipoic acid, when enzyme-bound, is convertedinto acetyl lipoate, which acetylates coenzyme A givingacetyl coenzyme A, the biological building block that sup-plies two carbon atoms.

P1: GOL Final Pages



NHHN

S

O

HHH

COOH

biotin

SS COOH

alpha-lipoic acid

C. Six-Membered Rings with One Heteroatom

Benzene, pyridine, pyridinium, and pyrylium cations forman isoelectronic series with decreasing electron densitiesat ring carbon atoms and have increasing susceptibility tonucleophilic attack.

N O

R

N

++

pyridinium pyrylium

benzene pyridine

Pyrylium salts are usually synthesized by condensationmethods.

Since they are very reactive, pyrylium salts undergomany trasnformations, giving pyridines with ammonia,pyridinium salts with primary amines, and even benzeneswith certain carbon nucleophiles, all by SN ANRORCmechanisms (nucleophilic substitution with addition ofnucleophile, ring opening, and ring closure).

O

NO2X_+

MeNO2

ButOK

OOH

+ _X

HX+ (MeCO)2O

Cl

CHO

Cl

CHO

O

O

O

O

+

+

+

CF3SO3H

2CF3SO3

_

Benzannelated pyrylium salts are naturally abundant;anthocyanin pigments of red and blue flowers are a com-

bination of sugars with anthocyanidins, of which flavyliumis the parent cation.

O+

flavylium cation

Pyrones, pyrans, and their benzannelated derivativescan be obtained from the corresponding pyrylium salts.Diels–Alder reactions are also useful routes to pyrans.Coumarins (benzo[b]pyrones) are often prepared by thevon

O OOH OEtO

O H2SO4+

synthesis of 4-methylcoumarin

Pechmann reaction, in which phenols are condensed withbeta-ketoesters in the presence of a catalyst. Coumarinitself is a flavor additive found in sweet clover. VitaminE (alpha-tocopherol) possesses properties of antisterilityand is a fused pyran. In the body, vitamin E acts as anantioxidant which after oxidation can be regenerated byvitamin C (ascorbic acid).

O

HO

(CH2CH2CH2CHCH3)3CH3HO

O

HO

HOOH

OH

alpha-D-glucosevitamin E

The hydroxypyran alpha-D-glucose is a main sourceof energy for living organisms. Cellulose and starch arepolymers of glucose.

Saturated or partially saturated pyran rings are con-stituents of many pheromones such as serricornin, a prin-cipal pheromone attractant secreted by the cigarette beetle,a major pest of cured tobacco leaves, and (−)-alpha-multistriatin, an important component of the pheromonebouquet of the elm bark beetle. One route that al-lows highly stereocontrolled placement of substitutentsinvolves a mercury(II)-catalyzed rearrangement of anacetylenic epoxy alcohol.

OOH

serricornin

O

O

multistriatin

O

O

Ph

OH

OPh HgSO4

H2SO4

P1: GOL Final Pages



1. Pyridines and Pyridinium Salts

A simple route to pyridinium salts is by direct alkylationof pyridines, the Menschutkin reaction.

N N

Me

_I

+MeI+

The pellagra-preventive nicotinic acid and the powerfulweedkiller paraquat are examples of important pyridinederivatives.

N

COOH

N N MeMe+ +

nicotinic acid paraquat

The antagonist nicotinamide adenine dinucleotide(NAD+) also contains a pyridinium ring.

O

HO OH

NO

HO OH

CONH2

O

OP

O

O

N

N

N

N

NH2

PO

HO

O

_

nicotinamide adenine dinucleotide (NAD+)

+

Pyridine is present in bone pyrolyzates and coal tar,from which it was isolated in 1846 by Anderson. It is nowprepared industrially from tetrahydrofurfuryl alcohol andammonia. Nicotine (in tobacco), the antibacterial agentsulfapyridine, and the benzene cofactor pyridoxine (vita-min B6) are three important pyridines.

N

HO

OH

OH

N

NMe

N NH

SO2

NH2

sulfapyridine pyridoxinenicotine

As for five-membered rings, common routes to six-membered rings involve condensation, ring closure, andcycloaddition reactions. In the Hantzsch pyridine synthe-sis, 2 moles of a beta-dicarbonyl compound are condensedwith 1 mole of an aldehyde in the presence of ammoniato give, after oxidation, a pyridine.

N

MeOOC COOMe

Ph

CHOMeOOC COOMe

Ph

O O2. HNO3

1. NH3, AcOH, ∆

Hydrogenated derivatives of pyridine are often used inalkaloid synthesis; heterocyclic enamines undergo vari-ous cyclization and annelation reactions. Complete hy-drogenation of pyridine leads to piperidine. Coniine, oralpha-propylpiperidine, is a poisonous principle in hem-lock, from which Socrates died. The synthesis of coniineby Ladenburg in 1886 is the earliest recorded synthesis ofan alkaloid.

NH

N

O

NH

N

O

acid

an enaminoketone

NH

NH

α

coniinepiperidine

2. Quinolines, Isoquinolines,and Their Derivatives

Formal benzannelation of pyridine gives quinoline andisoquinoline; bisbenzannelation gives phenanthridine andacridine. Whereas the former is numbered systematically,it should be noted that acridine is not, the central ring beingnumbered finally.

NN 6

5

9

acridine

8

7

4

2

10

110

phenanthridine

3

9

6

5

2

1

8

7

3

4

Acridine dyestuffs, such as acridine orange L, owe theircolor to the extensive delocalization of cationic charge.

N

H

NMe2Me2N N

H

NMe2Me2N+

_ _Cl Cl

acridine orange L

+

There are several important routes to quinolines. In theSkraup reaction, an aniline is heated with a mixture ofglycerol, nitrobenzene, and sulfuric acid to give a quino-line.

NNH2 OH

OH

OH

PhNO2

Skraup synthesis of quinoline

H2SO4+

P1: GOL Final Pages



In the Pfitzinger reaction, isatic acids are condensedwith alpha-methylene carbonyl compounds.

NH2

O

COONa

NH

O

O

N

COOH

R∆

RCOMe

isatin

NaOH

isatic acid(sodium salt)

In the Doebner–Miller reaction, quinolines are preparedby condensing anilines with unsaturated aldehydes or theirequivalent.

NNH2

OHC

+2. PhN=CHMe

1. H3O+

Doebner-Miller synthesis of 2-methylquinoline

Isoquinolines, abundant in nature, can be prepared bythe cyclodehydration of beta-phenethylamides, a proce-dure known as the Bischler–Napieralski reaction.

N

R

NH

O

R

P2O5

Bischler-Napieralski synthesis of dihydroisoquinolines

Isoquinolines are also formed in the Pictet–Spenglercondensation of beta-arylethylamines with carbonyl com-pounds.

NH

R'

NH2

MeO

MeO

MeO

MeO

R'CHO HCl

Pictet-Spengler synthesis of tetrahydroisoquinolines

+

Quinoline and isoquinoline alkaloids constitute a vastand important array of nitrogenous bases. The antimalarialquinine, a major alkaloid found in cinchona bark, containsa methoxy-substituted quinoline ring.

N

MeO

HON

quinine

Morphine, codeine, and papaverine (L. papaver =poppy) are all found in opium; they are all isoquinolinealkaloids. Morphine and codeine are well-known anal-gesics, while papaverine is a potent coronary and cerebalvasodilator.

N

MeO

MeO

OMe

OMe

O

RO

NMeH

HO papaverine

αβ

R = H morphineR = Me codeine

The Emde and Hofmann degradations have been usedin the structural elucidation of numerous naturally oc-curring substances, including many alkaloids. In theEmde degradation, quaternary ammonium salts are reduc-tively cleaved by sodium amalgam, N -allyl and N -benzylgroups being preferentially cleaved.

NMe2N+_

Na/Hg

Cl

Emde degradation of a quaternary ammonium salt

In the Hofmann degradation, pyrolysis of a quaternaryammonium hydroxide induces beta-elimination. Often, anunknown base is alkylated with methyl iodide, then treatedwith base, and the procedure is repeated exhaustively. Thefragments so obtained often help to elucidate the structureof the nitrogen base.

N

OH_

+ ∆CH2=CH2

An example of the Hofmann degradation

+ Me2NCH2CH2CH3

When the Hofmann degradation is applied to codeine,by quaternization with dimethyl sulfate and subsequentelimination of the quaternary salt with base, methylmor-phenol is formed.

O

MeO

methylmorphenol

The power of this degradative method is shown bythe formation of the relatively simple and identifiablefour-ring skeleton from the complex structure of codeine.In early work, the structures of many alkaloids wereelucidated by a combination of degradative and syntheticmethods.

Berberine is an antimalarial isoquinolinium salt that wasfirst isolated from Berberis vulgaris in 1837.

P1: GOL Final Pages



N

O

O

OMe

OMe

+

berberine

D. Five- and Six-Membered Ringswith Two or More Heteroatoms

An important pyrazolone is the developing agent“phenidone,” widely used in black-and-white photogra-phy. Histamine is an imidazole connected with many

NNH

O

Ph

N

N

H

H2N

"phenidone" histamine

pathological conditions, including allergies—hence thedevelopment of antihistamine drugs. Derivatives of thebiologically important thiamine (vitamin B1) assist in de-carboxylation and transketolase reactions by forming anucleophilic zwitterion based on the thiazolium ring.

N

N

NH2

N

S OH

N

S R3

R2R1

+

_

+

conversion of thiamine into a biologically active zwitterion

The sydnones are aromatic heterocyclic betaines thathave been called “mesoionic.” They have found use asantimalarial agents.

N

NO O

Ph

N

NO O

Ph

N

NO O

Ph

_

+

+

_ +_

canonical forms of a sydnone

A well-known derivative of isothiazole is saccharin, 500times sweeter than sugar.

N

NO O

O_

an antimalarial sydnone

+

S

NH

O

O O

saccharin

The three parent monocyclic diazines are pyridazine,pyrimidine, and pyrazine.

NN

N

N

N

N

pyridazine pyrimidine pyrazine

Luminol is a pyridazine that upon oxidation gives astriking emission of light. Veronal is a synthetic barbi-turate that acts as a sedative and hypnotic. However, the

NH

NH

O

O

NH2

HN

NH

O

O O

veronalluminol

pyrimidines are the most important class of compoundsin this group; the nucleic acids (see next subsection) arebased on the pyrimidine ring. Other important nitrogenheterocycles whose derivatives occur in nature are pteri-dine, alloxazine, and uric acid (a metabolic product).

N

N

N

N

N

N

NH

HN O

O

HN

HNO

O

NH

HN

O

uric acidpteridine alloxazine

There are many dyes based on cationic six-memberedheterocyclic rings, such as Meldola’s blue and methyleneblue, a biological staning agent.

O

N

Me2N S

N

Me2N NMe2+

Meldola's blue

+

methylene blueCl

_

1. Nucleic Acids

Nucleotides comprise a D-ribose or 2-deoxy-D-ribosesugar, which is linked by C-5′ to a phosphate unit andby C-1′ to a nitrogen of one of six heterocycles, of whichthree are pyrimidines—uracil, cytosine, and thymine—and three are purines—adenine, guanine, and hypoxan-thine.

N

NN

NH

O

HO OH

O N

N

N

N

NH2

PO

OHRO

4

3

15

65'

purineR = H adenosine 5'-monophosphate

7

3'

8

2'

1'

6

9 24'

Adenosine 5′-monophosphate is a typical nucleotide; itcontains the heterocyclic adenine 6-aminopurine. The di-and triphosphates (ADP and ATP) play central roles inbiological phosphorylation.

P1: GOL Final Pages



Nucleic acids are polynucleotides. They are of two maintypes: deoxyribonucleic acid (DNA) and ribonucleic acid(RNA). Both are present in every living cell; they direct thesynthesis of proteins and are responsible for the transferof genetic information.

In 1953, Watson and Crick postulated that DNA wasformed by the twisting of two polynucleotide chains intoa right-handed double helix. The chains are held togetherby the hydrogen bonding of two base pairs: thymine– oruracil–adenine and cytosine–guanine. No other combina-tion of those base pairs is possible.

N

N

R O

O

H

H

N

N

N

N

N

H

N

N

N

O

H

H

N

O

N

N

N

NH

H

H

thymine (R = H)uracil (R = H)

adenine

strand A strand B

cytosine

combination of base pairs in nucleic acids

guanine

strand Bstrand A

A segment of a DNA phosphodiester chain with thepositions of attachment of the heterocyclic bases is shown.

baseO

O

O

O

O

O basePO

HO

OP

HOO

PO

HO

segment of a DNA phosphodiester chain

5-Fluorouracil, used to treat breast and stomach can-cers, resembles uracil, without which the cell cannot makepyrimidine nucleotides. 5-Fluorouracil enters the biosyn-thetic chain but forms an inactive complex which blocksthe synthesis of pyrimidine nucleotides and hence inhibitsDNA synthesis in cancer cells. Methotrexate is a derivativeof folic acid and interferes with folic acid metabolism byinhibiting the enzyme dihydrofolate reductase, essentialfor the synthesis of purine and pyrimidine, and thereforeessential to the synthesis of DNA. Methotrexate finds usein treating childhood leukemia and other cancers, particu-larly in combination with other drugs. 6-Mercaptopurineblocks purine nucleotide synthesis and is a purine an-timetabolite in clinical use.

N

N

N

H2N N

MeN

CONHCH(COOH)CH2CH2COOHNH2

methotrexate

HN

N NH

N

S

6-mercaptopurine

E. Rings with Seven or More Members

1. Synthesis

The famous rules of Woodward and Hofmann govern,among many other reactions, the contractions and expan-sions of heterocyclic rings with seven or more members. Aseven-membered heterocyclic ring can give a 4,5-systemon photolysis, as well as a benzene ring on thermolysis.

X

X

X

XH

hν

transformations of seven-membered heterocycles

∆

X = NR azepineX = O oxzepineX = S thiepin

∆

∆

A nine-membered heterocyclic ring can give a 3,8-ringsystem on photolysis, and also a 5,6-ring system on ther-molysis. Some of these reactions are of considerable syn-thetic importance.

X

X X

H

H

∆

transformations of nine-membered heterocycles, X = O, NR

hν

2. Properties

Azepine, oxepine, and thiepin have eight pielectrons (in-cluding the lone pair on the heteroatom). Thus, they havelittle aromatic character and are very reactive. The tran-quilizing properties of certain diazepines, such as Libriumand Valium, have prompted much research in this area.

N

N

Cl

NHMe

OPh

N

MeN

Cl

O

Ph

HCl

librium

+_

valium (diazepam)

P1: GOL Final Pages



Many heterocycles possessing rings with more thanseven atoms are known, some of which are natural prod-ucts. A recently discovered antibiotic, LL-Z1220, under-goes reversible valence isomerization to a 1,4-dioxocin,an eight-membered heterocycle.

O

O

O

OO

O

O

O

boil, EtOAc

Ac2O, 60 °C

interconversion of 1,4-dioxocin and antibiotic LL-Z1220

F. Rings with Uncommon Heteroatoms

Nitrogen, oxygen, and sulfur are common, but by nomeans the only known heteroatoms in heterocyclic rings.For example, iodonium salts can be prepared by treatingaminobiphenyls with nitrous acid, followed by ring clo-sure of the diazonium salt.

INH2 I

N2 I

iodonium salts via diazotization

++

The corresponding chloronium and bromonium ions arenot very stable.

Phosphorus, arsenic, and antimony form heterocycliccompounds fairly readily, especially the first two.1-Phenylphospholane was prepared as early as 1916, andwith ethyl iodide gives a quaternary salt, just as do nitrogenheterocycles.

P

Ph

P

Ph Et

EtI

quaternization of 1-phenylphospholane

+I_

Phosphorins, in which phosphorus replaces one carbonatom in a benzene ring, can be prepared from pyryliumsalts.

OPh

Ph

Ph PPh

Ph

Ph

+

I+ _

P(SiMe)3

preparation of 2,4,6-triphenylphosphorin

Examples of arsenic and antimony heterocycles areisoarsindoline and stibabenzene.

AsH

Sb

isoarsindoline stibabenzene

Like sulfur, selenium and, less commonly, telluriumform heterocyclic rings, such as selenophene and diben-zotellurophene.

Se Te

selenophene dibenzotellurophene

Silicon, tin, and germanium, although in the same pe-riodic group as carbon, do not form stable double bonds;consequently, their heteroaromatic derivatives are un-known. However, many saturated heterocyclic derivativesare known, including 1,1-dimethylsilacyclohexane, ger-macyclohexane, and 6-stannaspiro[5,5]undecane.

Si

Me Me

Ge

H H

Sn

1,1-dimethylsilacyclohexane

germacyclohexane 6-stannaspiro [5,5] undecane

Finally, many boron heterocycles have been reported.The procedure of hydroboration, chiefly for which H. C.Brown received the Nobel Prize, is illustrated below; sub-sequent treatment with carbon monoxide affords replace-ment of boron by carbon.

B

OH HH

H

2. NaOH, H2O2

1. CO, 150 °CB2H6

carbonylation of a boron heterocycle

A variety of boron heterocycles are known; one such isprepared by reacting a borazine with biuret.

NHB

HN

NH

Me

O OH2N N

HNH2

OO

preparation of a boron heterocycle

(MeBNH)3 3 3NH3+ +

It can be expected that many new boraheterocycles willbe prepared in the near future.


ALKALOIDS • ANALYTICAL CHEMISTRY • BIOPOLYMERS

• ELEMENTAL ANALYSIS, ORGANIC COMPOUNDS • OR-GANIC CHEMICAL SYSTEMS, THEORY • ORGANIC CHEM-ISTRY, COMPOUND DETECTION • ORGANIC CHEMISTRY,SYNTHESIS • ORGANIC MACROCYCLES

P1: GOL Final Pages



BIBLIOGRAPHY

Acheson, R. M. (1976). “An Introduction to the Chemistry of Hetero-cyclic Compounds,” 3rd ed., Wiley, New York.

Coffey, S. (ed.), (1977). “Rodd’s Chemistry of Carbon Compounds,”Elsevier, New York.

Gilchrist, T. L. (1997). “Heterocyclic Chemistry,” Longman, Harlow,U.K.

Joule, J. A., Mills, K., and Smith, G. F. (1995). “HeterocyclicChemistry,” Chapman and Hall, London.

Katritzky, A. R. (ed.) (1963–1972). “Physical Methods in HeterocyclicChemistry,” Academic Press, New York.

Katritzky. A. R. (ed.), (1963–2000). “Advances in Heterocyclic Chem-istry,” Academic Press, New York.

Katritzky, A. R., and Rees. C. W. (eds.). (1984). “Comprehensive Hete-rocyclic Chemistry,” Pergamon Press, Oxford.

Katritzky, A. R., Rees, C. W., and Scriven, E. F. V. (1996). “Compre-hensive Heterocyclic Chemistry II,” Pergamon Press, Oxford.

Meyers. A. I. (1974). “Heterocycles in Organic Synthesis,” Wiley, NewYork.

Pozharskii, A. F., Soldatenkov, A. T., and Katritzky, A. R. (1997). “Het-erocycles in Life and Society,” Wiley, New York.

Suschitzky, H., and Scriven, E. (1989). “Progress in Heterocyclic Chem-istry,” Pergamon Press, Oxford.

Van der Plas, H. C. (1973). “Ring Transformations of Heterocycles,”Academic Press, New York.

Weissberger, A., and Taylor, E. C. (eds.). (1950–1996). “The Chemistryof Heterocyclic Compounds,” Wiley, New York.

P1: LLL/GVX P2: GQT Final Pages Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology EN011G-539 July 14, 2001 21:48

Organic Chemical Systems,Theory

Josef MichlUniversity of Colorado

I. Classical Bonding TheoryII. Qualitative Molecular Orbital Model of Electronic

StructureIII. Quantitative Aspects of Molecular StructureIV. Reaction Paths

GLOSSARY

Electron affinity Energy released when an electron isbrought from infinity and added to a molecule (it maybe negative if energy actually has to be provided).

Electronegativity Measure of the tendency of an atom oran orbital to attract and accommodate electron density.

Gradient of a surface The gradient at a point is a vectordirected up the steepest slope of the surface at thatpoint; its length is a measure of the slope.

Improper rotational axis of symmetry (of order n) Amolecule is said to possess such an axis located in aparticular direction if rotation by an angle 2π/n aboutthat direction, followed by mirroring in a plane perpen-dicular to the direction, converts the molecule back toitself.

Inner-shell electrons Electrons in orbitals of energylower than that of the valence shell (closer to thenucleus).

Interaction matrix element The interaction matrix ele-ment of the Hamiltonian H between wave functions

φ1 and φ2 is given by∫

φ1 Hφ2 dτ , where integrationis over all space.

Ionization potential Minimum energy that must be pro-vided to a molecule in order to remove one of its elec-trons to infinity.

Ortho, meta, para Designation of relative positions oftwo substituents on a benzene ring: located on adjacentring carbons (ortho), on next-nearest-neighbor ring car-bons (meta), and on carbons across the ring from oneanother (para).

Pauli principle A wave function describing the state ofa system containing two or more electrons is antisym-metric with respect to the exchange of all coordinatesof any two electrons (i.e., is converted to minus itselfon such an exchange). One of the consequences is thattwo electrons of the same spin cannot reside in the sameorbital.

THE THEORY of organic chemistry deals with the funda-mental concepts that underlie and unify the experimental

435

P1: LLL/GVX P2: GQT Final Pages


436 Organic Chemical Systems, Theory

observations made by chemists working with organicmolecules. It is believed that the behavior of moleculescan be understood, in principle, in terms of a few basiclaws of physics and that it is only the mathematical com-plexity of the resulting equations that limits the accuracywith which the behavior of organic molecules can be pre-dicted a priori. Despite the largely approximate nature ofthe theoretical treatments applicable to large molecules,the theory has made substantial contributions, primarilyby providing the language through which the various ob-served phenomena can be interrelated. It permits the ra-tionalization of trends and at times of individual observa-tions concerning the reactivity and properties of organicmolecules, and in some instances it has provided usefulpredictions.

I. CLASSICAL BONDING THEORY

Although the theory of organic chemistry has now beencast in terms of quantum theory, most of the older quali-tative concepts of the classical bonding theory remainuseful. The classical theory thus represents a suitable in-troductory level for the subject at hand.

A. Structural Formulas

In the classical description an organic molecule is rep-resented by a structural formula. This is a collection ofatomic symbols (C, H, O, N, etc.) representing atoms(including their inner-shell electrons but excluding theirvalence-shell electrons; the d electrons of a transitionmetal atom are included, although they may participate inbonding). At least one, but usually many, of these atomsmust be carbon in order for the molecule to qualify asorganic. The atomic symbols are connected by a networkof single, double, and triple lines, which stand for single,double, and triple covalent bonds, respectively. These areformed by the sharing of electron pairs between atoms,with one bond representing one pair. The number of near-est neighbors to which an atom is attached is called itscoordination number.

In addition, short lines (see 2a) or pairs of dots canbe used to represent unshared (lone) electron pairs on anatom, but these are frequently omitted. Single dots rep-resent odd (unpaired) electrons if such are present (see2a–2c).

The symbols C and H for the carbon atom and the hy-drogens attached to it, respectively, are also frequentlyomitted. Commonly occurring groups of atoms of well-known internal structure are often indicated by giving thekind and number of atoms involved (e.g., C2H5 for ethyl)or by an abbreviation (in this case, Et). A few examplesare given in 1a through 2c:

1a

C C

H

H

H C

H

H H

C

C

H

HC2H5

HH

H1b

1c

Et

C C

H

H

H

C

HH H

C

HH H

O O (CH3)3C O O

2a

2b 2c

t-BuO2

The number of bonds formed by an atom (its “cova-lency”) is dictated by the rules of valence. These state thatin order for an organic molecule to have reasonable stabil-ity under ordinary conditions rather than to appear only asa transient reaction intermediate, if at all, the valence shellsof all atoms in the structure have to contain a certain num-ber of electrons: 2 for hydrogen, 8 for other main-groupelements, and 18 for transition metal elements. The groupof 8 electrons in the valence shell of an atom is oftenreferred to as a valence octet. In order to determine thenumber of electrons in the valence shell of an atom, onecounts all the unpaired electrons or electrons present inlone pairs on that atom, plus two electrons for each singlebond in which the atom is participating (four for a doublebond, six for a triple bond). In structures 1a through 1c allatoms satisfy the rules of valence; in structures 2a through2c the terminal oxygen atom does not.

Each type of bond is associated with a contribution tothe total energy of the molecule, and these contributionsare approximately additive. Typical bond strengths arepresented in Table I. These are to be taken only as a roughguide since the immediate environment of the bond, stericstrain (Section I.B), and resonance (see Section I.C) canhave significant effects.

Atoms with valence shells that contain fewer electronsthan demanded by the valence rules are said to be co-ordinatively unsaturated and usually are carriers of highchemical reactivity (terminal oxygen in 2, the central car-bon in 3). Atoms with valence shells that contain a largernumber of electrons than dictated by the rules are said tobe hypervalent. This situation is rare for atoms of the ele-ments of the second row of the periodic table (presumablydue to their small size and the resulting steric crowding)but fairly common for those of the third and lower rows,where the number of valence-shell electrons can be 10, 12,or even higher. Molecules containing hypervalent atomsare often stable, particularly if the hypervalent atom isof lower electronegativity than its neighbors (e.g., the tin



Organic Chemical Systems, Theory 437

TABLE I Typical Bond Energies in Organic Moleculesa

X X H X C X N X O X C X C

C 100 81 69 84 148 194

N 93 69 38 43 148 213

O 110 84 43 33 172

F 135 105 65 50

Si 72 69 103

P 77 63

S 83 65 128

Cl 103 79 48 50

Br 88 67 53

I 71 57 57

N N 38 N N 100 N N 226

O O 33 N O 145

S S 54 O O 96

a In kilocalories per mole; 1 kcal = 4.184 kJ.

atom in 4; the electronegativity of an element increases asone moves up and to the right in the periodic table).

Those molecules that contain an odd number of elec-trons cannot satisfy the rules for all of their atoms and areknown as free radicals (e.g., 2).

In addition to atomic symbols and symbols for bonds,lone pairs, and unpaired electrons, the classical struc-tural formulas of organic chemistry also indicate atomiccharges (e.g., 3–5):

3

C

CH3

CH3

CH3�

4

Sn� CH3

CH3

CH3

H3C

H3C

5

N C

H

H

H C

H

H O�

O�

The way to determine the charge on an atom is to countthe valence-electron ownership of an atom in the moleculeand compare it with the number of valence electrons on aneutral isolated atom of the same element. If the two agree,the formal charge is zero. If there is one more electron onthe atom in the molecule than on an isolated atom, theformal charge is minus one and so on.

In order to determine the valence-electron ownership ofan atom in a molecular structure, one counts all electronsindicated as lone pairs as well as unpaired electrons on theatom, plus one for each bond in which the atom partici-pates. Thus, one assumes that the two electrons of a bondare shared equally between the two atoms that it joins. Toindicate that this is unrealistic when the two atoms differin their electronegativity, the charges are referred to as for-mal. For molecules containing transition metal elements,the individual formal charges are frequently not indicatedat all.

The sum total of formal charges on atoms in a moleculeis equal to its net charge, and this is always indicated.Negatively charged molecules are called anions, positivelycharged ones cations. The electrostatic force of attrac-tion between two oppositely charged ions is sometimesreferred to as an ionic bond.

Typical bonding situations in which atoms of elementsthat are most commonly found in organic molecules findthemselves in molecular structures are listed in Table II.Analogous bonding situations are found throughout eachcolumn of the periodic table, except that atoms of second-row elements resist hypervalency.

The hydrogen atom does not suffer from steric con-straints in its ordinary univalent state, in which it makesonly one bond. It can enter into a special kind of weak hy-pervalent interaction known as the hydrogen bond, whichattaches it to a lone-pair-carrying second atom. The hy-drogen bond is indicated by a dotted line. As usual forhypervalent interactions, hydrogen bonding is particularlyimportant if the neighbors of the hydrogen atom are highlyelectronegative.

B. Molecular Geometries

Classical structural formulas imply molecular geometries.These are determined by bond lengths, valence angles, anddihedral angles and describe the average nuclear positionswhen the molecule is at equilibrium.

In real molecules, at least some vibrational and inter-nal rotational motion is always present. In many organicmolecules, this can be neglected in the first approxima-tion, and the molecules can be considered rigid. Some,particularly those lacking rings and multiple bonds, aredefinitely floppy at room temperature due to nearly freerotation around single bonds but can be viewed as rigid atsufficiently low temperatures.

1. Bond Lengths

Bond lengths are generally determined by the nature ofthe two atoms bonded, with minor variations dependingon the environment. Each kind of atom can be associatedwith the value of its “covalent radius.” A bond length isapproximately equal to the sum of the covalent radii of theparticipating atoms. Typical lengths of the most commonbonds in organic molecules are listed in Table III.

2. Valence Angles

Valence angles are the angles between two bonds on thesame atom, generally dictated by the coordination numberof the atom. However, if lone pairs are present on an atom,each of these counts for yet another neighbor. On the other




TABLE II Common Building Blocks of Organic Molecules

Number ofelectrons in Coordination

Building blocks valence shell number a

In agreement with the rules of valence

�

�B C �N

C N�

O

N O�F

�

O � F

H 2 1

8 4

8 3

8 2

8 1

Coordinatively unsaturated

�

�B C

�N

C N�O

O FN � �

B C

C N

�ON

�

�

7 3

7 2

7 1

6 3

6 2

6 1

Examples of hypervalent atoms

�Si P

S

t

10 5

10 4

10 3

H (4) (2)

a Parentheses are for hydrogen atom in a hydrogen bond, not always considered hypervalent.

hand, the presence of a single unpaired electron on an atomusually has only a minor influence on its valence angles.

For atoms forming two single or multiple bonds andcarrying no lone pairs, the valence angles normally arein the vicinity of 180◦; for atoms forming three single ormultiple bonds and carrying no lone pairs they are ∼120◦;and for atoms forming two such bonds and carrying onelone pair, they are usually a little smaller. For atoms form-ing four bonds and carrying no lone pairs they are ∼109◦,and for atoms carrying a lone pair plus three bonds or twolone pairs plus two bonds they are a little smaller still.

The angles of 120◦ correspond to the center of an equi-lateral triangle being connected to its vertices, so that allthree bonds are coplanar. The angles of 109◦ correspondto the center of a regular tetrahedron being connected to itsvertices. The steric arrangement around an atom carryinga total of five bonds and lone pairs usually corresponds toa trigonal bipyramid and that around an atom carrying atotal of six bonds and lone pairs to a regular octahedron,with the atom in the center in each case.

The valence angles given provide only a rough guide.Their exact values in any real molecule depend on the




TABLE III Typical Bond Lengths in Organic Moleculesa

X X H X C X C X C

B 1.21 1.56

C 1.09 1.54 1.34 1.20

N 1.00 1.47 1.30 1.16

O 0.96 1.43 1.22

F 0.92 1.38

Si 1.48 1.84

P 1.42 1.87

S 1.34 1.81 1.56

Cl 1.27 1.76

Br 1.41 1.94

I 1.61 2.14

N O 1.36

N O 1.21

N F 1.36

N Cl 1.75

a In A; 1 A = 100 pm.

environment and, in particular, on the number of lone pairson the atom.

The bending motions are generally relatively easy at180◦ valence angles, so that large excursions from thenormally preferred angle are possible at room temperature.Similarly, out-of-plane vibrations around atoms character-ized by 120◦ valence angles are also relatively easy, as isthe interchange of positions of a lone pair and a singlebond. Tetrahedral bond arrangements around an atom arerelatively rigid. However, a flipping motion (“umbrella in-version”) of a lone pair from one to the other side of thethree bonds present is very facile on atoms of the secondrow (not those of lower rows). The interchange of ligandpositions is easy on pentacoordinate atoms and difficulton hexacoordinate ones.

The presence of small rings in the molecule may intro-duce very large deviations from the usual valence anglevalues. For example, in cyclopropane the carbon atomsform an equilateral triangle so that the CCC angle is 60◦,not much more than half of the normally expected value.Such deviations from the normally preferred angles areenergetically unfavorable, and the molecule is said to ex-hibit angular strain.

3. Dihedral Angles

A dihedral angle is defined as the angle between twoplanes, both of which pass through the same bond. Oneof the planes also contains one of the additional bondsformed by one of the bond termini, and the other planecontains one of the additional bonds formed by the otherterminus.

The preferred dihedral angles around a double bond are0◦ and 180◦, once again counting a lone pair on a terminusas another nearest neighbor. Thus, the usual geometriesaround C=C and N=N double bonds are planar (6–9):

C

H3C

H

6

C

CH3

H

C

H3C

7

C

H

H3C H

C

H3C

8

C

H

CH3

H

N

H5C6

9

N

C6H5

However, small twisting distortions from planarity are rel-atively easy at room temperature.

There is a much weaker preference for particular val-ues of the dihedral angle around single bonds, and rotationaround such bonds is nearly free. Usually, the value of 0◦

(“eclipsed”) is avoided, and values of around 60◦ (“stag-gered”) to 90◦ are somewhat preferred, depending on thenumber of lone pairs on the termini.

The general rules just stated for bond lengths andangles permit the construction of mechanical molecularmodels either from balls and sticks or on a computerscreen. The size of the balls that represent the volume ofindividual atoms is given by their van der Waals radii. Thesum of the van der Waals radii of two atoms represents thedistance of most favorable approach of these two atoms ifthey are not mutually bonded (e.g., atoms on neighboringmolecules in a crystal). Values for these quantities arecompiled in Table IV. Molecules in which two or moreatoms that are not bonded to one another and are locatedat distances shorter than the sum of the van der Waals radiiare strained by steric crowding and are less stable thanotherwise expected. Often, it is possible to avoid someof this unfavourable interaction by a distortion of thevalence angles.

4. Molecular Mechanics

It is possible to augment the set of bond energies by a setof energy increments for deviations from optimum bondlengths, valence angles, dihedral angles, and van der Waalsdistances as a function of the magnitude of each and tocompute the energy of a molecule as a function of its ge-ometry within the framework of such a “springs and balls”model. Equilibrium geometry can then be found by energy

TABLE IV Atomic van der Waals Radiia

Atom Radius (A) Atom Radius (A) Atom Radius (A)

H 1.2 O 1.4 F 1.4

N 1.5 S 1.9 Cl 1.8

P 1.9 Se 2.0 Br 2.0

As 2.0 Te 2.2 I 2.2

Sb 2.2

a 1 A = 100 pm.




optimization. This approach is known as molecular me-chanics and provides a good approximation, particularlyfor hydrocarbons, for which extensive and carefully opti-mized parameter sets are available. It runs into difficultieswith molecules in which more than one classical valencestructure is important.

C. Resonance (Mesomerism)

An important concept in classical structural theory is res-onance (mesomerism). This is related to the fact that morethan one classical structure can normally be written for amolecule. A double-headed arrow is usually placed be-tween the structures to indicate that all of them contributeto the description of bonding in the molecule. It is im-portant to note that all of the structures refer to the samemolecular geometry.

The situation is most easily exemplified in the case ofmultiple bonds (e.g., 10–12). The existence of two or morevalence structures can also be indicated by curved arrows,as in 10c to 12c.

C O

H3C

H3C

C

H3C

H3C

O� � Oor

10a 10b 10c

B

H5C2

11aH5C2

N

C2H5

C2H5

B

H5C2

11bH5C2

N

C2H5

C2H5

� ��or B

11c

N�

CH

CHH2NCH

O�

CH

CHH2NCH�

O �CH

CHH2NCH

O�

or

12a 12b 12c

C

H3C

H3C

H5C2

H5C2

C2H5

C2H5

The relative importance of several contributing struc-tures is dictated primarily by their energies (i.e., the ener-gies of hypothetical molecules in which only the structurein question would contribute). The energies can be esti-mated qualitatively. Low energy is favored by the pres-ence of a large number of bonds in the structure (a highlytwisted double bond does not yield much stabilization—steric inhibition of resonance), by the absence of unfavor-able charge separations, and by the presence of negativecharges on electronegative atoms and positive charges onelectropositive atoms.

Two structures contribute equally when they are equiv-alent by symmetry, as in the allyl radical 13:

CH

CH2H2CCH

CH2H2CCH

CH2H2C13a 13b 13c

or

A bond that is single in some and absent in other con-tributing resonance structures is called a partial bond andis often drawn with a dashed line.

The existence of several contributing structures is asso-ciated with a more or less significant effect on the thermo-dynamic stability of the molecule relative to that expectedfrom the simple rules for any one of the contributing struc-

tures. Usually, it results in a stabilization, referred to asresonance energy.

Resonance energy plays a particularly important role incyclic systems. Those cyclic systems with two equivalentdoubly bonded structures that contain 4N+2 electrons inthe perimeter (N is an integer) exhibit a large stabilizationand are known as aromatic. The stabilization is known asthe aromatic resonance energy. The archetypical exampleis benzene (14). Those containing 4N electrons in theperimeter are actually destabilized (antiaromatic); a goodexample is cyclobutadiene (15):

or

14

15

In the case of polycyclic ring systems of conjugateddouble bonds it is more difficult to specify the degree ofaromatic stabilization or antiaromatic destabilization. Agood rule of thumb is to count the number of possiblestructures, ignoring those with an even number of doublebonds located inside any single ring. The higher the num-ber, the larger the stabilization. These kinds of problemsare far more efficiently handled by the more advancedquantum mechanical theories of molecular structure.

In order to simplify notation and to avoid writing a largenumber of contributing structures, it is customary to drawa circle inside an aromatic ring, as shown for 14.

Ambiguities in the writing of molecular structures existeven in compounds containing only single bonds. In prin-ciple, it is possible to write the structures H+H−, H−H+,and H H for molecular hydrogen. The charge-separatedstructures are ordinarily not written, and their existenceis tacitly understood when H H is written; this is true ofall other bonds as well. In some cases the need to writemore than one equivalent classical structure for a moleculecontaining only single bonds is not so easily avoided, andsuch structures are often called nonclassical (e.g., struc-tures 16a–16d for the CH+

5 cation and structures 17a–17dfor the norbornyl cation):

16a

H C

H

HH

H�

H� H

� Hor

H

H�

16b 16c 16d

�

�

�

�or

17a 17b 17c 17d

H C

H

H

H

H C

H

H

H C

H

H

Just as in the case of resonance involving double bonds,such systems with several equivalent structures are




frequently written with auxiliary symbols such as circles,often with dashed lines, particularly to indicate the delo-calization of charge.

D. Isomerism

Molecules that do not differ in overall molecular formula(specified by the number of atoms of each kind and the netcharge) but only in internal structure are called isomers.

1. Constitutional Isomers

Constitutional isomers are compounds that differ in con-nectivity, that is, in the way in which the constituent atomsare connected to one another. Graph theory is an importanttool for their enumeration.

An important class of constitutional isomers are posi-tional isomers, in which the functional groups are the samebut differ in their location within the molecule (e.g., 6 and7, or the ortho, meta, and para isomers 18–20):

18

Cl

Cl

19Cl

Cl

20

Cl

Cl

2. Stereoisomers

Stereoisomers have the same connectivity but differ in theway in which the constituent atoms are oriented in space.They can be divided into configurational stereoisomersand conformational stereoisomers. The precise specifica-tion of the spatial arrangement of the groups in a con-figurational isomer is called its configuration, and in aconformational isomer, its conformation.

a. Configurational stereoisomers. These stereo-isomers cannot be made superimposable by any rotationsabout single bonds. In order to make them superimpos-able, rotation about a double bond or a dissociation of oneor more single bonds, or both, is necessary (e.g., 6 and8). Since these processes normally require considerableenergy, they usually do not occur at a measurable rate atroom temperature. Configurational stereoisomers can nor-mally be isolated from one another and stored essentiallyindefinitely at room temperature.

b. Conformational stereoisomers. Often calledconformers for short, these stereoisomers can be madesuperimposable by rotations about single bonds. Exam-ples are the axial (21) and equatorial (22) conformers of amonosubstituted

H

CH3

H3C

H21 22

cyclohexane. Since rotations about single bonds are nor-mally very facile, it is usually impossible to separateconformers from one another and to handle them sepa-rately at room temperature.

c. Chirality. Another important classification ofstereoisomers into two groups is related to optical activity,manifested by the rotation of a plane of polarized light onpassage through a sample.

A pair of stereoisomers that are related to one another inthe same way as an object and its mirror image are calledenantiomers (e.g., 23 and 24):

F H

H F

H F

F H

23 24

Any pair of stereoisomers that are not related in this wayare called diastereomers (e.g., 6 and 8 or 21 and 22).

A molecule that is not identical with its mirror imageis called chiral and occurs as a pair of enantiomers. Thenecessary and sufficient condition for chirality is the ab-sence of an improper rotational axis of symmetry in themolecule (this includes a center of inversion and mirrorplane symmetry elements).

A mixture of equal amounts of two enantiomers isknown as a racemic modification and is optically inac-tive. A pure enantiomer or an unbalanced mixture of twoenantiomers is optically active; the two enantiomers haveopposite handedness and cause the plane of polarizationto rotate in opposite directions.

II. QUALITATIVE MOLECULAR ORBITALMODEL OF ELECTRONIC STRUCTURE

Although the classical structural theory as outlined so faraccounts for much of organic chemistry, it leaves quitea few observations unexplained. In order to proceed andrelate the chemical behavior of molecules to fundamentalphysical principles it is necessary to use quantum theory.

This can be done at a more or less rigorous quan-titative level. A fairly direct translation of the struc-tural concepts just outlined into mathematical terms thenleads to the quantum mechanical valence–bond theoryof electronic structure. Although conceptually appealing,computationally this method is quite unwieldy and has




not seen much use. An approach that is almost universallyadopted nowadays is the quantum mechanical molecularorbital (MO) theory of electronic structure. This is compu-tationally manageable and is described in some detail later.

Even the mathematically simpler MO approach to elec-tronic structure requires the use of large computers forany quantitative applications. Although large-scale com-putations have been extremely valuable in enhancing theunderstanding of organic molecules, they are in them-selves not appropriate for the day-to-day thinking of benchchemists. However, they have had a great influence on themuch simpler qualitative models of electronic structureadopted for daily use by organic chemists and even someeffect on that ill-defined body of knowledge generally re-ferred to as the organic chemist’s “intuition.” The currentqualitative model of molecular electronic structure basedon the qualitative notions of MO theory represents a sig-nificant advance over the structural theory outlined in thepreceding sections, but is not easy to describe unequivo-cally, since its form varies among individuals.

Current qualitative thinking about the electronic struc-ture of organic molecules is based on the independent-particle model, in which it is assumed that the motion ofeach electron is dictated by the field of stationary nucleiand the time-averaged field of all the other electrons. Inthis model, any correlation of the instantaneous positionsof the many electrons present is neglected.

Only in several well-recognized and more or less ex-ceptional situations are correlation effects explicitly in-troduced. This is particularly true in the treatment of bi-radicals, which represent an important class of reactionintermediates but which are not discussed here, in the treat-ment of several other unusual bonding situations, and inthe treatment of photochemical processes (e.g., in the con-sideration of differences in the reactivity of excited singletand triplet states).

A. Atomic Orbitals

The independent-particle model is well known from thequantum mechanical description of atomic structure. Eachelectron in an atom is assumed to reside in an atomic or-bital (AO) with a maximum of two electrons (of oppositespin) in any one orbital (Pauli principle). The AO is afunction of the coordinates of one electron. The squareof its magnitude at any point in space gives the proba-bility density for finding the electron at that point. Themagnitude itself can be positive or negative, with zerovalues at the boundaries of the positive and negative re-gions. The boundaries are referred to as nodal surfaces(planes, spheres, etc.). The energy of an electron residingin an orbital increases with the increasing number of nodalsurfaces in the orbital.

Atomic orbitals are characterized by a principalquantum number n (there are n − 1 radial nodes in anAO) and a letter indicating their shape as dictated by theangular nodes: An ns orbital has no angular nodes andis of spherical symmetry; each of the three equienergeticnp orbitals has one angular mode (usually taken to bea plane through the nucleus, with respect to which theorbital is antisymmetric); each of the five nd orbitals hastwo angular nodes; and so on. Shorthand abbreviationsfor orbital shapes are shown in Fig. 1A. These are meantto indicate the regions of space in which the numericalvalue of the AO is the largest, also known as the lobes ofan AO, as well as their signs.

In the ground state of an atom, the AOs are assumedto be occupied by the available electrons in the order ofincreasing energy of the subshells: 1s, 2s, 2p, 3s . . . . Thelast at least partially occupied subshells and the more sta-ble ones of the same principal quantum number representthe valence shell [in transition metals this contains thend, (n + 1)s, and (n + 1)p subshells]. Only valence-shellAOs are normally considered important for bonding: the1s orbital in hydrogen and the 2s in lithium and beryllium(2p are very close in energy and could be included as well),2s and 2p in boron through neon, 3s in sodium and mag-nesium (3p could be included), 3s and 3p in aluminumthrough argon, and so on.

The qualitative theory of bonding also makes use ofcombinations of valence AOs known as hybrid orbitals.Combining an s with a p orbital produces two equivalentsp hybrids pointing in opposite directions. Combining ans with two p orbitals produces three equivalent sp2 hy-brids pointing to the corners of an equilateral triangle.And combining an s with three p orbitals produces fourequivalent sp3 hybrids pointing to the corners of a regular

FIGURE 1 (A) Atomic orbitals. (B) Hybrid orbitals.




tetrahedron (Fig. 1B). Inequivalent hybrids can be usedfor other desired valence angles.

B. Molecular Orbitals

The independent-particle model can be extended tomolecules by assuming that each electron again residesin an orbital of well-defined energy. Now, however, ex-cept for the inner-shell AOs, the orbital is spread over thespace spanned by the molecular framework and is referredto as a molecular orbital. To a very good approximation,the inner-shell electrons behave in exactly the same wayin the molecule as they would in an isolated atom, andwe will not be concerned with them further. Instead, weconcentrate on the valence MOs.

The delocalized form of MOs that we have just de-scribed is known as their canonical form. It can be shownthat the wave function describing the ground electronicconfiguration of a molecule, containing a certain num-ber of doubly occupied orbitals, does not change at allwhen these occupied MOs are mixed with one anotherin an arbitrary way. This degree of freedom in the wavefunction permits the construction of MOs that have beenmixed in such a way that each one is localized in thesmallest amount of space possible, using one of severalpossible criteria. The price one pays is that these localizedMOs no longer have well-defined individual energies, al-though the total energy of the system is just as well definedas before.

It turns out that each of the localized MOs tends to befairly well contained within the region of one bond or onelone pair; only in molecules with several important classi-cal structures is this localization poor. On close inspection,the localization is actually never perfectly complete, andeach localized orbital possesses weak “tails” extending toother parts of the molecule. Except for the exact nature ofthese tails, such a bond orbital often looks very much thesame in all molecules that contain that particular bond,say, C C. It is along these lines that one can begin tounderstand bond additivity properties and their failure inthe case of molecules in which several classical resonancestructures play an important role.

The standard qualitative model of molecular electronicstructure requires the construction of a number of valenceMOs sufficient to hold all the valence electrons. This isnormally done separately for the electrons responsible forthose bonds that are present in all important classical struc-tures of the molecule (“localized” bonds, hence “local-ized” electrons, although strictly speaking, they are notreally localized) and separately for the electrons responsi-ble for partial bonds that are present in some and absent inother important classical structures (“delocalized” bonds,hence “delocalized” electrons).

The usual approach is based on the recognition of thefact that the formation of bonds in molecules representsonly a small perturbation of electronic structure of theconstituent atoms, bonding energies being of the order of1% of total atomic energies. This is because the electricfield in the vicinity of any one atom, where an electronspends most of its time, is dominated by the nucleus ofthis atom, since the force of attraction of an electron bya positive charge grows with the second inverse power ofthe distance from that positive charge.

This situation is acknowledged by assuming that eachMO can be built by mixing AOs or hybrid orbitals.

C. Orbital Interactions

A brief consideration of the ways in which orbitals interactwill be useful not only for a description of the procedurein which AOs are mixed to produce MOs, but also forsubsequent reference to the mutual mixing of MOs, forinstance, those of two molecules reacting with one another.

In the mixing of two orbitals, the important factors aretheir energies and the interaction matrix element (this isfrequently referred to as the resonance integral β betweenthe two orbitals). As a result of the mixing of n orbitals, nnew orbitals result.

It is by far easiest to consider the case of only two mutu-ally interacting orbitals φ1 and φ2 (Fig. 2). Let the energyof φ1 be below that of φ2. After the interaction, the energy

FIGURE 2 Orbital energies. (A) Interaction of orbitals φ1 and φ2 toproduce linear combinations ψ1 and ψ2. (B) Interaction of atomicorbitals (outside levels) to produce a bonding and an antibondingcombination (inside levels, orbital shape indicated). The effect onthe net energy of the system is indicated for occupation with oneto four electrons. Arrows represent electrons with spin up andspin down. At the bottom, the classical structural representationis shown for comparison.




of the more stable of the two new resulting orbitals ψ1 willlie below that of φ1 and that of the less stable one ψ2 willlie above that of φ2, as if the two original energies repelledone another. The amount by which each orbital energy isshifted will depend on the strength of the interaction ele-ment β and on the difference ε in the original energies. Ifε � |β|, the shifts are approximately inversely propor-tional to ε (this approximation is known as first-orderperturbation theory). When the two orbitals are originallydegenerate (ε = 0), the shifts are the largest.

An important characteristic of the interaction be-tween two orbitals φ1 and φ2 is their overlap integralS12 = ∫

φ1φ2 dτ , where the integration is over all space.For normalized orbitals (S11 = S22 = 1) the value of S12

can vary between −1 and +1. Usually if S12 is positive,the interaction element β is negative, and if S12 is negative,β is positive. If S12 = 0, orbitals φ1 and φ2 are said to beorthogonal; they can still have a nonvanishing resonanceintegral. The energy shifts caused by the interaction oftwo orthogonal orbitals are opposite in direction but equalin magnitude. When two nonorthogonal orbitals interact,the less stable new orbital ψ2 is destabilized more than themore stable new orbital ψ1 is stabilized.

The more stable of the two new orbitals is referred to asbonding and the less stable as antibonding with respect tothe interaction considered. An orbital with the same en-ergy as before the interaction is called nonbonding; suchorbitals often result when more than two AOs are mixed.If two electrons are available to fill each bonding orbital,maximum stabilization will result relative to the situationbefore orbital interaction. If too many electrons are avail-able and some must be placed into antibonding orbitals,some or all of the stabilization is lost and net destabi-lization may result. For this reason, the interaction of adoubly occupied with an unoccupied orbital leads to a netstabilization and the interaction of two doubly occupiedorbitals to a net destabilization.

These general concepts can now be specialized to thecase of mixing of AOs or hybrid orbitals to produce bondorbitals. Two valence-shell AOs or properly constructedhybrid orbitals located at the same atom are always orthog-onal. For two orbitals located at the same center, the inter-action element β is zero if one or both are pure AOs. How-ever, two hybrid orbitals at the same center have a nonvan-ishing mutual interaction element. Its magnitude is relatedto the promotion energy (i.e., the energy difference be-tween the AOs from which the hybrids were constructed).

Two AOs or hybrid orbitals located at different atomscan be, but do not need to be, orthogonal. A nonvanishingoverlap integral between two p AOs can be produced intwo geometrically distinct ways. When the axes of the twoorbitals are aimed at one another, the overlap is said to beof the σ type; when they are parallel, it is said to be of the

FIGURE 3 Overlap between orbitals.

π type (Fig. 3). Intermediate situations are also possible.The interaction matrix element β between two AOs orhybrid orbitals located at different atoms is approximatelyproportional to the negative of their overlap integral S12.

The two new orbitals ψ1 and ψ2 that result from theinteraction are linear combinations of the two old orbitals(Fig. 2). In the bonding orbital ψ1, the overlapping portionsof the entering old orbitals ψ1 and ψ2 have the same sign.In the antibonding orbital ψ2, the overlapping lobes of theoriginal partners are of opposite signs, and its sign changesas one goes across the interaction region from the centerof φ1 to the center of φ2. The surface where the sign of ψ2

changes is called a nodal surface (plane).In summary, the bonding combination ψ1 lacks a nodal

plane and has an increased electron density between theatoms (in-phase mixing if S12 > 0), while the antibondingcombination ψ2 has a nodal plane and a reduced elec-tron density between the atoms (out-of-phase mixing ifS12 < 0). The bonding orbital ψ1 will accommodate twoelectrons and thus produce a net stabilization (a chemicalbond). This is the origin of the classical “shared electronpair.” For maximum bonding stabilization, it is desirableto maximize the absolute value of the overlap integral S12.

A third electron would have to enter the antibondingorbital φ2, causing a loss of much and possibly all of theoverall stabilization. A fourth electron would also have toenter φ2, and now interaction would definitely lead to anet destabilization. For this reason, filled AOs (lone pairs)repel and avoid one another (“lone-pair repulsion”). Forminimum destabilization, it is desirable to minimize |S12|.For instance, if the lone pairs are of the p type and are onadjacent atoms, a dihedral angle of close to 90◦ will bepreferred (e.g., in H2O2).




If the original orbitals φ1 and φ2 are degenerate, ε = 0,they will enter the new orbitals ψ1 and ψ2 with equalweights. If the initial energies are unequal, the more stableinitial orbital φ1 will enter with a numerically larger coef-ficient into the more stable resulting orbital ψ1 and with asmaller coefficient into the less stable resulting orbital ψ2.The opposite will be true for the less stable initial orbitalφ2. This result is quite logical: In the more stable of thetwo orbitals, the electrons spend more time on the moreelectronegative atom.

D. Construction of Molecular Orbitals

1. “Localized” Bonds

In the first step, AOs on the participating atoms are mixedto produce hybrids pointing in the directions of the desiredbonds. These are then combined pairwise, each pair pro-ducing a bonding and an antibonding orbital strictly local-ized at the atoms to be bound. Occupancy of the bondingorbital by two electrons then produces a localized bond,while the antibonding orbital remains unused. If electronsare left over after all bonding orbitals have been occupiedtwice, they are placed into remaining AOs or hybrid or-bitals that have not been used in the above-mentioned pro-cedure (nonbonding lone pairs or unpaired electrons). Sofar, the picture resembles the classical description, withoutresonance.

At a more advanced level, it is recognized that the sev-eral bonding and antibonding orbitals located at a singleatom have nonvanishing interaction elements with one an-other. If they are allowed to mix, they will produce thefully delocalized canonical MOs. This usually has a pre-dictable effect on the total energy, which can be absorbedin bond additivity schemes. Exceptions are the case ofcyclic delocalization (e.g., in cyclopropane) and certaincases of stabilization of radicals, carbenes, and biradi-cals. A well-known exception in which such delocalizationmust be considered involves the interaction of a lone pairon an atom such as oxygen with the orbitals correspondingto bonds leading to its neighbors, which dictates certainstereochemical preferences known as the anomeric effect.

2. Delocalized Bonds

Although delocalized bonds could be obtained in the samemanner in principle, it is more efficient to proceed moredirectly. All n AOs (and possibly hybrid orbitals) that par-ticipate in the “delocalized” system are mixed at once,producing n canonical delocalized MOs directly. Typi-cally, these are sets of orbitals characterized by mutual π

overlap and referred to as the π -electron system. A goodexample are the six p orbitals of the sp2 hybridized car-

FIGURE 4 (A) Hydrogen Is atomic orbitals and carbon sp2 hybridorbitals (major lobes only) needed for the description of “localized”σ bonds in benzene. (B) Carbon 2ρz atomic orbitals needed forthe description of “delocalized” π bonds in benzene.

bons of benzene that are left over after the localized bondframework has been constructed (Fig. 4).

On the one hand, the process of simultaneous mixing ofmore than two orbitals is more difficult to visualize, but onthe other hand, there is usually only one participating or-bital on an atom, simplifying matters somewhat. Usually,about half of the resulting MOs are stabilized (bonding),perhaps one or a few are nonbonding, and about a half aredestabilized and antibonding. Their approximate formscan be guessed from the requirement that a nodal surfaceis added each time that one goes to the energetically nexthigher MO. Mnemonic devices for the resulting energypatterns are shown in Fig. 5 for linear and cyclic polyenes.In the latter case, the origin of the special stability (“aro-maticity”) of 4N + 2 electron systems is apparent: This isthe electron count needed to produce a closed shell. In gen-eral, however, at least a back-of-the-envelope numerical

FIGURE 5 Conjugated π systems; orbital energies and shapes(top view, white represents positive, black represents negativesign). (A) Cyclic conjugation (annulenes). (B) Linear conjugation(polyenes).




calculation is necessary. Here, Dewar’s perturbational MO(PMO) method is particularly useful.

The interaction of fully or partly localized orbitals isfrequently referred to as conjugation. Interactions betweenseveral AOs that belong to a π system can be referred toas π conjugation, or conjugation for short. Interactions ofan orbital of the π type with one or more suitably alignedbond orbitals of the σ type are referred to as hyperconjuga-tion. Interactions between adjacent bond orbitals of the σ

type or one such orbital and a lone-pair orbital are referredto as σ conjugation. As noted above, they occur throughthe interaction between two hybrid orbitals located on thesame atom.

3. Hypervalent Bonding

The procedures described so far will not work for hyperva-lent atoms that do not have enough valence AOs to providea sufficient number of localized bond orbitals. Withoutgoing into detail, we note that in this case three-center or-bitals can be constructed instead, and a satisfactory simpledescription results.

The expansion of the electron count in the valence shellof an atom of a main-group element beyond eight is areflection of the inadequacy of the way in which the clas-sical rules of valence assign electrons into a valence shell.In reality, the true total electron density in the region ofspace corresponding to an atomic valence shell does notreach these high formally assigned numbers. Fundamentallimitations on the latter are given by the Pauli principle,which demands that any electron density in excess of anoctet has to occupy orbitals of the next higher principalquantum number (Rydberg AOs). It is a gross oversimpli-fication, however, to pretend that the two electrons of abond contribute full occupancy of two to the valence shellof each of the atoms connected by the bond, particularlywhen the two atoms differ greatly in electronegativity. Thedegree to which the valence shell of the atom of the moreelectropositive element is actually filled is overestimatedby the simple rules, and this accounts for the ease withwhich it enters into hypervalency.

III. QUANTITATIVE ASPECTS OFMOLECULAR STRUCTURE

A. Potential Energy Surfaces

1. Construction of Potential Energy Surfaces

Before attempting a quantitative quantum mechanicaltreatment of molecules it is customary to separate themotions of nuclei from those of electrons. This separa-tion is known as the Born–Oppenheimer approximation,

and it underlies all of the current thinking about molec-ular structure. It is justified by the much larger mass ofnuclei compared with that of electrons, which causes thenuclei to move much more slowly than electrons. Thus,electrons adjust their motions essentially instantaneouslyto any change in the location of the nuclei, as if they hadno inertia.

From the viewpoint of quantum theory a molecule isa quantum mechanical system composed of atomic nu-clei and electrons. It has an infinite number of stationarystates, that is, states whose measurable properties do notchange with time. Each of these states is characterized byan energy and a wave function. A wave function is a pre-scription for assigning a numerical value (“amplitude”) toevery possible choice of coordinates for all particles in thesystem. A square of the number assigned to any choice ofthese coordinates represents the probability density that ameasurement will find the system at that particular collec-tion of coordinates.

In order to find the stationary states of a quantum me-chanical system, their energies, and wave functions, onemust solve the Schrodinger equation Hψ = Eψ , whereH is the Hamiltonian operator of the system, ψ the wavefunction, and E the energy of the stationary state.

In order to obtain a quantum mechanical descriptionof a molecule within the Born–Oppenheimer approxima-tion, at least in principle, one proceeds as follows. Fixedmolecular geometry is assumed. Mathematically, this cor-responds to choosing a point in the nuclear configurationspace. As soon as this is done, the Hamiltonian operatorfor electronic motion is fully defined so that the corre-sponding Schrodinger equation can be solved for ψ andE . An infinite number of solutions exist, differing in theirenergies, E , and wave functions, ψ .

One of the characteristics of an electronic wave func-tion is the number of unpaired electrons it contains. Thosewave functions in which this number is zero describe sin-glet states. Those that contain two unpaired electrons de-scribe triplet states, and so on. Wave functions of radicalscan have one unpaired electron (a doublet state), three (aquartet state), and so on.

Of the infinite number of stationary wave functions thatare solutions to the Schrodinger equation for a chosen nu-clear geometry, one is of lowest energy. Almost alwaysthis is a singlet wave function if the organic molecule hasan even number of electrons and a doublet wave functionif it has an odd number of electrons. In the following, weassume an even number of electrons. We label the lowestenergy singlet wave function ψ(S0) and its energy E(S0).The next higher energy singlet wave function is then iden-tified and labeled ψ(S1), and its energy is identified andlabeled E(S1). This is the wave function of the first excitedsinglet state. Similarly, the wave functions of the second




and higher electronic excited singlet states can be identi-fied. Among the triplet wave functions the one with thelowest energy is called ψ(T1) and its energy E(T1). Simi-larly, higher triplet state wave functions and their energiesare identified.

A different molecular geometry is then chosen and theprocess repeated. While this is difficult to do in practice,one can at least imagine performing this kind of opera-tion for all possible molecular geometries. In a plot ofE(S0) against the values of the geometrical parametersthat describe the molecular structure, a surface will thenresult. This is the potential energy surface for this particu-lar electronic state of the molecule (the potential energy ofthe molecule is its total energy minus the energy of overalltranslational, rotational, and vibrational motion).

The resulting surface is easy to visualize if only one ortwo geometrical variables are used to describe the molec-ular structure. As shown in Fig. 6, in the former case theset of points E(S0) represents a line; in the latter case itrepresents a two-dimensional surface, often displayed inthe form of a contour diagram. For all organic molecules ofreal interest, the number of independent geometrical vari-ables necessary for the description of the internal geometryis large (3N − 6, where N is the number of atoms). Theresulting surfaces are multidimensional and difficult to en-visage. Frequently, they are referred to as hypersurfaces.

What can still be visualized readily are one-dimensionalor two-dimensional cross sections through these hyper-surfaces, which correspond to only a limited variation ofmolecular geometries, particularly to specific kinds of in-tramolecular motion, such as rotation around a bond.

2. Motions on the Surfaces

The gradient of the potential energy surface defines theforces acting on the nuclei. The resulting changes ofmolecular shape can be represented by a point that moves

FIGURE 6 Display of a one-dimensional (A) and a two-dimensional (B) cut through a potential energy surface. In (A),energy is plotted against a geometrical variable Q. One minimumis present. In (B), contour lines connect points of equal energyin the Q1, Q2 geometrical space. Two minima are present. Theseparation into two catchment basins is shown by a dashed line;the transition state structure is indicated by a double dagger.

through the nuclear configuration space. It is possible tovisualize the vibrational motions of the molecule as wellas its internal rotations as the motions of a marble rollingon the potential energy surface.

If the molecule is isolated, its total energy will remainconstant and the marble will perform endless frictionlessmotion on the surface, trading the potential energy againstthe kinetic energy of nuclear motion and vice versa. If themolecule exchanges energy with its environment, it willtend to lose any excess energy it may have and settle inone of the valleys or minima on the surface.

A proper description of the motion that correspondsto vibrations and internal rotations again must be quan-tum mechanical since even the relatively heavy nuclei re-ally obey quantum rather than classical mechanics. Onceagain, one can find the stationary states of the vibrationalmotions and their wave functions and energies by settingup the appropriate Schrodinger equation and solving it.The Hamiltonian operator that enters into this equationnow contains the information on the potential energy em-bodied in the shape of the potential energy surface. An in-finite set of possible solutions again exists. A finite numberof solutions have energies corresponding to bound states,that is, those with energies below the dissociation limit forthe molecule (energy required to break the weakest bondand separate one of the atoms to infinity). The wave func-tion of lowest energy represents the vibrational groundstate of the molecule. In this state the kinetic energy ofthe nuclear motion is not zero since this would violatethe uncertainty principle. It is referred to as the zero-pointenergy.

Most molecules have a well-defined equilibrium ge-ometry that corresponds to a minimum in the E(S0) sur-face. The wave function of the lowest vibrational station-ary state is heavily localized near this minimum. If theshape of the potential energy hypersurface in this vicin-ity can be approximated by a paraboloid, the vibrationalmotion in the lower vibrational states is harmonic and canbe described as a product of 3N − 6 normal mode mo-tions (3N − 5 for a linear molecule), each characterizedby a frequency ν. The zero-point energy is obtained bysumming the contribution 1

2 hν from each of the normalmodes. The value hν is equal to the energy separationfrom the lowest to the next higher energy level in eachmode. The transitions between these individual levels liein the region from several hundred to several thousandwave numbers and are commonly studied by infrared andRaman spectroscopy.

In most cases more than one local minimum is found onthe E(S0) hypersurface (Fig. 6B). This means that a givencollection of nuclei and electrons has more than one possi-ble equilibrium geometry. Usually, this means that severalisomers of the molecule exist, but some of the minima may




also correspond to dissociation products, the sets of two ormore smaller molecules formed from the same collectionof nuclei and electrons. Some of the minima may haveequal energies (e.g., a pair of enantiomers). In general,one minimum is of lower energy than all the others andcorresponds to the most stable isomer of the molecule.

Each minimum is surrounded by its catchment basin(Fig. 6B). Within a given basin, the potential energy sur-face slopes toward the minimum in question. Each catch-ment basin defines the range of geometries that correspondto a given chemical species.

3. Chemical Reactions

Under ordinary conditions molecules are frequently notin stationary vibrational states. Vibrations and internalrotations are affected by collisions with neighboringmolecules, which add or subtract small amounts of vibra-tional energy more or less randomly. At thermal equilib-rium the vibrational energy content is 1

2 kT for each degreeof freedom. It is this supply of random thermal energy thatpermits molecules to escape from one catchment basin toanother in a thermal reaction.

In order to move from one minimum to another it isnecessary to overcome ridges that separate them (Fig. 6B).This is done most easily by travel through saddle points,which correspond to transition structures and are usuallycalled transition states. Some of these lie only a littlehigher in energy than a starting minimum. Travel overthese saddles is easy even at low temperatures, and a chem-ical species corresponding to a shallow catchment basinmay not be isolable at room temperatures or even below.Conformational isomers are a good example of such chem-ical species. Other catchment basins may be deeper andthe surrounding saddle points difficult to reach. They oftencorrespond to configurational or constitutional isomers.

An important attribute of a saddle point connecting twocatchment basins is the width of the saddle. Very nar-row saddles are difficult for the rolling marble to findand decrease the probability of travel from one catchmentbasin to another. Wide saddles accommodate a much largerflux under otherwise identical conditions and lead to fastmotion.

For a reaction whose rate is not limited by the rate ofdiffusion, the reaction rate is well described by Eyring’stransition state theory. For the rate constant k, we have thefollowing:

k = κK T

hexp

(S‡

R− H ‡

RT

)

,

where κ is the transmission coefficient, k the Boltzmannconstant, T absolute temperature, h Planck’s constant, Rthe gas constant, H ‡ the activation enthalpy (i.e., the

difference from the lowest vibrational level in the origi-nal catchment basin to the lowest level available over thesaddle point), and the activation entropy S‡ is a functionof the width of the saddle relative to that of the startingminimum.

Isotopic substitution does not have any effect on thepotential energy surfaces as usually defined. However, itdoes affect the dynamics of the vibrational process and thedynamics of the motion over saddle points by changingthe magnitude of the energy of zero-point motion and thespacing of the vibrational stationary states and thus thedensity with which these states are packed. Isotopic sub-stitution thus leads to changes in vibrational spectra andalso to changes in reaction rate constants.

Some vibrational stationary states have energies that lieabove some of the lower saddle points. In such states themolecule is free to travel from one catchment basin toanother and in that sense has no fixed chemical structure.This sort of situation occurs, for instance, for rotationsaround single bonds at elevated temperatures.

The perfect quantum mechanical analogy to the rolling-marble description given earlier is described in terms ofthe motion of a wave packet, that is, of a nonstationarywave function initially more or less strongly localized ina particular region of nuclear geometries. Since the vibra-tional levels are spaced quite closely together, however,the classical description in terms of the rolling marble isoften quite adequate.

In the Born–Oppenheimer approximation, the overallrotation of a molecule can also be uncoupled from otherkinds of motion. For an isolated molecule, the solutionof the Schrodinger equation for rotational motion leads toa set of very closely spaced stationary levels. Transitionsbetween these levels occur in the microwave region of theelectromagnetic spectrum. In solution, the quantization ofrotational motion is usually destroyed by intermolecularinteractions, so that it has little importance in theoreticalorganic chemistry.

So far we have concentrated on the lowest singlet hy-persurface E(S0). However, motion can be studied simi-larly on other hypersurfaces if they can be calculated tostart with. Such electronically excited states are usuallyproduced either by absorption of light in the ultravioletand visible regions, as studied by electronic spectroscopy,or by energy transfer from another electronically excitedmolecule. The latter is particularly useful in the case of ex-citation into a triplet state, since excitation from a groundsinglet to a triplet state by direct absorption of light has avery low probability.

The study of the processes involving the higher poten-tial energy surfaces is the domain of photochemistry andphotophysics. These are very difficult or impossible tounderstand in terms of the classical picture described in




Section I, and it is thus easy to see why photochemistsare more likely than any other organic chemists to discusschemistry in terms of potential energy surfaces. Furtherdiscussion of photochemically induced reactions can befound below.

4. Reaction Coordinate Diagrams

For many purposes it is useful to condense the multidimen-sional complexity of molecular motion from one catch-ment basin to the next into a one-dimensional reactioncoordinate diagram in which the degree of progress fromthe first minimum over the transition state to the secondminimum is plotted horizontally as the so-called reactioncoordinate. The quantity plotted vertically can be the po-tential energy E , which we have been discussing so far,but then all information that has to do with the propertiesof the surface along dimensions other than the reactioncoordinate is lost. It is more common to plot instead theGibbs free energy G = H − T S corresponding toall degrees of freedom other than the reaction coordinate(H is enthalpy, S is entropy). In this manner, informa-tion on the entropic constraints dictated by the shape ofthe potential energy surface in directions perpendicular tothe reaction path is preserved.

The effects of structural perturbations on such reactiondiagrams can be relatively easily envisaged. For instance,in many cases a structural factor that will stabilize theproduct relative to the starting material will also tend tostabilize the transition state, albeit to a smaller degree, asis indicated schematically in Fig. 7. Reactions of this kindare said to follow the Bell–Evans–Polanyi principle. Suchreactions tend to obey linear free energy relationships,

FIGURE 7 Reaction coordinate diagram for two similar reactionsdiffering in G. Location of the transition states is indicated bydouble daggers.

which tie thermodynamic quantities such as equilibriumconstants to rate quantities such as kinetic rate constants:G‡

1 − G‡2 = α(G0

1 − G02), where G‡ is the free

energy of activation and G0 the free energy of reac-tion. Perhaps the best known example is the Brønsted law,which relates equilibrium and kinetic acidity. In general,linear free energy relationships interrelate changes in freeenergies or free energies of activation for a series of reac-tants, usually differing by substitution. They are useful inmechanistic studies.

Cases are also known in which structural perturbationsact quite differently on the transition state and the product.Then, reactions with larger equilibrium constants do notnecessarily proceed faster than those with smaller ones.

Another consequence of a parallel stabilization of aproduct and of the corresponding transition state is thedisplacement of the transition structure toward the start-ing materials along the reaction coordinate (Fig. 7). Thosereactions that obey the Bell–Evans–Polanyi principle thuswill have earlier transition states relative to other, similarreactions if they are more favorable thermodynamicallyand a later transition state if they are less favorable. Thisstatement is known as the Hammond postulate.

At times it is useful to separate motion not only alongthe reaction coordinate, but also along one other directionselected from all the others and to plot the free energyof the reacting system against two geometrical variables.Diagrams of this kind are particularly popular in the studyof substitution and elimination reactions and are known asMore O’Ferrall diagrams. The geometries correspondingto the starting material and the product lie at diagonallyopposed corners of a square (Fig. 8) and are connected byan energy surface that rises up to a saddle point and thendescends to the other corner. The effects of a change inthe relative stability of the starting materials and products

FIGURE 8 More O’Ferrall contour diagram for a reaction requir-ing two types of nuclear motion that can proceed more or lesssynchronously. The reaction coordinate is shown as a dashedline, and the structure of the transition state is indicated by a dou-ble dagger. The two motions involved could be, for instance, theapproach of a nucleophile and the departure of the leaving groupin a nucleophilic substitution reaction.




on the position of the transition state are given by theHammond postulate: The transition structure moves awayfrom the corner that has been stabilized. The shape of thesurface along the other diagonal is just the opposite: Thereaction path is of low energy relative to the two corners.For this reason a change in the relative stability of thegeometries represented by the two corners has an effectopposite to that discussed in the Hammond postulate. Astabilization of one of the corners will move the transitionpoint closer to that corner.

While it can be generally assumed that the molecules inthe initial catchment basin are almost exactly at thermalequilibrium, which is being perturbed quite insignificantlyby the escape of the most energetic molecules over a saddlepoint or saddle points, this is not always true. In hot groundstate reactions a molecular assembly is initially generatedwith a vibrational energy content much higher than wouldcorrespond to the temperature of the surrounding medium,perhaps as a result of being born in a photochemical pro-cess or in a very highly exothermic thermal process. Suchan assembly can then react at much higher rates.

A study of the reaction rates of molecules starting inindividual vibronic levels provides much more detailedinformation than the study of molecules that are initiallyat thermal equilibrium but has so far remained a domainof small-molecule chemical physicists and has had littleimpact on the theory of organic reactions.

An important point to consider is the possibility that amolecule has a choice of more than one saddle over whichit can escape from an original catchment basin. The natureof products isolated from such competing reactions candepend on the choice of reaction conditions. Under kineticcontrol the molecules of the two products are not givenan opportunity to travel back over their respective saddlesinto the initial catchment basin. The preferred product willbe that which is being formed faster. On the other hand,if the molecules are provided with an opportunity to returnto the initial catchment basin so that an equilibrium iseventually approached, the reaction is said to be run underthermodynamic control and the thermodynamically morestable product will be isolated.

B. Electronic Wave Functions

Once a wave function of a quantum mechanical state isavailable, it is possible to calculate its observable prop-erties in a straightforward manner. Unfortunately, theSchrodinger equation is very difficult to solve, and thishas not been achieved exactly for any but the simplestmolecular systems. The fundamental mathematical prob-lem can be traced to the fact that even for fixed nuclearpositions the molecule still represents a many-body sys-tem if it contains more than one electron. Nevertheless, it

is possible to obtain useful approximate solutions. Thesebecome gradually less accurate as the molecule in questionincreases in size.

Essentially all of the approximate methods for the calcu-lation of electronic wave functions for an organic moleculestart with a basis set of AOs mentioned above in the qual-itative discussion. In a so-called minimum basis set de-scription, the AOs assigned to an atom in a molecule arethose that would be occupied in a neutral isolated atomof that element plus the remainder of its valence shell.The AOs can be used in two ways for the construction ofmolecular electronic wave functions.

1. The Valence–Bond Method

In the conceptually simple but computationally very diffi-cult valence–bond (VB) method one first assigns all avail-able electrons to individual AOs in a way that satisfies thePauli principle; that is, placing no more than two elec-trons in any one AO and assigning opposite spins to elec-trons that share the same AO. The method of doing thisis not described in detail here, but it is possible to ensurethat the resulting many-electron wave function is properlyspin-adapted and represents a singlet, a doublet, a triplet,and so on, as appropriate. An example is the H+H− VBstructure for the H2 molecule, in which the two availableelectrons are both assigned to the 1s orbital of the hy-drogen on the right. Since they are then necessarily spin-paired this represents a singlet structure.

It is generally possible to write a very large number ofsuch VB structures for any given molecule. For instance,for the H2 molecule we could equally well have writtenH−H+ or four additional structures in which one electronwas assigned to each of the available 1s AOs and whichdiffered only in the assignment of electron spin. In orderto obtain the desired final electronic wave function ψ it isnecessary to allow these VB structures to mix.

Mathematically, this corresponds to taking a linear com-bination of all structures of the same spin multiplicity. Inorder to determine the coefficients with which each of thestructures enters into the final mixture, one uses the vari-ational theorem, which states that the best approximationto the ground wave function ψ(S0) is given by those co-efficients that lead to a wave function of lowest energy.Determining the coefficients requires the diagonalizationof a matrix the dimension of which is equal to the num-ber of VB structures of a given multiplicity possible forthe molecule. This increases very rapidly with the size ofthe molecule, making a full calculation impractical evenfor quite small organic molecules. Another important dif-ficulty in the calculation is the complicated nature of thecomputation of each of the elements of the matrix, due ulti-mately to the fact that the AOs are not orthogonal.Because




of these mathematical difficulties the VB method has nothad much numerical use, although it remains important asa conceptual tool.

2. The Molecular Orbital Method

The way in which almost all computations of molecularelectronic wave functions are performed nowadays is theMO method already mentioned. In this method it is rec-ognized from the outset that in a molecule electrons aredelocalized over the whole region of space spanned by thenuclear framework. Accordingly, AOs are first combinedinto MOs. Unlike the VB structures mentioned above,these are still one-electron wave functions.

Mathematically, the MOs are written as linear combi-nations of atomic orbitals. From n AOs it is possible toconstruct up to n linearly independent MOs. This is morethan is usually needed since in the next step, in which elec-tron occupancies are assigned to these molecular orbitals,each can be occupied by up to two electrons of oppositespins. The unused MOs are referred to as unoccupied (orvirtual). Such an assignment of electron occupancies toMOs results in a many-electron wave function known asa configuration. The number of configurations possibleis usually very large and is in fact equal to the numberof VB structures possible for the same spin multiplicity(singlet, doublet, etc.). In order to obtain the final elec-tronic wave function ψ it is now necessary to mix allpossible MO configurations in a fashion analogous to thatused to mix the VB structures before. This configurationmixing is referred to as configuration interaction (CI). In-deed, the final wave function ψ obtained by the VB pro-cedure in which all VB structures are included and thatobtained by the MO procedure in which all configurationsare included will be one and the same. It is referred toas the full configuration interaction (FCI) wave function.It cannot be obtained in practice for any but the small-est organic molecules because the number of possibleconfigurations increases astronomically with molecularsize.

The MO path to the FCI wave function ψ involves onemore step than the VB path, and this appears to be morecomplicated at first sight. However, the availability of theadditional step endows the MO method with more flexibil-ity than the VB method, which permits a mathematicallymuch simpler formulation. This flexibility arises in thestep of combining AOs in MOs. First, it is easy to ensurethat the MOs are mutually orthogonal, and this leads toan immense simplification in the calculation of matrix ele-ments. Second, it is also possible to optimize the choice ofthe MOs in such a way as to speed up the convergence ofthe final summation in which configurations are combinedto obtain the state wave function.

Indeed, in the most commonly used form of the MOprocedure this final step is omitted altogether, and a singleconfiguration built from optimized molecular orbitals isused as an acceptable, if poor, approximation of the FCIelectronic wave function ψ . These optimum MOs areknown as the self-consistent field (SCF) or Hartree–Fock(HF) MOs. They are obtained using the variational prin-ciple and demanding that the energy of the one config-uration under consideration be as low as possible. Theenergy difference between the HF description and the FCIdescription is referred to as the correlation energy. It isimportant to note that weak intermolecular interactions(van der Waals interactions), important in processes suchas molecular recognition and complexation, cannot be cal-culated at the SCF level. In calculations of these effects,inclusion of correlation effects is essential.

One of the advantages of the MO method is that it isrelatively easy to improve the SCF solution partially with-out having to go all the way to the unreachable limit ofFCI. Several methods for making such an improvementare available, such as (1) limited configuration interac-tion in which the CI expansion is truncated in some syste-matic fashion well before the FCI limit is reached and (2)methods in which correlation effects are viewed as a smallperturbation of the SCF solution and are treated by per-turbation theory. The use of these methods is particularlyimportant (1) when the electronic wave function ψ is beingcalculated for a geometry far removed from the molecularequilibrium geometry, (2) when the molecule has very lowlying excited electronic states, (3) when the molecule is abiradical, (4) if the calculation is performed for an excitedelectronic state, or (5) if intermolecular forces are to becalculated.

So far we have discussed only the minimum basis setapproximation. This might be quite adequate if the AOswere chosen in a truly optimal manner, but it would requirea nonlinear optimization and this is normally not done.Even so, in order to obtain a truly accurate solution for theelectronic wave function ψ within the Born–Oppenheimerapproximation it would be necessary to increase the num-ber of basis set AOs used in the calculation (in principle,to infinity).

The form of the AOs usually adopted in numericalwork is normally chosen for computational convenience(Gaussian-type orbitals) in a way that makes them quitedifferent from the optimum orbitals of an isolated atom.Thus, in practice most computations are not performedwith a minimum basis set but with extended basis sets,and several such standard sets are in common use.

The electronic wave functions resulting from suchlarge-scale calculations are usually not easy to visualize.Frequently, it helps to draw the resulting electron densitiesin the form of contour maps. Alternatively, electrostatic




potential contour maps can be constructed, indicating thenature of the electric fields to be expected in the vicinityof the molecule.

3. Ab Initio and Semiempirical Methods

The approach described so far is of the so-called ab ini-tio type, in which the whole computation is done fromfirst principles, taking from experiment only the valuesof fundamental constants such as electron charge. As al-ready indicated, accurate results cannot be obtained bythese methods for most molecules of interest in organicchemistry, but the approximate solutions obtained at theSCF or improved SCF level provide much useful informa-tion. For certain properties, such as molecular geometries,dipole moments, and the relative energies of conformers,the agreement with experiment is excellent.

An alternative approach to the problem of molecularelectronic structure is provided by semiempirical mod-els. In these no attempt is made to derive the propertiesof atoms from first principles. Rather, they are taken asdescribed by a set of parameters obtained by fitting ex-perimental data, and an attempt is made to find a modelHamiltonian that will provide a good description of inter-atomic interactions. The form of the model Hamiltonianis patterned after the ab initio analysis. In the most com-mon semiempirical methods it is still a fairly complicatedmany-electron Hamiltonian, so that its exact stationarywave functions cannot be found for molecules of interest,and only approximate solutions are obtained. One almostinvariably starts with a minimum basis set of AOs and pro-ceeds to an SCF type of wave function, possibly followedby a limited amount of improvement toward the FCI wavefunction. The parameters that enter the Hamiltonian areoptimized so as to bring about close agreement betweenthe molecular properties computed from the approximatewave function (usually SCF) and those observed experi-mentally. Most commonly, the properties fitted are heatsof formation, molecular equilibrium geometries, or suit-able spectroscopic properties. In this way one attempts toincorporate intraatomic correlation energies and a largepart of interatomic electron correlation energies into themodel through parameter choice, although one works onlyat the SCF level or at least not much beyond it.

The agreement of the calculated properties with exper-iment is roughly comparable to the agreement obtainedby extended basis set ab initio methods at the level of theSCF approximation or slightly better, at least for thoseclasses of molecules for which the semiempirical pa-rameters were originally optimized. However, even formolecules quite different from those on which the originaloptimization was performed, the agreement is frequentlystriking considering that orders of magnitude less com-

puter time is needed for the semiempirical computations.This great reduction in computational effort is to a largedegree due to the almost universal use of the so-called zerodifferential overlap approximation, which greatly reducesthe number of electron repulsion integrals needed in thecomputations.

Some of the best known examples of semiempiricalmethods are those developed for the treatment of elec-tronic ground states by Dewar and collaborators (MNDO,AM1, and MINDO/3) and the methods developed by Jaffe(INDO/S), Zerner (INDO/S), and their respective collab-orators for calculations involving electronically excitedstates. A very simple procedure is the extended Huckelmethod popularized by Hoffmann. Others are the oldermethods developed for the treatment of π electrons only:the PPP method of Pariser, Parr, and Pople and the ex-tremely crude but also extremely simple HMO method ofHuckel.

These semiempirical models should not be confusedwith approximate models that are designed to mimic theresults of ab initio calculations in a simpler manner ratherthan to mimic the results of experiments. The best knownapproximate MO methods are the CNDO and INDO meth-ods developed by Pople and collaborators.

C. Molecular Properties

Although we have indicated how the electronic wave func-tion of an electronic state and its energy are calculated, wehave said very little about the calculation of other mole-cular properties once the wave function is known.

The calculation of molecular equilibrium geometry in agiven electronic state, usually S0, is performed by varyingthe assumed nuclear geometry and repeating the calcu-lation of the energy by one of the methods referred toearlier until a minimum is found. This search is normallyperformed by computer routines that compute surface gra-dients in order to speed up convergence toward a local min-imum in the (3N − 6)-dimensional nuclear configurationspace. From the computed curvatures of the surface at theminimum one obtains the force constants for molecularvibrations and the form of the normal modes of vibration.For a true local minimum, all the force constants must bepositive. A similar type of procedure, minimization of thenorm of the gradient, can be used for finding transitionstates. After a transition point is found, it is essential toconvince oneself that a normal mode analysis producesonly one vibration with a negative force constant. Thismode corresponds to the path from one catchment basinto the other, and the corresponding vibration has an imag-inary frequency (the restoring force is negative). Othermodes of vibration are ordinary and permit the evaluationof the entropy of the transition state.




Most other molecular properties are normally evalu-ated only at the equilibrium geometry, although strictlyspeaking they should be calculated at a large number ofgeometries and averaged over the vibrational wave func-tion of the state in question. Generally, they are obtainedby representing the observable by a quantum mechani-cal operator and computing the expectation value of thisoperator over the wave function.

At the HF level, ab initio or semiempirical, some ofthese properties can be obtained in an approximate mannermore simply. Thus, the lowest ionization potential of themolecule is approximately equal to the negative of the en-ergy of the highest occupied molecular orbital (HOMO).The electron affinity of the molecule is similarly approx-imated by the negative of the energy of the lowest unoc-cupied molecular orbital (LUMO). The energies of elec-tronic excitation are usually more complicated to obtainin that the introduction of CI may be quite necessary.

Those electron excitations that can be well describedas a promotion of an electron from one single occupiedMO to one unoccupied MO are rare but exist in somemolecules. An example is the so-called La band in theabsorption spectra of aromatic hydrocarbons and the firstintense band of polyenes. In the SCF approximation theelectronic excitation energy is equal to the energy dif-ference between the orbital out of which the promotionoccurs and the orbital into which it occurs, minus the re-pulsion energy of two electron densities, each provided byan electron in one of these two MOs. This approximatesthe energy of the triplet excited state. The energy of thesinglet excited state is higher by twice the exchange in-tegral between the two MOs involved (the self-repulsionenergy of a charge density produced by taking a productof the two orbitals).

Due to the approximate nature of the SCF treatmentand to the additional approximations involved in the state-ments just made, the results are usually more useful foran interpretation of trends within a group of compoundsrather than the absolute value for any one compound. Thesemiempirical methods in particular have had much usein this kind of application.

A property related to this is the formation of intermo-lecular complexes characterized by a charge-transfer tran-sition, which frequently occurs in the visible region andin which an electron is transferred from one molecule toanother. The energy of this transition is related to the ion-ization potential of the donor and the electron affinity ofthe acceptor moiety and once again can be correlated withthe computed MO energies.

Not only the energies but also the coefficients of theMOs computed in the SCF picture are approximately re-lated to observable properties. Thus, for a given MO, thesquare of the coefficient on a particular AO is related to

the probability that an electron in that MO will be foundin that particular AO. This relation is particularly simplein those semiempirical methods that use the zero differ-ential overlap approximation. Then, the AOs are mutuallyorthogonal and the relation between the square of the co-efficient and the probability is a simple proportionality.In methods that use nonorthogonal AOs, such as the abinitio ones, it is more difficult to define electron popu-lations for AOs and atoms. The procedure usually usedis known as the Mulliken population analysis. This per-mits a calculation of electron densities in AOs and of totalelectron densities on atoms. These in turn can be relatedto molecular dipole moments, infrared spectral intensi-ties, and, much more approximately, to nuclear magneticresonance (NMR) shielding constants.

The distributions of unpaired spin obtained in an anal-ogous fashion from the squares of coefficients of a singlyoccupied orbital are related to the hyperfine coupling con-stants in electron spin resonance spectroscopy.

There is another class of molecular properties that arenot related in a simple way to the expectation value ofan operator: the so-called second-order properties. Someof the most important of these are molecular polarizabil-ity, Raman intensities, and chemical shielding in NMRspectroscopy. They can be computed by introducing anoutside perturbation such as an electric or magnetic fieldexplicitly into the calculation of the molecular wave func-tion. These calculations are more difficult and less reliable,particularly with respect to magnetic properties, where theincompleteness of the basis sets used tends to make theresults dependent on the choice of origin of coordinates.Good progress has been made in the calculation of NMRchemical shielding constants and their anisotropies, whilethe calculation of intensities in Raman spectra still leavesmuch to be desired.

IV. REACTION PATHS

A. Thermal Reactions

A reaction involving motion through a single transitionstate is referred to as an elementary reaction step. Mostreactions of organic compounds involve a sequence ofsuch reaction steps in which the reacting system passesthrough a series of intermediates and transition states thatseparate them. The reaction intermediates are usually un-observed. An important part of the investigation of organicreactivity is the determination of these individual steps inan overall sequence, usually accomplished using the toolsof chemical kinetics.

In terms of transition state theory an understandingof the reactivity of organic molecules under conditions




usually employed in the laboratory requires the knowledgeand understanding of activation enthalpies and activationentropies for elementary reaction steps; that is, the shapesof the relevant potential energy surfaces in the vicinity ofthe local minimum in the reactant catchment basin and inthe vicinity of the transition state saddle point in the ridgeseparating it from the product catchment basin.

The reaction path of an elementary step is usually de-fined as the steepest descent path from the transition pointto the minima in the starting and the final catchment basins.It describes the geometry of the approach of the reactingmolecules and, specifically, of the reaction centers towardone another. This approach is frequently described in pic-torial terms, such as “face-to-face” approach of an olefin toa diene in a Diels–Alder reaction or a “backside attack” bya nucleophile on a carbon atom carrying a leaving group.Frequently, the geometrical path is described in even morequantitative terms—for example, by giving an angle of ap-proach of a nucleophile attacking a carbonyl group.

Note, however, that within the constraints of the transi-tion state theory only the geometry at the transition stateand in its immediate vicinity matters, and the precedingpath taken by the molecules to approach this point is irreve-lant. Molecules do not react by following a well-describedstraight path of minimum energy toward the transitionstate. Rather, they undergo random excursions in theirshape, mutual orientation, and distance that have little todo with any particular path and are dictated by collisionswith the surrounding medium. This chaotic motion contin-ues until the molecules happen to acquire enough energyand the appropriate direction of motion to reach the transi-tion state, and then passage to the product catchment basin(i.e., a chemical reaction) follows.

Although the concept of a reaction path therefore cannotbe taken as literally representing the actual movements of areacting molecule, it is useful in that it helps to understandthe geometries at which transition states occur. In mostreactions other than mere conformational changes, bondsare broken and made (either directly or by means of elec-tron transfer). If old bonds had to be broken completelybefore new ones were made, the enthalpies of activationwould be given by bond strengths. In reality, they are fre-quently lower. The search for favorable energy paths andlow-energy transition states is thus equivalent to a searchfor ways in which the transition state can be stabilized byintroducing new bonding interactions before the old oneshave been completely lost.

Two important ways of introducing favorable interac-tions between the reacting centers can be visualized fromthe knowledge of the electronic wave functions of the re-actants. One of these is to bring unlike charges on the twopartners close together and keep like charges well sepa-rated. The other is to introduce favorable interactions of

occupied MOs of one reacting center with the unoccupiedones of the other and vice versa. Often a reaction can beviewed as an interaction of a Lewis acid with a Lewisbase. Then the former kind of favorable interactions pre-vails in the reaction of a so-called hard acid with a hardbase, while the latter kind prevails in the interactions of aso-called soft acid with a soft base. Interactions of a hardpartner with a soft one are generally less favorable.

In the frontier MO theory of organic reactivity, intro-duced by Fukui, attention is limited to the HOMO andthe LUMO of the reacting partners. Usually, their energyseparations are such that only one of the HOMO–LUMOinteractions has to be considered, namely that between thehigher lying of the two HOMOs (that of the better donormolecule) and the lower lying of the two LUMOs (that ofthe better acceptor molecule). A favored orientation forthe reaction will be one in which the interaction of thesetwo MOs is maximized.

Very helpful tools for identifying favorable reactionpaths and recognizing superficially similar unfavorableones are correlation diagrams. In their most useful formthese resemble crosscuts through potential energy hyper-surfaces and indicate how electronic states of starting ma-terial correlate with those of a product. Usually, the easiestway to construct such state correlation diagrams is to pro-ceed in several steps starting with an MO correlation dia-gram, proceeding to a configuration correlation diagram,and finally introducing configuration mixing to obtain thedesired state correlation diagram.

The use of molecular symmetry is helpful in the con-struction of these diagrams because this frequently permitsready identification of the nodal patterns in the wave func-tions involved, which determine the behavior of the energyalong the reaction path. Since nodal patterns are not par-ticularly sensitive to minor perturbations, it is often possi-ble to ignore those secondary perturbations that lower thesymmetry of the reacting system and still obtain useful re-sults. For example, the face-to-face interaction of ethylenewith propylene can be understood on the basis of consid-ering the interaction of two ethylene molecules instead.

The use of correlation diagrams has been particularlyhelpful in the case of the so-called pericyclic reactions.These are elementary reaction steps the transition stateof which is characterized by a cyclic array of mutuallyoverlapping and interacting AOs on the reacting centers.It is not necessary and usually not possible for all theinteractions along the periphery of the cycle to be equalin magnitude.

Such transition states are isoelectronic with cyclic con-jugated π systems and can therefore be classified as aro-matic or antiaromatic. Not surprisingly, other factors beingequal, the aromatic transition states are favored energeti-cally over the antiaromatic ones. The differences are the




largest for small cycles and decrease as the number ofAOs in the cycle increases. It has become customary torefer to the pericyclic reaction paths proceeding througharomatic transition states as “allowed” and to those pro-ceeding through antiaromatic transition states as “forbid-den.” Pericyclic reactions of both types can occur, but theformer are normally strongly favored if all other factorsare the same. Many different theoretical approaches, ofwhich we mention only two here, have been used to derivesimple rules that enable one to predict whether a particu-lar pericyclic process will be of the allowed or forbiddenkind. The rules have become known as the Woodward–Hoffmann rules, although many other workers have alsomade fundamental contributions to their formulation.

Although it is possible to distinguish allowed fromforbidden pericyclic reaction paths readily using thearomaticity criterion for the transition state, developedoriginally by Dewar and Zimmerman and just outlined,correlation diagrams are also often used for this purpose.They are particularly advantageous in the considerationof photochemical processes (see Section IV.B). In orderto construct an MO correlation diagram, the MO energiesof the reactant are plotted vertically on one side and thoseof the products on the other side. They are identified assymmetric (S) or antisymmetric (A) with respect to thosesymmetry elements that are preserved through the wholeassumed reaction path and that cut through the bonds be-ing formed or broken. The energies of those MOs that haveequal symmetries on both sides are then connected, takingaccount of the noncrossing rule. This rule states that linescorresponding to wave functions of like symmetries mustnot cross. The result is shown in Fig. 9 for the face-to-face cycloadditions of two ethylenes and of ethylene withbutadiene.

In the latter case (Fig. 9B), all occupied and bonding or-bitals of the reactant electronic ground state remain bond-ing throughout the reaction path and in the product as well.Similarly, all antibonding unoccupied MOs of the startingelectronic ground state material remain antibonding andunoccupied. Clearly, bonding is preserved throughout thereaction path, and one would expect the transition stateto be of relatively favorable energy. Indeed, the transitionstate is of the aromatic type, containing six electrons inthe cyclic area of orbitals, with positive overlaps of all theorbitals (isoelectronic with the π system of benzene).

On the other hand, in the former case of two ethylenesone of the originally occupied bonding MOs of the reac-tant becomes antibonding and unoccupied in the groundelectronic state of the product, and one of the antibondingand unoccupied orbitals of the reactant becomes bondingand occupied in the ground state of the product. In theregion of transition state geometries halfway through thereaction path, both orbitals are approximately nonbond-

FIGURE 9 Molecular orbital correlation diagrams for the con-certed face-to-face cycloaddition of two ethylene molecules (A)and for the concerted Diels–Alder cycloaddition of ethylene withbutadiene (B). Symmetry planes preserved throughout the re-action path are indicated by dashed lines at the bottom. Theirnumbering on the left is keyed to the subscript on the symmetrysymbols S (symmetric) and A (antisymmetric) on the molecularorbitals.

ing. Between them, they contain two electrons, and thesetwo electrons do not contribute to bonding in the moleculeat the transition state geometry. In effect, the molecule isa biradical and contains one less bond than its number ofvalence electrons would in principle allow it to have. Thetransition state is unfavorable and is of the antiaromatictype, containing four electrons in an array of four AOs withall positive overlaps (isoelectronic with cyclobutadiene).

Although it is already apparent which of the two reac-tions chosen as examples is allowed and which is for-bidden, it is useful to consider the construction of theconfiguration correlation diagram as well (Fig. 10). Here

FIGURE 10 Configuration (thin lines) and state (thick lines) corre-lation diagrams for the concerted face-to-face cycloaddition of twoethylene molecules (A) and for the concerted Diels–Alder cycload-dition of ethylene to butadiene (B). Full lines, singlets; dashedlines, triplets. See legend to Fig. 9.




low-energy configurations are constructed by consideringall suitable occupancies of the MOs of the reactants onthe left-hand side and of the products on the right-handside. The symmetry of each is again identified as antisym-metric or symmetric with respect to each of the abovesymmetry elements, using the rule S × S = A × A = Sand A × S = A. Correlation lines are now drawn fromleft to right by keeping the occupancy of each MO ineach configuration constant. This often produces cross-ings of configurations of the same symmetry. Accord-ing to the noncrossing rule these must ultimately beavoided.

This is accomplished by introducing configuration in-teraction, which converts the configuration correlation di-agram to the desired state correlation diagram, as indicatedin Fig. 10. Clearly, the crossing of correlating MOs in thecase of the ethylene + ethylene cycloaddition causes asimilar crossing of lines in the configuration correlationdiagram. Since the effects of configuration mixing, whichproduces the final state diagram, are generally relativelysmall, a memory of the crossing at the geometry of thetransition state survives and results in a large barrier in theenergy of the ground state in the middle of the correlationdiagram. It is then concluded that the transition state isunfavorable relative to the case of the ethylene + butadi-ene process, in which no such barrier is imposed by thecorrelation.

B. Photochemical Reactions

In photochemical reactions, initial electronic excitation isintroduced by the absorption of a photon or by an energytransfer from another molecule. It is normally followedby a very rapid, radiationless conversion to the lowest ex-cited singlet or the lowest triplet energy surface, dependingon the multiplicity of the initial excited state. Also, anyvibrational energy in excess of that dictated by the tem-perature of the surrounding medium, whether generatedby the initial excitation or by the radiationless process,is rapidly lost to the solvent, unless one works in a gasphase at low pressure. Thus, in a matter of a few picosec-onds or less the molecule ends up in one or another ofthe local minima in the S1 or T1 surface. Further motionon the surface may follow, depending on the tempera-ture and the height of the barriers surrounding the localminimum. Also, radiationless conversion from the S1 tothe T1 state, known as intersystem crossing, can occur.This often happens on a nanosecond time scale. Sooner orlater a radiationless return to the S0 state ensues. The finalfate of the molecule is further loss of excess vibrationalenergy and thermal equilibration at the bottom of one oranother catchment basin in the S0 surface, depending onwhere on the S0 surface the molecule landed. If this is the

same minimum from which the initial excitation occurred,the process is viewed as photophysical. If it is not, a netchemical reaction has occurred and the process is labeledphotochemical.

At times the excited S1 surface may touch or nearlytouch the S0 surface, in which case the return to S0 is veryfast. Such areas in S1 are often referred to as funnels sincethey very effectively return molecules to the ground state.

In order to understand photochemical reaction paths itis thus important to have an understanding of the locationof barriers as well as minima and funnels in the S1 andT1 surfaces, plus a sufficient understanding of the S0 sur-face to allow a prediction or rationalization of the fate ofa molecule that lands in a known region of this surface.Correlation diagrams are often useful for this purpose. Forinstance, the diagram for the face-to-face cycloaddition oftwo ethylenes shown in Fig. 10 shows the presence of aminimum in the S1 surface in the general area of geome-tries at which the pericyclic transition state occurred in theground state. While the latter was energetically unfavor-able in the S0 state, making the reaction highly unlikelysince the molecules will probably find other reaction paths,the minimum in the S1 state provides an efficient drivingforce for the photochemical cycloaddition to proceed ef-ficiently. Thus, reactions that fail to occur in the groundstate are often smooth when performed photochemicallyand vice versa.

In general, by virtue of molecules landing at otherwiseimprobable and highly energetic areas on the S0 surface,photochemical processes are capable of producing veryhighly energetic ground state products. Yet, frequentlythe same perturbations, such as substituent effects, thatincrease the stability of a molecule in the ground statealso facilitate its photochemical reactions by loweringthe barriers encountered along the way. The interplay ofthese two aspects of the excited-state surfaces—minimaand barriers—make the consideration of photochemicalprocesses far more complex than the study of thermalreactions.


HYDROGEN BOND • INFRARED SPECTROSCOPY • KINET-ICS (CHEMISTRY) • MOLECULAR ELECTRONICS • OR-GANIC CHEMISTRY, SYNTHESIS • PERIODIC TABLE (CHEM-ISTRY) • PHOTOCHEMISTRY, MOLECULAR • POTENTIAL

ENERGY SURFACES • QUANTUM MECHANICS • RAMAN

SPECTROSCOPY

BIBLIOGRAPHY

Albright, T. A., Burdett, J. K., and Whangbo, M.-H. (1985). “OrbitalInteractions in Chemistry,” Wiley, New York.




Burkert, U., and Allinger, N. L. (1982). “Molecular Mechanics,” Ameri-can Chemical Society Monograph, Washington, DC.

Carsky, P., and Urban, M. (1980). “Ab Initio Calculations,” Springer–Verlag, Berlin.

Chemical Reviews, Special Issue on Reactive Intermediator 91(3) May1991.

Dewar, M. J. S., and Dougherty, R. C. (1975). “The PMO Theory ofOrganic Chemistry,” Plenum, New York.

Dewar, M. J. S. (1984). J. Am. Chem. Soc. 106, 669.Hehre, W. J., Radom, L., Schleyer, P. V. R., and Pople, J. A. (1986). “Ab

Initio Molecular Orbital Theory,” Wiley, New York.

Lowry, T. H., and Richardson, K. S. (1981). “Mechanism and Theory inOrganic Chemistry,” 2nd ed., Harper & Row, New York.

Michl, J., and Bonacic-Koutecky, V. (1990). “Electronic Aspects ofOrganic Photochemistry,” Wiley, New York.

Salem, L. (1982). “Electrons in Chemical Reactions,” Wiley, New York.Schaefer, H. F., III, and Segal, G. A. (eds.) (1977). “Modern Theoretical

Chemistry,” Vols. 3, 4, 7, and 8, Plenum, New York.Simons, J. (1983). “Energetic Principles of Chemical Reactions,” Jones

and Bartlett, Boston, MA.Simons, J. (1991). An experimental chemist’s guide to ab initio quantum

chemistry. J. Phys. Chem. 95, 1017.

P1: GPJ/GLT P2: GQT Final Pages Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology en011-542 July 26, 2001 15:33

Organic Chemistry, SynthesisJohn WelchState University of New York, Albany

I. Functional Group ManipulationII. Carbon Carbon Bond Forming ReactionsIII. Natural Product SynthesisIV. Asymmetric SynthesisV. Biomimetic Synthesis

GLOSSARY

Asymmetric synthesis Stereoselective preparation of asingle enantiomer.

Biomimetic synthesis Construction of molecules bymimicking biosynthetic processes.

Diastereoselectivity Tendency of a reagent or reactant toform a single diastereomer in a reaction.

Electrophilicity Tendency of a reactant that is electrondeficient to satisfy this deficiency in a reaction.

Enantioselectivity Tendency of a reagent to form a singleenantiomer in a reaction.

Functional group Structural feature of a larger moleculewhich imparts to that molecule a particular chemicalreactivity. Functional groups often contain heteroatomsbut may also result from changes in hybridization orbonding.

Natural products Molecules formed by biological or-ganisms. Natural products are frequently chiral and of-ten complex.

Nucleophilicity Tendency of a reactant that has an elec-tron pair available for bonding to contribute this elec-tron pair in a reaction.

Regioselectivity Ability of a reactant to discriminate andprincipally form a single regioisomer.

Retrosynthetic analysis Systematic bond disconnec-tion of a synthetic target for purposes of planning asynthesis.

ORGANIC SYNTHESIS, the science of the preparationof organic molecules, is a crucial element of organic chem-istry. It is synthesis that makes possible the preparationof materials with novel structures and facilitates the studyof naturally occurring substances otherwise available onlyafter tedious and exacting isolation. The synthetic chemistis often called upon to test the latest advances in theory bypreparing new molecules; yet the observations of the syn-thetic chemist frequently contribute to the development ofnew theories.

Organic chemistry, as the chemistry of carbon com-pounds, requires the functionalization of these compoundsto prepare new substances. Once functionalized, construc-tion of new carbon molecular frameworks is possible. His-torically, the preparation of natural products has been themost important application of synthetic methods. Con-temporary chemists not only seek to prepare the chiralnatural compounds efficiently by asymmetric synthesisbut, by utilizing biomimetic strategies, to copy the biosyn-thetic pathways.

497

P1: GPJ/GLT P2: GQT Final Pages


498 Organic Chemistry, Synthesis

Limitations of space restrict this article to highlightsfrom this diverse field. Throughout the discussion, namereactions and procedures are cited in parentheses.

I. FUNCTIONAL GROUP MANIPULATION

As discussed in the introductory section, a key elementof directed organic synthesis is the selective introductionand modification of functional groups. Manipulation of afunctional unit may provide discrimination in chemical re-activity. Functional groups typically contain heteroatomssuch as halides, oxygen as in alcohols or carbonyls, ornitrogen as in amines. Variation of the hybridization andbonding of carbon as in alkenes, alkynes, and arenes alsoconstitutes functionality. Recently, carbon bonds to mostof the elements of the periodic chart have been exploited byorganic chemists; however, the chemistry of most of thesemore esoteric molecules is illustrated by those groups dis-cussed earlier.

A. Halogenation

Halogenation is one of the most effective means to intro-duce functionality in saturated or unsaturated hydrocar-bons. The relatively low reactivity of alkanes and alkenesrequires reactive reagents for substitution. A large numberof well-established methods for the interconversion of or-ganic halides to other functionalities increases the utilityof halogenated hydrocarbons as intermediates in the intro-duction of other functional groups. Conversely selectivepreparation of halides from alcohols carbonyls, or aminesis also an important process.

1. Reactions of Alkanes and Alkenes

Reaction of an alkane with a halogen usually proceedsvia a free radical pathway. The reactions of chlorine andbromine are more useful than those of fluorine or iodine.Direct fluorination of hydrocarbons with molecular flu-orine is a very exothermic process resulting in complexproduct mixtures and extensive decomposition. Recently,it has been possible to improve the selectivity of the directreaction of fluorine by dilution with an inert gas and byusing very efficient cooling in a special apparatus.

Reactions with bromine or chlorine must be initiatedeither photochemically or by the use of a free radicalinitiator:

X2hν→ 2X·

RH + X· → R· + HXR· + X2 → RX + X·R· + X· → RX

(1)

As with fluorination, a mixture of products is formed. Pho-tochemically induced iodination suffers from side reac-tions resulting from the hydrogen iodide formed.

Although arenes will undergo radical addition reactionswith loss of aromaticity, the Lewis acid promoted elec-trophilic halogenation of arenes is a substitution reactionwhere aromaticity is not lost.

� Br2

H

H

Br

Br

� Br2FeBr3

Br

� HBr

(2)

As an electrophilic substitution reaction, halogenation isgoverned by directing effects and can be regioselective.

In contrast to alkanes, alkenes readily add bromine andchlorine in electrophilic addition reactions:

C C � Cl2 C C C C

Cl�

C C

Cl

C C

Cl

Cl�

Cl

Cl� �

(3)

Hydrogen halides also add to alkenes followingMarkovnikov’s rule, where the halide attaches to themore highly substituted carbon:

C C � HX C C

CH3R CH3

HCH3

H

HXCH3

R

(4)

Hydrogen halide addition is occasionally accompanied byside reactions resulting from cationic rearrangements. Ad-ditions of hydrogen halides to alkynes are generally slow,forming, after addition of 2 moles of hydrogen halide, thegeminal dihalides.

Alkenes may be treated with reagents such as N -bromosuccinimide in the presence of an initiator such asperoxide or light:

CH3 CH � Nhv

CH2Br CH CH2CH2

O

O

Br

NBS

� � � �

(5)

Under these conditions, where products of free radicalattack are favored, the reaction can be very specific forallylic halogenation.



Organic Chemistry, Synthesis 499

2. Reactions of Alcohols

Alcohols are easily converted to halides by displacementreactions. It is necessary to improve the leaving groupability of the hydroxyl by protonation or by conversion toa sulfonate or phosphate ester for successful displacement:

CH3CH2CH2CH2 � CH3OH

CH3

SO2Cl

CH3CH2CH2CH2OS

CH3CH2CH2CH2 CH3CH2CH2CH2CH3OS

O

O

NaBrBr �

CH3

SO3Na

O

O

(6)

In reactions of primary and secondary alcohols with hy-drogen fluoride, hydrogen chloride, or hydrogen bromide,competing rearrangements of the cationic intermediatemay lessen the selectivity of the reaction:

CH CH

CH3

CH3

OH

CH3

� HBr CBr CH2CH3

CH3

CH3

(7)

Treatment of an alcohol with thionyl chloride, sulfurtetrafluoride, phosphorous tribromide, or phosphorouspentachloride can yield the unrearranged halide regios-electively.

3. Haloalkylation Reactions

Halogenation can also be effected simultaneously withcarbon carbon bond forming reactions employing halo-genated reactants. The chloromethylation reaction of aro-matics with chloromethyl methyl ether and a Lewis acid isprobably the best known example of such a reaction. It ispossible to avoid the use of the toxic chloromethyl methylether by in situ formation of the reagent with formalde-hyde and hydrogen chloride:

� HCHOHCl/ZnCl2

CH2Cl

(8)

Alpha haloketones may be prepared by the reaction ofcarboxylic acid halides with diazomethane. Direct treat-ment of a diazoketone with a hydrogen halide, in a relatedreaction, also yields alpha haloketones or, under modi-fied conditions, dihaloketones substituted with differenthalogens:

�CH2FHF

N2

O O

CHN2

(9)

Dihalomethylation may also be effected by reaction ofdihalocarbanions with carbonyl compounds or by the alky-lation of a carbanion with chloroform:

�

O

RR LiCHCl2

R OH

R CHCl2(10)

Halomethylenation is possible by reaction of an appro-priately halogenated Wittig reagent with an aldehyde orketone.

B. Hydroxylation

The selective direct hydroxylation of alkanes is not apractical general laboratory transformation although en-zymatic hydroxylation of unactivated carbons is a majormetabolic reaction in animals. Preparatively hydroxyla-tion is more easily accomplished by hydrolysis of a halide.The ease of hydrolysis, iodide > bromide > chloride, par-allels the synthetic utility of the reaction:

CH3CH2CH2CH2IH2O

CH3CH2CH2CH2OH (11)

Hydrolysis of fluorides is of little consequene. In order tominimize side reactions, such as elimination to an alkeneor rearrangement, SN2 reaction conditions are desirable.However, under some conditions elimination reactions canbe useful or desired. The synthesis of phenol from treat-ment of chlorobenzene with hydroxide ion first proceedsby elimination to form benzyne followed in a second stepby the addition of water to form phenol.

1. Addition of Water to Alkenes

The direct hydroxylation of alkenes with aqueous acidresults in the Markovnikov addition of water. Under theseconditions, rearrangements of the intermediate cation arepossible with a corresponding loss of regiospecificity.

C CH2

CH3

CH3

C CH3

CH3

CH3

�(CH3)3COH

H� H2O/

(12)

Regiospecific addition of water in an anti-Markovnikovsense is readily possible via hydroboration of an olefin.The addition of borane or alkylborohydrides can be highly




regioselective. Oxidation of the boron carbon bond withhydrogen peroxide generates the desired alcohol:

H2O2/OH�

3CH3CH CH2 � BH3 (CH3CH2C2)3B

3CH3CH2CH2OH (13)

When addition in a Markovnikov sense is required,oxymercuration can be very effective and is not accom-panied by rearrangements. Addition of an electrophilicmercury species to an olefin, followed by trapping ofthe intermediate cation, leads to the formation of anoxygenated organomercurial. The carbon mercury bondmay be cleaved reductively with sodium borohydride:

CH2

�

Hg(OAc)2 CH3CH2CH CH2HgOAcCH3CH2CH

OAc

NaBH4 CH3CH3CH2CH Hg�

OH�

(14)

2. Reduction of Carbonyl Compounds

Aldehydes, ketones, carboxylic acids, and carboxylic acidderivatives can be reduced to alcohols. Historically, cat-alytic reduction with hydrogen and a catalyst such as plat-inum oxide or Raney nickel was employed to form the hy-droxylic product in high yields. More recently, the moreconvenient metal hydride reducing agents such as lithiumaluminum hydride or sodium borohydride and their deriva-tives have been employed to form alcohols at ambient orsubambient temperatures:

�NH4Cl/H2O

R R

O

LiAlH4 R H

OH

R (15)

A variety of additional methods exist for the reductionof carbonyl compounds to alcohols. A typical exampleis the Meerwein–Pondorf–Verley reduction of carbonylswith aluminum alkoxides:

� Al(OCH(CH3)2)3H�

OH

CH3

O

(16)

3. Electrophilic Oxygenation

Contemporary work has focused on the development ofmethods for the electrophilic oxygenation of alkanes,alkenes, and arenes:

H�O3 HO3

�

OH

� H� � O2

(17)

Treatment of ozone or hydrogen peroxide with very strongacids, such as hydrogen fluoride–antimony pentafluorideor fluorosulfuric acid–antimony pentafluoride, is proposedto lead to the formation of protonated ozone or hydrogenperoxide, which are highly electrophilic reagents.

4. Hydroxymethylation

By simple analogy with haloalkylation chemistry, alcoholscan also be introduced by carbon carbon bond formingreactions. The addition of formaldehyde to a nucleophiliccarbon such as the alpha carbon of an enol or organometal-lic reagent leads to hydroxymethylation:

� HCHO

CH2OHMgBrH /H2O�

(18)

In a different approach, treatment of an alkyl boron com-pound with carbon monoxide leads to carbon monoxideinsertion in the alkyl carbon boron which can be reducedto form the hydroxymethyl group:

C C

R

H H

HBH3 (RCH2 CH2)3B

CO(RCH2CH2)2BCCH2CH2R

O

LiBH4 H2ORCH2CH2CH2OH

3

(19)

The Williamson ether synthesis, in which treatment ofan alkyl halide with the alkali metal salt of an alcoholyields an ether, appends an alkoxy chain to an alkyl halide.

CH3I �CH3CH2CH2CH2ONa

CH3CH2CH2CH2OCH3 � NaBr (20)

Alkoxymethylation is possible by reaction of substitutedethers with an appropriate nucleophile.

C. Amination

Selective formation of amines is extremely important forthe synthesis of many natural products and biologically ac-tive compounds. The introduction of amines in the pres-ence of other functionality is a particularly challengingsynthetic problem. The selectivity required for practicalutility is not generally associated with the more rigorous




conditions required for direct amination; therefore, func-tional group transformations are employed.

1. Displacement of Halides

The reaction of an alkyl halide with ammonia often leads toa mixture of primary, secondary, and tertiary amines wheremore than one alkyl halide molecule has reacted with asingle ammonia. Selective formation of the primary aminemay be possible if the alkyl halide is soluble in the presenceof a large excess of ammonia. On a laboratory scale, theGabriel synthesis or one of its modifications may be amore practical approach. The alkyl halide is treated withphthalimide or an alkali metal salt of phthalimide. Theproduct alkylated phthalimide is hydrolyzed to selectivelyform the primary amine:

Na � CH3CH2CH2BrN

O

O

CH2CH2CH3 � NaBrN

O

O

NaOH

O

O

ONaONa

� H2NCH2CH2CH3 (21)

Selective formation of secondary or tertiary amines byreaction of an alkyl or dialkyl amine is generally ineffec-tive for the same reason that the reaction of ammonia isnot selective. However, direct reaction of an amine withan excess of an alkyl halide can be a useful method to thepreparation of quaternary ammonium salts.

2. Derivatization of Carbonyl Compounds

An especially useful way to prepare secondary and ter-tiary amines is via an intermediate imine or iminium ion.For example, reaction of formaldehyde with a secondaryamine in the presence of formic acid leads directly to anew tertiary amine:

NH �HCHO

R

R�

N CH2

R

R�

�� HCO2H

N CH3

R

R�

� CO2 � H� (22)

Dimethyl amines are conveniently prepared by reactionof formaldehyde with a primary amine in the presence offormic acid:

CH3CH2CH2NH2 �HCHO � HCO2H

CH3CH2CH2N(CH3)2 (23)

Acids may be converted to amides through reaction ofthe acid chloride with an amine:

RCOCl � R�CH2NH2 R

O

NHCH2R�

RR

O

NHCH2R�LiAlH4 CH2NHCH2R�

(24)

Reduction of the acid amide to a new amine is possibleby catalytic hydrogenation using a metal catalyst (suchas platinum oxide, palladium oxide, palladium on bariumsulfate, or Raney nickel). Metal hydrides such as lithiumaluminum hydride, aluminum trihydride(alane), or dibo-rane will also conveniently reduce amides to amines.

The direct conversion of acids to amines via an acylnitrene intermediate is also possible by the Hofmann,Schmidt, or Curtius rearrangement:

RNR

O

N3 C O � N2H2O

RNH2 �CO2

(25)

3. Allylic Amination

Allylic amination of alkenes via sulfur and seleniumreagents is also possible. In the latter case, the reagentis prepared by treatment of selenium with chloramineT, in the former, by reaction of thionyl chloride withp-toluenesufonamide:

2CH3 SO2NH2 � SOCl2 SO2NCH3 S NSO2 CH2

RCH2 CH CH2R CH CH CH2

NH

SO2 CH3

(26)

4. Mannich Reaction

Condensation of formaldehyde with an amine forms animinum ion which can act as an electrophile in further syn-thetic transformations. One result is the addition of a car-bon bearing an amine to the nucleophile. Commonly thereaction is effected with enolizable carbonyl compounds;however, the reaction will work with any compound withan activated hydrogen:

HCHO � NH

CH3

CH3

CH2CH2 CH3

OH

N

CH3

CH3

CH2 CH2 CH3

O

N

CH3

CH3

�

(27)




The intermediate iminium ion is even electrophilic enoughto react with arenes.

D. Carbonylation and Carboxylation

One of the most useful functional groups in organic syn-thesis is the carbonyl group. The carbonyl function re-acts at the carbonyl carbon as an electrophile or mayenolize and react at the alpha carbon as a nucleophile.Commensurate with the importance of carbonyl com-pounds to synthetic chemistry, a large number of meth-ods have been developed for the preparation of thesecompounds.

1. Oxidation of Alcohols

A tremendous variety of oxidants are known for the con-version of alcohols to carbonyl containing compounds.The most difficult of these transformations is the oxidationof a primary alcohol to an aldehyde, the aldehyde oftenbeing more susceptible to oxidation, than the primary alco-hol. Typical of the reagents employed to effect these oxida-tions are chromium and manganese oxides, dimethyl sul-foxide and dicyclohexylcarbodiimide (Pfitzner–Moffat),and dimethyl sulfoxide and oxalyl chloride (Swern):

RCH2OH[Ox]

RCHO (28)

Oxidation of secondary alcohols to ketones is simpler ifonly because ketone products are more stable to furtheroxidation. Typically chromium reagents (CrO3) or per-manganate salts (KMnO4) are used:

[Ox]R

CHR

OHR

CR

O��

(29)

The conversion of primary alcohols or aldehydes to car-boxylic acids is normally a facile reaction using KMnO4.Aldehydes can be easily oxidized by air to acids, a resultof the lability of the aldehydic hydrogen.

2. Carbonylation

The direct addition of a carbonyl group to olefins is pro-moted by metal catalysts such as rhodium, iridium, andcobalt. The oxo reaction, the commercial route to thepreparation of aldehydes which are reduced to alcoholsimportant as plasticizers, relies on such a carbonylation.

HCo[CO]4CH2 CH2 � CO � H2 CH3CH2CHO

(30)

The formylation of arenes by carbon monoxideand Lewis acids (Gatterman–Koch) is an importantwell-established method for the synthesis of arenealdehydes:

HCl/AlCl3� CO

CH3 CH3

CHO

(31)

The use of carbon monoxide can be avoided by reaction ofdimethylformamide in the presence of phosphorous oxy-chloride (Vilsmeier–Haack). The electrophilic intermedi-ate chloroiminium ion adds to activated arenes. The sub-stitution product is readily hydrolyzed to an aldehyde:

POCl3� HCN(CH3)2

N

CH3 CH3

OH2O

N

CH3 CH3

CHO

(32)

3. Nucleophilic Reagents

Carbonyl compounds may also be prepared by the reac-tion of a nucleophilic reagent such as an organolithiumor Grignard reagent with formamide (Bouveault) of withan orthoester. In these examples, the initial product of re-action is a hemiaminal or acetal, respectively, which isunreactive to additional equivalents of the nucleophilicreagent but is readily hydrolyzed under acid conditions toreveal an aldehyde:

� RMgBrCH(OR�)3 � ROMgBrR CH(OR�)2

� H�� H2ORCH(OR)2 RCHO (33)

The reactions are not limited to Grignard or organolithiumreagents; any activated hydrogen compound will give sim-ilar results. In related reactions an organometallic com-pound will react directly with carbon dioxide to form onhydrolysis a carboxylic acid.

E. Introduction of Sulfur

The importance of sulfur in natural products and in prod-ucts of commerce should be mentioned. Sulfur-containingcompounds also have significant utility in directing furthersynthetic transformations.




Sulfonation of arenes is one of the oldest and best-known reactions for the introduction of sulfur into organicmolecules. Treatment with sulfuric acid and oleum orchlorosulfonic acid leads to the preparation of sulfonatedaromatics that are useful as surfactants, dye constituents,or chemical intermediates:

H2SO4

SO3

SO3H

(34)

Sulfonation under acidic conditions is of little utility inthe preparation of aliphatic sulfonates. Alkyl sulfonatesare selectively prepared by reaction of alkanes with sulfurdioxide and chlorine. When the reaction is conducted inthe presence of ultraviolet light, side reactions such aschlorination are effectively suppressed:

h�RCH3 � SO2 � Cl2 RCH2SO2Cl (35)

The initial product of these reactions, as well as the prod-uct of the reaction of arenes with chlorosulfuric acid, aresulfonyl chlorides which may be hydrolyzed to the desiredacids.

1. Sulfide Formation

Sulfides may be prepared by the reaction of alkali metalsalts such as sodium sulfide with alkyl halides. Dialkyl sul-fides are prepared by the reaction of the alkali metal salt ofan alkyl sulfide with a second molecule of an alkyl halide:

RSNa � R�CH2X R�CH2SR � NaX (36)

Alternatively alkyl halides may be reacted with thioureato form sulfides upon hydrolysis. Sulfides are alsoformed by the Markovnikov addition of hydrogensulfide to alkenes. Alkyl sulfides will add in a conjugatemanner to α,β-unsaturated carbonyl compounds, givingβ-ketodialkyl sulfides.

II. CARBON CARBON BONDFORMING REACTIONS

The preparation of the functional groups most importantin organic synthesis has been described, but in the de-sign of a synthesis, the construction of the carbon skele-ton is often the greatest challenge. Frequently the desiredfunctionality is either protected to prevent undesired sidereactions or is carried through a synthetic sequence in amasked form, to be liberated after other transformationshave been accomplished. These protected or masked func-tional groups have been described as synthons, structural

units which can be formed or assembled by known or con-ceivable synthetic transforms. The retrosynthetic analysisof the carbon skeleton determines which carbon carbonbonds should be formed and in what order they should beassembled. After analyzing the molecule and determiningconvenient fragments, the emphasis now focuses on de-termining which carbon carbon bond forming reactionscan be employed.

Carbon carbon bond forming reactions can be de-scribed as being ionic processes, in which one fragmentis electron deficient (electrophilic) and the other fragmentis electron rich (nucleophilic), or as radical processes inwhich each fragment contributes a single electron. Tradi-tionally, ionic processes have been the better understoodand more effectively manipulated. However on occasion,reactions with radical intermediates have been describedand effectively employed as well. Ionic processes will bedescribed as resulting from nucleophilic or electrophiliccomponents. Such a distinction is somewhat artificial, butparallels traditional descriptions for ionic carbon carbonbond forming reactions in which one component mustbe electrophilic and the other nucleophilic. Typicalelectrophilic reactions such as alkylation and acylationand typical nucleophilic reagents such as organometalliccompounds, enamines, and deprotonated imines will bediscussed.

A. Alkylation

As mentioned earlier alkylations are described as elec-trophilic from the perspective of the alkylating agent. Theelectrophilic alkylation of alkanes or alkenes is known butis only of limited utility because of the difficulty of achiev-ing selectivity. Nonetheless, the electrophilic alkylation ofarenes is a reaction of substantial significance.

1. Alkyl Halides

Reports of the reaction of alkyl halides with arenes inthe presence of Lewis acids appeared in the literatureas early as 1877. Although alkyl fluorides are the mostreactive of the alkyl halides, alkyl chlorides and bro-mides are the most widely used. Alkyl iodides are lesscommonly used as a result of accompanying side reac-tions and decomposition. Tertiary and benzylic halidesare the most reactive, with secondary halides less reactivebut more reactive than primary halides. Dialkyl haloniumions prepared by treatment of excess alkyl halide withantimony pentafluoride are especially reactive alkylatingreagents. However, the most reactive alkylating agentsare the methyl fluoride–antimony pentafluoride and ethylfluoride–antimony pentafluoride complexes:




RCH2Cl �AlCl3

CH2R

(37)

2CH3I � SbF5 CH3ICH3�

CH3ICH3 ��

CH3

� CH3I

(38)

CH3CH2FSbF5 �

CH2CH3

(39)

These powerful electrophiles will even react with poornucleophiles such as alkanes.

2. Alkenes, Alcohols, Esters, and Ethers

In addition to alkyl halides, alkenes will act as elec-trophiles in the presence of Lewis or Brφnsted acids. Alco-hols, esters, and ethers also will act as carbon electrophilesunder acidic conditions.

B. Acylation

Acylation is another important electrophilic reaction forthe substitution of arenes. The electrophilic reagent, anacyl cation, is stabilized by resonance. Frequently usedas acylating reagents in the presence of a Lewis acid areacylchlorides, anhydrides, acids, and esters. Generally thereaction is performed by the addition of the Lewis acidcatalyst to a mixture of the reactant arene and acyl halide:

CH3

� CH3COCl

CH3

CH3

O

�

CH3

O CH3

AlCl3

(40)

�CH3CF SbF5

O

CH3C

O

� SbF6�

CH3C O�

CH3C�

O (41)

Similar to the reactions of dialkyl halonium ions ormethyl fluoride antimony pentafluoride complexes, sta-ble acylium ions, prepared by reaction of an acid fluoride

with antimony pentafluoride, are extremely reactive acy-lating agents.

C. Alkylation of Organometallic Reagents

Organometallic reagents often react as carbanions andas such are nucleophilic. The electrophilic component ofthese reactions is generally aliphatic, but reactions of aro-matic compounds are known. Although the alkylation ofanionic carbon is well known, the mechanism of the reac-tion is still not clearly understood.

1. Alkyl Halides

The reactions of organometallic reagents with alkylhalides are coupling reactions. The reactions of alkylchlorides, bromides, or iodides proceed well when thenucleophilic partner is a cuprate or a Grignard reagent(organomagnesium halide) in the presence of a copper,iron, or nickel catalyst:

� RI R CH3(CH3)2CuLi (42)

�CH3MgBr RICu

�

RCH3 (43)

Allylic and propargylic halides also have been employedto alkylate organometallic reagents.

2. Alkyl Sulfonates

Often the alkylation of organometallic reagents is more ef-fectively accomplished with an alkyl sulfonate. Sulfonateswill undergo displacement reactions by both Grignardreagents and cuprates, frequently more cleanly and in bet-ter yield than the corresponding halides.

�

MgBr

R OS

O

O

CH3

R

(44)

3. Epoxides

Epoxides may serve as a source of electrophilic carbonfor reactions with organometallic reagents. The epoxideis opened, alkylating the organometallic while simultane-ously forming a new alcohol.

�RLi CH2 CH2

OH�/ H2O

R CH2CH2OH

(45)




4. α,β-Unsaturated Carbonyl Compounds

Nucleophilic organometallic reagents, particularlycuprates, will also add in a conjugate manner to α,β-unsaturated carbonyl compounds to form an enolate,stabilized by resonance. Bulky alkyl groups (R) enhancethis reaction relation to carbonyl addition. Similarly,α,β-unsaturated sulfones will add organometallics toform a stabilized anionic product:

R � (CH3)2 CuLiH�/H2O

O

R

O

CH3

(46)

5. Alkylation of Active Hydrogen Compounds

Active hydrogen compounds are those in which the sub-stituents present are capable of stabilizing the conjugatebase formed on deprotonation of the starting material.Stabilization of the resultant anion significantly enhancesthe acidity of the proton, hence “active hydrogen” com-pounds. Common substituents which in this way stabi-lize negative charge are esters, aldehydes, ketones, acids,nitriles, nitro groups, sulfoxides, sulfones, sulfonamides,and sulfonates. Even further increased acidity is found ondisubstitution, as in malonic esters. Deprotonation by evena weak base is sufficient to form the anion which may thenbe trapped by an alkyl halide. The reaction is a general oneand applies to the other groups described as well:

OR

O

OR

O

CH2B� OR

O

OR

O

�CH R�X OR

O

OR

O

R CH

(47)

Imines, prepared by condensation of a primary aminewith an aldehyde or ketone, may be deprotonated with astrong base to form the nitrogen equivalent of an enolate.The ready alkylation of these anions has found utility inenantioselective processes to be discussed later:

N

R�X

RN

RN

R

R��

(48)

In addition to the oxidized forms of sulfur, dithioacetals,and ketals, such as 1,3-dithianes, also have activated hy-drogens. Dithiane is readily deprotonated and easily alky-lated. The aldehyde or ketone functionality masked by the

dithioacetal or ketal is easily released under a variety ofconditions:

B� R�X

S S S S S S

R�

.(49)

6. Alkylation of Organoboron Reagents

Trialkyl boron reagents react readily with alpha halocar-bonyl compounds, displacing the halide with an alkylgroup. The reaction is of broad, general utility. Relatedreactions occur with diazoalkanes, esters and ketones:

R3B� BrCH2CR�

O

RCH2CR�

O

(50)

D. Reaction of Organometallic Reagentswith Carbonyl Compounds

1. Grignard and Related Reagents

The addition of Grignard reagents to carbonyl compoundsto form alcohols is an extraordinarily useful and generalreaction. The reaction is facile and relatively simple. Ke-tones, aldehydes, and esters which require two equivalentsof the Grignard reagent) react well:

�R H

O

CH3MgBrH�/ H2O

H

CH3

R OH

�R OCH2CH3

O

2CH3MgBr

CH3

R

OH

CH3

(51)

Other organometallic compounds, such as organolithiumorganozinc, and organoaluminum reagents, add tocarbonyls. Noteably, organocadmium and organomer-cury reagents do not add. Selectivity is possible withorganocuprate reagents which generally add to aldehydesbut not to ketones:

(CH3)2CuLiRCHO

xO

R R

R

CH3

H OH(52)

As mentioned earlier, addition of Grignard reagents toα,β-unsaturated carbonyl compounds often results in1,4-rather than 1,2-addition reactions. In closely related




chemistry, α-bromo esters, acids, amides, and nitrileswill add to carbonyl compounds in the presence of zincdust. Presumably an organozinc reagent (Reformatsky),formed in situ, reacts like a Grignard reagent and adds tothe carbonyl compound:

BrCH2CO2CH2CH3 �

CHOZn�

OH O

OCH2CH3

(53)

Organolithium reagents will add to the lithium salt of car-boxylic acids to yield, on acidification, ketones:

�

OLiCO2H

2CH3Li

OLi

CH3 H�/H2O CH3

O

(54)

2. Addition of Active Hydrogen Compoundsto Carbonyls

On deprotonation, active hydrogen compounds such asthose substituted with carbonyl groups (as in esters, alde-hydes, ketones, or acids) as well as nitriles, nitro groups,sulfoxides, sulfones, sulfonamides, and sulfonates willadd to carbonyl compounds. The best known of these con-densations is the aldol reaction. An enolate adds readilyto ketones or aldehydes to form a beta hydroxy carbonylcompound:

�CH3CHO CH2

CH

OLi

H�/ H2OCH3CHCH2CHO

OH

CH3CH CHCHO (55)

Under more vigorous reaction conditions this product isreadily dehydrated to give the α,β-unsaturated carbonylcompound. The reaction may involve self-condensation ormay be a directed (crossed) aldol condensation betweendifferent carbonyl components (Claisen–Schmidt). An al-dol condensation with a disubstituted enolate, e.g., mal-onate anion, is known as the Knoevenagel reaction:

�

CHOOH

�CH OROR

O

O

CO2R

CO2R

(56)

If substituted with an α-carboxylic acid, the aldol productis especially susceptible to decarboxylation.

3. Wittig Reactions

One of the most useful reactions of a stabilized anion witha carbonyl compound is the condensation of a phospho-rous ylide with an aldehyde or ketone to form an olefin.The ylide is usually prepared from an alkyl triphenylphos-phonium salt by deprotonation with a strong base. Ad-dition of the ylide to the carbonyl compound initiallyforms a betaine intermediate which collapses to the olefinand triphenylphosphine oxide. The phosphonium salts arecommonly prepared with triphenylphosphine and an alkylhalide; however, other less common phosphines have beenemployed as well:

�Ph3P RCH2X Ph3PCH2RX�

� (57)

�Ph3PCH2R B Ph3P��

� CH R_ (58)

�Ph3PCHR R��

R� P CHR�

Ph3

R�

R�O�

betaine

O

Ph3P R

R�

R�O

�

H

Ph3P �

R� H

R� RO

(59)

Ylides may also be prepared form phosphonates to yielda slightly more reactive ylide, in a variation known as theHorner–Emmons reaction:

(RO)2PCH2R�

O

(RO)2PCHR�

O

� PhCHOPhCH CHR�

(RO)2PO

O

�

�

(60)

The olefin product stereochemistry may be controlled bychoice of reaction conditions to form selectively either thecis or trans product.

It is also possible to form reactive ylides with sulfurcompounds. Typically, dimethyl sulfonium methylide willadd to carbonyl compounds to form epoxides rather thanalkenes:

CH3S

CH2

PhCHO CH3SCH3�CH3�

B�

CH3

S

CH3

CH2� �

PhCH

O CH2

(61)

III. NATURAL PRODUCT SYNTHESIS

The application of synthetic methods to the preparationof naturally occurring compounds has always been an im-portant part of organic chemistry. From the early days




when separation and structural elucidation was the princi-pal object, organic chemists have been interested in naturalproducts. Rapidly, however, the synthesis of these mate-rials came to be of importance, in part to verify proposedstructural assignments and, of increasing importance to-day, to prepare quantities of natural materials with interest-ing pharmacological properties otherwise available onlyin very limited amounts. The preparation of natural com-pounds has also had a significant impact on the theory oforganic chemistry, for example, the principle of the conser-vation of orbital symmetry may have resulted in part frominsights developed during the synthesis of vitamin B-12.In this section, several examples have been taken from thesynthesis of steroids and prostaglandins to illustrate thedevelopment of synthetic strategy.

A. Steroids

The early work of Robinson on the synthesis of the steroidnucleus may be contrasted with the elegant Woodward ap-proach to the total synthesis of steroids. Robinson’s ap-proach requires the masterful manipulation of function-ality so that through the course of the synthesis, relaycompounds may be reached. Relay compounds are ma-terials reached by alternate pathways, frequently by thedegradation of a readily available natural product. TheRobinson synthesis of androsterone, which spanned sev-eral years, required four stages, each stage comprised ofnumerous steps. The first stage was the transformationof 2,5-dihydroxynapthalene to the Reich diketone. Thesecond stage was the transformation of the Reich dike-tone to the Koester and Logemann (KL) ketone. The thirdstage was the transformation of the KL ketone to dimethylaetioallobilianate benzoate. The fourth required the con-version of dimethyl aetioallobilianate benzoate to andros-terone. Both the KL ketone and aetioallobilianic acid wereavailable from natural materials (Fig. 1).

In contrast to the lengthy manipulation of functional-ity to facilitate relay synthesis, Woodward’s synthesis ofmethyl d,l-3-keto-�4,9(11). 16-etiocholatrienoate is con-siderably shorter and illustrates an advance in syntheticstrategy. This efficient approach relies strongly on an un-derstanding of stereochemical relationships and the effi-cient choice of functionality (Fig. 2).

B. Prostaglandins

The prostaglandins are a closely related family of com-pounds discovered as early as 1930 but whose structurewas not determined until the 1960s. These compoundshad a variety of physiological effects, but were only avail-able in very minute quantities. As such they were idealtargets for synthesis. Biosynthetic analysis has shownthat fatty acids, in particular arachidonic acid, are the

FIGURE 1 Robinson synthesis of androsterone. (a) Sodiummethoxide, dimethyl sulfate; (b) sodium, ethanol; (c) sodiummethoxide, methyl iodide; (d) diethylaminobutanone, methyl io-dide, potassium; (e) hydrogen iodide; (f) platinum oxide, hydro-gen; (g) palladium–strontium carbonate, hydrogen; (h) sodiumtriphenyl methide, carbon dioxide; (i) diazomethane; (j) ethyl bro-moacetate, zinc; (k) platinum oxide, hydrogen; (l) phosphorousoxychloride; (m) platinum oxide, hydrogen; (n) potassium hydrox-ide, methanol; (o) oxalyl chloride; (p) diazomethane; (q) silvernitrate; (r) potassium hydroxide, methanol; (s) acetic anhydride;(t) heat.




FIGURE 2 Woodward synthesis of methyl d ,l -3-keto�4.9(11).16-etiocholatrienoate; (a) Lithium aluminum hydride; (b) acid, diox-ane; (c) acetic anhydride, zinc; (d) ethyl vinyl ketone, potassiumt-butoxide; (e) base, dioxane; (f) osmium tetroxide; (g) acid, ace-tone; (h) palladium–strontium carbonate, hydrogen; (i) acryloni-trile, Triton-B; (j) base; (k) methyl magnesium bromide, base;(l) periodic acid; (m) aqueous dioxane; (n) potassium dichromate,diazomethane.

precursors of the prostaglandins. A common feature of theprostaglandins was the cyclopentane ring, with as manyas four adjacent stereocenters (Fig. 3). It was demon-strated by Corey and widely adopted by others that acommon intermediate, “Corey’s lactone,” could be usedto synthesize a number of the prostaglandins (Fig. 4).Prostaglandin synthesis from this common intermediatecan therefore be convergent, the side chains being pre-pared by separate synthetic procedures and coupled intactto the cyclopentanoid system. This is in sharp contrast toboth the Robinson and Woodward syntheses, which werehighly linear, and demonstrates another advance in syn-thetic strategy.

Because of its utility, a number of syntheses for theCorey lactone have been developed. The variety of ap-proaches which may be employed to reach this key inter-mediate demonstrates the power of convergent syntheticstrategies. A few of the many routes to this compound arediscussed in the following illustrations.

FIGURE 3 Prostaglandins.

A functionalized cyclopentadiene was allowed to reactin a Diels–Alder reaction to yield a bicyclo-2.2.1-heptanemolecule, which on further elaboration contained the cor-rect relative stereochemistry and functionality (Fig. 5).The lactone was carried on to PGF2α by Horner–Emmonsreaction, by Wittig reaction, reduction, then deprotection(Fig. 6).

Bicyclo-2.2.1-heptadiene may also serve as a startingmaterial. Under acidic conditions the functionalized nor-tricyclic intermediate was prepared. After a few addi-tional manipulations the prescribed lactone was revealed(Fig. 7).

The lactol formed on reduction of the Corey lactonewas also the target of several synthetic approaches.1,3,5-Cyclohexanetriol was converted into the lactol via aring contraction sequence. This approach could be madeenantiospecific via resolution of an intermediate alcohol(Fig. 8).

Ring contraction was also the key step in the transforma-tion of 1,3-cyclohexadiene to the lactol. An ene reaction

FIGURE 4 Synthetic tree demonstrating the utility of the Coreylactone in preparing prostaglandins.




FIGURE 5 Synthesis of protected Corey lactone by cycloadditionto a functionalized cyclopentadiene. (a) Thallium sulfate, benzylchloromethyl ether; (b) 2-chloroacryloyl chloride; (c) sodium azide;(d) aqueous acetic acid; (e) m-chloroperbenzoic acid; (f) base,carbon dioxide; (g) potassium iodide–iodine; (h) p-phenylbenzoylchloride, tri-butyl tin hydride.

was employed in this synthetic scheme to introduce theappropriate substitution pattern (Fig. 9).

An asymmetric synthesis of the lactol from S malicacid required a homologation sequence followed by anintramolecular aldol condensation (Fig. 10).

IV. ASYMMETRIC SYNTHESIS

As was clear from the preceeding discussions of steroidsand prostaglandins, natural products are often opticallyactive. Their biological effects may be dependent on a spe-cific configuration. Traditionally enantiomers were sep-arated by resolution, i.e., the separation of the pair ofdiastereomers formed by reaction of the racemic productand an optically pure auxiliary. The use of resolution to

FIGURE 6 Elaboration of the Corey lactone to PGF2α .(a) (CH3O)2POCHCOC5H11; (b) lithium triethyl borohydride;(c) potassium carbonate, dimethoxyethane; (d) diisobutylalu-minum hydride; (e) Ph3P CH (CH2)3CO2Na.

FIGURE 7 Bicyclo-2.2.1-heptadiene route to the Corey lactone.(a) Formaldehyde, formic acid; (b) chromic acid; (c) hydrochloricacid; (d) m-chloroperbenzoic acid; (e) ethyl chloroformate, zincborohydride, dihydropyran; (f) base, hydrogen peroxide.

isolate the desired enantiomer can be feasible, but with thedrawback that one-half of the material separated will bethe undesired enantiomer as well as requiring an elementof luck in the crystallization of the desired material.

Synthethc chemists have made remarkable strides inenantioselective reactions, where the desired enantiomeris the principle product. Asymmetric syntheses have suc-cessfully employed a variety of techniques, but we willlimit discussion to enantioselective reducing agents, alky-lations, and directed aldol reactions, which are typical ofmany other reactions and reagents.

For asymmetric synthesis it is necessary to employ achiral component to direct the further transformations ofthe substrate. It is most desirable that this component bereadily available and inexpensive, therefore common natu-ral products such as terpenes or amino acids are frequentlyemployed. The synthetic chemist may opt to recover thechiral auxiliary, in the case of reductions, or may chose to

FIGURE 8 Woodward synthesis from 1,3,5-cyclohexanetriol.(a) Glyoxylic acid; (b) sodium borohydride, methanesulfonylchloride; (c) potassium hydroxide; (d) potassium carbonate,dimethoxyethane; (e) methanesulfonyl chloride; (f) potassium hy-droxide; (g) hydrogen peroxide; (h) aqueous ammonia, methanolichydrogen chloride; (i) sodium nitrite, acetic acid.




FIGURE 9 Preparation of prostaglandin precursor via an enereaction. (a) Dichloroacetyl chloride; (b) zinc, acetic acid;(c) hydrogen peroxide; (d) diisobutylaluminum hydride; (e) N-phenyltriazolinedione; (f) sodium hydroxide, methyl iodide; (g) os-mium tetroxide; (h) potassium hydroxide, methanol; (i) platinumoxide, hydrogen; (j) sodium nitrite, acetic acid.

incorporate the chiral fragment in the carbon skeleton, asin directed aldol reactions.

A. Enantioselective Reductions

Among the most efficient and well-developed enantios-elective transformations are asymmetric reductions ofalkenes and carbonyl compounds.

1. Hydroboration

Selective hydroborating reagents have been developedfrom readily available terpenes such as α-pinene andlongifolene. The most useful reagents would react in highchemical and optical yields, yet permit recycling of thechiral auxiliary. One successful reagent is diisopinocam-

FIGURE 10 Chiral synthesis of the Corey lactone from malicacid. (a) Acetyl chloride; (b) dichloromethyl methyl ether, zincchloride; (c) HOOCCH2CO2CH3, base; (d) hydroxide; (e) hydro-genation; (f) potassium hydroxide, methanol; (g) acetic anhydride;(h) dichloromethyl methyl ether, zinc chloride: (i) sodiumborohydride.

pheyl borane, prepared by reaction of two equivalents ofα-pinene with borane:

BH3 �

BH)2

(62)

Reaction of the diisopinocampheyl borane with 2-methyl-1-alkenes, leads on oxidation to a disappointing 21% enan-tiomeric excess (i.e., the excess of one enantiomer relativeto the other):

�

B)2BH)2

HHO

H

(63)

However treatment of cis-2-butene leads to e.e.’s as highas 98%. The general reaction of cis olefins appears to belimited only by the optical purity of the pinene startingmaterial:

�

BH)2 HO(64)

Dilongifolylborane, prepared by partial hydroboration ofthe longifolene, the most abundant sesquiterpene in theworld, leads to only poor asymmetric induction in thereaction of 2-methyl-1-butene:

HB(2

(65)

However, reaction of the cis alkenes is again more selec-tive, e.e.’s as high as 78% are possible.

Limonylborane, a boraheterocycle with nonequivalentalkyl groups bound to boron, is prepared from limoneneand has led only to disappointing 60% e.e.’s:

HB

H

(66)

The most reactive of the hydroborating agents, mono-isopinocampheyl borane reacts with cis olefins with 70%e.e. Also, the high reactivity of the reagent has led to itsuse with less reactive alkenes, with high e.e.’s having beenreported:




BH2

(67)

2. Chiral Borohydrides

The enantiospecific reduction of carbonyl compounds toalcohols is an extremely useful reaction. Aldehydes havebeen successfully reduced with chiral trialkylboranes togive products with very high optical purity. The hydrob-oration of α-pinene with 9-borabicyclononane (9-BBN)forms a chiral trialkylborane which reduces aldehydeswith e.e.’s from 85 to 99%:

9-BBN

HB

CD

O

C

OH

D

H

(68)

The reduction of normal ketones does not proceed well.However, the less sterically demanding acetylenic ketonescan be reduced with 72–98% e.e.’s:

B

� R C C C R�

O

R C C C R�

HO H

(69)

Complexed borohydrides have not yet lived up to theirpotential as enantiospecific reducing reagents. When thehydride reagent, lithium B-3-pinananyl-9-BBN-hydride,prepared by treatment of the α-pinene-9-BBN reagentwith t-butyl lithium, was allowed to react with aldehydesthe asymmetric induction was a disappointing 17–36%:

B

H

Li�

�

(70)

However, a related reagent prepared from nopol benzylether and 9-BBN gave e.e.’s as high as 70%. The presenceof the benzyl ether side chain was predicted to improveasymmetric induction by improving coordination of thecation:

BOBZ

(71)

3. Complexed Lithium Aluminum Hydrides

Enantioselective reduction of aldehydes and ketones hasalso been possible with lithium aluminum hydride re-agents complexed in a chiral environment. The labilityof the ligands around lithium aluminum hydride reagentshas so far limited the effectiveness of these reductions. Themost effective ligands have been those containing at leastone nitrogen. N -substituted amino methyl pyrrolidines,prepared from proline, have been used in the reduction ofketones to give optical yields as high as 96%:

N

H

N

H R

� LAHR

R�

O

HO H

R R�(72)

1,2-Amino-alcohols, such as ephedrine, also have beensuccessfully used to reduce ketones with e.e.’s from 88 to90%:

OHCH3

HN

R CH3

(73)

1,3-Amino-alcohols, such as Darvon alcohol, have beensuccessfully employed in the enantioselective reductionof acetylenic ketones:

(CH3)2N OH

CH2PhPh

CH3

(74)

Although asymmetric reductions may be very usefulprocedures, it is important to recognize that the reduc-tions are generally carried out at very low temperatureswith excess reagents in order to maximize both the op-tical and chemical yields. These conditions often makethese reagents prohibitively expensive for larger scaleoperations.

B. Asymmetric Syntheses with Enolates

The carbonyl group is one of the most useful functionalgroups because of its ability to act as an electrophile or, inthe derived enolate, as a nucleophile. As described earlier,the enolate can react with alkylating agents or with car-bonyl compounds in two very useful synthetic reactions.The stereoselectivity of an enolate is directly related to thestereochemistry of the enolate which is in turn affected bythe stereochemistry of the parent molecule.




1. Alkylation Reactions

Stereoselective formation of the enolate is essential forstereoselective alkylation. An enolate may be either E orZ using the convention that the highest priority is alwaysassigned to the OM group:

R

OM

H

R�

Z

R

OM

R�

E

(75)

Enolate geometry may be controlled by careful choiceof the deprotonation conditions. Use of sterically de-manding lithium 2,2,6,6-tetra-methylpiperidide (LiTMP)favors formation of the E enolate. Addition of hexam-ethylphosphorous triamide (HMPT) to the reaction willreverse this selectivity to favor formation of the Z eno-late. Use of the bulky base lithium hexamethyldisilazide(LiHMDS) will also favor formation of the Z enolate.

With control of the enolate geometry, in cyclic systemsalkylation tends to follow the pathways which minimizesteric interactions:

CO2R

MeLDAMeI

MeMe CO2R

�

MeRO2C Me

major (76)

Employment of these interactions, with optically activefunctional groups, leads to control of the stereochemistryof alkylation.

In acyclic systems, even with control of enolate geom-etry, control of the alkylation reaction requires additionalinteractions. It is possible to create steric interactions com-parable to those seen in cyclic systems by chelation effectswith the enolate. If the chelating functionality is asym-metric, then stereochemical control will be possible. Theresult of such chelation is obstruction of one face of theenolate during the alkylation reactions:

RO

N O

OR

O

N O

O

Li

RO

N O

O

R

(77)

A number of chiral auxiliaries derived from available chi-ral substrates such as valine, norephedrine, or proline havebeen successfully employed.

FIGURE 11 Zimmerman–Traxler transition state hypothesis forthe directed aldol reaction.

2. Directed Aldol Reactions

The stereochemical selectivity of directed aldol reactionsis also dependent on enolate geometry. Diastereoselectiv-ity has been postulated to result from steric interactions ina six-membered cyclic transition state (Fig. 11). Diastere-ofaceselectivity results from addition to a chiral aldehyde,where the two new asymmetric centers may be controlledby a third center present in the starting aldehyde.

As illustrated in Fig. 11, as a general rule Z enolates tendto form syn aldols and E enolates tend to form anti aldols.The rule holds better for the reactions of Z enolates thanE enolates. Selectivity increases with the steric demand ofthe enolate substituents. Diastereoselectivity can be pooreven with good control of the enolate geometry if the stericinteractions are not significant. Selectivity is also effectedby the enolate counterion. The tighter the transition state,the more effective the steric repulsions will be in effectingdiastereoselectivity. Metals with shorter bonds to oxygen,therefore, will be more effective. Boron enolates are gen-erally more selective for a given carbonyl compound thanother metal enolates.

Diastereofaceselectivity describes the preference of anenolate to react with one face of a chiral aldehyde over theother. This problem has been analyzed many ways; how-ever, Cram’s rule is adequate in many cases. Cram’s rulestates that a nucleophile will react with a carbonyl from thesame face as the smallest substituent when the carbonylgroup is flanked by the largest substituent. Cram’s rule isillustrated by a Newman projection below.




O MS

Nu R L

(78)

More complete analyses, taking into consideration the na-ture of the flanking substituents, which employ orbitaloverlap arguments, dipole effects, and chelation, havebeen advanced to explain results which cannot be under-stood by the simple Cram model. As might be deducedfrom this discussion, diastereofaceselectivity is remark-ably dependent on the nature of the reactants.

Asymmetric induction is also possible in the reactionof chiral enolates with achiral aldehydes and ketones.This approach has been successfully employed with chiralimides, amides, and alpha hydroxyketones:

RO

N O

OR

O

N

OR�

CH3 Ph

R�O

OSiMe2t-Bu

(79)

The asymmetric inductions realized with these chi-ral reagents may be amplified remarkably by their reac-tion with chiral aldehydes which have a complimentaryasymmetry that tends to induce formation of the sameenantiomer:

R CHO �

O

H

OSiMe2tBu

RHOSiMe2tBu

OOH

�

RHOSiMe2tBu

OOH

100

(80)

V. BIOMIMETIC SYNTHESIS

As discussed earlier natural products have stimulated thedevelopment of synthetic chemistry by providing chal-lenging targets since the nineteenth century. In attemptingto prepare known natural substances, it is only recentlythat chemists have turned to mimicking the synthetic ap-proaches employed by nature. This discussion is limitedto two different types of biomimetic synthesis, one basedon the polyene cyclization reaction and the other the useof biomimetic reagents, functionalized cyclodextrins, asguest–host enzyme models.

A. Biomimetic Polyene Cyclization

The independent proposal of groups at Columbia Uni-versity and the ETH in Zurich that the stereoselectivityobserved on the cyclization of polyenes is a consequenceof the olefin stereochemistry constitutes the basis of theStork–Eschenmoser hypothesis. The epoxide-initiated cy-clization of squalene results in the stereospecific formationof dammaradienol. The all-trans double bond geometry ofsqualene results in the trans–anti-stereochemistry of thering junctions in dammaradienol:

Squalene

HO

(81)

This postulate led W. S. Johnson to study the prepara-tive utility of polyene cyclizations. The Johnson groupstudied numerous functional groups to both initiate andterminate cyclization of trans polyenes. The treatment ofallylic alcohols under acidic conditions to form an al-lylic cation proved to be one of the more useful initiat-ing functional reactions. Termination of cyclization by analkyne, to form an intermediate vinyl cation which wasthen trapped, proved to be exceptionally efficient:

CH3

HOCH3

CH3

O

(82)

The efficiency of this process results in a 78% yield oftetracyclic material with very good control of the ringjunction stereochemistry. This can be contrasted to thework of Robinson and Woodward described earlier. Par-ticularly exciting was the observation that asymmetric cy-clization was possible even when the asymmetric centerwas not involved in the bond forming process. Cycliza-tion of a pro-C-11 hydroxy polyene with one of the bestterminators, a vinyl fluoride, resulted in a 79.5% yield ofthe compound shown.




CH3HO

CH3

HOF F

HO

HOO

�

(83)

This material, with the correct configuration at C-11, maybe efficiently converted to 11-hydroxy-progesterone, anintermediate in a commercial synthesis of hydrocortisoneacetate. Johnson has shown that excellent asymmetric in-duction in polyene cyclizations is possible by treating ac-etals derived from (2s,4s)-pentanediol with Lewis acids:

O O

SiEt

O

HO

(84)

B. Guest–Host Inclusion Enzyme Models

Enzymes have unique catalytic, regulatory, and transportproperties. The development of reagents that mimic theseproperties would be a significant synthetic advance. Inearly work it was possible to prepare model compoundsthat could duplicate simplistically some of the transportproperties but with little substrate selectivity. Currentlyenzyme models are being prepared which not only havethe ability to recognize the substrate but also can effectsite specific reactions.

To promote substrate selectivity several types of inclu-sion host molecules have been developed. Remarkablesuccesses in chiral recognition have been achieved withcyclic polyether compounds known as crown ethers:

O O

O O

O

O

(85)

Cyclodextrins, oligosaccharides forming a distortedtorus whose cavity may be occupied by other molecules,have also been explored as selective host molecules:

(86)

Cyclodextrins can have very regular cavity sizes. The cav-ity will incorporate aromatics to a uniform depth, oftentilting the substrate away from the axis of the truncatedconical cyclodextrin, most probably to maximize van derWaals contact. The limited size of the cavity results in sub-strate selectivity. It has also been possible to selectivelyfunctionalize cyclodextrins as illustrated by the prepara-tion of a cyclodextrin substituted with imidazole units:

ClSO2 SO2Cl

OO O

SO2 SO2

O

KI

I I

N NH

N

N

N

N(87)

After the development of methods to prepare thesemolecules, the preparation of compounds more closelyresembling enzymes was possible. Typical of these are cy-clodextrins bearing a histamine-zinc complex and foundto function as anhydrase models or polyamine metalcomplexes which act as carboxylic hydrolases:

H2N�NH2

�

N

N

N

N

N

N

N

O

C

O

M�(88)

There are numerous other examples of biomimetic syn-thesis. This narrow selection has been chosen to illustratethe progress and potential of this area.





CATALYSIS, HOMOGENEOUS • ELECTRON TRANSFER

REACTIONS • HALOGEN CHEMISTRY • HETEROCYCLIC

CHEMISTRY • ORGANIC CHEMICAL SYSTEMS, THEORY •ORGANIC CHEMISTRY, COMPOUND DETECTION • ORGA-NIC MACROCYCLES • ORGANOMETALLIC CHEMISTRY •PHARMACEUTICALS • POLYMERS, SYNTHESIS • RUBBER,SYNTHETIC

BIBLIOGRAPHY

Boeckman, R. K., ed.-in-chief (1996). “Organic Synthesis,” John Wiley& Sons, New York.

Burke, S. D., ed. (1999). “Handbook of Reagents for Organic Synthesis,Oxidizing and Reducing Agents,” John Wiley & Sons, New York.

Coates, R. M., and Scott, E. D., eds. (1999). “Handbook of Reagentsfor Organic Synthesis, Reagents, Auxiliaries, and Catalysts for C CBond Formation,” John Wiley & Sons, New York.

Diederich, F., and Stang, P. J., ed. (2000). “Templated Organic Synthe-sis,” John Wiley & Sons, New York.

Fieser, M. (1999). “Fiesers’ Reagents for Organic Syntheses,” Vol. 18,John Wiley & Sons, New York.

Ho, T.-L. (2000). “Fiesers’ Reagents for Organic Synthesis,” Vol. 20,John Wiley & Sons, New York.

Lednicer, D. (1997). “Strategies for Organic Drug Synthesis and Design,”John Wiley & Sons, New York.

Lednicer, D., and Mitscher, L. A. (1998). “The Organic Chemistry ofDrug Synthesis,” Vol. 6, John Wiley & Sons, New York.

Liska, F., and Volke, J. (1994). “Electrochemistry in Organic Synthesis,”Springer-Veriag, Berlin/New York.

Mattay, J., ed. (1994). “Photochemical Key Steps in Organic Syn-thesis: An Experimental Course Book,” John Wiley & Sons, NewYork.

Paquette, L. A., ed.-in-chief (1995). “Encyclopedia of Reagents for Or-ganic Synthesis, Vol. 8, John Wiley & Sons, New York.

Pearson, A. J., ed. (1999). “Handbook of Reagents for Organic Synthesis,Activating Agents and Protecting Groups,” John Wiley & Sons, NewYork.

The University of Liverpool, U.K. (2001). “Catalysts for Fine ChemicalsSynthesis, Vol. 2, A Manual for Organic Synthesis,” John Wiley &Sons, New York.

Ward, R. S. (1999). “Selectivity in Organic Synthesis,” John Wiley &Sons, New York.

Zaragoza-Dorwald, F. (2000). “Organic Synthesis on Solid Phase: Sup-ports, Linkers, Reactions,” John Wiley & Sons, New York.

P1: GNH Final Pages Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology EN011A-543 February 12, 2002 12:40

Organic MacrocyclesJ. Ty ReddSouthern Utah University

Reed M. lzattJerald S. BradshawBrigham Young University

I. Macrocyclic Structure and Metal CationComplexation

II. Complexation of Organic CationsIII. Complexation of Anions and Neutral MoleculesIV. Applications of Macrocyclic LigandsV. Synthesis of Macrocyclic Compounds

GLOSSARY

Binding constant Equilibrium constant associated witha reaction in which a ligand binds to a substrate.

Biomimetic That which mimics biological systems.Calixarene Class of compounds consisting of a ring of

phenol moities connected by methylene bridges.Crown ether Cyclic organic compound consisting of a

number of connecting ethylene oxide units.Cryptand Organic compound consisting of two or more

bridges connecting nitrogen atoms to give multiplerings.

Cyclic polyether See crown ether.Cyclodextrin Class of cyclic polysaccharide com-

pounds.Enantiomeric recognition Differentiation of one enan-

tiomer of the guest from the other by a chiral host.Guest Chemical species that can be trapped by a host

compound.Host Compound that can trap a guest species.Macrobicyclic effect Extra stability of complexes of

macrobicyclic (cryptands) over those of crown ethers.

Macrocycle Cyclic ligand large enough to accommodatea substrate in the central cavity.

Macrocyclic effect Extra stability of complexes of cyclicligands over those of analogous acyclic ligands.

Molecular recognition Ability of molecules to discernone another through molecular interactions that conse-quently form stable organized structures.

Self-assembly A spontaneous organization of moleculesor objects into stable aggregates by noncovalent forces.

Spherand Type of macrocycle (see compound 11, Fig. 1).Supramolecular chemistry A discipline that exploits

fundamental concepts such as self-assembly, organi-zation, and replication.

MACROCYCLES comprise a large group of heterocyclicorganic compounds that can bind cationic, anionic, or neu-tral substrates by entrapment within the cavity created bythe macrocyclic structure. The selectivity of the macrocy-cle for certain chemical species is a function of many pa-rameters, an important one being the match between sub-strate size and macrocycle cavity size. When the substrate

517

P1: GNH Final Pages


518 Organic Macrocycles

is surrounded by the macrocycle structure, it is partly orcompletely isolated from the solvent. By this means, itis possible to solubilize bound substrates into solvents ormembranes in which the unbound substrate is not solu-ble. Furthermore, the change in chemical reactivity of thebound substrate may be exploited to yield catalytic andbiomimetic substrate transformation.

I. MACROCYCLIC STRUCTURE ANDMETAL CATION COMPLEXATION

Macrocycles, also called macrocyclic compounds, macro-cyclic ligands, macropolycycles, and so forth, include a va-riety of basic structures. The major headings below groupcurrently known macrocycles into broad, general classes.However, the term macrocycle is not confined to the lim-ited number of representative compounds presented in thisarticle.

A. Crown Ethers

The name crown ether was applied to cyclic polyethermolecules such as compounds 1–3 in Fig.1, by Pederson,who first reported their preparation in 1967. A trivial non-rigorous nomenclature is commonly used to streamlinenaming of these complex molecules. Names are structuredas follows: (1) principal ring substituents, (2) heteroatomssubstituted for oxygen, (3) number of atoms in the princi-pal ring, (4) the name crown, and (5) the number of het-eroatoms in the principal ring. Thus, compound 1 is named15-crown-5, compound 2 is 1,10-dithia-18-crown-6,and compound 3 is dibenzo-30-crown-10.

Crown ethers are particularly interesting ligands for tworeasons: They measurably bind alkali metal cations in wa-ter solution, and they demonstrate size-based selectivityof metal ions. These features are illustrated for the lig-ands 15-crown-5, 18-crown-6, and 21-crown-7 in Fig. 2,where the thermodynamic equilibrium constant K for thereaction in methanol

Mn + + 18-crown-6 = (M − 18-crown-6)n +

is plotted versus cation radius. Of the monovalent metalcations, K+ is bound most strongly by 18-crown-6.X-ray crystallographic determination of the structure ofthe K+–18-crown-6 complex shows that the K+ ion sits atthe center of the ligand cavity surrounded by the six ligandoxygen atoms as shown schematically in Fig. 3. The K+

ion is nearly the correct size to fill the ligand cavity andis bound most strongly. The Na+ ion is smaller than the18-crown-6 ligand cavity, so the ligand must fold slightlyto permit all six oxygens to associate with the cation. BothRb+ and Cs+ are too large to fit into the ligand cavity. Thus,the relative sizes of cation and ligand cavity explain the se-

lectivity of 18-crown-6 for K+. Likewise, among alkalineearth cations, the Ba2+ ion both fits best in the 18-crown-6ligand cavity and is bound most strongly. Table I, whichlists the ionic radii of a number of metal cations, showsthat K+ and Ba2+ are of almost equal size.

Figure 2 shows that 21-crown-7, like 18-crown-6, bindsthe monovalent cation whose size matches that of theligand cavity, that is, Cs+. However, the selectivity of15-crown-5 is not easily explained on the basis of relativesize. While the cation Na+ best matches the cavity in size,K+ is bound slightly more strongly. This case illustratesthat size is not the only factor and is often not the deter-mining factor that controls cation selectivity. The problemarises from the fact that Na+ ion is slightly too large to fitinto the 15-crown-5 cavity. For such a case, cation solva-tion energies dominate in the free energy cycle

Mg + Lg −→ MLg∣∣∣∣∣

∣∣∣∣∣

∣∣∣∣∣

Ms + Ls −→ MLs

for complex formation. (The subscripts g and s indicatespecies in the gas and solvent phases, respectively.) Com-pared to Na+, the larger K+ is less strongly solvated be-cause of its lower charge-to-radius ratio, so less energy isexpended in removing solvent molecules in the complex-ation process. Because the range of macrocycle sizes ismuch larger than the range of cation sizes, it is relativelyrare that selectivity is governed by size predominantly.It is more often the case that solvation, ligand flexibility,and the effective charge on the binding sites play thedominant roles.

Crown ethers have affinity for metal ions besides thoseof the alkali and alkaline earth series. Figure 4 shows thebinding constants of 18-crown-6 with the series of triva-lent lanthanide cations, which decrease in size across theseries. Table II shows the binding constants of several sim-ple crown ethers with Pb2+, Ag+, Tl+, and Hg2+.

When the oxygen heteroatoms of crown ethers are re-placed by nitrogen or sulfur, the selectivity of the ligandschanges markedly. For example, sulfur-containing analogsof 18-crown-6 have lower affinity for alkali and alka-line earth cations and greater affinity for more polarizablecations such as Tl+ and Hg2+. When nitrogen is substi-tuted, the affinity for alkali and alkaline earth cations alsodrops, while that for Pb2+ and Ag2+ increases.

Substitution of aliphatic or aromatic groups onto theheterocyclic backbone of crown ethers has a destabilizingeffect on complex stability. Table III shows that the stabili-ties of complexes of dicyclohexano-18-crown-6 are muchmore like those of 18-crown-6 than are those of dibenzo-18-crown-6. In the latter case, the benzene rings with-draw electron density from the oxygen atoms, lowering the

P1: GNH Final Pages


Organic Macrocycles 519

FIGURE 1 Structures of compounds discussed in this article.

P1: GNH Final Pages



FIGURE 2 Log K for the reaction in CH3OH at 25◦C of severalunivalent cations with 15-crown-5, 18-crown-6, and 21-crown-6versus cation radius.

energy of the ion–dipole interaction in the complex. Thisexplanation is borne out in the observation that if electron-withdrawing substituents are added to the benzene rings,cation complex stability constants drop even further.

Bound metal ions show markedly different redox prop-erties from unbound ions. It is possible, using macrocyclessuch as crown ethers, to stabilize oxidation states of metalions such as Eu2+.

B. Cryptands

Cryptands, or macrobicyclic ligands, are similar in struc-ture to crown ethers, differing in the addition of a bridgethat reaches across the ring to give football-shaped struc-tures such as compound 4 in Fig. 1. The trivial nomen-clature used for these ligands is given by the number ofethylene oxide units in each of the three bridges connecting

FIGURE 3 Structure of the potassium thiocyanate-18-crown-6complex based on an X-ray crystallographic determination.

TABLE I Radii of Cations and of Crown EtherCavities

Cation Radius ( A)

Li+ 0.74

Na+ 1.02

K+ 1.38

Rb+ 1.49

Cs+ 1.70

Mg2+ 0.72

Ca2+ 1.00

Sr2+ 1.16

Ba2+ 1.36

Ag+ 1.15

Tl+ 1.50

Pb2+ 1.18

Hg2+ 1.02

Cd2+ 0.95

Macrocycle Radius ( A)

15-crown-5 0.85

18-crown-6 1.3

21-crown-7 1.7

the two nitrogen heteroatoms. Thus compound 4 with twooxygen atoms in each bridge is designated 2.2.2. The ad-ditional bridge facilitates more effective encapsulation ofmetal ions. Consequently, the complexes of these ligandsare in general more stable than those of crown ethers. Fig-ure 5 shows the binding constants of a series of cryptandsfor alkali metal ions. Comparison of Fig. 5 with Fig. 2shows the higher stability of the cryptand complexes. Italso shows that the cryptands have a high degree of se-lectivity and that there is a cryptand of the correct sizeto be selective for each of the ions in the alkali metalsequence. The selectivity in all these cases is largely aresult of the match between cation and cavity sizes. Theeffects of substituting other heteroatoms for oxygen andof organic substitution on the ring backbone are similar tothose for crown ethers.

C. Naturally Occurring Macrocycles

Before crown ethers or cryptands were synthesized,macrocyclic ligands of various types had been isolatedas natural products from microbial species. Compoundssuch as valinomycin and enniatin B have been studiedextensively because of their ability selectively to bindalkali and alkaline earth metal ions and to transportsuch ions through biological membranes. Valinomycin isused in K+ ion selective electrodes because of its high(10,000:1) K+ : Na+ selectivity.

P1: GNH Final Pages



FIGURE 4 Log K for the reaction M3+ + L = ML3+ (L = 18-crown-6) in methanol at 25◦C versus reciprocal of cation radius. No re-action was observed with Tb3+, Dy3+, Ho3+, Er3+, Tm3+, Yb3+,or Lu3+.

Valinomycin (compound 5, Fig. 1) is similar to a cyclicprotein in structure. In the unfolded configuration, it is toolarge to accommodate metal ions. However, intramolecu-lar hydrogen bonds cause the ring to tighten in on itself,providing a nearly octahedral arrangement of the oxygenatoms of the correct size to accommodate a K+ ion. Theexterior of the ligand is hydrophobic, making it soluble

TABLE II Stability Constants (log K ) of Crown Ethers withHeavy Metal Ions in Water at 25◦C

Stability constant

Ligand Pb2++ Hg2++ Tl++ Ag++

15-Crown-6 1.85 1.68 1.23 0.94

18-Crown-6 4.27 2.42 2.27 1.50

Dibenzo-18-crown-6 1.89 — 1.50 1.41

Dicyclohexano-18-crown-6 4.43 2.60 1.83 1.59(cis-anti-cis)

1,10-Dithia-18-crown-6 3.13 >5 0.93 4.34

1,10-Diaza-18-crown-6 6.90 17.85 — 7.8

TABLE III Effect of Substituent Groups on Cation Com-plex Stability (log K ) with 18-Crown-6 and Its Analogs inMethanol at 25◦C

Stability constant

Crown ether Na++ K++ Ba2++

18-Crown-6 4.36 6.06 7.04

Cyclohexano-18-crown-6 4.09 5.89 —

Dicyclohexano-18-crown-6 3.68 5.38 —(cis-anti-cis)

Benzo-18-crown-6 4.35 5.05 5.35

Dibenzo-18-crown-6 4.36 5.00 4.28

in lipid membranes. Careful kinetic studies demonstratethat the selectivity of the ligand for K+ is a function of therate of cation release from the complex, there being littledifference in the rate of cation uptake into the ligand.

D. Tetraaza Macrocycles

Macrocyclic ligands containing four nitrogen heteroatomsseparated by various organic bridges have been studied

FIGURE 5 Variation of equilibrium constant K in 95 volume per-cent methanol for the reaction of several cryptands (see com-pound 4, Fig. 1) with the alkali metal ions Li+–Cs+ (plotted ac-cording to increasing metal ion radius).

P1: GNH Final Pages



for many decades. Examples are porphyrin (compound 6,Fig. 1) and its analogs, which are the metal-binding sitesin many metalloenzymes, hemoglobin, and other natu-rally occurring compounds. There are also many synthetictetraaza macrocycles that have affinity for divalent (andother) transition metal ions. Their binding constants withthese cations are generally much higher than those typicalof crown ethers and cryptands, so the metal ion is heldvirtually irreversibly. For example, the binding constant(log K ) of compound 7, Fig. 1, with Hg2+ is 23 and withNi2+ is 23.5. The importance of these complexes lies inthe binding of additional ligands at the axial sites of thebound ion, which serves as the enzyme’s active site. Thedegree and type of aromaticity in the tetraaza ligand struc-ture has a profound influence on the electronic propertiesof the bound metal ion, which in turn affects the strengthand nature of binding to additional ligands.

E. Other Macrocycles

A wide variety of macrocycle types has been reported inaddition to the general categories discussed above. Thecalixarene ligands (compound 8, Fig. 1), which are waterinsoluble, have a strong, selective affinity for Cs+. Theyform neutral complexes through loss of a proton. The lariatethers (compound 9, Fig. 1) form neutral complexes bythe same mechanism, resembling a crown ether with anarm that can reach around to provide ligation at the axialposition. Macrotricyclic cryptands (compound 10, Fig. 1)provide essentially spherical or cylindrical neutral trapsfor metal ions. Spherands (compound 11, Fig. 1) likewiseoffer elegant binding geometries in which metal ions arebound. The list of macrocycles is far greater than can bepresented in this limited space.

II. COMPLEXATION OF ORGANIC CATIONS

Ammonium and organosubstituted ammonium cationsbind to crown ethers and other macrocycles by the for-mation of hydrogen bonds to the ligand heteroatoms. Anexample is the complex of an alkylammonium cation with18-crown-6 shown in Fig. 6a. The stability of such com-plexes is influenced by the number of hydrogen bondsthat can form and by the degree of steric hindrance forapproach of the substrate to the ligand. Table IV lists thebinding constants for a number of ammonium cations with18-crown-6. The stability drops dramatically as the num-ber of available hydrogen bonds is reduced from 3 to 2to 1. Furthermore, anilinium ions, which contain orthosubstituents, form weak or no complexes because the sub-stituents sterically hinder approach of the −NH+

3 group tothe ligand.

FIGURE 6 Diagrammatic representation of mode of binding of(a) anilinium and (b) benzenediazonium cations to 18-crown-6.

Unlike ammonium cations, diazonium cations complexto crown ethers by insertion of the positive moiety into thecavity, as in Fig. 6b. Table V shows that complex stabilitydeteriorates markedly with ortho substitution in benzene-diazonium cation because of steric hindrance. Figure 7

TABLE IV Stability Constants (log K ) for Reac-tion of 18-Crown-6 and with Several Organic Am-monium Cations in Methanol at 25◦C

Cation Log K

RNH+3 cations

NH+4 4.27 ± 0.02

HONH+3 3.99 ± 0.03

NH2NH+3 4.21 ± 0.02

CH3NHNH+3 3.41 ± 0.02

CH3NH+3 4.25 ± 0.04

CH3CH2NH+3 3.99 ± 0.03

CH3CH2OC(O)CH2NH+3 3.84 ± 0.04

CH3(CH2)2NH+3 3.97 ± 0.07

CH3(CH2)2NH+3 3.90 ± 0.04

CH2CHCH2NH+3 4.02 ± 0.03

CHCCH2NH+3 4.13 ± 0.02

(CH3)2CHNH+3 3.56 ± 0.03

CH3CH2OC(O)CH(CH3)NH+3 3.28 ± 0.02

(CH3)3CNH+3 2.90 ± 0.03

PhCH(CH3)NH+3 3.84 ± 0.01

PhNH+3 3.80 ± 0.03

2-CH3C6H4NH+3 2.86 ± 0.03

4-CH3C6H4NH+3 3.82 ± 0.04

2,6-(CH3)2C6H3NH+3 2.00 ± 0.05

3,5-(CH3)2C6H3NH+3 3.74 ± 0.02

R2NH+2

NH2C(NH2)NH+2 21.7 ± 0.02

(CH3)2NH+2 1.76 ± 0.02

(CH3CH2)2NH+2

R3NH+ cations

(CH3)3NH+ No complex

R4N+ cations

(CH3)4N+ No complex

P1: GNH Final Pages



TABLE V Stability Constants (log K ) for Reac-tion in Methanol at 25◦C of Arenediazonium andAnilinium Cations with 18-Crown-6

Cation Log K

PhNH+3 3.80

2-CH3C6H4NH+3 2.86

2,6-(CH3)2C6H3NH+3 2.00

PhNN+ 2.50

2-CH3C6H4NN+ a

2,6-(CH3)2C6H3NN+ a

a No measurable reaction.

shows that complex stability is a regular function of theelectron density in the N+

2 moiety. As electron densityincreases, stability declines.

The binding of organoammonium-type cations tomacrocycles is the subject of intense interest due to theability of such systems to mimic enzymes. Specifically,it is possible to add functionalities to the macrocyclicstructure to permit chiral recognition in substrates. Chiralmacrocycle 12 in Fig. 1, in the (S,S)-form, for example,binds one enantiomer of α-(1-naphthyl)ethylammoniumperchlorate more strongly than the other isomer [log K =2.47 ± 0.01 for the (R)-ammonium salt and 2.06 ± 0.01for the (S)-salt]. The difference in binding constants re-sults from a greater steric hindrance to the approach of thesubstrate for one isomer due to the presence of the bulkyfunctionalities.

III. COMPLEXATION OF ANIONSAND NEUTRAL MOLECULES

Considerably less attention has been given to the bindingof anions to macrocycles than to that of cations. Basically

FIGURE 7 Plot of log K for formation in methanol at 25◦C of the18-crown-6 complex of p -RC6H4NN+ versus Hammett σ+

p valuesof R.

TABLE VI Stability Constants (log K ) for Reac-tion in Anion with Macrocyclic Ligand 10

Log K , reaction withAnion ML 10 (protonated)

Cl− (in water) >4.0

Br− (in water) <1.0

Br− (in 90% methanol) 1.75

two types of anion binding are known. The first involvesmacrocyclic structures containing basic sites that providepositively charged binding sites when protonated. Exam-ples of these are compounds 10 (in protonated form) and14 in Fig. 1, and binding constants with typical anions arefound in Table VI. Structure 14 is based on cyclodextrin,which is a large cyclic polysaccharide.

Cyclodextrins are able to accommodate neutral mole-cules as well as anions. The large cavity contains numer-ous hydrogen-bonding sites if needed. In general, the cav-ity simply provides a comfortable microenvironment formany neutral species, especially when the solvent envi-ronment is less than ideal.

IV. APPLICATIONS OF MACROCYCLICLIGANDS

A. Extractants

One of the first identified uses for crown ether and cryptandligands was as phase transfer agents in catalyzing syn-thetic organic reactions. The macrocycle can be used tosolubilize salts having oxidizing or reducing anions intohydrophobic solvents. For example, KMnO4 can be solu-bilized into benzene by 18-crown-6. The “naked” MnO−

4ion that accompanies the K+–18-crown-6 complex in so-lution is a very powerful oxidizing agent in this medium.Use of naked anions of this type has provided a method toenhance the efficiency of many synthetic reactions.

Macrocycles have also been proposed as metal ion ex-tractants in separation processes. It has been shown byJ. McDowell and his co-workers at Oak Ridge NationalLaboratory that synergistic effects occur when crownethers are used as coextractants with traditional extrac-tants like diethyl hexylphosphoric acid (HDEHP). Specif-ically, the degree of metal ion extraction is greater whenboth crown and HDEHP are used together than the sumof extraction efficiencies when each is used separately.Furthermore, by using the crown, the selectivity of ex-traction processes can be altered in this manner.

B. Selective Ion Separations

Macrocyclic crown ethers can be attached to silica gelthrough stable C Si and Si O Si bonds (see Fig. 8) by

P1: GNH Final Pages



FIGURE 8 Chiral pyridino-18-crown-6 macrocyclic host attachedto a solid support (silica gel).

heating a mixture of silica gel and crown ether-containingethoxy silanes. These solid-supported macrocycles arecurrently finding success in the selective separation ofmetal cations. The macrocycle is permanently bound tothe solid support and cannot be removed unless the silicagel is destroyed, as in concentrated aqueous base. Thisprovides avenues for reusable separation systems.

Equilibrium constants for the association of metalcations with the silica gel-bound crown ethers are compa-rable to log K (H2O) values for the association of the samecations with unbound ligands. These bound macrocyclesystems can completely retain the desired metal ions whileallowing the other metal ions to pass through the system.This method of using solid supported macrocyclic ligandsto remove, separate, and concentrate specific metal ionsfrom aqueous solutions has been commercialized for usein many analytical, environmental, metallurgical, preciousmetal, nuclear, and industrial operations. For example, thismolecular recognition technology has been used in a pilotplant to separate and purify desired metal ions at a su-perfund site in Butte, Montana. The composition of metalcontaminants in the water at this site includes Al, As, Cd,Ca, Cu, Fe, K, Mg, Mn, Ni, Na, Pb, Zn, and Si. The an-ions are chloride and sulfate. The pH of this solution isapproximately 3 and the reduction potential is approxi-mately 620 mV. The pH and ion concentrations exceeddischarge regulations as promulgated by the U.S. Envi-ronmental Protection Agency (EPA) Gold Book standardsfor pristine water supplies by large factors, often severalorders of magnitude. Table VII illustrates that this envi-ronmentally friendly solid supported macrocyclic systemdeveloped by IBC Advanced Technologies, Inc., of Amer-ican Fork, Utah, was capable of achieving high metal re-covery and effluent water qualities that meet EPA GoldBook and/or regulatory drinking water standards. The tar-get metals were recovered without the addition of anyundesirable ion to the system and were of sufficient purity

TABLE VII Silica Gel-Supported Macrocyclic SeparationSystem Results from Superfund Site

Influent Treated(Source) effluent

concentration concentrationa Recovery EluentMetal (mg/L) (mg/L) (%) purity (%)

Cu 180 <0.02 >99 96–98

Fe 994 <0.1 >99 99.5

Al 270 <1 >99b 97–98

Zn 554 <0.05 >99 99–99.5

Mn 194 <1 >99 75–80b

Cd 2 <0.02

As 0.3 <0.1

a All below detection by the analytical methods used.b Some Al slipped past the Al system but was recovered with the Mn.

Total Al recovery was >99%.

to allow them to be marketed as concentrated solutions forfurther refining.

There will be increased emphasis in the future on de-veloping macrocyclic ligands that are highly selectivefor particular target cations and anions in many matri-ces and conditions. The impetus for these developmentswill be increased environmental awareness and the needfor greater cost effectiveness in the separation and purifi-cation procedures.

C. Supramacrocycles: Self-Assemblyand Molecular Machines

Tremendous advances in synthetic methodologies andmacrocyclic structure exegesis have allowed the creationand characterization of macrocycles and surpramacrocy-cles which, until recently, were sequestered in the mindsof creative chemists. Examples of molecular topology, in-tertwining, and self-assembly are ubiquitous in the nat-ural world. For example, the double helix, lipids, vi-ral capsids, and the secondary, tertiary, and quaternarystructures of many biomolecules all posses varied levelsof molecular topology, intertwining, and self-assembly.Macrocyclic chemistry of the 1970s has fathered thesupramolecular chemistry of the 1980s and 1990s. Itis now a new millennium and it is becoming feasibleto construct large (nanometer to millimeter dimensions)and intricate supramolecular entities with machine-likeproperties in order to investigate the processes centralto nature’s forms and functions. Rotaxanes, pseudorotax-anes, and catenanes (Fig. 9) are interesting examples ofthis vast supramolecular class because the relative posi-tions of their component parts can be induced to change asa result of an external stimulus. In most cases this mechan-ical mobility occurs between well-defined states that canbe switched on or off by external stimuli such as chemical,photochemical, electrochemical, or electrical energy.

P1: GNH Final Pages



FIGURE 9 Examples of supramacrocycles.

Rotaxanes derive their name from the Latin word rota,meaning wheel, and axis, meaning axle. Rotaxanes con-sist of a dumbbell-shaped moiety, in the form of a rod andtwo bulky stopper groups around which there are encir-cling macrocyclic component(s). The sterically impedingstoppers of the dumbbell prevent the macrocycle(s) fromdisassociating from the rod portion of the assemblage. Ifthese stopper groups are absent from the ends of the rodmolecule, or of insufficient size to provide a steric barrier,the supramolecules are termed a pseudorotaxane.

Catenanes, from the Latin word catena, meaning chain,are molecules containing two or more interlocked ringsthat are inseparable unless a covalent bond is broken. Thenomenclature used for these systems involves placing thenumber of components involved in square brackets prior tothe name of the molecule. Thus, compound (a) in Fig. 9 isa [2]catenane and compound (c) in Fig. 9 is a [2]rotaxane.

Clearly, the first steps have been taken toward creatingsimple molecular machines. Figure 10 shows a molecularshuttle made from a [2]rotaxane capable of translationalisomerism controlled by electrochemical and chemicalswitching. The dumbbell-shaped component in this ex-ample incorporates two different recognition sites, one abiphenol and the other a benzidine unit. The cyclophanering system with its four formal positive charges prefer-entially locates at the more π -electron-rich benzidine unitin CD3CN at −44◦C (84:16) (middle of Fig. 10). Elec-

trochemical oxidation of the benzidine unit converts it toa radical monocation state, and the tetracationic macro-cyclic ring moves to the biphenol site (top of Fig. 10).This redox situation is completely reversible. Chemicalswitching also occurs by protonation (trifluoroacetic acid,TFA) to form a diprotonated benzidine species causingthe tetracationic cyclophane ring to move to the biphenolsite (bottom of Fig. 10). This chemical process is reversedupon addition of pyridine to remove the protons from thebenzidine. Thus, a controllable molecular shuttle systemhas been developed. It must be emphasized that this is butone example in an incredibly large and extremely interest-ing field of supramolecular chemistry. Future research willsurely focus on arrays and higher assemblies of moleculesin this area.

V. SYNTHESIS OF MACROCYCLICCOMPOUNDS

The synthetic organic macrocyclic ligands come in allshapes and varieties. Early work was with the polynitro-gen macrocyclic compounds such as the cyclams (com-pound 7, Fig. 1). More recent innovations have been withthe macrocyclic polyether (crown) compounds. CharlesJ. Pedersen of duPont first reported these compoundsin 1967. Pedersen was preparing the bis-phenol substi-tuted polyether shown in Eq. (1). He isolated a good

P1: GNH Final Pages



FIGURE 10 A molecular shuttle controllable by chemical and electrochemical switching.

yield of his intended product but persisted in purifyingthe by-product to obtain dibenzo-18-crown-6 (compound13, Fig. 1) which proved to be a remarkable complexingagent for cations. Compound 13 was produced from thecatechol impurity in the starting phenol shown in Eq. (1).When the same reaction was carried out with catechol, agood yield of compound 13 was isolated [Eq. (2)]. Thisreaction is a Williamsen ether synthesis.

(1)

(2)

The synthesis of the crown compounds has been ac-complished by a number of different cycloaddition meth-ods. The basic 2 unit plus 2 unit addition as shown inEq. (2) has been most used for the simple crowns. Often

part of the macrocycle is first synthesized and then a sec-ond part added on, as in Eq. (3). This is particularly usefulwhen preparing unsymmetrical crowns such as dibenzo-21-crown-7. The last part of the synthesis shown in Eq. (3)is a simple 1 unit to 1 unit cycloaddition. This processhas been used to prepare other types of macrocyclic com-pounds such as pyridino diester-18-crown-6 (compound12, Fig. 1) [Eq. (4)]. This is a transesterification reactionand gives excellent yields. Indeed, most of these cycload-dition reactions exhibit a template effect in that greateryields are realized when a cation, or in some cases, aneutral molecules, that fits into the macrocyclic cavity isused in the reaction.

(3)

P1: GNH Final Pages



(4)

Compound 12 has shown chiral recognition for theenantiomers of various organic ammonium salts in thatenantiomerically pure 12 forms a stronger complex withone enantiomer of a chiral organic ammonium salt thanwith the other enantiomer. Chiral crown 15 (Fig. 1) syn-thesized in a similar manner has been used to effect anNADH type of asymmetric reduction of certain aromaticketones.

There are a great number of ways to prepare the aza-crown ethers which include the cyclic polyamines suchas cyclam (compound 7, Fig. 1). Some of these synthesesinclude the cycloaddition reaction of a diacid dichlorideand a diamine followed by reduction of the macrocyclic di-amide and treatment of a dihalide with a bistosylamide fol-lowed by removal of the N -tosyl groups. A more recent in-novation to prepare a cyclam is the treatment of a diaminewith a bis(α-chloroamide) as shown in Eq. (5) for thepreparation of trimethylcyclam. The bis(α-chloroamide)is readily prepared from a diamine and chloroacetylchoride or chloroacetic anhydride. The two amide nitro-gen atoms can contain hydrogens since the amide nitrogencannot act as the nucleophile. Cryptand 2.2.2 (compound4, Fig. 1) was initially prepared by Lehn and co-workersin a multistep process. A recent advance in the synthe-sis of macrobicyclic compounds is the one-step reactionof the appropriate bisprimary amine with two moles of aditosylate as shown in Eq. (6). The ditosylate is the key tothis one-step synthesis which results in the preparation ofthe cryptands in relatively high yields.

(5)

(6)

Synthetic supramolecular chemistry is currently receiv-ing considerable attention. The creation of multicom-ponent macrocyclic architectures utilizing noncovalentbonding interactions and the synthesis of discrete molec-ular entities interlocked or intertwined by covalent bondsand mechanical associations assisted by intermolecu-lar, noncovalent interactions is driven by the desire tounderstand and engineer well-defined, self-assembledstructures with desirable properties. In the beginning,these supramolecular compounds were prepared in lowyield via “statistical” methods. Advances in syntheticmethodologies that take advantage of complexation andnoncovalent association properties make the synthesisof these molecules more available. Self-assembly as asynthetic tool shows promise for supramacrocyclic com-pounds. A single representation of supramolecular syn-thesis is given in Eq. (7). The first step is the associationof the organic ammonium cation with the crown ether ina process called threading. Then large bulky groups arelinked to each end which prevents dethreading of the or-ganic ammonium ion.

(7)

Thirty years after Pedersen’s discovery of crown ethers,the host–guest concept has firmly established itself inthe realm of synthetic chemistry, as evidenced by theexpansion of the macrocyclic into the self-assembledsupramolecular synthesis.

P1: GNH Final Pages




INCLUSION (CLATHRATE) COMPOUNDS • LIGAND FIELD

CONCEPT • MEMBRANES, SYNTHETIC (CHEMISTRY)

BIBLIOGRAPHY

Amabilino, D. B., and Stoddart, J. F. (1995). Chem. Rev. 95, 2725–2828.Balzani, V., Gomez-Lopez, M., and Stoddart, J. F. (1998). Acc. Chem.

Res. 31, 405–414.Bradshaw, J. S., Krakowiak, K. E., and Izatt, R. M. (1993). “Aza-Crown

Macorcycles,” Wiley, New York.

Bradshaw, J. S., and Stott, P. E. (1980). Tetrahedron 36, 461–510.Dietrick, B. (1996). “Cryptands,” in “Comprehensive Supramolecular

Chemistry,” Vol. 1 (Gokel, G. W., ed.), Pergamon, Oxford, New York,Tokyo.

Izatt, R. M., Pawlak, K., Bradshaw, J. S., and Bruening, R. L. (1995).Chem. Rev. 95, 2529–2586.

Lehn, J.-M., Atwood, J. L., Davies, J. E. D., MacNicol, D. D., andVogtle, F. (eds.) (1996). “Comprehensive Supramolecular Chemistry,”Volumes 1–10, Pergamon, Oxford, New York, Tokyo.

Lindoy, L. F. (1989). “The Chemistry of Macrocyclic Ligand Com-plexes,” Cambridge Univ. Press, Cambridge.

Reinhoudt, D. N. (ed.) (1999). “Supramolecular Materials and Technol-ogy,” Wiley, New York, Seoul, Tokyo.

Zhang, X. X., Bradshaw, J. S., and Izatt, R. M. (1997). Chem. Rev. 97,3313–3361.

P1: GJC/GLE P2: GNB Final Pages

Encyclopedia of Physical Science and Technology EN011A-544 July 25, 2001 18:30

Organometallic ChemistryRobert H. CrabtreeYale University

I. Main Group Organometallic CompoundsII. Transition Metal Organometallic Compounds

GLOSSARY

Back donation Donation by the metal of one or more ofits electron pairs into the ligand. Since an empty ligandorbital is required to allow this to happen, the ligandmust usually be unsaturated, that is, have double ortriple bonds.

Main group metals Metals from groups 1, 2, and 11–15of Mendeleev’s periodic table, such as Li, Mg, Al, Pb.

Organometallic compound A compound containing ametal-carbon bond.

Transition metals Metals from groups 3–10 ofMendeleev’s periodic table, such as Ti, Fe, W, Pt.

ORGANOMETALLIC CHEMISTRY is the study ofsubstances that contain an organic compound or frag-ment bound to a metal atom or ion by a metal-carbon(M–C) bond. By tradition, compounds containing a metal-hydrogen (M–H) bond are also included in organometallicchemistry. In certain non-English-speaking nations, suchas Russia, the scope of the subject is extended beyond themetals and is called “organo-element chemistry.”

Organometallic compounds behave somewhat differ-ently depending on the type of metal involved. The maingroup metals are groups 1, 2, and 11 12, and the heavierelements of 13–15, and the transition metals are groups3–10. The transition metals can bind a wider range of

ligands and have more types of reactions available to themthan do the main group metals.

I. MAIN GROUP ORGANOMETALLICCOMPOUNDS

A. With Metal Carbon Bonds

We will look at the main group cases first, the most im-portant organometallic derivatives of which are the alkyls,such as LiMe, MeMgBr, PbEt4, and similar species (Me =methyl, CH3; Et = ethyl, C2H5). They are usually preparedeither by treating an alkyl halide with the metal, or by thereaction of a metal halide with an alkylating agent, oftenan organometallic compound such as LiMe.

MeI + Li = LiMe + LiI (1)

LiMe + CuCl = LiCuMe2 + LiCl (2)

The most electropositive metals, such as Na and K, tendto form ionic organometallic species, and therefore thebest known are ones in which the anion of the organicfragment is unusually stable because the negative chargeis delocalized over many atoms of the organic anion. Forexample, Na[Ph2CO] contains the [Ph2CO]− radical an-ion; the unpaired electron carrying the negative charge isdelocalized over the whole Ph2CO molecule. There is noM C bond per se, because the compound is ionic. These

529



530 Organometallic Chemistry

species are good reducing agents; that is, they readily addan electron to a suitable acceptor molecule. They are veryunstable to moisture and air.

Slightly less electropositive metals such as Li, Mg, Aldo form covalent organometallic compounds. Many ofthese alkyls are very reactive, depending on the electroneg-ativity of the metal involved, and are often rapidly decom-posed by air and water. Aluminum compounds such asAl2Me6 burst into flames in air. The electropositive met-als, such as Li, Mg, Zn, and Al, form alkyls that are ex-ceptionally useful in synthetic organic chemistry becausethey introduce an R− group into an organic compound, asshown in Scheme 1.

C O

Me

Me

RLiC O

Me

Me

R LiH2O

C OH

Me

Me

R

Scheme 1.

This reaction is strongly promoted by the binding of themetal to the oxygen end of the C O bond, which polarizesit and so prepares this bond for R− attack at the carbon.

The electronegative metals form much more stable andless reactive alkyls, for example, HgMe2. These are alsouseful in organic chemistry, because the R group in anR2Hg or RHgX (X = halide) compound can have any ofa wide variety of functional groups, such as COOMe,

CONH2, organic carbonyl, OH, etc. This is not truefor RLi, because the C Li bond reacts and is thereforeincompatible with these groups.

Some of the group 1, 2, and 13 alkyls are dimericor polymeric. The reason is that these elements havefewer than four electrons and so cannot reach the sta-ble octet of electrons by forming alkyls MRn , where nis the valency of the metal. Bridging allows the metal toreach the stable 8e configuration, as shown for Al2Me6 inScheme 2. Main group organometallics tend to be tetrahe-dral, 4-coordinate, as shown for Al2Me6, but the heaviermetals (At. No. > 10) can also attain higher coordinationnumbers.

Al

Me

Me

Me

MeAl

Me

Me

Scheme 2.

Some of the later element organometallics have spe-cial applications. Organocopper compounds, for exam-ple, have a strong tendency to add in the unusual 1,4-fashion to α,β-unsaturated ketones, rather than in the

usual 1,2-fashion (to give CH2 2CMe(OH)R) as wouldLiMe.

CH2 CH CORi)LiCuMe2, ii)H2O−−−−−−−−−→ MeCH2CH2COR (3)

The interest in new ways of making semiconductors,such as GaAs and InP, has led to the development ofMOVD (metal-organic vapor deposition). In this appli-cation, two organometallic compounds are co-pyrolyzedin a ratio suitable for the deposition of the desired semi-conductor. This low-temperature synthesis allows layersof semiconductors or insulators to be deposited in a con-trolled fashion and is being used for the fabrication ofcomplex electronic components.

AsH3 + GaMe3 = GaAs + 4MeH (4)

Tetraethyl-lead, PbEt4, was used in many countries as anantiknock agent in gasoline, but concerns about potentialenvironmental problems led to its abandonment in favorof using a more highly branched chain and aromatics-richgasoline.

Methyl mercury cation, HgMe+, is a water-soluble andunusually toxic form of this heavy metal. It can be formedfrom Hg2+ by bacterial action and was the causative agentin Minimata disease. Japanese families were poisoned be-cause they were eating seafood caught in a bay wheremercury wastes were being dumped. Mercury is normallypresent in the environment, although at much lower lev-els than those reached at Minimata. Certain bacteria havedeveloped sophisticated enzymatic pathways to detoxifymercury compounds by reduction to metallic mercury,which is far less dangerous and will eventually evaporatefrom the cell. Higher animals, including humans, may alsohave proteins that protect from mercury and heavy-metaltoxicity in general. The metallothioneins have been pro-posed to fill this role by binding heavy metals, but theirfunction in vivo is still a matter of dispute.

B. With Metal Hydrogen Bonds

Main group hydrides are numerous. Once again, there is agradation of properties across the periodic table, from thevery reactive ionic compound KH to the modestly reac-tive, covalent SiH4. The early hydrides, such as KH itself,are useful as nonnucleophilic bases, which only abstracta proton and do not add H− to an organic compound.The later ones, especially in the anionic “-ate” complexform, are nucleophilic, for example, NaBH4 and LiAlH4.These contain the MH−

4 anion, which achieves its 8econfiguration without bridging [M = 3e, H = 1e, anioniccharge = 1e].

The neutral hydrides can have very interesting bridgedstructures. The boranes, Bx Hy , are the most extreme cases.



Organometallic Chemistry 531

Hundreds of H-bridged species of this type are known, butthe simplest is B2H6 (Scheme 3).

BH

HB

H

H H

H

Scheme 3.

Diborane is one of the most useful main group organ-ometallic species, because it adds to alkenes in the unusualanti-Markownikov direction to give organoboron deriva-tives that can be converted into any of a number of usefulorganic compounds.

RCH CH2 + 1/2B2H6 = RCH2 CH2BH2 (5)

RCH2 CH2BH2 + H2O2 = RCH2 CH2OH (6)

RCH2 CH2BH2 + Br2 = RCH2 CH2Br (7)

Almost all other types of reagents that add to alkenes doso in the opposite direction to give the commercially lessuseful branched derivatives.

A development of interest in the area of main grouporganometallics was the isolation of compounds with mul-tiple bonds between elements such as P and Si, e.g.,RP PR or R2Si SiR2. Such heavy elements had beenthought to be incapable of double bonding, and attemptsto make compounds like these had always led to the forma-tion of polymers such as (PR)n . The solution turned out tobe the use of very bulky R groups, which prevented poly-merization. The resulting double bonds prove to have elec-tronic structures interestingly different from those presentin such long-known compounds as R2C CR2.

II. TRANSITION METALORGANOMETALLIC COMPOUNDS

A. With Metal Carbon Bonds

The transition metals, the elements from groups 3–10 ofthe periodic table, have been more intensively studied inrecent years than the main group elements. Transition-metal organometallic chemistry grew out of coordinationchemistry, of which it is a subfield. Coordination chem-istry was founded in its modern form by Alfred Wernerfrom 1896 and deals with compounds of metals and metalions with ligands (symbolized in a general fashion as L).From the point of view of coordination chemistry, it isimmaterial whether or not these ligands are bound by anM C (or M H) bond, and indeed the majority of co-ordination compounds have M N or M O bonds to theligands. Metal ions are Lewis acids; that is, they can acceptone or more pairs of electrons from one or more ligands L

to form a complex or coordination compound [MLn]m+,such as [Co(NH3)6]3+. The square brackets denote thatthe complex molecule (if m = 0) or ion (if m �= 0) is an en-tity that retains its identity and can be regarded as a singlemolecule or ion. The M L bonds are relatively strong. Akey feature of coordination chemistry is that the presenceof the ligands modify the properties of the central atomor ion, and the presence of the atom or ion modifies theproperties of the ligand. The extent of these mutual modi-fications of metal and ligand can vary from minor to veryprofound. When we consider organometallic complexes,we find that this same mutual effect of metal on ligandsand vice versa is also a very marked feature of the chem-istry. One additional general property of complexes bestillustrated by organometallic compounds is the possibil-ity that ligands, usually of chemically different types, canreact with one another within the coordination sphere ofthe metal (i.e., while still attached to the metal) in waysthat are not observed for the free ligands L in the absenceof a metal.

The commonest structural arrangements of the ligandsaround the metal in organometallic complexes in octahe-dral (1), square planar (2) or tetrahedral (3). Less regulararrangements can often be described as distortions of 1–3.

M

L

L LL

L

M

LL

M

L

L

1 2 3

L L L

L L

A ligand such as Cl− or H2NCH2CH2NH2 has morethan one lone pair of electrons (four on Cl−, one on eachof the two N atoms). In such a situation, two M L bondscan be formed. Chloride ion commonly bridges two met-als (4), while H2NCH2CH2NH2 more commonly chelatesto a metal to form a ring (5) (LnM refers to a metal and un-specified ligands). Chelate complexes are often more sta-ble (thermodynamically and kinetically) than analogousspecies with monodentate ligands (i.e., ligands that haveonly one point of attachment to a metal).

LnM MLn

Cl

LnM

NH2

NH2

CH2

CH2

4 5

Cl

Complexes that differ in the arrangement of the ligandsaround the central metal are different compounds anddiffer in properties. For example, cis-[PtCl2(NH3)2] (6) isan active anti-tumor drug, but trans-[PtCl2(NH3)2] (7) isnot. Both 6 and 7 are square planar complexes, but they




are shown from the top in these pictures not from the sideas in 2.

Pt

Cl NH3

Cl NH3

Pt

Cl NH3

H3N Cl

6 7

The simplest carbon ligand is the methyl group, CH3.Almost all metals form methyl complexes MMen , wheren is the valency of the metal, which varies in a regular wayacross the periodic table (e.g., Na: n = 1. Mg: n = 2.). TheM CH3 bond consists of a single shared pair of electronsand is not very different from the case of M NH3. Sincecarbon is more electronegative than the early metals (i.e.,the ones on the left-hand side of the periodic table), themethyl group bears a negative charge in the methyl com-plexes of these elements (e.g., LiMe, mentioned earlier).

The next simplest carbon-based ligand is carbon mono-xide, CO. Although superficially similar to the M Mesystem in having a metal-carbon bond, M CO is a verydifferent type of bond. This molecule has a lone pairof electrons on the carbon atom, which can be sharedwith a metal to give a metal carbonyl complex, such asFe(CO)5. A new feature of the bonding, which differsfrom the situation in NH3 or CH3, is that the CO is anunsaturated compound (i.e., has double or triple bonds).In such a case, there is always the possibility of an ad-ditional interaction not present to any significant extentin complexes of NH3 or the methyl group. The metalmay donate some of its electron pairs to the ligand (anempty orbital allowing this to happen is almost alwayspresent in an unsaturated ligand); this is called back do-nation and accounts in part for the large modificationboth in metal and ligand properties on binding. For CO,the back bonding has the effect of weakening the C Obond (this can be detected by infrared spectroscopy) andpreventing the M L bond from having the M−−L+ po-larity of M NH3 (because for CO the ligand-to-metalelectron donation is balanced by metal-to-ligand backbonding).

So important is this back donation in stabilizing theM CO bond that metals which do not have any elec-trons to back donate form either no CO complexes orform very unstable ones. A metal will only bind COstrongly if it has d-electrons and therefore comes fromgroups 2–11 ( = N ) of the periodic table and has a va-lency state ( = v), such that N − v ≥ 1. Only unsaturatedligands (said to be soft ligands) have the power to ac-cept back donation. Metals are able to engage in backbonding if they have a low valency state (e.g., v ≤ 2). Soft

metals include Pt(II), Mo(0), Fe(0), Ir(I), Hg(II), Cu(I),and Ag(I). Soft ligands include CO, ethylene, benzene,allyl, cyclopentadienyl.

A remarkable feature of the zerovalent metal carbonylsis that there is a regular change in their formulas andtherefore in their constitution as we move from left to rightacross the periodic table: Cr(CO)6, (CO)5Mn Mn(CO)5,Fe(CO)5, (CO)3Co(CO)2Co(CO)3, Ni(CO)4. Alternatecompounds are mononuclear (i.e., contain one metal)but have one less CO for every two steps to the right.This regularity is embodied in the Eighteen Electron rule,which states that special stability often accompanies avalence electron count of 18 electrons per metal (and insome cases, such as Pd(II), 16 electrons). To take thecase of Cr(CO)6, the valence electron count is obtainedas follows: Cr is in group 6 of the periodic table, so has6e; CO donates the 2e of its C lone pair to the metaland so:

6 + (2 × 6) = 18 (8)

In the general case of a complex of formula [MXaLb]c+

(where M is a metal of group N , X is an anionic ligandsuch as Cl or CH3, and L is a neutral ligand, such as NH3

or CO), the count is given by Eq. (9)

e count = N + a + 2b − c (9)

Useful concepts are the oxidation state (O.S.), given by Eq.(10), and the dn configuration, or number of d-electrons,given by Eq. (11).

O.S. = c + a (10)

n = N − c + a (11)

The (II) (i.e., + 2) of Pd(II) is the oxidation state; Pd(II)is said to have the d8 configuration.

In metal carbonyls, the CO removes metal electrondensity so efficiently that the lower oxidation states ofthe metal, which have more electrons, are strongly stabi-lized. Methyl complexes and carbonyls tend to have verydifferent oxidation states for this reason, such as WMe6

(O.S. = 6) and W(CO)6(O.S. = 0).Early metal alkyls tend to be reactive and nucleophilic

(i.e., tend to transfer an alkyl anion to a reagent), but thelate metals, such as Pt, Ir, Hg, can form very stable alkyls;the very high stability of the water-soluble [HgCH3]+ con-stitutes one of the major hazards of mercury pollution,because this species, formed by microbial action, can en-ter the food chain. Metal alkyls that have a β-H (i.e., ahydrogen at the second carbon from the metal) can un-dergo the β-elimination reaction provided they are coor-dinatively unsaturated (i.e., have an electron count of lessthan 18) or, if coordinatively saturated (18e count), canlose a ligand so as to generate the necessary unsaturation.




The reaction is much faster if the metal has a dn config-uration between 2 and 10, because d-electrons stabilizethe transition state (or most unstable species along thereaction pathway; the more stable this is, the faster thereaction).

[(Indenyl)L2IrCH2CH3] is an 18e d6 species, but theindenyl can slip sideways on illumination so as to dis-engage 2e and generate unsaturation. As expected on theprinciples discussed above, illumination of [(Indenyl)-L2IrCH2CH3] with ultraviolet light leads to β-elimination,in which the β-H is transferred to the metal and (in thiscase) CH2 CH2 is released (Scheme 4).

IrL2Et�

�IrL2H

�

(L � PPh3)18e

18e 16e

IrL2Et�

hν

CH2 CH2

Scheme 4.

We next examine unsaturated ligands, such as ethylene,C2H4. These are very soft ligands and only bind well tod2–d10 metals. Ziese made the first one in 1837, but thestructure (8) was only established in the 1950s.

Pt CH2

CH2

Cl

Cl

Cl

8

�

The ethylene acts as a neutral, 2e donor via its C Cπ -bonding electrons, back donation (into the π∗ levels)stabilizes the complex. Two extreme situations can be dis-tinguished. If the back bonding is very restricted in itsextent, the ethylene acts essentially only as a donor, andthe ligand becomes depleted of electrons (9a, the Chatt–Dewar model). In such a situation the ligand can be at-tacked at the carbon atom by nucleophiles [e.g., Eq. (3)],a reaction that does not occur in the case of free ethylene.This is an example of the metal exerting a modifying effecton the reactivity of the ligand. This type of effect makesorganometallic chemistry very useful in organic synthesis,the art of constructing organic molecules. A metal com-plex can be erected like a temporary scaffold around an

organic compound so as to carry out a reaction that wouldnot otherwise take place.

LCl2Pt(C2H4) + Me2NH

= [LCl2PtC2H4NMe2]− + H+ (12)

In the opposite bonding extreme, back bonding is veryimportant and the resulting complex can be considered ashaving two M C single bonds, as shown in the metalacy-clopropane model 9b. Alkynes such as acetylene (HCCH)bind to metals in a similar way.

MCH2

CH2

CH2

CH2

M

9a 9b

The allyl group is an interesting ligand because it caneither bind as a monodentate ligand (10a), rather like themethyl group, or it can bind via this bond and via theimmediately adjacent C C group as well (10b). If it bindsin the first way it is denoted η1-allyl; if in the second, η3-allyl. In each case, the superscript indicates the number ofligand atoms bound to the metal, otherwise known as thehapticity. Variable hapticity is shown by a large numberof organometallic ligands (e.g., Scheme 4).

M

10a 10bM

1,3-Dienes normally bind in an η4-form, 11a; both C Cbonds are effectively bound to the metal. When this hap-pens an electronic redistribution takes place in the diene,so that the C2 C3 bond of the complexes diene becomesshorter than the C1 C2 and C3 C4 bonds, as illustratedby the resonance form 11b; the C2 C3 bond is longerthan the C1 C2 and C3 C4 bonds in the free ligand. Thisis another example of metal-induced change in a ligand.In this case the reason is that back bonding populates amolecular orbital, which leads to greater C2 C3 bondingin the complex than in the free diene.

11aM M

11b

Cyclopentadienyl, C5H5 ( Cp) is one of the most cele-brated ligands in organometallic chemistry. Unlike the un-saturated ligands mentioned up to now, it does not requireback donation to bind strongly, and it forms a wide varietyof complexes spanning the whole periodic table. η1- (12a),




η3- (12b), and η5- (12c) structures are found, but the latterare by far the most numerous, as in ferrocene, 12d.

12a

M H

12b

M

12c

M

Fe

12d

The great advantage of this ligand is that it is unaffectedby the usual reagents and so can be used as a spectatorligand to stabilize a metal fragment while other chemistrytakes place. For example, the fragment CpFe(CO)2, ( Fp)binds alkyl groups very efficiently, and it is used for thispurpose in organic synthesis.

Arenes such as benzene normally bind in an η6-mode,13, but only to soft metals; even then the arene dissociateseasily. This difference with respect to Cp might at firstappear puzzling, but the reason is that Cp is a poor leavinggroup; because Cp− or Cp• are relatively unstable, an arenecan dissociate as the free arene, a very stable entity.

M

13

The preparation and characterization of organometal-lic compounds uses methods similar to those employed inorganic chemistry. Notable in organometallic chemistryis the more common use of X-ray crystallography. Alsonotable are the strong spectral features due to metal car-bonyls in the infrared spectrum and metal hydrides in theproton nuclear-magnetic-resonance (NMR) spectrum.

B. With Metal Hydrogen Bonds

Metal hydrides play an important role in organometal-lic chemistry. They add to a variety of unsaturated or-ganic compounds by a reaction that is the reverse of theβ-elimination step of Eq. (6) to give species with M Cbonds. Hieber made the first example of a transition-metal

hydride in 1931, but his claim was not generally accepteduntil around 1960. Hydrides are now very easy to de-tect, thanks to the appearance of a characteristic peakin the nuclear magnetic resonance spectrum. Hydridescan be terminal, as in (CO)5Mn H, or bridging, as in(CO)5Cr H Cr(CO)5. H2 can also be bound as a neutral2e ligand to a metal, as in [(CO)5Mo (H2)], in which theH H bond of the hydrogen molecule is retained.

C. Typical Reaction Types

The simplest reaction of an organometallic compound isa substitution in which one ligand replaces another at ametal center. A very common example is the replacementof a CO by a phospine, PMe3 or PPh3. Phosphines are veryuseful ligands because their bulk and electronic propertiescan be varied over a wide range in a controlled way; thisallows us to explore how these affect the types of reactionswe can bring about. In an 18e complex, such as Mo(CO)6,a CO usually has to dissociate to give the 16e reactiveintermediate {Mo(CO)5 } before a new ligand can bind, asshown in Eq. (13), a dissociative substitution.

Mo(CO)6 + PMe3 = {Mo(CO)5 } + PMe3 + CO

= [Mo(CO)5(PMe3)] + CO (13)

On the other hand, a 16e complex will usually give an asso-ciative substitution, in which the incoming ligand attacksto give an 18e intermediate, as shown in Eq. (14).

IrCl(PMe3)3 + CO = {IrCl(CO)(PMe3)3 }= [IrCl(CO)(PMe3)3] + PMe3 (14)

Illumination sometimes causes a ligand to dissociate,and under these circumstances a photochemical substitu-tion is often observed; carbonyls are especially prone togive this reaction.

A variety of compounds with reactive X Y bondscan give the oxidative-addition reaction with any of alarge number of reduced-metal species, as illustrated inEqs. (15–17).

PPh3 + Cl2 = Cl2PPh3 (15)

IrCl(PPh3)4 + H2 = IrH2Cl(PPh3)4 (16)

Pd(PCy3)2 + MeI = MePdI(PCy3)2 (17)

This reaction is a very versatile way of making M Lbonds. In spite of the similarity of the overall transfor-mations, oxidative addition is a mechanistically diversereaction class. In the reverse process, reductive elimina-tion, an X Y molecule is formed from a metal complexLnM(X)(Y).

A related process, oxidative coupling, is illustrated inScheme 5.




Fe(CO)5C2F4

(OC)4FeCF2

CF2CF2

CF2

Scheme 5.

Insertion and elimination are mutually inverse reac-tions, in which a ligand, shown as AB, inserts into anM X bond, illustrated in Schemes 6 and 7. Ligands thatbind end-on usually give 1,1-insertions in which both Mand X are attached to the same atom in the product. Lig-ands that bind side-on usually give 1,2-insertions. Somespecific examples are shown: Scheme 8 is a 1,1 insertionto give an acetyl complex; Scheme 9 is a 1,2-insertion togive a vinyl complex; and Scheme 10 is a 1,2-insertion togive a sulfinate complex. β-Elimination, mentioned ear-lier, can now be seen to be a 1,2-deinsertion. It is notablethat although β-elimination of an H is very rapid, thereare very few examples of β-elimination of an alkyl group.

X

M A BM A

X

B

Scheme 6.

X

MA

B

M A

B X

Scheme 7.

CO

Mn

CO

OC CH3

OC

COCO*

CO

Mn

CO

OC C

OC

*O

CH3

CO

Scheme 8.

PtCl HL

LCF3 CF3

PtClL

H

CF3

CF3L

Scheme 9.

Cp(CO)2Fe MeSO2

Cp(CO)2Fe OS

Me

O

Scheme 10.

Nucleophilic addition to alkenes has been mentioned(Section I.B), but similar additions occur for otherorganometallic ligands. Especially facile are reactionswith allyl, diene, and arene groups. For example, the

arene complex (C6H6)Cr(CO)3 reacts with MeLi asshown in Scheme 11.

RLi

Cr(CO)3

R H

Cr(CO)3

Scheme 11.

An alternative way that a nucleophile can react is toabstract an X+ fragment, as illustrated for H+ abstractionin Eq. (18).

Cp2TaMe+2 + R3P CH2

= Cp2Ta( CH2)Me + R3P CH+3 (18)

An electrophile, in contrast, tends to add to uncoordi-nated parts of an unsaturated ligand (electrophilic addi-tion) or abstract an X− fragment (electrophilic abstrac-tion). Equation (19) shows protonation of an η1-allyl togive an η2-alkene cation, and Eq. (20) shows the protona-tion of a hydroxyethyl complex.

Fp CH2 CH2 CH2 + H+ = [Fp(CH2 CH2 CH3)]+

(19)

Fp CH2 CH2 OH + H+ = [Fp(CH2 CH2)]+ + H2O

(20)

Nucleophilic addition to coordinated CO is an importantway of making carbenes, as we will see later.

D. Applications in Catalysis

Homogeneous catalysis is an important commercial ap-plication of organometallic chemistry; many millions ofpounds of a wide variety of organic compounds are madein this way each year. The object is to find an organometal-lic complex which, when present in small, even trace,amounts, will bring about the conversion of a given reagentinto a desired product. To do this, the metal has to coordi-nate to the substrate (i.e., organic reagent) or substrates andactivate them for the desired reaction. In this area of chem-istry, the intention is to mimic certain of the properties ofenzymes, but in contrast to many enzymes, organometal-lic catalysts tend to be robust and accept a wide range ofsubstrates. Catalysis is an example of “green,” or envi-ronmentally conscious, chemistry, because it reduces theamount of waste products formed.

Alkene hydrogenation is an area in which several dif-ferent catalysts have been shown to be useful for differentpurposes. For example, RhCl(PPh3)3 is selective for un-hindered C C double bonds and causes very little isomer-ization in the substrate. [Ir(cod)(PCy3)py]+, on the otherhand, is very active for very hindered C C groups, and if




a binding group such as OH is present in the alkene, thecatalyst will bind to this group and then add the H2 to theC C bond. This alters the stereochemistry of the productin a useful way. A simplified mechanistic scheme for thereaction is shown for ethylene and H2 in Eqs. (21–24). Anoxidative addition of H2 leads to the dihydride. The ethy-lene then binds and inserts into one M H bond to give themetal ethyl. The ethyl group then reductively eliminateswith the remaining H ligand to give ethane. The final re-ductive elimination is the reverse of the initial oxidativeaddition, but involves a C H rather than an H H bond.Note that the catalyst, symbolized by LnM, is regeneratedin the final step so that it can react with a second andsubsequent molecules of the reagents. Many thousands ormillions of catalytic turnovers can be carried out, whichmeans that very little catalyst needs to be present.

LnM + H2 = LnMH2 (21)

LnMH2 + H2C CH2 = LnMH2(H2C CH2) (22)

LnMH2(H2C CH2) = LnMH(CH2 CH3) (23)

LnMH(CH2 CH3) = LnM + H3C CH3 (24)

Hydroformylation, Eqs. (25) and (26), is useful for thepreparation of aldehydes and, by hydrogenation of thealdehydes in situ, alcohols.

RCH CH2 + CO + H2 = RCH2 CH2 CHO (25)

RCH2 CH2 CHO + H2 = RCH2 CH2 CH2OH(26)

The mechanism is similar to that for hydrogenation, but atthe alkyl hydride stage a carbonyl insertion reaction takesplace to lead to the sequence shown in Eqs. (27)–(29).

LnMH(CH2 CH3) + CO

= L(n −1)M(CO)H(CH2 CH3) (27)

L(n −1)M(CO)H(CH2 CH3)

= L(n −1)MH(COCH2 CH3) (28)

L(n −1)MH(COCH2 CH3)

= L(n −1)M + HCOCH2 CH3) (29)

Some hydroformylation catalysts are also alkene iso-merization catalysts; that is to say they move the positionof the C C double bond along a chain, as shown in Eq.(30). This can happen much faster than hydroformylationitself. Advantage can be taken of this effect if one of theisomers reacts faster in the hydroformylation sequencethan the others. In the case shown in Eqs. (30) and (31)the terminal alkene reacts faster and forms the linearaldehyde shown. Of three possible aldehydes, only oneis formed in significant quantities if the right catalyst isused. This high reaction selectivity is a useful propertyof homogeneous catalysts.

CH3CH CHCH3 = CH3CH2CH CH2 (30)

CH3CH2CH CH2 + H2 + CO

= CH3CH2CH2CH2CHO (31)

Alkene hydrocyanation, the addition of HCN across analkene, is useful in the industrial preparation of adiponi-trile [Eq. (32)], the key intermediate in the manufacture ofnylon.

CH2 CHCH CH2 + 2HCN

= NC CH2CH2CH2CH2CN (32)

Hydrosilation, the addition of a silane, R3Si H acrossan alkene is useful in the industrial preparation of siliconepolymers and materials.

RCH CH2 + R3Si H = RCH2CH2SiR3 (33)

Polymerization of alkenes is an important area in whichhomogeneous catalysts, in the form of Ziegler–Natta andmetallocene systems, have been very important. These cat-alysts are formed from an early metal compound, such asCp2ZrCl2 or TiCl4, an aluminum alkyl, such as Me2AlCl,and a trace of water. They are believed to operate by suc-cessive insertion reactions, as shown in Eq. (34).

LnM(CH2CH3) + H2C CH2 = LnM(CH2CH2CH2CH3)(34)

Metal-carbon multiple bonds are found in a numberof situations. The simplest is illustrated in Scheme 12,which shows the synthesis of a carbene complex. Whena heteroatom is present, α to the carbene carbon, theM C bond is not a full double bond because both res-onance structures (Scheme 12) are thought to contribute.Carbenes without heteroatom substituents have been pre-pared, and these seem to have a true double bond. Oneexample is Cp2Ta( CH2)Me, formed by deprotonation ofthe Cp2TaMe+

2 cation.

(CO)5M CO (CO)5M C

Me

OLi

(CO)5M C

Me

OMe

(CO)5M C

Me

OMe�

�

MeLi � MeI

Scheme 12.

M C double bonded intermediates are believed to beimportant in alkene metathesis Eq. (35), a useful com-mercial process that has been applied to organic synthe-sis, polymerization, and to changing the molecular weightdistribution of a mixture of alkenes (Eq. 35).

RCH CH2 = RCH CHR + CH2 CH2 (35)




The key cyclic intermediate in this process is shown inScheme 13.

M CH2M

R

R R

M

R RR

M

R

R

Scheme 13.

The Shell Higher Olefins Process (SHOP) is an ex-ample of a large-scale commercial process that relies ona metathesis step. In SHOP, ethylene is oligomerized toalkenes having a chain length of around C16 by a homo-geneous nickel catalyst. The chain lengths of the resultingmixture can usefully be manipulated via metathesis.

E. Compounds with Metal–Metal Bonds

Some compounds containing a metal–metal bond occur inorganometallic chemistry. M M bonds are important in“clusters,” such as Os3(CO)12, which contains a triangleof metal atoms. The CO ligands are shown only as linesradiating from the triosmium core in the diagram. Clusterscontain three or more metals with at least two M M bonds(see Scheme 14).

Os Os

Os

Scheme 14.

The second situation in which M M bonds are impor-tant is the case of dinuclear species. Multiple bonding ismore often found in dinuclear cases, for example Re2Cl2−

8(Scheme 15), but there are not as yet very many metal–metal multiply bonded organometallic compounds.

Re Re

ClCl

ClCl

ClCl

ClCl

2�

Scheme 15.

A number of useful electron counting rules, analogousto the 18e rule for mononuclear compounds, enables usto predict the structures to be expected among metal clus-ters. Nobelist Roald Hoffmann has developed a series ofanalogies between metal fragments and organic fragmentsthat is another useful structural and conceptual tool. Forexample, Os(CO)4 is said to be isolobal with CH2, and sothe triosmium cluster shown in Scheme 14 is isolobal withcyclopropane.

F. Organometallic Species in Biology

M C bonds are proving to be important in biology. Thebest-known case is vitamin B-12, one derivative of whichis shown in Scheme 16. The form shown is one of severalderivatives that contains a metal–carbon bond. In one othernaturally occurring organometallic derivatives of B-12, anadenosyl group can replace the methyl group. Lack ofB-12 leads to the fatal disease pernicious anemia. Thereason is that coenzyme B-12 is a cofactor in a number ofvital steps in human biochemistry, for example synthesisof deoxy-ribose for DNA, and the synthesis of the aminoacid methionine. B-12 contains cobalt, and its biochem-ical role involves making and breaking of Co C bonds.It has also been found that Ni C and Ni H bonds areimportant in bacterial biochemistry, where they allow cer-tain bacteria to live on H2 and CO as energy and carbonsource and to produce methane. Methane, or “marsh gas,”is a common natural product; almost all of it appears to beformed by a bio-organometallic route.

P

O�O

O

O

HN

N

OO

HON

N

CH2OH

CO

N

H2NOCN

CONH2

H2NOC

CH3

N

CONH2

CONH2

CONH2

Scheme 16.




G. Recent Advances and Current Problems

One of the areas of greatest current interest is the organ-ometallic chemistry of alkanes. Alkanes are abundant andcheap, yet methods for their conversion into useful deriva-tives are lacking. For example, methane is flared off inthe Sahara because of transport problems for this gaseousfuel. A way to combine it with air to give the easily trans-portable fuel methanol would be of value. It is hopedthat organometallic catalysts will be found that can trans-form methane into methanol. Although certain bacteriacan carry out this reaction, the chemist cannot yet do thesame thing efficiently in the laboratory.

CH4 + O2 = CH3OH (36)

It is not so much that CH4 is unreactive, because itwill easily burn in air to give CO2. Rather it is the in-termediates in the air oxidation that would be valuable ifthe oxidation could be stopped before the carbon diox-ide stage. The reason this is exceptionally difficult is thatthese intermediates are much more reactive than methaneitself.

Nature solves the problem by designing an enzyme thatbinds CH4 strongly but rejects CH3OH from the reactionsite. This is done by arranging for the active site to be verynonpolar, a circumstance favorable to the binding of thenonpolar methane but not to binding of methanol, a polarmolecule.

Another area of current interest is the attempt to pre-pare hydrides with H H bonds in them. H2 is the simplestmolecule, and so its binding to metals is of unusual inter-est. The resulting structure is shown in Scheme 17. Scoresof similar cases have now been discovered, and chemistsare interested to see whether larger assemblies of hydro-gen atoms can be stabilized by binding to a metal; forexample, could an H3 complex be made? One of the morechallenging features of this search is that it is exceedinglydifficult to distinguish between the well-known polyhy-drides LnMHx , where there are M H bonds only, and theinteresting “nonclassical” form containing an Hx ligand.

This is because the properties of the two might be very sim-ilar, and the method usually used to determine structures,X-ray crystallography, is very poor at determining H po-sitions, because X-rays interact poorly with light atomslike H. The much more arduous and difficult neutron-diffraction experiment can locate hydrogen atoms pre-cisely, because neutrons interact relatively strongly withall nuclei, whether of light or of heavy atoms. This tech-nique can only be used sparingly, however, because thereare only a very few neutron-diffraction facilities aroundthe world, and so suitable candidate molecules must belocated by other means.

H

H

M

Scheme 17.


CATALYSIS, HOMOGENEOUS • MAIN GROUP ELEMENTS

• METALORGANIC CHEMICAL VAPOR DEPOSITION • OR-GANIC CHEMISTRY, SYNTHESIS • POLYMERS, INORGANIC

AND ORGANOMETALLIC

BIBLIOGRAPHY

Albert, M. R., and Yates, J. T. (1987). “A Surface Scientist’s Guide toOrganometallic Chemistry,” American Chemistry Society, Washing-ton, D.C.

Crabtree, R. H. (1994). “Organometallic Chemistry of the TransitionElements,” Wiley, New York 2nd ed.

Jenkins, P. R. (1992). “Organometallic Reagents in Organic Synthesis,”Oxford, New York.

Schlosser, M. (1994). “Organometallics in Synthesis,” Wiley, New York.Wilkinson, G., ed. (1982–1994). “Comprehensive Organometallic

Chemistry, I and II,” Pergamon, Oxford.Yamamoto, A. (1986). “Organometallic Chemistry: Fundamental Con-

cepts and Applications” (Engl. Trans.), Wiley, New York.

P1: ZBU Final Pages Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology EN011J-141 July 31, 2001 15:14

Pharmaceuticals, ControlledRelease of

Giancarlo SantusRecordati Industria Chimica e Farmaceutica S.p.A.

Richard W. BakerMembrane Technology and Research, Inc.

I. Introduction/HistoryII. Methods of Achieving Controlled Release

III. Important Controlled Release ProductsIV. Future Directions

GLOSSARY

Biodegradable polymers Materials that undergo slowchemical degradation in the body, finally degrading tolow-molecular-weight fragments that dissolve or aremetabolized.

Controlled release systems Devices that meter deliveryof drugs or other active agents to the body at a ratepredetermined by the system design and are largelyunaffected by the surrounding biological environment.

Enteric coatings pH-Sensitive materials used to delaydissolution of tablets in the acid environment of thestomach but allow rapid dissolution when the tabletreaches the more neutral pH environment of the gas-trointestinal tract.

Microcapsules Small, drug-containing particles with di-ameters between 50 and 1000 µm.

Targeted drug delivery Delivery of a drug directly to thebody site where the drug has its biological effect.

Transdermal patches Drug-containing laminates at-tached to the skin with an adhesive. Drug containedin the laminate migrates from the patch through

the skin and is absorbed into the blood circulatorysystem.

Zero order delivery (of drug) Constant drug deliveryover a certain period of time.

THE OBJECTIVE OF CONTROLLED drug deliverydevices is to deliver a drug to the body at a rate predeter-mined by the design of the device and independent of thechanging environment of the body. In conventional medi-cations, only the total mass of drug delivered to a patient iscontrolled. In controlled drug delivery medications, boththe mass of drug and the rate at which the drug is deliveredis controlled. This additional level of control enhances thesafety and efficacy of many drugs. Often a membrane isused to moderate the rate of delivery of drug. For example,in some devices, a membrane controls permeation of thedrug from a reservoir to achieve the drug delivery rate re-quired. Other devices use the osmotic pressure producedby diffusion of water across a membrane to power minia-ture pumps. In yet other devices, the drug is impregnatedinto a polymer material, which then slowly dissolves or

791

P1: ZBU Final Pages


792 Pharmaceuticals, Controlled Release of

degrades in the body. Drug delivery is then controlled bya combination of diffusion and biodegradation.

I. INTRODUCTION/HISTORY

The pharmaceutical industry has produced long-actingoral medications since the 1950s. Enteric coated tabletswere the first long-acting medication to be widely used.In 1952, Smith Kline French (SKF) introduced Spansules,or “tiny time pills,” an over-the-counter cold medicationconsisting of millimeter-sized pills containing the drugand covered with a coating designed to delay its disso-lution. By varying the thickness of the coating, differentdrug release profiles were achieved. These early productsare best called sustained release systems, meaning that therelease of the drug, although slower and more controlledthan for a simple, fast-dissolving tablet, was still substan-tially affected by the external environment into which itwas released. In contrast, the release of drug from con-trolled release systems is controlled by the design of thesystem and is largely independent of external environmen-tal factors.

The founding of Alza Corporation by Alex Zaffaroniin the 1960s gave the development of controlled releasetechnology a decisive thrust. Alza was dedicated to de-veloping novel controlled release drug delivery systems.The products developed by Alza during the subsequent25 years stimulated the entire pharmaceutical industry.The first pharmaceutical product in which the drug regis-tration document specified both the total amount of drugin the device and the delivery rate was an Alza product,the Ocusert, launched in 1974. This device was designedto delivery the drug pilocarpine to control the eye dis-ease glaucoma. The device consisted of a thin, elliptical,three-layer laminate with the drug sandwiched betweentwo rate-controlling polymer membranes through whichthe drug slowly diffused. It was placed in the cul de sacof the eye, where it delivered the drug at a constant ratefor 7 days, after which it was removed and replaced. TheOcusert was a technical tour de force, although only a lim-ited marketing success. Alza later developed a number ofmore widely used products, including multilayer transder-mal patches designed to deliver drugs through the skin.The drugs included scopolamine (for motion sickness),nitroglycerin (for angina), estradiol (for hormone replace-ment), and nicotine (for control of smoking addiction).Many others have followed Alza’s success, and more than20 transdermal products, delivering a variety of drugs, arenow available. Alza also developed the first widely usedosmotic controlled release drug delivery systems under thetrade name Oros. The first billion-dollar controlled releaseproduct was an osmotic product, Procardia XL, delivering

the drug nifedipine, a widely prescribed antihypertensive.Other important products launched in the last 15 yearsinclude Prilosec, a diffusion-controlled microparticle oralformulation system for the drug omeprazole, and Lupron(leuprolide) and Zoladex (goserelin), polypeptide drugsdelivered from biodegradable intramuscular and subcuta-neous implants. Since the first controlled release productwas launched, the controlled release industry has grownto a $10–20 billion industry with more than 30 major con-trolled release products registered with the U.S. Food andDrug Administration. A time-line showing some of theimportant milestones in the growth of controlled releasetechnology is shown in Fig. 1. Development of the tech-nology has been reviewed in a number of monographs.

Controlled slow release of drugs to the body offers sev-eral important benefits to the patient.

• More constant drug blood levels. In controlled re-lease products, drug delivery to the body is metered slowlyover a long period, avoiding the problems of overdos-ing and underdosing associated with conventional med-ications. These problems are illustrated in Fig. 2, whichshows the blood levels achieved when a relatively rapidlymetabolized drug is given as a series of conventional fast-dissolving tablets. Immediately after the drug is ingested,blood levels begin to rise, reaching a peak value, afterwhich the concentration falls as drug is metabolized. InFig. 2 two important concentration levels are shown: theminimum effective concentration, below which the drugis ineffective and the toxic concentration, above which un-desirable side effects occur. Maintaining the concentrationof the drug between these two levels is critical for safetyand effectiveness. Controlled release systems meter deliv-ery of the drug to the body at a rate equal to the rate ofdrug removal by metabolization, so that a prolonged con-stant blood level is maintained. Because controlled releaseproducts use the drug more efficiently, the total amountrequired to produce the desired effect is often very muchreduced, frequently by a factor of 10 or more.

• Improved patient compliance. A second benefit ofcontrolled release formulations is improved patient com-pliance. Many studies have shown that few patients taketheir medications in complete accordance to physicianinstructions, and that the degree of noncompliance in-creases significantly as the complexity of the instructionsincreases. In one study of patients given once-per-day es-trogen/progesterone contraceptive tablets in a controlledfashion, the pregnancy rate was less than 1 per 1000patient-years. The same tablets prescribed to a normalpopulation of women resulted in a pregnancy rate of 50per 1000 patient-years due to noncompliance. Deliveryof the similar contraceptive steroid levonorgestrel from asustained release implant resulted in a pregnancy rate of

P1: ZBU Final Pages


Pharmaceuticals, Controlled Release of 793

FIGURE 1 Milestones in the development of controlled release pharmaceuticals.

0.5 per 1000 patient-years. This product has the lowestpregnancy rate, that is, the highest degree of contraceptiveefficiency, of any steroidal contraceptive because it elim-inates poor patient compliance, the principal reason fordrug failure.

• Targeted site of action delivery. A final benefit of con-trolled release formulation is the ability to deliver drugdirectly to the systemic circulation. Drugs taken orally areabsorbed into the blood stream in the gastrointestinal (GI)tract. The drug then enters the portal circulation, so is firsttaken to the liver before entering the systemic circulationand being transported to the site of drug action. The liver’sfunction is to deactivate toxic agents that enter the bodythrough the GI tract. Unfortunately, the liver often treatsdrugs as toxic agents and metabolizes a large fraction be-fore the drug reaches the blood circulatory system; this iscalled the first-pass effect. For example, in the first passthrough the liver, 70–90% of the hormone estradiol is lost.Therefore, when used for hormone replacement therapy, itmust be administered as tablets of 1–2 mg/day to achieveeffective systemic blood levels. When the same drug is

delivered directly to the systemic blood circulation from atransdermal patch, a delivery rate of only 50 µg of estra-diol/day is sufficient to achieve the required effect.

II. METHODS OF ACHIEVINGCONTROLLED RELEASE

A. Membrane Diffusion-Controlled Systems

In membrane diffusion-controlled systems, a drug isreleased from a device by permeation from its interiorreservoir to the surrounding medium. The rate of diffusionof the drug through the membrane governs its rate ofrelease. The reservoir device illustrated in Fig. 3 is thesimplest type of diffusion-controlled system. An inertmembrane encloses the drug, which diffuses through themembrane at a finite, controllable rate. If the concentrationof the material in equilibrium with the inner surface of theenclosing membrane is constant, then the concentrationgradient, that is, the driving force for diffusional releaseof the drug, is constant as well. This occurs when the inner

P1: ZBU Final Pages



FIGURE 2 Simplified blood level profile illustrating the differ-ence between repeated treatment with a conventional instant de-livery formulation and a single, controlled release, long-actingformulation.

reservoir contains a saturated solution of the drug, pro-viding a constant release rate for as long as excess solidis maintained in the solution. This is called zero-orderrelease. If, however, the active drug within the device isinitially present as an unsaturated solution, its concentra-tion falls as it is released. The release rate then declinesexponentially, producing a first-order release profile. Thedrug release profile for a simple membrane reservoirsystem is also shown in Fig. 3.

The drug release rate from a membrane reservoir devicedepends on the shape of the device, which may be a sim-ple laminate, a cylindrical device, or a spherical device.For the simple laminate, or sandwich geometry, the drugrelease rate can be written

FIGURE 3 Schematic and drug release profile of the simplest form of membrane diffusion-controlled drug deliverysystem.

d Mt

dt= AP�c

l, (1)

where d Mt/dt is the device release rate, A is the totalarea of the laminate, P is the membrane permeability, l isthe membrane thickness, and �c is the difference in drugconcentration between the solution at the inside surface ofthe membrane and the drug concentration in the solution atthe outer surface of the membrane, usually close to zero.When the solution inside the enclosure is saturated, thedrug concentration is cs , and Eq. (1) reduces to

d Mt

dt= APcs

l. (2)

Drug release is then constant as long as a saturated solu-tion is maintained within the enclosure. The total durationof constant release depends on the initial volume of theenclosure V , the mass of encapsulated drug M0, and thesolubility of the drug cs . The mass of agent that can bedelivered at a constant rate is M0 − csV . Thus, it followsthat the duration of constant release t∞ is

t∞ = M0 − csV

d Mt/dt. (3)

A second type of diffusion-controlled system is a mono-lithic or matrix device in which the drug is disperseduniformly throughout the rate-controlling polymericmedium. The drug release rate is then determined by itsloading in the matrix, the matrix material, the shape of thedevice (flat disk, cylinder, or sphere), and the permeabilityof drug in the matrix material. Equations describing re-lease from all the various device types and geometries canbe found elsewhere. As an example, desorption of druguniformly dissolved in a simple disk (slab)-shaped devicecan be expressed by either of the two series

Mt

M0= 4

(Dt

l2

)1/2[

π−1/2 + 2∞∑

n=0

(−1)n ierfc

(nl

2√

Dt

)]

(4)

P1: ZBU Final Pages



or

Mt

M0= 1 −

∞∑

n =0

8 exp[−D(2n + 1)2 π2t / l2]

(2n + 1)2 π2 , (5)

where M0 is the total amount of drug sorbed, Mt is theamount desorbed at time t , and l is the thickness of thedevice.

Fortunately, these expressions reduce to two much sim-pler approximations, reliable to better than 1%, valid fordifferent parts of the desorption curve. The early-time ap-proximation, which holds over the initial portion of thecurve, is derived from Eq. (4):

Mt

M0= 4

(Dt

πl2

)1/2

for 0 ≤ Mt

M0≤ 0.6. (6)

The late-time approximation, which holds over the finalportion of the desorption curve, is derived from Eq. (5):

M1

M0= 1 − 8

π2 exp

(−π2 Dt

l2

)

for 0.4 ≤ Mt

M0≤ 1.0.

(7)

The release rate is easily obtained by differentiatingEqs. (6) and (7) to give

d Mt

dt= 2M0

(D

πl2t

)1/2

(8)

for the early time approximation and

d Mt

dt= 8DM0

l2exp

(

−π2 Dt

l2

)

(9)

for the late time approximation.These two approximations are plotted against time in

Fig. 4. For simplicity, M0 and D / l2 have been set to unity.The release rate falls off in proportion to t −1/2 until 60%of the agent has been desorbed, after which the decay isexponential. Although the release rate from monolithicdevices is far from constant, this defect is often offset bytheir ease of manufacture.

FIGURE 4 Drug release rates as a function of time, showingearly- and late-time approximations.

B. Biodegradable Systems

The diffusion-controlled devices outlined above are per-manent, in that the membrane or matrix of the device re-mains in place after its delivery role has been completed.In applications in which the controlled release device is animplant, the drug-depleted device shell must be removedafter use. This is undesirable; such applications requirea device that degrades during or after completion of itsdelivery role.

Several polymer-based devices that slowly biodegradewhen implanted in the body have been developed. Themost important materials are those based on polylacticacid, polyglycolic acid, and their copolymers, as shownin Fig. 5. Other, less widely used biodegradable materialsinclude the poly(ortho esters), polycaprolactone, polyan-hydrides, and polycarbonates.

In principle, the release of an active agent can be pro-grammed by dispersing the material within such polymers,with erosion of the polymer effecting release of the agent.One class of biodegradable polymers is surface eroding:the surface area of such polymers decreases with time asthe cylindrical- or spherical-shaped device erodes. Thisresults in a decreasing release rate unless the geometry ofthe device is appropriately manipulated or the device is de-signed to contain a higher concentration of the agent in theinterior than in the surface layers. In a more common classof biodegradable polymer, the initial period of degradationoccurs slowly. Thereafter, the degradation rate increasesrapidly and the bulk of the polymer then erodes in a com-paratively short time. In the initial period of exposure tothe body, the polymer chains are being cleaved but themolecular weight remains high. Therefore, the mechani-cal properties of the polymer are not seriously affected. Aschain cleavage continues, a point is reached at which thepolymer fragments become swollen or soluble in water;at this point the polymer begins to dissolve. This typeof polymer can be used to make reservoir or monolithicdiffusion-controlled systems that degrade after their de-livery role is complete. A final category of polymer hasthe active agent covalently attached by a labile bond tothe backbone of a matrix polymer. When placed at the siteof action, the labile bonds slowly degrade, releasing theactive agent and forming a soluble polymer. The meth-ods by which these concepts can be formulated into actualpractical systems are illustrated in Fig. 6.

C. Osmotic Systems

Yet another class of delivery devices uses osmosis asthe driving force. Osmotic effects are often a problem indiffusion-controlled systems because imbibition of waterswells the device or dilutes the drug. However, several

P1: ZBU Final Pages



FIGURE 5 Biodegradation of poly(lactic acid-co-glycolic acid) to its simple acid precursors. The rate of biodegradationis a function of the copolymer composition. The pure homopolymers are both relatively crystalline and biodegradeslowly, but the more amorphous copolymers biodegrade more rapidly.

devices that actually use osmotic effects to control the re-lease of drugs have been developed. These devices, calledosmotic pumps, use the osmotic pressure developed bydiffusion of water across a semipermeable membrane intoa salt solution to push a solution of the active agent fromthe device. Osmotic pumps of various designs are widelyapplied in the pharmaceutical area, particularly in oraltablet formulations.

The forerunner of modern osmotic devices was theRose–Nelson pump. Rose and Nelson were two Australianphysiologists interested in the delivery of drugs to the gutof sheep and cattle. Their pump, illustrated in Fig. 7, con-sists of three chambers: a drug chamber, a salt chambercontaining excess solid salt, and a water chamber.The drugand water chambers are separated by a rigid, semiperme-able membrane. The difference in osmotic pressure acrossthe membrane moves water from the water chamber into

FIGURE 6 Various types of biodegradable drug delivery sys-tems.

the salt chamber. The volume of the salt chamber increasesbecause of this water flow, which distends the latex di-aphragm separating the salt and drug chambers, therebypumping drug out of the device.

The pumping rate of the Rose–Nelson pump is givenby the equation

dMt

dt= dV

dt· c, (10)

where d Mt/dt is the drug release rate, dV/dt is the vol-ume flow of water into the salt chamber, and c is the con-centration of drug in the drug chamber. The osmotic waterflow across a membrane is given by the equation

dV

dt= Aθ�π

l, (11)

where dV/dt is a water flow across the membrane ofarea A, thickness l, and osmotic permeability θ (cm3 · cm/

cm2 · hr · atm), and �π is the osmotic pressure differencebetween the solutions on either side of the membrane.This equation is only strictly true for completely selective

FIGURE 7 Mechanism of action of a Rose–Nelson osmoticpump, the precursor of today’s osmotic controlled release deliverysystems.

P1: ZBU Final Pages



membranes, that is, membranes permeable to water butcompletely impermeable to the osmotic agent. However,this is a good approximation for most membranes. Sub-stituting Eq. (11) for the flux across the membrane gives

d Mt

dt= A θ�πc

l. (12)

The osmotic pressure of the saturated salt solution is high,on the order of tens of atmospheres, and the small pres-sure required to pump the suspension of active agent isinsignificant in comparison. Therefore, the rate of waterpermeation across the semipermeable membrane remainsconstant as long as sufficient solid salt is present in thesalt chamber to maintain a saturated solution and hence aconstant-osmotic-pressure driving force.

Variations of the Rose–Nelson pump have been devel-oped as tools for drug delivery tests in animals. How-ever, the development that made osmotic delivery a majormethod of achieving controlled drug release was the in-vention of the elementary osmotic pump by Theeuwes in1974. The concept behind this invention is illustrated inFig. 8. The device is a simplification of the Rose–Nelsonpump, and eliminates the separate salt chamber by usingthe drug itself as the osmotic agent. The water requiredto power the device is supplied by the body, so a sep-arate water chamber is also eliminated. The Theeuwesdevice is formed by compressing a drug having a suit-able osmotic pressure into a tablet using a tableting ma-chine. The tablet is then coated with a semipermeablemembrane, usually cellulose acetate, and a small hole isdrilled through the membrane coating. When the tablet isplaced in an aqueous environment, the osmotic pressure ofthe soluble drug inside the tablet draws water through thesemipermeable coating, forming a saturated aqueous so-lution inside the device. The membrane does not expand,so the increase in volume caused by the imbibition of wa-

FIGURE 8 The Theeuwes elementary osmotic pump. This sim-ple device consists of a core of water-soluble drug surrounded bya coating of a water-permeable polymer. A hole is drilled throughthe coating to allow the drug to escape.

ter raises the hydrostatic pressure inside the tablet slightly.This pressure is relieved by a flow of saturated agent so-lution out of the device through the small orifice. Thus,the tablet acts as a small pump, in which water is drawnosmotically into the tablet through the membrane wall andthen leaves as a saturated drug solution through the ori-fice. This process continues at a constant rate until all thesolid drug inside the tablet has been dissolved and only asolution-filled shell remains. This residual dissolved drugcontinues to be delivered, but at a declining rate, until theosmotic pressures inside and outside the tablet are equal.The driving force that draws water into the device is thedifference in osmotic pressure between the outside en-vironment and a saturated drug solution. Therefore, theosmotic pressure of the dissolved drug solution has tobe relatively high to overcome the osmotic pressure of thebody. For drugs with solubilities greater than 5–10 wt%,this device functions very well. Later variations on thesimple osmotic tablet design use water-soluble excipientsto provide part of the osmotic pressure driving force; thisovercomes the solubility limitation.

III. IMPORTANT CONTROLLEDRELEASE PRODUCTS

The controlled release sector is a rapidly expanding partof the pharmaceutical industry. Growth has occurred at theremarkable annual rate of 15% over the past decade, fueledby an explosion of new technologies. The value of thepharmaceuticals using controlled drug delivery reached$20 billion in 1999, and while only modest growth in theoverall pharmaceutical market is projected for the next fewyears, the drug delivery share of the market is expected tocontinue to grow at least 15% per annum. As much as20% of the U.S. pharmaceutical market is projected to becontrolled release products by 2005.

The majority of drug delivery products reach the marketas a result of a strategic alliance between a drug deliverycompany, which supplies the technology, and a pharma-ceutical company, which supplies the drug and the re-sources needed for full development. A good example ofsuch a collaboration is provided by protein and peptidedrugs, an increasingly important area of pharmaceuticalsdriven by recent advances in biotechnology. Currently, amajor factor limiting the use of these drugs is the needfor their administration by repeated injections. This is be-cause peptides and proteins are poorly absorbed in theGI tract, so cannot be delivered as oral tablets, and havevery short biological half-lives, so cannot be delivered asa single, long-acting injection. Several specialized drugdelivery companies are developing innovative techniquesthat will allow these drugs to be delivered orally, by nasal

P1: ZBU Final Pages



inhalation, or in a long-acting injectable formulation. Al-liances between these technology providers and the phar-maceutical companies producing the protein/peptide ac-tive agents increase the likelihood that these novel drugformulations will reach the marketplace.

Controlled drug delivery has a record of success inthe pharmaceutical industry and can offer a high returnon capital investment. Some relatively low-budget devel-opment programs, which have enabled existing, effectivedrugs to be administered to the patient by controlled re-lease technology, have been very successful; examples in-clude Lupron Depot (leuprolide, a hypothalamic releasinghormone used for the suppression of testosterone in thetreatment of malignant neoplasms of the prostate), Pro-cardia XL (nifedipine, a calcium channel blocker used forthe treatment of hypertension and angina), and CardizemCD (diltiazem, a calcium channel blocker with propertiessimilar to nifedipine). These are all billion-dollar productsthat employ controlled release or delivery techniques. Todevelop a new chemical entity through to regulatory ap-proval in the mid-1990s took, on average, 10–12 yearsand cost about $300–600 million. In contrast, an existingdrug can be reformulated into an innovative drug deliverysystem in 5–7 years at a cost of about $20–$100 million.Interestingly, there are also recent examples in which con-trolled release is no longer a simple reformulation of anold drug. For example, new drugs are being developed andmarketed for the first time as controlled release products.Some of these drugs might not have reached the marketexcept for controlled release technology—felodipine andomeprazole are examples.

A. Oral Formulations

The oral route is by far the most common and convenientmethod of delivering drugs to the body. Unfortunately,the method has a number of problems that interfere witheffective drug delivery. First, a drug taken by mouth isimmediately exposed to low-pH stomach acids containinghigh concentrations of digestive enzymes. Many drugs arechemically degraded or enzymatically metabolized in thestomach before they are absorbed. Drugs that are absorbedthen enter the portal circulation and may be destroyedby the first-pass metabolism in the liver described earlier.Controlled release is a method of avoiding these problems.

The typical transit time of material through the GI tractis 12–18 hr, so most controlled release oral formulationsare designed to deliver their loading of drug over a 6- to15-hr period. In this way the action of short-half-life drugsor rapidly absorbed drugs can be spread over a prolongedperiod. Drugs that might require dosing two or three timesa day to achieve relatively uniform and nontoxic bloodlevels can then be dispensed as a single once-a-day tablet.

This simple change, by improving patient compliance andproducing a more controlled constant blood level, has pro-duced measurable improvements in efficacy and reducedtoxicity for many drugs. Many oral controlled release for-mulations are designed to produce a relatively low drugdelivery rate for the first 1–3 hr while the formulation isin the stomach, followed by prolonged controlled releaseof the drug once the formulation has reached the GI tract.This avoids chemical degradation of the drug in the aggres-sive environment of the stomach. This type of delivery is,for example, particularly important for polypeptide drugswhich are rapidly and completely destroyed if delivered tothe stomach. Delivery to the GI tract is also done to achievelocal delivery of the drug, such as the anti-inflammatoryMesalazine for irritable bowel disease and ulcerativecolitis.

The precursors of today’s controlled release oral for-mulations were enteric tablets based on various wax ma-trices designed to circumvent degradation in the stomach.Enteric formulations were later improved by using new,more reliable polymers. By the mid-1970s, the first oralcontrolled drug delivery systems began to appear. Two im-portant delivery technologies developed at that time wereAlza’s Oros osmotic controlled release system and Elan’sSodas multiparticulate system. Elan’s Sodas system con-sisted of large numbers of micropellets, each designed torelease a microdose of drug by diffusion from a matrixat a predetermined rate. By blending pellets with differ-ent release profiles, the overall target rate was achieved.Since then, a wide variety of other oral formulations us-ing osmosis and diffusion have been produced, as wellas slow-release bioerodible tablets, ion exchange beads,multiple-layer tablets, and others.

If the drug is relatively water-soluble, osmotic or sim-ple table formulations are often used to achieve controlleddelivery. However, with more-insoluble drugs, release ofthe complete dosage from a single tablet in an 8- to 12-hrperiod may be difficult. For such drugs, a microencapsu-lated or granulated form of the drug is enclosed in a gelatincapsule. Microencapsulation exposes a much greater sur-face area of the device to interact with the body, so drugsthat dissolve and diffuse slowly can still be completely re-leased in an 8- to 12-hr period. Drugs can be microencap-sulated by physical and chemical methods. Physical meth-ods include encapsulation by pan coating, gravity flow,centrifugation, and fluid bed coating. Chemical microen-capsulation normally involves a two-step process calledcoacervation. Drug particles or droplets of drug solutionare first suspended in a polymer solution. Precipitationof the polymer from solution is then caused by, for ex-ample, changing the temperature or adding a nonsolvent.The polymer then coats the drug particles to form themicrocapsule.

P1: ZBU Final Pages



The leading developers of oral drug delivery formu-lations are Alza and Elan. Other important producersare Skyepharma, which has developed a technologycalled Geomatrix (a multilayer tablet with each layerreleasing the drug at a different rate), R. P. Scherer, whichhas several different technologies including Zydis (alyophilized tablet), and Eurand, with several technologiesfor controlled release, taste masking, and improvedbioavailability.

B. Transdermal Systems

Scopolamine, for control of motion sickness, was the firstdrug to be marketed in the form of a transdermal patchsystem. Since then the market for transdermal patcheshas grown steadily. However, the number of drugs thatcan be delivered through the skin is more limited thanwas anticipated when the first patches were introduced inthe 1980s. The main problem limiting widespread use isthe low permeability of most drugs through the skin. De-pending on the drug, skin permeabilities are in the range0.01–10 µg of drug/cm2 · hr. Because the maximum ac-ceptable size of a transdermal patch is limited to about50 cm2, drugs delivered through the skin must be effec-tive at doses of 0.01 mg/day for a poorly skin-permeabledrug and 10 mg/day for a highly skin-permeable drug.Very active skin-permeable drugs are required to maketransdermal drug delivery possible. Despite the enormousinterest in transdermal delivery in academia and industry,the number of drugs delivered transdermally is currentlylimited to nitroglycerin, nicotine, estradiol, clonidine, fen-tanyl, testosterone, and isorbide nitrate. Nitroglycerin andnicotine are currently the most important drugs, but thearea of hormone replacement therapy is growing as a resultof improved estradiol and testosterone delivery patches.

The three main types of diffusion-controlled transder-mal devices are shown schematically in Fig. 9. The simpleadhesive device (Fig. 9, top) has a two-layer “Band-aid”configuration comprising the backing layer coated withadhesive. The drug is mixed in the adhesive layer usedto fix the bandage to the skin. These medicated bandagesbring a known quantity of drug to a known area of skinfor a known period of time, but have no mechanism forcontrolling the rate at which the drug is delivered to thepatient.

The second type of device is a monolithic system (Fig. 9,middle) incorporating a backing layer, a matrix layer, andan adhesive layer. The matrix layer consists of a polymermaterial in which the solid drug is dispersed; the rate atwhich the drug is released from the device is controlledby this polymer matrix. With this type of system, the drugrelease rate falls off with time as the drug in the skin-contacting side of the matrix is depleted.

FIGURE 9 Transdermal patch designs.

The third type of device is the reservoir system (Fig. 9,bottom). In this case, the drug—usually in liquid or gelform—is contained in a reservoir separated from the skinby an inert membrane that controls the rate at which drugis delivered to the skin. These devices offer an importantadvantage over the monolithic geometry: as long as thedrug solution in the reservoir remains saturated, the drugrelease rate through the membrane is constant.

The pattern of drug release from the device is important.If drug is delivered to the skin at less than the maximumrate at which it can be absorbed, the device is the primarydosage-controlling mechanism. When the drug is deliv-ered faster than the skin can absorb it, the skin surface isthen saturated with drug at all times, and the limiting fac-tor for systematic dosage is the rate of absorption throughthe skin. Thus, at least in principle, devices for which thedosage-controlling mechanism is either the skin or the de-vice can be designed.

To reach the systemic blood circulatory system, drugfrom a transdermal patch must cross several layers of skin,as shown in Fig. 10. The top surface layer of skin, calledthe stratum corneum, represents the main barrier to drugpermeation. The stratum corneum is only 10–15 µm thick,but it consists of layers of flattened, cornified cells that arequite impermeable. The interspace between these cellsis filled with lipids, giving the structure a “bricks-and-mortar” form, with the cells being the bricks and the lipidsbeing the mortar. The most important pathway for drugabsorption is through the lipid (mortar), which dictates

P1: ZBU Final Pages



FIGURE 10 Membrane-moderated transdermal drug delivery system (not to scale). Drug permeates from the patchreservoir through the protective outer layers of the skin and is absorbed into the underlying capillaries of the generalsystemic blood circulation.

the characteristics of usefully permeable drugs. These aredrugs with a molecular weight of less than 1000 dalton,a low melting point, an octanol/water partition coefficientbetween 0 and 2, and few polar centers.

Many drugs do not have the chemistry required toachieve high skin permeabilities, so various methods ofenhancing drug permeability have been developed. Onemethod is to soften and swell the stratum corneum by dis-solving or dispersing the drug in a simple solvent. For ex-ample, the first estradiol-delivery transdermal patch usedethanol to enhance the skin’s permeability to estradiol. Inthe absence of ethanol, estradiol flux through the skin isvery low, on the order of 0.01 µg/cm2 · hr. However, if asuspension of estradiol in ethanol is applied to the skin,the estradiol flux increases 10- to 20-fold. A second wayto enhance skin permeability is to use a compound suchas a long-chain fatty acid, ester, or alcohol. These com-pounds penetrate the stratum corneum more slowly thansmall molecules such as ethanol but have a more prolongedplasticizing effect. A third method combines elements ofthe first two. An enhancer formulation containing both asimple solvent and a fatty component is used to combinethe rapid onset of solvent effect with the prolonged actionof the fatty component.

Another method of enhancing drug permeation is toincrease the driving force for drug permeation by usinga small electric current. This last approach, called ion-tophoresis, has been widely studied. The principle of ion-tophoresis is illustrated in Fig. 11. In this method, a bat-tery is connected to two electrodes on the skin. An ionized

drug placed in contact with one electrode will migrate un-der the influence of the voltage gradient through the skinand enter the system circulation; very large enhancementscan be obtained. The earliest devices having the essen-tial features of iontophoresis date back to the 1890s, al-though apparently their objective was to shock their sub-jects rather than to administer drugs to them. The firstmodern device appeared in 1972, and advances in elec-tronics have since allowed smaller and smaller devices tobe built. The newest devices have a built-in battery layerand are comparable in size to a normal transdermal patch.Iontophoresis appears to be a particularly promising toolfor the delivery of very active peptides, small proteins, oroligonucleotides, which are otherwise almost completely

FIGURE 11 Mechanism of an iontophoresis patch. The poles ofa small battery are connected to the skin. A solution of an ionizeddrug placed at one electrode migrates through the skin into thesystemic circulation under the action of the voltage driving forceof the battery.

P1: ZBU Final Pages



skin-impermeable. Under the action of a small voltagegradient, the skin permeation rate of the drug increases10- to 100-fold. Another advantage of iontophoresis isthat the voltage driving force can be easily regulated; thisallows pulsatile drug delivery or variable delivery accord-ing to the patient’s needs. This control is useful in thedelivery of analgesics such as fentanyl for the control ofpain. Despite the many patents in this field, no commercialproducts have yet reached the market.

The main advantages of transdermal over oral drug de-livery are the ability to maintain constant plasma lev-els with short-half-life drugs and to avoid the hostileconditions of the gastrointestinal tract and consequentdrug deactivation because of the hepatic first-pass effect.Moreover, patient compliance tends to increase, reducingthe number of administrations required. Patches are nowavailable for once-a-day, twice-a-week, and once-a-weektreatment.

C. Nasal Spray/Inhalers

Drug delivery by inhalation has a long history and is anobvious method of administering agents that act on therespiratory system. However, it is now being used in-creasingly to deliver systemically active drugs. The firsthand-held, pressurized, metered inhaler was launched in1960; several other products have been introduced since.Intranasal delivery is currently employed in treatmentsfor migraine, smoking cessation, acute pain relief, noc-turnal enuresis, osteoporosis, and vitamin B12 deficiency.In 1999, Aviron’s intranasal influenza vaccine, FluMist,was filed with the FDA, and several other vaccines arebeing considered for nasal administration. Other applica-tions for nasal delivery under development include cancertherapy, epilepsy control, antiemetics, and treatment ofinsulin-dependent diabetes.

The advantages of the nasal cavity over other drug ad-ministration routes are rapid, direct access to the sys-temic circulation and complete avoidance of first-passmetabolism. The nasal cavity is also a far less aggressiveenvironment than the gastrointestinal tract, and so is par-ticularly useful for the delivery of peptides and proteins,which are easily degraded in the stomach. Patient compli-ance is also improved, particularly if the alternative treat-ment is intravenous injection. However, the nasal route isnot without its problems. Issues that need to be consideredare the rate of drug absorption through the nasal mucosa,the residence time of the formulation at the site of delivery,local toxicity and tolerability, and degradation of the drugin the nasal cavity. The rate of absorption of the drug can becontrolled by including carriers and enhancers to modifypermeation, or by chemical modification of the drug into

a more readily absorbed form. Most nasal devices deliverthe drug as a fine liquid or solid spray; the average par-ticle size significantly affects the rate of drug absorption.Intranasal delivery of systemic drugs requires sophisti-cated delivery devices to ensure accurate, repeatable dos-ing and minimal irritation. Dosing accuracy is particularlyimportant for delivery of potent and expensive drugs suchas the new generation of peptide and protein products.Long-term nasal delivery may also require dose-countersand lock-out systems to prevent overdosing as, for exam-ple, in the pain relief area.

Prior to 1985, chlorofluorocarbons were widely used asinert nontoxic inhaler propellants; these compounds arenow all but banned. Consequently, much recent researchhas centered on the development of dry powder inhalers,which deliver the active ingredient as ultrafine particles di-rectly to the lungs with minimal deposition in the mouthand trachea. Some of these devices are activated directlyby the inspiration of the patient without the need to coor-dinate activation and inspiration.

Although dry powder inhalers have a number of ther-apeutic benefits, they also have problems. For example,contact of the powder formulation with moisture can leadto agglomeration and inaccurate dosing, and dose unifor-mity is hard to achieve. The potential for nasal delivery isgreatest in two areas: local delivery to the lung for respi-ratory treatment diseases and systemic delivery of a broadvariety of drugs via the alveoli, which have highly absorp-tive properties. Its greater long-term potential is in the de-livery of macromolecules, but further research is needed todetermine the long-term immunogenicity, reproducibility,and stability of the delivery systems.

D. Targeted Drug Delivery

Targeted or directed drug delivery is a relatively newmethod of delivering drugs. The objective is to alter the ef-fectiveness of drugs by targeting their delivery directly tothe site needing drug action. Promising techniques includethe use of liposomes, polyethylene glycol (PEG)-coatedmolecules, blood–brain barrier transfer agents, and severalantibody conjugate approaches.

The most advanced targeted delivery technology usesliposomes, which are ultrafine water/oil/water emulsionsin which the emulsion droplets consist of lipid vesiclescontaining an aqueous drug solution. The surface of thevesicles is sometimes modified to promote targeting of thelipid-drug-containing vesicle to selected tissues. The firstliposomes were developed in 1965 as a model of biologi-cal membranes. Their potential as a drug delivery systemwas recognized later; now, after many years of gestation,liposomes are finally being introduced in commercial

P1: ZBU Final Pages



products. The mean diameter of liposomes is less than0.1 µm. This allows them selectively to extravasate intotissues characterized by leaky vasculature, such as solidtumors, achieving targeted delivery to the diseased organwith low adverse effects on normal tissues.

The first liposome product on the market was Ambi-some, containing amphotericin B, an antifungal encapsu-lated in liposomes to reduce its toxicity. Since then, liposo-mal preparations of other drugs have been developed, mostimportantly Daunoxome, an anticancer product contain-ing the antitumoral drug daunorubicin for the treatmentof Kaposi’s sarcoma. Liposomes are also being investi-gated as vehicles for gene therapy. Their main advantagesover viral carriers is a higher loading capacity and a lowerrisk of evoking an immune response. Surface-coating li-posomes with antibodies to allow active targeting to a par-ticular disease site is also being studied in clinical trials. Ifthese products are successful, targeted drug delivery willbe a breakthrough in the treatment of a number of majordiseases.

A second approach to targeted drug delivery usespolyethylene glycol (PEG), a water-soluble polymer cova-lently linked to proteins, which alters their properties andextends their potential use. The main product using this ap-proach is for α-interferon for the treatment of hepatitis C.This technique has been applied to several proteins in-cluding adenosine deaminase, cytokines, and granulocyte-macrophage colony-stimulating factor. Generally the re-sults obtained by the modification with PEG are increasedcirculating life, reduced immunogenicity and antigenicity,and increased stability and solubility with a minimal lossof biological activity.

E. Implants

Three types of implantable drug delivery devices are cur-rently in use or being developed. The first type is an im-plantable polymeric capsule most commonly made fromsilicone rubber, which when placed under the skin deliversthe drug load at a constant rate for as long as 2–5 years.Upjohn’s Depo-Provera and American Homes’ Norplant,used to deliver contraceptive steroids for long-term con-traception, are examples of this approach. Implantable de-vices achieve long-term, controlled delivery of the drugand patient compliance is no longer an issue, both signif-icant advantages. The key disadvantage is the size of thedevice, which means minor surgery is required to insertthe implant and later to remove the device once its drugdelivery role has been completed. There is also a potentialto release all of the drug in a few days if the capsule be-gins to leak. For these reasons this type of nondegradabledevice has not become a major product.

A second type of device uses implants made from bio-degradable polymers. Because the device is biodegrad-able, device retrieval once drug delivery is complete isno longer necessary. This makes the device much moreacceptable to patients. Unfortunately, it is technically dif-ficult to create implantable, biodegradable devices that de-liver drug reliably for more than 1 or 2 months. Therefore,these devices are generally restricted to delivering drugsfor 4–6 weeks. Examples include Abbot’s Lupron Depotformulation, a single monthly injection of leuprolide forendometriosis, and Zoladex, Zenaca’s 4-week implantableformulations of goserelin for endometriosis in women andprostrate cancer in men.

A final category of implantable devices, in late-stagedevelopment, is miniature pumps driven by osmosis orfluorocarbon propellants. These pumps are designed todeliver essentially any drug at a predetermined rate forweeks or even months. The problem to be solved in thiscase is to make the pump small enough and reliable enoughthat the trauma of device removal is outweighted by itsdrug delivery benefits.

IV. FUTURE DIRECTIONS

The era of modern controlled drug delivery started with thelaunching of the first product registered at the U.S. Foodand Drug Agency in terms of both the total amount of drugdelivered and the rate of drug delivery. This first product,Alza’s Ocusert, was launched in 1974. By 1990 the totalcontrolled release market was approximately $1 billion.Since then the market has grown 20-fold, and the numberof such products is expected to continue to grow rapidlyin the next few years.

The largest growth area will be the extension of al-ready developed controlled release technologies to a widernumber of drugs. A long list of oral and some transder-mal controlled release products are in late-stage clinicaltrials and should appear on the market in the next fewyears.

A second area of significant future growth will bethe use of controlled release systems to deliver the newgeneration of peptide and protein drugs. As a conse-quence of the sequencing of the human genome, the in-formation required to design protein and peptide drugsto treat important diseases is at hand. However, many ofthese new drugs are essentially ineffective if administeredas conventional pharmaceutical formulations. New, well-designed controlled release systems that can deliver thedrugs intact at a prolonged steady rate close to the siteof action are required; such systems are being activelydeveloped.

P1: ZBU Final Pages



A third potential growth area for controlled release tech-nology is the development of systems for targeted drugdelivery. Drug delivery to the ultimate site of action is amultistep process. Conventional controlled drug deliverysystems generally only address the first step, the rate ofdelivery of drug to the body. The path of the drug from thedosage site to the diseased cells is largely uncontrolled.Targeted drug delivery systems attempt to improve thisstep in the delivery process, in effect, to improve the aimof Ehrlich’s magic bullet. Antibody-coated liposomes andthe PEG-peptide products described above are the firstsimple prototype product designs to tackle this problem.Over time, more sophisticated and more effective tech-niques will be developed.

Finally, all the controlled release products describedin this article are preprogrammed, with the rate of de-livery being established by the designer of the device.No subsequent adjustment of the drug delivery rate inresponse to patient need is possible. However, many dis-ease conditions are episodic, for example, diabetes, stroke,migraines, heart attacks, and epilepsy. Controlled drugdelivery systems that could sense the body’s need andautomatically deliver the drug at the rate required wouldbe a huge benefit. In the intensive care facilities of modernhospitals, this type of control is achieved through contin-uous electronic sensors and monitoring by the attendingnurses and physicians. In the future, development of mi-croelectronic/micromechanical patient-portable machinesto produce the same level of control on ambulatory, at-riskpatients can be imagined.


AEROSOLS • BIOPOLYMERS • ELECTROPHORESIS • ION

TRANSPORT ACROSS BIOLOGICAL MEMBRANES • MEM-BRANES, SYNTHETIC, APPLICATIONS • PHARMACEUTI-CALS • PHARMACOKINETICS

BIBLIOGRAPHY

Baker, R. (1987). “Controlled Release of Biologically Active Agents,”Wiley, New York.

Benita, S. (1996). “Microencapsulation: Methods and Industrial Appli-cations,” Marcel Dekker, New York.

Chasin, M., and Langer, R. (1990). “Biodegradable Polymers as DrugDelivery Systems,” Marcel Dekker, New York.

Chien, Y., Su, K., and Chang, S. (1989). “Nasal Systemic Drug Delivery,”Marcel Dekker, New York.

Clark, A. (1995). “Medical aerosol inhalers: Past, present and future,”Aerosol Sci. Technol. 22, 374–391.

Friend, D. (1992). “Oral Colon-Specific Drug Delivery,” CRC Press,Boca Raton, Fl.

Katre, N. (1993). “The conjugation of proteins with polyethylene glycoland other polymers,” Adv. Drug Deliv. Rev. 10, 91–114.

Lasic, D., and Papahadjopoulos, D. (1998). “Medical Applications ofLiposomes,” Elsevier Science, New York.

Potts, R., and Guy, R. (1997). “Mechanisms of Transdermal Drug De-livery,” Marcel Dekker, New York.

Santus, G., and Baker, R. (1995). “Osmotic drug delivery: A review ofthe patent literature,” J. Controlled Release 35, 1–21.

Smith, E., and Maibach, H. (1995). “Percutaneous Penetration En-hancers,” CRC Press, Boca Raton, FL.

Wise, D. (2000). “Handbook of Pharmaceutical Controlled Release Tech-nology,” Marcel Dekker, New York.

P1: LDK/GJK P2: GQT/Final Pages Qu: 00, 00, 00, 00


Physical Organic ChemistryCharles L. PerrinUniversity of California, San Diego

I. Molecular Structure: BondingII. Molecular Structure: StereochemistryIII. Chemical ReactivityIV. Methodology of Mechanistic StudiesV. Illustrative Examples

VI. Quantitative Relationships

GLOSSARY

Aromaticity Unusual stability of cyclic conjugated sys-tems containing 4n + 2 pi electrons.

Brønsted relationship Linearity between logarithmsof the catalytic constants for acid or base andpK .

Chiral Not superimposable upon its mirror image.Cis/trans isomerism Isomerism arising from sidedness

on a double bond or ring.Conformations Structures interconvertible by rotation

about single bonds.Diastereomers Stereoisomers that are not mirror images

of each other.Enantiomers Stereoisomers that are (nonsuperimpos-

able) mirror images of each other.General acid catalysis Catalysis where each acid present

functions as a catalyst, rather than H3O+ alone.Hammett equation Quantitative comparison with sub-

stituent effects in benzoic acid dissociation.Intermediates Energy minina between reactant and

product on a reaction coordinate.

Linear free energy relationship Logarithmic linearitybetween rate and equilibrium constants.

Nucleophile Atom or reactant that uses an electron pairto form a bond to an electrophile.

Orbital One-electron wave function.Order (kinetic order) Power of the concentration of a

chemical species entering into a rate equation.Pericyclic reaction Reorganization of electrons around a

cycle of nuclei, with simultaneous bond breaking andbond making.

Product-development control Predominant formationof the more stable product because its transition statepartakes of that stability.

Racemic mixture A 50:50 mixture of two enantiomers.Reaction coordinate Geometric parameter that measures

progress from reactants to products.Stereoisomers Different molecules with the same con-

nectivity (bonding relationships between atom pairs).Torsional strain Destabilization of eclipsed conforma-

tions.Transition state Position of maximum energy along

reaction coordinate.

211

P1: LDK/GJK P2: GQT/Final Pages


212 Physical Organic Chemistry

PHYSICAL ORGANIC CHEMISTRY undertakes theinvestigation of the phenomena of organic chemistry byquantitative and mathematical methods. This was the orig-inal approach, as presented in L. P. Hammett’s influen-tial book, “Physical Organic Chemistry: Reaction Rates,Equilibria, and Mechanisms,” published in 1940. Prior toHammett’s studies, organic chemistry had largely beenviewed as a collection of empirical observations, withoutany underlying sense. Hammett’s contribution was to takethe methodology of physical chemistry, apply it to organicchemistry, and derive regularities that placed organic re-activity on a quantitative basis.

Since then, physical organic chemistry has broadenedits focus to the structure, properties, and reactions of or-ganic molecules, and especially to the relationship be-tween structure and reactivity. Physical organic chem-istry asks how chemical reactivity depends on molecularstructure. Part of the answer comes from detailed andquantitative studies of some reactions. Part comes fromrecognizing analogies among classes of reactions. Thislatter is a particularly powerful method since it permitsthe extension of understanding from one well studiedclass of reactions to another, less well understood one.The underlying principle is that small perturbations ofmolecular structure are unlikely to lead to major changesin reactivity. This principle does not always hold, butit holds often enough to provide a broad general the-ory of chemical reactivity. In recent years physical or-ganic chemists have been successful in codifying theunderlying principles of chemical structure and reactiv-ity. The results have provided not only a deep under-standing of organic reactions, but also applications toorganic synthesis, biochemical processes, and materialsscience.

I. MOLECULAR STRUCTURE: BONDING

A. Covalent Bonding

Since structure determines chemical reactivity, it is neces-sary to have a thorough understanding of molecular struc-ture. Two atoms are held together by the sharing of valenceelectrons between them. The interaction that arises fromsharing a pair of electrons is called a covalent bond. Inearly illustrations the two electrons were shown as dots.Now the covalent bond is symbolized by a line connectingthe two atoms. The simplest case is the hydrogen molecule(1), formed from two hydrogen atoms, each with one va-lence electron. The line symbolizes the two electrons thatare shared.

A covalent bond represents a stabilization of themolecule relative to the separated atoms. Through sharingof electrons, each atom achieves an “octet configuration”analogous to that of the noble gases. For example, each ofthe hydrogen atoms in 1 is associated with two electrons,just as in a helium atom. Thus the hydrogens achieve theinertness or unreactivity of a noble gas. Hydrogen and he-lium are unique in that only two electrons are necessaryto complete a filled shell. For the next rows of the peri-odic table, eight valence electrons are necessary, hencethe designation “octet.” For example, carbon in methane(2) achieves its octet by sharing its four valence electronswith four hydrogens, each contributing another electron,and each of the carbons in ethane (3) shares a pair of elec-trons with each of three hydrogens and also with eachother. In ethene (4) each of two carbons achieves its octetby sharing two pairs of electrons with each other and alsotwo additional electron pairs with two hydrogens. The fourelectrons that are shared between the two carbons form adouble bond, symbolized by two connecting lines. Simi-larly, ethyne (5) has a triple bond, with six electrons sharedbetween the two carbons, each of which achieves an octet.

A system of graphic symbols, called Lewis structures,permits the illustration of molecular structures. Accord-ing to this system, every valence electron must be shown.Some valence electrons are in covalent bonds, whethersingle, double, or triple. Some valence electrons are un-shared and are symbolized as dots on the atoms to whichthey belong. Sometimes these are paired, as in ammonia(6) or water (7), where the unshared electrons are calledlone pairs. Sometimes an unshared electron is unpaired,as in methyl radical (8 ), where it is called an odd electron.This last is unusual in that the carbon lacks an octet. Innearly all stable molecules every atom has its octet, andnonoctet atoms are usually quite reactive (NO and NO2

are exceptions).

Exceptions to the octet rule are possible. Metallic ele-ments, at the left of the periodic table, often lack an octet.For example, the boron in BF3 has only six electrons. Morecommon exceptions are with atoms in the second full rowof the periodic table, such as the phosphorus in PCl5 orthe sulfurs in dimethyl sulfoxide (9) and sulfuric acid (10),which have 10, 10, and 12 electrons, respectively.



Physical Organic Chemistry 213

A further feature of Lewis structures is the designationof formal charge at each atom. The formal charge Zf isdefined as follows, where Nval is the number of valenceelectrons, Nbond is the total number of bonds in which theatom is engaged, Npair is the number of lone pairs on theatom, and Nodd is the number of its odd electrons (almostalways zero):

Zf = Nval − Nbond − 2Npair − Nodd. (1)

This equation reflects the loss or gain of electrons onforming covalent bonds, since Nval is the number ofelectrons that the atom originally had, 2Npair + Nodd

electrons remain on the atom, and the two electrons inany bond are considered to be shared equally, so thatonly one of them “belongs” to the atom. For example, theformal charges in hydronium ion (11), bicarbonate ion(12), and diazomethane (13) are as shown (omitted foratoms with Zf = 0), since Zf(O11) = 6 − 3 − 2 × 1 = +1,

Zf(Oright) = 6 −1 −2 × 3 = −1, Zf(Oother) = 6 −2 −2 × 2= 0, Zf(Nmiddle) = 5 − 4 = +1, and Zf(Nright) = 5 − 2 −2 × 2 = −1.

The assumption of equal sharing of electrons in a co-valent bond is an oversimplification for the purpose ofassigning formal charge. In fact, the electrons are heldmore tightly by whichever atom is more electronegative.Electronegativity is an elusive concept that is best treatedqualitatively. Electronegativity increases toward the rightof the periodic table (C < N < O < F, Si < P < S < Cl) andtoward the top (I < Br < Cl < F, Se < S < O). Thus fluo-rine is the most electronegative and the other halogens arealso electronegative since they need only one additionalelectron to complete their octet. Metals are not electroneg-ative, and both carbon and hydrogen are intermediate andof nearly equal electronegativity. In cases of unequal shar-ing, covalent bonds have a polar character, with partial pos-itive and partial negative charges symbolized by δ+ andδ−, as in methyl chloride (14) and methyl lithium (15).In the extreme case of a large difference in electronega-tivities, there may be no sharing. Instead the electron isfully transferred and the bond is said to be ionic, as insodium chloride (16). The imbalance of charge leads to alocal dipole moment µ given by the following equation,where q is the partial (or full) charge and d is the distancebetween the positive and negative charges:

µ = qd. (2)

It is unwieldy to display each hydrogen in a Lewis struc-ture. Often, condensed structures are used, with the n hy-drogens on an atom written as Hn following (or some-times preceding) the atom, and with single bonds omittedand instead merely implied by juxtaposing the bondedatoms. Thus methyl chloride (14) and methyl lithium (15)become CH3Cl and CH3Li, respectively. Also, parenthe-ses are used to indicate branching, as in CH3C( O)CH3

or CH3CH(OH)CH3, but they may be omitted, as inCH3COOH, which can be misleading.

B. Resonance

For many molecules the Lewis symbolism provides an ad-equate representation of the molecular structure. The gen-eral requirement for such molecules is that they must haveonly single bonds and lone pairs, and they also must haveneither nonoctet atoms nor conjugated multiple (doubleor triple) bonds. A conjugated multiple bond is one that isadjacent to a lone pair or multiple bond, meaning that it isseparated by only one bond. For example, 17 satisfies therequirement, but 12 and 13 do not, since they have a lonepair (on oxygen or nitrogen) adjacent to a C O (vertical)or C N (separated by C O or N N).

Molecules for which this requirement is not met havedelocalized electrons, electrons that cannot be assigned toa covalent bond between two atoms or to a lone pair or oddelectron on a single atom. One way to describe moleculeswith delocalized electrons is through the symbolism of res-onance. The phenomenon is quantum mechanical and notreadily understood in classical terms. A resonance hybridis composed of several resonance forms each of which isa single Lewis structure with localized electrons. The in-dividual resonance forms have no independent existence,except that they together contribute to the hybrid. The hy-brid is symbolized with a double-headed arrow associatingthe individual structures with each other. For example, di-azomethane is a hybrid of the previous resonance form 13and another one, 18.




There are three rules to be obeyed in dealing with reso-nance: (1) In each contributing resonance form every atommust satisfy the octet rule (as modified by the exceptionsabove) and resonance forms with a nonoctet atom canusually be ignored; (2) resonance expresses the delocal-ization of electrons, but the resonance hybrid has a singlegeometry, with fixed nuclei; and (3) resonance providesstabilization, such that the resonance hybrid is of lowerenergy than would be expected for any of its contributingresonance forms. This stabilization is called resonanceenergy.

Electron pushing, symbolized with curved arrows, pro-vides a convenient method for generating additional res-onance forms. The rule is to delocalize a lone pair (or apair of electrons in a multiple bond, or, more rarely, in asingle bond) toward an adjacent atom so as to form a mul-tiple bond, while also removing an electron pair from thatadjacent atom to avoid a violation of the octet rule. Forexample, delocalizing an electron pair from Oright of bi-carbonate ion (12) generates another resonance form (19),and delocalizing an electron pair from Oleft generates yetanother (20). Benzene (21) is an example of delocaliz-ing electrons in multiple bonds. The second resonanceform is not just the first one rotated by 60◦; the nucleiremain fixed, and the electrons of the double bonds aredelocalized.

It is tedious to draw all resonance forms, and a short-hand notation is often used. Bonds that are present in someresonance forms but not in all are shown as dotted, formalcharges that are present in some resonance forms but notall are shown as δ+ or δ− (same symbolism as before,different context), and lone pairs are not shown. Thus dia-zomethane (13, 18), bicarbonate ion (12, 19, 20), and ben-zene (21) become 22, 23, and 24, respectively. For this lasta further abbreviation, often applied to cyclic compounds,is to omit hydrogens attached to carbons and to symbolizeeach carbon as a vertex. Moreover, the six dotted lines areoften simplified to a circle, as in 24′.

C. Atomic Orbitals

Lewis structures do not describe the geometry ofmolecules. The set of bonds in a molecule describes onlythe connectivity—which atom is bonded to which otheratom or atoms. Indeed, most of the structures 2–20 arenot drawn with any attempt at correct geometry. To de-scribe the geometry, another approach is necessary. Orbitaltheory provides a complementary approach to molecularstructure.

Orbital theory takes account of the wave nature of elec-trons. The uncertainty principle of quantum mechanicsprohibits locating an electron exactly. Instead only prob-abilities can be specified, as described by a wave func-tion ψ . For a single electron that wave function is calledan orbital, and its square is the probability of finding theelectron “at” the point x, y, z in three-dimensional space:

P(x, y, z) = [ψ(x, y, z)]2. (3)

It is possible to express ψ as a mathematical formulawith a value at every point in space but it is often moreconvenient just to sketch surfaces of constant ψ . Forsome orbitals, called s orbitals, the surface is a sphere.For some other orbitals, called p orbitals, the surface re-sembles a pair of flattened spheres separated by a planewhere ψ(x, y, z) = 0. Such a plane is called a nodal planeand it divides the region of space where ψ(x, y, z) > 0from the region where ψ(x, y, z) < 0. There is no furthersignificance to positive or negative ψ(x, y, z) since onlythe probability in Eq. (3) is measurable. There are also dorbitals, with two nodal planes.

It takes a little imagination to reconstruct a three-dimensional surface of constant ψ from its illustration ona two-dimensional page. The sphere of an s orbital can beillustrated as a circle, as shown in Fig. 1. The near-spheresof a p orbital are rarely illustrated accurately. Insteadstylized (very approximate) versions are often used, asalso shown in Fig. 1.

FIGURE 1 Atomic orbitals.




D. Molecular Orbitals

For molecules the orbitals become even more complicated.In general, they are not simply s or p orbitals on a singleatom, but orbitals that spread over the entire molecule.In some cases they can be approximated in terms of twoatomic orbitals on two adjacent atoms. For example, inthe simplest case of the hydrogen molecule, the molecu-lar orbital ψMO can be expressed as a linear combination(“sum”) of the s orbitals on each of the two atoms, labeledA and B, as follows (where cA and cB are numerical co-efficients that can be evaluated by quantum mechanicalcalculation):

ψMO = cA ψA + cB ψB . (4)

According to the Pauli exclusion principle, only two elec-trons are permitted in this or any molecular orbital.

This molecular orbital could be illustrated by drawingits surface of constant ψ . It is more common just to showthe stylized surfaces of the two atomic orbitals and imag-ine that the surface of the molecular orbital is the boundaryenvelope of those atomic orbitals, as suggested in Fig. 2. Ifthe probability, Eq. (3), associated with ψMO is evaluated,it is found that the two electrons in this molecular orbitalare more likely to be found in the overlap region, the re-gion of space between the two nuclei where the atomicorbitals overlap. Since this is a region where the elec-trons are attracted to both positively charged nuclei, theprobability increase represents a stabilization. This is themolecular orbital counterpart of the stabilization associ-ated with each hydrogen’s achievement of its “octet.”

Molecular orbitals [Eq. (4)] can be formed between twoatomic orbitals on any adjacent atoms A and B. With p or-bitals there are two possible orientations of the atomicorbitals. They can overlap end-on, as in Fig. 3a, or theycan overlap sideways, as in Fig. 3b. In the former casethe molecular orbital is designated as sigma ( σ , “cylindri-cal”), and in the latter it is designated as pi (π ) because thenodal plane of the atomic orbitals is preserved. Pi molec-ular orbitals must be used for the second bond of a dou-ble bond and for the second and third bonds of a triplebond.

E. Hybrid Orbitals

Placing two electrons into any of these molecular orbitalsleads to stabilization since the electron probability is in-creased in the overlap region, where the electrons are at-

FIGURE 2 Molecular orbital of the H2 molecule, composed fromthe two individual atomic orbitals.

FIGURE 3 Two p orbitals forming (a) sigma and (b) pi molecularorbitals.

tracted to two positive charges. However, because s and patomic orbitals are symmetric, there is a limit to the ex-tent to which that probability can increase. For any s orp orbital on atom A half of its probability is on the sideopposite to atom B. To overcome this limitation, the sand p orbitals on an atom can be constructed into hybridorbitals, defined as follows, where ψs , ψx , ψy , and ψz arerespectively s, px , py , and pz orbitals (Fig. 1), and the c’sare coefficients:

ψhybrid = cs ψs + cx ψx + cy ψy + cz ψz . (5)

Figure 4 shows that adding s and py orbitals leads to anincreased probability of finding the electron to the rightof the atom because in that region of space the s and py

orbitals reinforce, whereas to the left the positive value ofthe s orbital tends to cancel the negative value of the py

orbital and reduce the value of ψhybrid and also its square,which is the probability. The surface of constant ψhybrid isreadily drawn, but usually a stylized version is adequate,as in Fig. 4.

These hybrids are then used to form the molecular or-bitals, as in Eq. (4) but with a ψhybrid formed from atomicorbitals on atom A or B in place of ψA or ψB. If thosehybrids point toward each other, as in Fig. 5, they canform a sigma bond of increased strength and stability be-cause of an increased probability for the electrons to befound in the overlap region. The surface of constant ψMO

is quite elaborate, and only the individual hybrid orbitalsare shown, to suggest their boundary envelope.

Many hybrids are possible, depending on how the coef-ficients are chosen. However, the number of distinct hybridorbitals must equal the number of atomic orbitals used toconstruct them. In a hybrid designated as spλ, the super-script λ represents the ratio of p character to s character interms of the coefficients that enter Eq. (5):

λ = (

c2x + c2

y + c2z

)1/2/cs . (6)

The most common hybrids are sp1 (customarily writtensimply sp), sp2, and sp3. No hybridization is possible toimprove sideways overlap, so that one p orbital must be

FIGURE 4 Hybrid atomic orbital formed from s and py orbitals.




FIGURE 5 Two hybrid orbitals forming a sigma molecular orbital.

dedicated to each pi bond. The remaining orbitals are hy-bridized to improve directionality. Then the hybridizationof any (octet) atom is given by the following equation, interms of the number of pi bonds or multiple bonds that theatom is forming:

Hybridization = sp(3−Npi) = sp(3−Ndouble −2Ntriple) . (7)

According to this formulation, the double bond ofethene (4) is composed of both sigma and pi molecularorbitals. The pi orbital is formed from sideways overlap oftwo p atomic orbitals on the carbons, as shown in Fig. 6a.According to Eq. (7), the remaining orbitals are of sp2

hybridization. One of those hybrids on each carbon formsthe C C sigma bond, as shown in Fig. 6b. The other twosp2 hybrids on each carbon, along with the s orbitals onthe hydrogens, are used to form two C H sigma bonds,shown in Fig. 6c.

Hybridization affects bond lengths and electronegativ-ity. Because a p orbital has a node at the nucleus, an elec-tron in that orbital has zero probability of being found atthe nucleus, but there is no such restriction on an electronin an s orbital. Consequently, an s electron is closer to thenucleus and is held more tightly. Moreover, the distanceof an electron from the nucleus and its ease of removal in-crease with the degree of p character. One manifestation ofthis feature is that C H bond lengths decrease from 1.11 Ain ethane (CH3CH3, sp3) to 1.10 A in ethene (CH2 CH2,sp2) to 1.08 A in ethyne (HC CH, sp). Another is that theelectronegativity of the carbon increases from sp3 to sp2

to sp.The advantage of the orbital approach is that it provides

a geometric picture. Since the sp2 hybrid orbitals in ethene(Fig. 6) are constructed from p orbitals perpendicular tothe one used for the pi bond, the molecule must be planar.However, for most purposes this approach is equivalent tothe Lewis approach, where a molecular orbital [Eq. (4)]is symbolized simply by a line joining atoms A and B,regardless of whether it is sigma or pi.

FIGURE 6 Molecular orbital description of ethene, CH2 CH2,viewed lying in a horizontal plane. (a) Atomic orbitals forming thepi molecular orbital. (b) Hybrid sp2 orbitals forming the C C sigmamolecular orbital. (c) Hybrid sp2 orbitals on carbon and s orbitalson hydrogens forming the C H sigma molecular orbitals.

F. Antibonding Orbitals

When two atoms A and B interact, there are two solutionsto Eq. (4), or two sets of coefficients cA and cB. One ofthose solutions is the bonding molecular orbital that allowselectron density to accumulate between A and B. The othersolution has a nodal surface between these atoms, andelectrons are excluded from this region. This molecularorbital is said to be antibonding.

Usually such orbitals are empty, but this conclusion de-pends on the number of electrons available. If two heliumatoms interact, there are a total of four electrons, only twoof which are permitted in the bonding molecular orbitalaccording to the Pauli exclusion principle. The other twomust occupy the antibonding orbital. According to calcu-lations, this orbital is more antibonding than the bondingorbital is bonding. The result is a net repulsion betweenthe two helium atoms whenever they approach so closelythat their orbitals overlap. Such repulsion between stableatoms or molecules is quite general. It is often called vander Waals repulsion or steric repulsion.

G. Molecular Orbital Theory

A unique advantage of molecular orbitals is that they pro-vide an alternative to resonance theory for describing de-localized electrons. Equation (3) is generalized as follows,where ψ1, ψ2, . . . , and ψN are atomic orbitals on atoms1 through N and the coefficients c1–cN are evaluated byquantum mechanical calculation:

ψMO = c1ψ1 + c2ψ2 + · · · + cN ψN . (8)

Usually only p orbitals are treated this way, and the sigmabonds are treated by one of the previous approaches. Thereare actually N different molecular orbitals, each with itsset of coefficients. Some of these carry two electrons eachand some are empty. For any of these the surface of con-stant ψMO is quite elaborate. Figure 7 suggests the bound-ary envelopes of two of the three filled pi molecular or-bitals of benzene (21). Because these molecular orbitalsextend over the entire molecule, they automatically allowfor the electrons to be delocalized.

Calculations lead to the conclusion that cyclic sys-tems containing 4n + 2 (n = 0, 1, 2, . . .) pi electrons are

FIGURE 7 Two filled pi molecular orbitals of benzene, viewedlying in a horizontal plane and showing the p atomic orbitals. Hy-drogens are omitted and carbons are not labeled.




unusually stable. Benzene (21 or 24; Fig. 7), with sixpi electrons in the three pi bonds, is an example. Theconclusion is very general, and other examples includecyclopentadienyl anion (25; 4n + 2 = 6), pyrrole (26;4n + 2 = 6, including the lone pair), cyclopropenyl cation(27; 4n + 2 = 2), and naphthalene (28; 4n + 2 = 10 aroundperiphery). In contrast, cyclic systems containing 4n(n = 1, 2, . . .) pi electrons, such as cyclobutadiene (29;4n = 4) are unusually unstable. The unusual stability ofcyclic systems containing 4n + 2 pi electrons is called aro-maticity, and the instability of cyclic systems containing4n pi electrons is called antiaromaticity. This is a dis-tinction not accessible from considerations of resonancehybrids.

Owing to the power of modern computers, it is now pos-sible to calculate electronic structures, geometries, ther-modynamic properties, and dynamic behavior of largemolecules with a reliability of ±1 kcal/mole. The meth-ods are quantum mechanical, of greater generality thanEq. (8). Software programs are widely available for per-forming such computations and for visualizing the results.

II. MOLECULAR STRUCTURE:STEREOCHEMISTRY

Stereochemistry is that aspect of molecular structure thatdeals with the geometry of molecules in three dimensions.Isomers are different molecules with the same atomiccomposition (same number of each kind of atom). Forexample, both CH3OCH3 and CH3CH2OH have two car-bons, six hydrogens, and an oxygen but they are assembleddifferently. Likewise, CH3CH2CH2CH3 and (CH3)3CHhave the same atoms but different connectivity. Isomersthat differ in connectivity are often called constitutionalisomers. They have different Lewis structures. In contrast,stereoisomers have the same connectivity and the sameLewis structures but they are still different molecules.

Even the simplest aspects of stereochemistry often re-quire a method for displaying the three-dimensional struc-ture of molecules on a two-dimensional page. One suchmethod is to use ordinary lines for bonds that are in theplane of the page, a wedge for a bond that comes forward,and a dotted line for a bond that goes behind the page.Another possibility is to draw a perspective view of themolecule, with thicker lines for bonds at the front (ratherthan shortening distant bonds, as in true perspective), orwith bonds in front “concealing” bonds behind. Still an-other possibility is to sight along a single bond, show thebonds at the front atom as emanating from that atom, and

show the bonds at the rear atom as partially hidden by thesigma bond, depicted as a circle. All of these methods areused below.

A. Bond Distances and Bond Angles

Bond distances are governed by the interplay between theaccumulation of the electrons in the overlap region andthe repulsions of the electrons for each other and of thenuclei for each other. For each bond there is an optimumdistance that minimizes the total energy. That distanceshortens with the number of electrons bonding the twoatoms together. For example, the C C distances in ethane(CH3 CH3), ethene (CH2 CH2), and ethyne (HC CH)are 1.54, 1.34, and 1.20 A , respectively.

The optimum values for bond angles are governedlargely by the mutual repulsion of electron pairs. Foran octet atom it might be expected that its four electronpairs would be as far apart from each other as possible.The optimum arrangement then is to direct them towardthe vertices of a regular tetrahedron. Methane (CH4; 30)is exactly such a molecule. The HCH angle is tetrahe-dral, 109.5◦. Likewise, sp3-hybridized carbons in othermolecules have angles that are tetrahedral or nearly so.Angles at nitrogen and oxygen, as in NH3 (6), H2O (7),and related molecules, are also close to tetrahedral, to min-imize the repulsions of the four electron pairs, regardlessof whether they are in bonds or lone pairs. The deviationsfrom the idealized 109.5◦ are due to a greater s characterfor lone pairs, which are held closer to the nucleus, or agreater p character for electrons that are unequally sharedwith a more electronegative atom.

Since every octet atom has four electron pairs, it mightbe thought that every such atom is tetrahedral. However, anunhybridized p orbital must be used for each pi bond andthe remaining electron pairs then minimize their mutualrepulsions. For an sp2-hybridized atom, with three suchpairs, the angle between them is 120◦, and with two, onan sp-hybridized atom, it is 180◦. Thus the HCH anglein ethene (CH2 CH2) and the CNC angle in CH2 NCH3

(31) are both close to 120◦ and the HCC angle in ethyne(HC CH) is 180◦.

B. Conformations about Single Bonds

Conformations are different molecular structures relatedby rotation about single bonds. Since a sigma bond iscylindrically symmetric, there is little restriction to ro-tating around it. The restriction does not come from the




FIGURE 8 Energy of ethane versus dihedral angle (defined aszero for a staggered conformation).

bond itself but from the interactions of the adjacent bondswith each other. In the simplest case, ethane, the moststable conformation is 32. This is called the staggeredconformation. The least stable is called the eclipsed con-formation (33) because one hydrogen is directly behindthe other (even though the drawing offsets them so thatthe rear hydrogens can be seen). The eclipsed conforma-tion is 2.9 kcal/mole less stable than the staggered one.Figure 8 shows a graph of the conformational energy ver-sus dihedral angle, the angle of rotation of one methylgroup relative to the other. There are other staggered con-formations (32′ and 32′′) at 120◦ and 240◦ as well as othereclipsed conformations (33′ and 33′′) at 180◦ and 300◦.

Eclipsed conformations are destabilized by torsionalstrain. In ethane this arises from a subtle interaction of theelectrons in one C H bond with the electrons in the C Hbonds on the adjacent carbon. Those electrons repel, andthe repulsion is minimized in the staggered conformation.

For butane there are two different kinds of staggeredconformations, one called anti (34), with the two methylgroups as far apart as possible, and the other called gauche(35), with the methyls adjacent. Neither of these has tor-sional strain. The anti is most stable. The gauche is desta-bilized by steric repulsion, which arises because one of thehydrogens of one methyl is so close to one of the hydro-gens of the other methyl that their orbitals overlap. Thisdestabilization amounts to 0.9 kcal/mole. The conforma-tional energy is graphed in Fig. 9. At a dihedral angle of 60◦

there is an eclipsed conformation (36) that is 3.4 kcal/moleless stable than the anti and 2.5 kcal/mole less stable than

FIGURE 9 Energy of butane versus dihedral angle (defined aszero for the anti conformation).

the gauche owing to torsional strain. At 180 ◦ there is an-other eclipsed conformation that is ∼5 kcal/mole less sta-ble than the anti owing to both torsional strain and stericrepulsion.

A conformation that is a minimum on such an en-ergy diagram is called a conformational isomer, or con-former. However, it is not customary to consider dif-ferent conformers as isomeric or stereoisomeric becausethey interconvert very quickly. The energy required to ro-tate one group relative to another comes from the ran-dom thermal energy of molecular motion. If only 2.9 or3.4 or 2.5 kcal/mole must accumulate, this happens veryoften (within picoseconds). Thus a chemical species isoften present as a mixture of conformers. In some specialcases conformers are interconverted “slowly” and theseare called atropisomers.

C. Isomerism about Double Bonds

In contrast to single bonds, rotation about double bonds isseverely restricted. The second bond of a double bond isa pi bond, from sideways overlap of p orbitals on adjacentatoms. Figure 6 illustrates ethene, redrawn as 37. To rotateone end of the double bond requires twisting one p orbitalso that it does not overlap with the other, as in 38. Thisstructure is less stable by the energy of the pi bond, or∼60 kcal/mole. Random thermal energy cannot providethis much, so rotation around double bonds ordinarily doesnot occur.




Consequently some molecules with double bonds canexist as a pair of stereoisomers, two different moleculeswith the same connectivity. For example, 2-butene existsas both (Z)-39 and (E)-39. These designations rely on theCahn–Ingold–Prelog priority rules. Older terminology isto call these cis and trans, respectively, and this sort ofisomerism, arising from sidedness on a double bond, isoften called cis/trans isomerism.

The (Z ) isomer is 1.0 kcal/mole less stable than the (E)owing to steric repulsion between the two methyl groups.With cycloalkenes, such as cyclooctene and smaller rings,the (Z ) or cis isomer is much more stable because bondangles are severely strained if a short chain must span thecarbons on opposite sides of the double bond of the (E)isomer (40; n = 5–8). Likewise, bicyclic alkenes like 41are severely strained by a double bond that can be (Z ) orcis in one (six-membered) ring but must be (E) or trans inthe other. The instability of such molecules is often calledBredt’s rule.

Amides (42) represent an intermediate situation. Be-cause of the additional resonance form 43 there is partialC–N double-bond character. Rotating the NHR′ relativeto the RC( O) prevents the lone pair on the N from over-lapping with the p orbital on carbon, similar to 38. Thetwisted structure is less stable by the energy of the partialdouble bond, or ∼20 kcal/mole. Consequently there is an-other stereoisomer, 44. The two stereoisomers interconvertrapidly at room temperature but they can be recognized asseparate species by NMR techniques.

D. Stereoisomerism from Stereo Centers

A tetrahedral atom with four different groups attached isnot superimposable upon its mirror image. Such an object

is said to be chiral, and the atom is called a stereocen-ter. Lactic acid (2-hydroxypropionic acid, 45, or its mirrorimage, 45′) is such a molecule. No matter how 45′ is ro-tated, it cannot be placed on top of 45 so that all atomscorrespond. They have different configurations, or spe-cific three-dimensional arrangements of the four groupsaround the central carbon. According to designations thatrely on the Cahn–Ingold–Prelog priority rules, 45 is the(S) configuration and 45′ is (R).

A molecule that is superimposable on its mirror im-age is said to be achiral. Propionic acid (46) is such amolecule. There are two simple structural features eitherof which guarantees achirality. One of these is a plane ofsymmetry through the middle of the molecule such thateach atom not on that plane is matched by an identicalatom reflected through that plane. The plane of the paperis a plane of symmetry for 46. The other feature is an in-version center, a point in the middle of the molecule suchthat each atom is matched by an identical atom equidis-tant from that point and in the exact opposite direction.A molecule with this feature is (R, S)-1,2-dibromo-1,2-dichloroethane (47). This has two stereocenters but is nev-ertheless achiral. A molecule that is achiral despite havingstereocenters is said to be meso.

A chiral molecule and its (nonsuperimposable) mir-ror image are two different molecules with the sameconnectivity. They are therefore stereoisomers. Suchstereoisomers are said to be enantiomers. Enantiomers arevery similar to each other, with the same melting point,boiling point, solubility, spectral characteristics, stabil-ity, and reactivity (except toward other chiral molecules).They differ in their ability to rotate the plane of polarizedlight, a property known as optical activity. One enantiomerwill be dextrorotatory, rotating the plane of polarization tothe right, whereas the other will be levorotatory, rotatingthe plane to the left. These are designated with prefixes(+)- and (−)-, respectively. The (S)-2-hydroxypropionicacid (45) is dextrorotatory (+)-2-hydroxypropionic acid.There is no universal relation between the R or S desig-nation and the rotatory power.

Chirality is a necessity for optical activity, but it is nota guarantee. A 50:50 mixture of two enantiomers is opti-cally inactive because the dextrorotatory and levorotatorypowers exactly cancel. Such a material is called a racemic




mixture, designated by a prefix (±)-. The process of sepa-rating the individual enantiomers from a racemic mixtureis called resolution. The conversion of an optically activematerial into its racemic mixture is called racemization.

Diastereomers are stereoisomers that are not mirrorimages of each other. Cis/trans isomers like 39 are di-astereomers. Another form of diastereomerism occurswith molecules that have two or more stereocenters.For example, a second stereoisomer of (R,S)-1,2-dibromo-1,2-dichloroethane (47) is (R,R)-1,2-dibromo-1,2-dichloroethane (48) and a third is (S,S)-1,2-dibromo-1,2-dichloroethane (48′). These last two are enantiomersof each other, but neither is the enantiomer of 47. Eachmust then be a diastereomer of 47. Even though diastere-omers have identical connectivity and therefore all thesame bonds, they are genuinely different from each other,with different melting points, boiling points, solubilities,spectral characteristics, stability, and reactivity.

This characteristic of diastereomers makes possible theresolution of a racemic mixture into its enantiomers. Theprocess also takes advantage of the fact that chiral naturalsubstances exist as single enantiomers. If the racemic mix-ture of (+) and (−) enantiomers can react with a naturalsubstance that is (for definiteness) the (R) stereoisomer,the products are (+, R) and (−, R). These are diastere-omers, with different solubilities, so they can be separatedby treatment with a suitable solvent. Then the (+) enan-tiomer can be recovered from the (+, R) product and the(−) enantiomer from the (−, R).

E. Cyclic Compounds

Cyclic compounds can show various combinations ofthese stereochemical features, including conformations,stereocenters, chirality, enantiomerism, diastereomerism,and another form of cis/trans stereoisomerism. The con-formational behavior of cycloalkanes, CnHn , depends onthe ring size. For n = 3, 4, or 5 the small ring constrainsthe CCC angle to <109.5◦, leading to an instability calledangle strain that arises from increased repulsion betweenelectron pairs and poorer overlap. Also, there is torsionalstrain along each C C bond, which is eclipsed in the pla-nar conformation (e.g., 49). For n = 4 or 5 small distor-tions from planarity reduce the torsional strain but at theexpense of an increase of angle strain.

A disubstituted cycloalkane can have both of the sub-stituents on the same face of the ring, as in 50, or theycan be on opposite faces, as in 51 (hydrogens omitted in

both). These are diastereomers. The former is called cisand the latter is trans. If R1 �= R2, both of these are chiral,so each can exist as a pair of enantiomers. The configura-tion at each stereocenter can be designated as (R) or (S).If R1 = R2, only 51 is chiral, since 50 has a vertical planeof symmetry between the two CHR carbons and is meso.

Cyclohexane can eliminate both angle strain and tor-sional strain by distorting from coplanarity to the chairconformation (52). This has the feature of axial and equa-torial hydrogens in two distinctly different environments,as depicted in 52′ and 52′′, respectively. Rotation aboutthe single bonds converts 52 into a different chair confor-mation, 53, whereby each hydrogen that was axial (e.g.,the boldface H at the rear of 52) becomes equatorial (bold-face in 53) and each hydrogen that was equatorial becomesaxial. This process, called ring inversion, occurs quickly(within microseconds at 25◦C).

A monosubstituted cyclohexane thus exists as a rapidlyequilibrating mixture of two conformers (54, 55, mosthydrogens omitted). The conformer with a bulky R sub-stitutent in the axial position (55) is destabilized by stericrepulsion with the two axial hydrogens nearby.

A disubstituted cyclohexane can be either cis or trans.If the substitution pattern is 1,4 (56, 57), these are diastere-omers. Both are achiral, as guaranteed by a vertical planeof symmetry through carbons 1 and 4. Each can undergoring inversion, so each is a mixture of two conformers,only one of which is shown. If the substitution patternis 1,2 or 1,3 and if R1 �= R2, the cis and trans moleculesare both chiral. If R1 = R2, the cis-1,3-disubstituted cyclo-hexane is achiral, but the trans-1,2, the trans-1,3, and thecis-1,2 (58) are all chiral. However, an unusual feature ofthis last is that ring inversion rapidly converts it to another




conformer (59) that is the enantiomer of 58, so that thiswill always be present as a racemic mixture.

F. Molecular Mechanics

Each molecule has its own preferred geometry, determinedby the various interatomic forces. If the preferences foroptimum bond distances and bond angles and for mini-mum torsional strain and steric repulsion are character-istic of the atoms involved and independent of the restof the molecule, then it becomes possible to extend ourunderstanding from one set of molecules to others. It isnecessary to express the dependence of energy on molec-ular geometry in a form like Eq. (9), where V ({xi , yi , zi })depends on the positions of all the individual atoms, andwhere Vstretch [Eq. (10)] represents the energy to distortthe distance between atoms i and j from the optimum d0

i j ,Vbend [Eq. (11)] and Vtorsion [Eq. (12)] represent the energyto distort the ijk bond angle or the ijkl dihedral angle fromits optimum, and Ves and VvdW are electrostatic and vander Waals (steric) energies, respectively, which depend onthe distance between atoms i and j :

V ({xi , yi , zi }) =∑

Vstretch(di j ) +∑

Vbend(θi jk)

+∑

Vtorsion(θi jkl) +∑

Ves(di j )

+∑

VvdW(di j ), (9)

Vstretch(di j ) = 12 ki j

(

di j − d0i j

)2, (10)

Vbend(θi jk) = 12 ki jk

(

θi jk − θ0i jk

)2, (11)

Vtorsion(θi jkl) = 12 ki jkl

(

θi jkl − θ0i jkl

)2. (12)

The sums are over all bonds, angles, and atom pairs. Theparameters that enter each of these terms can be cal-culated for model compounds by quantum mechanics.Alternatively they can be calibrated to give the best fitto an extensive set of experimental data on molecular ge-ometries, stabilities, and infrared vibrational frequencies.

These calculations are formidable, but they are easy ona computer. Moreover, programs are available, completewith parametrization. Thus it is possible to calculate theenergy for any arbitrary molecule in any geometry and toseek the geometry that minimizes that energy. It is furtherpossible to compare that minimum with the energies of

isomeric molecules and to calculate how the energy variesas the molecule is distorted to some other geometry.

III. CHEMICAL REACTIVITY

There are two aspects of chemical reactivity, static and dy-namic. The static aspect relates to the question of chemicalequilibrium, determined by the stability of reactants andproducts. The dynamic aspect relates to the question ofchemical kinetics, or rates of reactions. This is more com-plicated, but it can be converted into static terms.

A. Chemical Equilibrium

For a general reaction, written as Eq. (13), the equilibriumconstant is related to concentrations and to free energies byEq. (14), where �G ◦ = G ◦B − G ◦A , R is the gas constant,8.314 J/mole or 0.001987 kcal/mole, and T is the absolutetemperature:

A ⇀↽ B, (13)

K = [B]/[A] = exp(−�G ◦/RT ). (14)

Much of our understanding of chemical reactions comesfrom reasoning by analogy. If the equilibrium constant Kis known for some standard reaction, it is often possibleto predict, at least qualitatively, the equilibrium constantK ′ for a related reaction, involving some modification ofthe molecular structure. To do so, it is necessary to knowhow the modification affects energies.

For the modified reaction of Eq. (13′) if the modificationstabilizes B′ (distinguished with a prime) relative to B,then it follows from Eq. (14) that K ′ > K and that theequilibrium is shifted to the right:

A′ ⇀↽ B′. (13′)

Likewise, if the modification destabilizes A, or raises itsenergy relative to A, then K ′ > K and the equilibriumis again shifted toward the right. If the modification sta-bilizes A′ relative to A or destabilizes B′ relative to B,then K ′ < K and the equilibrium is shifted toward the left.These conclusions are made graphic in Fig. 10.

Bond strengths represent a simple case of energetics thataffect equilibrium. They are usually expressed as bond-dissociation energies (BDEs), positive numbers that cor-respond to the energy required to break a molecule intoits constituent fragments, or the energy released when thebond forms. Table I lists some representative values.

Such values can be used to understand equilibria. Forthe reaction of Eq. (15) the total bond-dissociation energyof the C H and Cl Cl bonds on the left is 162 kcal/mole(104 + 58), whereas that on the right is 187 kcal/mole




FIGURE 10 Effects of energy modifications on position of equi-librium A′ ⇀↽ B′ relative to A ⇀↽ B: (a) A′ destabilized or B′ stabilized(b) A′ stabilized or B′ destabilized.

(84 + 103). Therefore more energy is released when theproducts on the right are formed, meaning that this reac-tion is exothermic by 25 kcal/mole. In contrast, the corre-sponding reaction with iodine, Eq. (15′), is endothermicby 13 kcal/mole (104 + 36 − 56 − 71):

CH3 H + Cl Cl ⇀↽ CH3 Cl + H Cl, (15)

CH3 H + I I ⇀↽ CH3 I + H I. (15′)

The modification of changing chlorine to iodine shifts theequilibrium to the left because the C I and H I bondsof the modified product are less stable than the C Cl andH Cl bonds of the unmodified product.

Other bond-dissociation data that are widely tabulatedare acid-dissociation constants [Eq. (16), where HA is anacid in the reaction of Eq. (17)]. In contrast to the (ho-molytic) bond dissociations in Table I, these are heterolyticbecause when the H A bond is broken, both electronsremain with A. Because acid-dissociation constants canrange over many orders of magnitude, from the strongestacids, where Ka ≈1020, to the weakest, where Ka ≈10−50,the data are always tabulated in terms of pKa [Eq. (18)]:

TABLE I Some Important A B Bond-Dissociation Energies(kcal/mole)

H CH3 OR F Cl Br I

H 104 104 110 135 103 87 71

CH3 104 88 85 108 84 68 56

OR 35

F 38

Cl 58

Br 46

I 36

Ka = [H+][A−]/[HA], (16)

HA + H2O ⇀↽ H3O+ + A−, (17)

pKa = −log10 Ka. (18)

Acidity increases from left to right across the periodictable. For example, the pKa values of CH3 H, H2N H,HO H, and F H are ∼50, 32.5, 15.7, and 3.2, respec-tively. This is a consequence of increasing electronegativ-ity, which stabilizes the anions F− > OH− > NH−

2 > CH−3 .

Similarly, the acidity of hydrocarbons increases withcarbon electronegativity from ethane (CH3CH3, sp3) toethene (CH2 CH2, sp2) to ethyne (HC CH, sp). Ofcourse electronegativity and acidity also increase withincreasing positive charge, as for H2O < H3O+ andNH3 < NH+

4 .

B. Chemical Kinetics

Just because a reaction is exothermic or has a favorableequilibrium constant does not mean that it will occur. Thereaction of Eq. (15) is exothermic by 25 kcal/mole andhas a very large equilibrium constant, but if the reactantsare mixed in the dark at room temperature, no reactionoccurs. To understand such phenomena it is necessary tostudy chemical kinetics, the rates of chemical reactions.

For a general reaction [Eq. (19), slightly more elabo-rate than Eq. (13)] the rate of reaction can be defined byany of the forms of Eq. (20). The derivatives representthe decreasing concentration of reactants (hence the mi-nus signs) or the increasing concentration of products astime passes. It is often found that the rate of reaction isproportional to some power (nAB or nC) of the concen-trations of reactants (and perhaps to the concentrations ofother species, such as catalysts), as in Eq. (21):

A B + C → A + B C, (19)

rate = −d[AB]

dt= −d[C]

dt

= d[A]

dt= d[BC]

dt, (20)

rate = k[A B]nAB [C]nC · · · . (21)

The constant of proportionality k is called the rate con-stant. It is a measure of how fast the reaction is.

C. The Transition State

In order to understand rate constants, it is necessary tospecify a measure of the progress from reactants to prod-ucts. This measure is a geometric parameter called thereaction coordinate. It is a composite of bond distances,chosen to increase in value from reactants to products. For




the reaction of Eq. (19) an appropriate reaction coordinateQ is given as follows since the distance between B andC is long in reactants and short in products, whereas theA B distance increases as reaction proceeds:

Q = dA B − dB C . (22)

Therefore Q is a large negative number in the reactantsand rises to a large positive number in the products.

The key to understanding reaction rates is the depen-dence of energy upon Q. As C approaches A B, the en-ergy must increase, from two sources: van der Waals re-pulsion arising from overlap of the electron clouds, andan endothermicity from starting to break the A B bond.This energy continues to increase until eventually, at largeQ, the energy will decrease as the B C bond developsand as the van der Waals repulsion between B C and A isrelieved. Somewhere, in between, near Q = 0, when bothbonds are stretched about equally, the energy reaches amaximum, as sketched in Fig. 11.

The position of the maximum is called the transitionstate, symbolized with a double dagger, ‡. This is the keyconcept for understanding chemical reactivity. It is a spe-cific point along the reaction coordinate Q, or a uniquegeometry, with particular values of dA B and dB C. It isthe geometry at which the energy is maximum.

The structure of the transition state cannot be describedusing simple Lewis structures since there are delocalizedelectrons. A pair of electrons is not localized in either theA B bond, which is being broken, or in the B C bond,which is being formed. The transition state can be de-scribed as a resonance hybrid of two resonance forms,60 and 61. Alternatively, in the shorthand form of 22–24it can be symbolized as 62, where the dotted lines sym-bolize the partial bonds that are breaking or forming. Yetaccording to the rules of resonance, only the electrons aredelocalized. The transition state has its unique geometry,

FIGURE 11 Variation of energy along reaction coordinate Q forthe reaction A B + C → A + B C.

with nuclei fixed at the particular value of Q where theenergy is maximum.

That maximum energy of the transition state lies abovethat of the reactants by an amount �G‡, equal to G‡ −Greactants, as shown in Fig. 11. This energy difference iscalled the activation energy or the free energy of activation.The rate constant of Eq. (21) can be related to this energydifference through the following equation, where kB isBoltzmann’s constant (1.381 × 10−23 J/K), h is Planck’sconstant (6.626 × 10−33 J-sec), R is the gas constant, andT is the absolute temperature:

k = (kBT/h) exp(−�G‡/RT ). (23)

This equation arises because �G‡ is the amount of energythat the reacting molecules must acquire in addition to thatof the reactants. It is the barrier that must be overcome.This energy is used to push C into A B and to break theA B bond sufficiently for the B C bond to start forming.That energy is derived from the accumulation of randomthermal motions. The average thermal energy is RT , orabout 0.6 kcal/mole at room temperature. Some moleculeshave less than average and some have more. Very few haveas much as �G‡, but those that do will react. Conversionof Eq. (23) into Table II shows how sensitive the rate ofreaction is to both �G‡ and to T . A low barrier corre-sponds to a reaction that is very fast. A reaction with ahigh barrier is very slow, but increasing the temperaturemakes it go faster.

Much of our understanding of chemical reactivitycomes from reasoning by analogy. If the rate constant kis known for some standard reaction, it is often possibleto predict, at least qualitatively, the rate constant k ′ for arelated reaction involving some modification of molecular

TABLE II Rate Constants and Activation Energy at DifferentTemperatures

At 25◦◦ At 100◦◦∆G‡‡

(kcal/mole) k (sec−1) Half-life k (sec−1) Half-life

0 6.21 × 1012 0.112 psec 7.78 × 1012 0.089 psec

5 1.34 × 109 516 psec 9.17 × 109 75.6 psec

10 2.91 × 105 2.39 µsec 1.08 × 107 64.1 nsec

15 6.28 × 101 11.0 msec 1.28 × 104 54.4 µsec

20 1.36 × 10−2 51.0 sec 1.50 × 101 46.1 msec

25 2.94 × 10−6 65.5 hr 1.77 × 10−2 39.1 sec

30 6.35 × 10−10 34.5 years 2.09 × 10−5 9.21 hr

35 1.37 × 10−13 160 kyears 2.47 × 10−8 325 days

40 2.97 × 10−17 739 Myears 2.91 × 10−11 755 years




FIGURE 12 Effects of energy modifications on rate of reaction ofA′ relative to A via transition state ‡ or ‡′: (a) A′ destabilized or ‡′stabilized, (b) A′ stabilized or ‡′ destabilized.

structure. To do so, it is necessary to know how the mod-ification affects energies, just as for equilibria above. Fora general reaction written as follows, the rate constant isstill given by Eq. (23):

A → ‡ → B. (24)

For the following modified reaction (distinguished witha prime), if the modification stabilizes the transition state‡′ (not the product B′), then it follows from Eq. (23) thatk ′ > k and that the reaction is faster:

A′ → ‡′ → B′. (24′)

Likewise, if the modification destabilizes the reactant A′

relative to A, the reaction is again faster. If the modifica-tion stabilizes reactant A′ or destabilizes transition state‡′, then k ′ < k and the reaction is slower. These conclu-sions are made graphic in Fig. 12, which is very similar toFig. 10.

D. Electron Pushing

Because the transition state has an electronic structure de-scribable as a resonance hybrid, electron pushing, whichprovided a convenient method for generating additionalresonance forms, also provides a method for generatingtransition states. An electron pair is delocalized towardanother atom so as to form a new bond, while also remov-ing an electron pair from that adjacent atom if necessaryto avoid violating the octet rule. The electron pair cancome from a lone pair or from a multiple bond or a singlebond. For example, this method permits the generation ofthe transition states for the methoxide-induced E2 elimi-nation of HBr from ethyl bromide (63), the nucleophilicsubstitution of ammonia on methyl iodide (64), the elec-trophilic addition of H+ (from HBr) to propene (65), thehydride transfer from borohydride ion to methanal (66),and the cycloaddition of ozone to ethene (67).

In such reactions the atom that carries the electron pairis called a nucleophile since it uses its electron pair to forma bond to some other nucleus. The partner with which thenucleophile reacts is called an electrophile since it formsa bond with that electron pair. In some contexts these arecalled Lewis bases and Lewis acids, respectively. One ofthe unifying features of physical organic chemistry is therecognition that a very large proportion of all reactions canbe classified as ones where a nucleophile (or Lewis base)reacts with an electrophile (or Lewis acid). The vastness ofthe possible chemistry then arises from the great diversityof nucleophiles and electrophiles.

This method is also applicable to multistep reac-tions. For example, the hydrolysis of phenyl acetate,CH3C( O) OPh, with hydroxide to form CH3CO−

2 plusPhOH does not proceed via transition state 68, involvingsimultaneous breaking and making of the various bonds.Instead it is a three-step reaction, proceeding via threesequential transition states, 69–71, each of which can begenerated by electron pushing.

Between those transition states there are reaction inter-mediates 72 and 73. These are ordinary chemical species,not transition states, but they are not very stable and theydo not persist. Figure 13 shows how the energy varies dur-ing the course of the reaction, as measured by a reactioncoordinate that is a composite of the various bond dis-tances. The transition states are at local maxima and theintermediates are at local minima. In such a diagram thereis generally one transition state that is higher than the oth-ers. That one is called THE transition state and its step iscalled the rate-limiting step. In the hydrolysis of phenyl




FIGURE 13 Variation of energy along reaction coordinate Q forthe reaction RC( O) OPh + OH− → RCO−

2 + PhOH.

acetate it is the first transition state (69) that is highest, sothat the first step is rate-limiting.

One of the principal goals of physical organic chemistryis to determine the mechanisms of reactions. A mechanismis a detailed description of the sequence of steps, involvingbreaking and making of bonds, that converts reactants intoproducts, along with a description of the structure andenergy of reaction intermediates and transition states alongthe way.

A notational system for representation of mechanismshas been devised. Bond making is indicated as A (asso-ciation) and bond breaking as D (dissociation). A nucle-ophilic or nucleofugic (loss of a nucleophile) process ata core atom is indicated with a subscript N, or else E forelectrophilic or electrofugic. A subscript H indicates hy-drogen as electrophile or electrofuge at a core atom, but hif at a peripheral atom and xh if intermolecular. Simulta-neous processes are indicated by juxtaposition, and suc-cessive steps are separated by a +. Thus the nucleophilicsubstitution corresponding to 64 is DNAN, the eliminationcorresponding to 63 is AxhDHDN, and the sequence of69–71 is AN + DN + AxhDh. In the older (Ingold) systemthese were designated as SN2, E2, and BAc2, respectively.

E. Inductive Effects

One of the simplest modifications that can affect chemicalreactivity is the introduction of a substituent that produces

an electrostatic effect. The energy of interaction of twopoint charges q1 and q2 with each other or of one pointcharge with a dipole moment µ [Eq. (2)] is given as fol-lows, where r is the distance between the two, ε is thedielectric constant of the solvent, and θ is the angle be-tween the direction of the dipole and the direction towardthe charge:

Echarge-charge = q1q2

εr, (25)

Echarge-dipole = q µ cos θ

εr2. (26)

These correspond to an energy lowering when two op-posite charges are near each other or when the charge iscloser to the end of the dipole of opposite charge, and anenergy increase when these are reversed.

Table III lists acidity constants of some carboxylic acidsRCOOH. Entry 1 is the reference acid, and the othersdiffer because of electrostatic effects on the stability of theconjugate base, RCO−

2 (not the stability of the reactant).For entry 2 the negative charge on CO−

2 is stabilized bythe positive charge on the NH+

3 [Eq. (25)], so that theequilibrium of Eq. (17) shifts toward the right (decreasedpKa), according to the considerations in Fig. 10. For entry3 the identical negative charges on the two CO−

2 groupsdestabilize each other [Eq. (25)], so that the equilibriumshifts toward the left (increased pKa relative to entry 1).For entry 4 the positive end of the C Cl dipole is closer tothe negative charge on CO −2 and leads to stabilization ofthe products [Eq. (26)] so that the equilibrium of Eq. (17)shifts toward the right, as for entry 2. Entry 5 is similar toentry 2 but the distance between the positive and negativecharges is larger, so that the effect is smaller, according tothe dependence of Eq. (25) on r .

These electrostatic effects are often referred to as in-ductive effects because they can also be viewed as arisingfrom the substituents’ pulling or pushing the electrons ofthe molecule and making it easier or harder for H+ to beremoved from the COOH. Inductive effects can also op-erate on rates, and they can be interpreted according toFig. 12. For example, hydrolysis of ClCH2C( O)OPh isfaster than that of CH3C( O)OPh because transition state

TABLE III Acidity Constants of Some CarboxylicAcids

Entry Acid pK a

1 CH3COOH 4.75

2 +H3NCH2COOH 2.85

3 −O2CCH2COOH 5.38

4 ClCH2COOH 2.87

5 +H3NCH2CH2COOH 3.60




74 is stabilized relative to 69 by the interaction betweenthe C Cl dipole and the negative charge that is distributedover the O and the OH.

F. Solvent Effects

Solvent effects also arise from electrostatic interactions,but between solvent and solute. The energy of interactionof an ion or dipole is given approximately as follows,where q is the charge, µ is the dipole moment, r is itsradius (the closest distance of approach between solventand solute), and ε is the dielectric constant of the solvent:

Esolv,ion = q2

r

ε − 1

2ε, (27)

Esolv,dipole = µ2

r3

ε − 1

2ε + 1 . (28)

The dielectric constant is an empirical measure of the po-larity of a solvent, or of how well it stabilizes ions ordipoles. The second factor in each equation correspondsto a greater stabilization with increasing ε.

A polar solvent achieves that stabilization by clusteringits own dipoles around the solute, as illustrated in Fig. 14.An anion or the negative end of a dipole is especially wellstabilized by water because the H is so small that its δ+can approach quite close to the negative charge. The smallr in Eqs. (27) and (28) then allows a large stabilization E .This is a general aspect of protic solvents, those with OHor NH groups. It is often attributed to hydrogen bonding,as though there were a bond between the H and the solute,but the interaction is largely electrostatic. In contrast, theδ+ of polar aprotic solvents, such as dimethyl sulfoxide,(CH3)2S O, is buried in the center of the molecule andcannot stabilize anions as well as the exposed δ− on theoxygen stabilizes cations.

Solvents have a large effect on reactions that create ordestroy ions. For example, the acidity constant (Table III)

FIGURE 14 Solvation stabilization of solute cations, anions, ordipoles by water molecules.

of acetic acid is 4.75 in water, the standard solvent for pKa,but is 9.65 in methanol. The dielectric constant of wateris 81, but methanol is a less polar solvent, with a dielec-tric constant of only 33. Therefore, according to Eq. (27),methanol provides less stabilization of the ions H+ andCH3CO−

2 . In terms of Fig. 10, the change from water tomethanol represents a relative destabilization of the prod-ucts of ionization and therefore a decreased equilibriumconstant (from 10−4.75 to 10−9.65).

Similarly, the solvolysis of tert-butyl chloride (SOH =H2O or C2H5OH) is 3 × 105-fold faster in water than inethanol as solvent:

(CH3)3C Cl → [

(CH3)3Cδ+ · · · Clδ−]‡

→ (CH3)3C+ + Cl− (29)

(CH3)3C+ + SOH → (CH3)3CO(H)S+

→ (CH3)3COS + H+. (30)

In this three-step reaction the first step is rate-limiting andits transition state has a large dipole moment, which canbe stabilized by a polar solvent [Eq. (28)]. In terms ofFig. 12, the change from methanol to water represents astabilization of the transition state and therefore a fasterreaction.

Much current research involves the investigation ofions in the gas phase, free of the influence of solvation.A dominant influence is the size of the ion becausea positive or negative charge constrained to a smallvolume repels itself strongly. Thus the order of gas-phase acidities of alcohols is CH3OH < CH3CH2OH <(CH3)2CHOH < (CH3)3COH because the negative chargeis distributed over a larger volume in (CH3)3CO−. This isexactly the opposite order from solution, where solvationof (CH3)3CO− is most hindered.

G. Delocalization Effects

Delocalization of electrons generally leads to a stabiliza-tion. This is certainly true with both resonance and aro-maticity, and the only exception is with antiaromaticity.Then, according to Fig. 10 or Fig. 12, delocalization canaffect the position of equilibrium or the rate of reactionthrough stabilizing (or destabilizing) reactant, product, ortransition state.

Table IV lists C H bond-dissociation energies of somehydrocarbons. All the other bonds are weaker than theC H of methane and require less energy to break. Thisweakening does not arise from differences in the hydro-carbons since all these C H bonds are the same in thatthey are formed from an sp3 orbital on carbon and an s or-bital on hydrogen. Instead the weakening must come fromstabilization of the radical that results from removing the




TABLE IV Some C H Bond-DissociationEnergies (BDE)

BDEEntry Bond (kcal/mole)

1 CH3 H 104

2 CH3CH2 H 98

3 (CH3)2CH H 95

4 (CH3)3C H 92

5 CH2 CHCH2 H 89

hydrogen. For entry 5 that radical (75) is stabilized by anadditional resonance form (76) which is not available forCH3 ·. According to the rules of resonance, 76 is not just75 flipped left-for-right, but it expresses the delocalizationof the pi electrons in the double bond.

For the ethyl radical of entry 2 (77), resonance is less ob-vious, but three additional resonance forms are 78, 78′, and78′′. According to the rules of resonance, these do not im-ply that hydrogens are removed. They symbolize only thedelocalization of electrons from the adjacent C H bondsto relieve the nonoctet character of the radical center at theexpense of a hydrogen which loses its octet. In contrastto the delocalization of pi electrons in 75–76, electronsin 77–78 are delocalized from C H sigma bonds. Thisphenomenon is often called hyperconjugation. Similarly,for entries 3 and 4 there are hyperconjugative resonanceforms that stabilize the radical that results from hydrogenremoval. For the 2-propyl radical, (CH3)2CH·, of entry 3,there are a total of six additional forms, since electrons ineach of the six adjacent C H bonds can be delocalized.For entry 4 there are a total of nine additional forms. Thusthe number of resonance forms for the radical increasesand consequently the bond-dissociation energy decreasesfrom entry 1 to 2 to 3 to 4.

The heterolytic bond-dissociation energies of R H toR+ + H− (or of R Cl to R+ + Cl−) also decrease alongthe series R = CH3 > CH3CH2 > (CH3)2CH > (CH3)3C.Again this is due to hyperconjugation, which delocal-izes electrons in the adjacent C H bonds and relieves thenonoctet character of the carbocation at the expense of ahydrogen which loses its octet. In CH3CH+

2 there are threeadditional resonance forms, in (CH3)2CH+ there are six,

and in (CH3)3C+ there are nine, paralleling the increasedstabilization.

Similarly, the solvolysis of tert-butyl chloride is fasterthan that of isopropyl chloride. The transition state for theformer is given in Eq. (29). That for the latter is given asfollows:

(CH3)2CH Cl → [(CH3)2CHδ+ · · · Clδ−]‡

→ (CH3)2CH+ + Cl− → etc. (31)

However, those structures are incomplete. The transitionstate from isopropyl chloride has six additional hypercon-jugative resonance forms, corresponding to delocalizationof electrons in six adjacent C H bonds. That from tert-butyl chloride has nine such forms, which confer greaterstability. In terms of Fig. 12, the change from isopropyl totert-butyl represents a stabilization of the transition stateand therefore a faster reaction.

Another example is cyclopentadiene (pKa 16), whichis remarkably acidic for a hydrocarbon. This can be at-tributed to electron delocalization in the conjugate base(79, C’s, H’s, and lone pairs omitted), where there arefive equivalent resonance forms, since the formal nega-tive charge can be at any carbon. However, cyclohepta-triene does not show any unusual acidity compared toother polyenes, even though its conjugate base (80) hasseven equivalent resonance forms. This contrast is theconsequence of the aromaticity of 79, owing to the sixpi electrons delocalized around the five-membered ring,whereas 80 is antiaromatic, with eight pi electrons.

H. Steric Effects

A steric effect arises from the sheer bulk of substituents.Just as two helium atoms repel, any two substituents re-pel whenever their orbitals overlap. In contrast to res-onance, which is always stabilizing, steric repulsion isalways destabilizing, although it can destabilize reactant,product, or transition state.

A simple example is the destabilization of o-xylene (81,R = CH3) or o-di-tert-butylbenzene [81, R = C(CH3)3]by 0.3 and 22.3 kcal/mole, respectively, relative to theirmeta isomers (82). Another is the destabilization of ax-ial conformer 55. A quantitative measure of the effec-tive bulk of a substituent R is the free-energy differ-ence, −RT ln([55]/[54]), often called the A value ofthe substituent. Some representative values are listed inTable V. These values can be used to estimate the posi-tion of the equilibrium between the two conformers of




TABLE V Energy Difference A =−RT ln([55]/[54]) between Axial and Equatorial R

R A (kcal/mole)

F 0.15

Cl 0.43

Br 0.38

CN 0.17

CH3 1.74

CH2CH3 1.79

CH(CH3)2 2.21

C(CH3)3 >5.4

COOH 1.35

CO−2 1.92

OCH3 0.60

a disubstituted cyclohexane like 56. Still another exam-ple is given by the reactivity of alkyl bromides towardsubstitution by hydroxide, which decreases in the orderCH3Br > CH3CH2Br � (CH3)2CHBr. This is due to in-creasing destabilization of the transition state (83, 83′,83′′) by repulsion between OH or Br and zero, one, or twoCH3 groups on the central carbon.

I. Strain Effects

Strain effects arise from the distortion of bond or torsionalangles. The strain energy V can be expressed as in Eq. (11)or Eq. (12), where θ0 is the optimum angle, of minimumenergy.

Like steric effects, strain effects are always destabiliz-ing, although they can affect reactant, product, or transi-tion state. For example, cycloheptyne (84) is less stablethan ordinary alkynes because the seven-membered ringprohibits the 180◦ angles preferred by the sp-hybridizedcarbons of the triple bond. Another example is the un-usual behavior (compared to ordinary ketones) of cyclo-propanone (85), which reacts with water to form cyclo-propanediol (86). Both of these have strain energy becauseof the 60◦ bond angle in a three-membered ring. How-ever, according to Eq. (11), the strain energy is greater for85, where the preferred bond angle of the sp2 carbon is120◦ (Vbend = 1

2 kCCC(60 − 120)2 = 1800kCCC), compared

to 86 (Vbend = 12 kCCC(60 − 109.5)2 = 1225kCCC, which is

smaller regardless of the magnitude of kCCC).

Torsional strain is responsible for the greater reactiv-ity of cyclopentanone (87, n = 5) toward borohydride, ascompared to that of cyclohexanone (87, n = 6), to form thealcohol (88, after protonation of the oxygen). The transi-tion state (89, n = 5) leading to the five-membered ringis destabilized by increased torsional strain along all fiveC C bonds relative to the reactant, where bonds to the car-bonyl carbon are not eclipsed. Transition state 89 (n = 6) isstabilized by permitting all six C C bonds to be staggered,eliminating any torsional strain. Therefore the transitionstate in the six-membered ring is more stable, accountingfor the faster reaction.

J. Stereochemistry and Reactivity

How do stereoisomers differ in reactivity? The answerlies in the energetics of stereoisomeric transition states.Certainly enantiomeric transition states must have identi-cal energies, since they are merely mirror images of eachother. Therefore if two enantiomers react with an achi-ral reagent, they must react at identical rates. Moreover,if two enantiomeric products are formed from the sameachiral reactant, they must be formed at equal rates andtherefore in equal proportion. Only a racemic product canbe formed. For example, the possible transition states foraddition of borohydride to 2-butanone are enantiomers 90and 90′. The hydride is transferred to either of the twofaces of the carbonyl group, but the two faces must beequally reactive.

In contrast, diastereomeric transition states have dif-ferent energies. Therefore diastereomers show unequalreactivity and they can be formed at different rates.For example, addition of borohydride to (R)-3-methyl-2-pentanone, CH3C( O)CH(CH3)C2H5, produces both(2R,3R)- and (2S,3R)-3-methyl-2-pentanol, but in un-equal amounts. The reactant is chiral and there is an en-ergetic preference for borohydride to be transferred to




one of the two faces of the carbonyl. Much current re-search is devoted to understand the steric interactions thatgovern these preferences, with the aim of finding con-ditions to select one diastereomeric transition state overanother.

Even if the reactant is (±)-3-methyl-2-pentanone, theproducts are equal amounts of (2R,3R)- and (2S,3S)-3-methyl-2-pentanol (a racemic mixture), plus equalamounts of (2R,3S)- and (2S,3R)-3-methyl-2-pentanol(another racemic mixture). However, the amount of thefirst pair is not equal to the amount of the second be-cause the transition states are diastereomeric. The twotransition states for formation of the first pair are enan-tiomeric, hence the racemic mixture. It is a general rulethat only achiral products or racemic mixtures can be ob-tained from achiral or racemic reactants. It is not possibleto obtain optically active product from optically inactivestarting materials. The mechanism for the origin of thesingle enantiomers of chiral natural substances is not yetunderstood.

K. Relation between Kineticsand Thermodynamics

There is no universal relation between kinetics and ther-modynamics. Just because a reaction is themodynamicallyvery favorable does not mean that it must be fast. Thereare many reactions that are very exothermic but proceedextremely slowly. However, there is enough of a relationbetween kinetics and thermodynamics to make it usefulfor understanding.

Very often it is observed that in a pair of related reac-tions the one that is more favorable thermodynamicallyis the faster one. When a more stable product is formedbecause the transition state leading to it partakes of thestability of the product, this is said to be a case of product-development control. For example, the rates of reactionof bromine atoms with the hydrocarbons of Table IV fol-low the series CH3 H < CH3CH2 H < (CH3)2CH H <

(CH3)3C H < CH2 CHCH2 H. This order parallels thebond-dissociation energies, in that the weaker the bond is,the faster it cleaves. The transition state for any of thesereactions can be written as a resonance hybrid of 91 and92, adapted from the generic 60 and 61. However, thereare additional resonance forms. Because 92 is a contribut-ing resonance form, so are forms like 76 or like 78, 78′,and 78′′, which provide additional delocalization and ad-ditional stabilization. Consequently the transition statesbecome increasingly stabilized and the reaction becomesfaster.

It is cumbersome to consider transition states, with alltheir resonance forms and partial bonds. Instead it is eas-ier to focus on products and recognize their similarityto the transition states. Then one can say that the morestable product is formed faster because it is more sta-ble. This cannot be strictly true since rates depend ontransition-state stabilities. However, if some feature sta-bilizes both transition state and product, then this state-ment is a convenient simplification. Thus the reactivityorder of hydrocarbons toward bromine atoms may beattributed to the fact that radical stabilities increase inthe order CH3· < CH3CH2· < (CH3)2CH· < (CH3)3C· <CH2 CHCH2·. Another example is the borohydride re-duction of cyclopentanone (87, n = 5) and cyclohexanone(88, n = 6), where it is justifiable to say that the latteris faster because cyclohexanol is the more stable productowing to the absence of torsional strain.

In contrast, kinetic control is the case where one prod-uct is formed faster than another because of rate constantsunrelated to equilibrium constants for the overall reaction.Such a case is the faster solvolysis of tert-butyl chloride[Eqs. (29)–(30)] compared to isopropyl chloride [Eq. (31)]even though the equilibrium constants for the two reac-tions are nearly the same. Therefore this is not product-development control. Yet it is possible to understand thegreater reactivity of tert-butyl chloride because its transi-tion state has nine additional hyperconjugative resonanceforms. It is cumbersome to consider the transition states.Instead it is easier to focus on the carbocation interme-diate, which is more stable in the tert-butyl case. Thusthrough an understanding of the mechanism, it is possi-ble to understand this case of kinetic control in terms ofthe stability of intermediates that resemble the transitionstates.

IV. METHODOLOGY OF MECHANISTICSTUDIES

In the study of chemical kinetics one of the key piecesof information is how the reaction rate depends on reac-tion conditions. Among the variables are solvent, temper-ature, and the concentrations of the various reactants andcatalysts and possibly other chemical species. Also, care-ful attention to product structures can provide additionalinformation. However, it must be recognized that physicalorganic chemistry is an inductive science. It is never pos-sible to prove a mechanism. At best it may be possible toobtain experimental results that are inconsistent with allconceivable mechanisms save one. Even then, there maybe an “inconceivable” mechanism that was overlooked.Besides, for reasons of brevity experimental results areusually presented only as supporting a mechanism without




explicitly explaining how they are inconsistent with othermechanisms, and it is left to the critical reader to completethe logic.

A. Kinetic Order

It is often observed that ν, the rate of a chemical reaction,is simply proportional to a power of the concentration ofa chemical species:

v = −d[reactant]

dt= d[product]

dt= k[A]nA [B]nB · · · .

(32)

Usually these species are reactants or perhaps catalysts,but they may be products or other additives. Such a re-action is said to be nAth order in A, nBth order in B, and(nA + nB + · · ·)th order overall. Usually reactions are firstorder, but some show second-order or zeroth-order de-pendence, and the exponent n can even be fractional ornegative.

This is a differential equation that can be solved to ex-press the time dependence of concentrations. It is an ex-perimental task to determine each n. One way is to verifythat the observed time dependence of concentrations fol-lows that derived from solving Eq. (32). A better way isto vary the initial concentrations of the various chemicalspecies and verify that ν follows the power dependenceof Eq. (32). For example, if a reaction is second order ina species, then ν must quadruple if the concentration ofthat species is doubled. Another way is to hold all con-centrations but one fixed and measure the dependence ofν on that one variable concentration, using the followingequation, derived from Eq. (32) by taking logarithms andpartial derivatives:

(∂ ln v

∂ ln[A]

)

[B],...

= nA. (33)

For example, the reaction of 2-bromopropane (RBr)with hydroxide, to form a mixture of 2-propanol andpropene, shows a rate given by

v = {k1 + k2[OH−]}[RBr]

= k1[RBr] + k2[RBr][OH−]. (34)

This reaction is clearly first order in RBr. The order inhydroxide is undefined since it does not match the formof Eq. (32). However, the rate can be separated into twoterms, one zeroth order in hydroxide and the other firstorder.

A convenient experimental technique is to follow thedisappearance of one key reactant, called the substrate (S),while maintaining the concentrations of all other chemicalspecies constant during each individual reaction. This canbe accomplished (1) if the other species is in large excess,

so that it is not consumed to any appreciable extent, or(2) if it is a catalyst, or (3) if it is H+ or OH− and thesolution is buffered. The constancy of those other con-centrations simplifies Eq. (32) to the form of Eq. (35) (orperhaps another form that is zeroth order or second orderin substrate), where kobs is called a rate coefficient (not arate constant, because it varies with the concentrations ofthose other species). The solution is then Eq. (36), whereS0 is the initial concentration of substrate:

v = −d[S]

dt= kobs[S], (35)

[S] = S0 exp(−kobst). (36)

The value of kobs can thus be measured from the variationof [S] with time. By varying those other concentrations,the dependence of kobs on those concentrations can thenbe evaluated experimentally.

In the example of Eq. (34) comparison with Eq. (35)shows that kobs is given by

kobs = k1 + k2[OH−]. (37)

By running the reaction in excess OH− or in buffer, [OH−]can be kept constant during a reaction, to evaluate kobs.By running the reaction with different concentrations of[OH−], the variation of kobs with [OH−] can then be foundto follow Eq. (37). The dependence of a kobs on [OH−]or [H+] is often displayed as a pH–rate profile showinglog10 kobs versus pH. Figure 15 shows such a plot for thedependence of Eq. (37).

Kinetic orders give valuable information about mech-anism. The kinetic order in a chemical species gives thenumber of those molecules in the transition state (strictly,the number of each of the constituent atoms that are re-quired to form the transition state). For example, the first-order dependence on RBr in the reaction of Eq. (34) means

FIGURE 15 The pH–rate profile for a reaction where kobs =k1 + k2[OH−].




that the atoms of one molecule of 2-bromopropane (threecarbons, seven hydrogens, and a bromine) are present inthe transition state. The dependence on [OH−] means thatthere are two pathways occurring together, one with anoxygen, an additional hydrogen, and an electron containedin the transition state, and the other without these.

One fundamental limitation of this method is that it isimpossible to determine the kinetic order for the solventmolecules. It is not possible to vary the concentration ofsolvent without also introducing solvent effects of the sortarising via Eqs. (27) and (28). Therefore there is no wayto know how many solvent molecules are present in thetransition state, and the composition of the transition stateis always subject to an uncertainty of an arbitrary num-ber of solvent molecules. For example, the k1 pathway inEq. (34), which involves no hydroxide, might have a watermolecule (two hydrogens and an oxygen) present in thetransition state or it might not.

The first example of mechanistic inference from kinet-ics came from Lapworth’s 1904 observation that the rateof bromination of acetone to form bromoacetone is firstorder in acetone and first order in acid but zero order inbromine:

v = −d[CH3COCH3]

dt= d[CH3COCH2Br]

dt

= k[CH3COCH3][H+]. (38)

Even though bromine is a reactant, the rate of reaction isindependent of its concentration. These results mean thatthe transition state is composed of three carbons, sevenhydrogens, an oxygen, and a positive charge (and an un-known number of the atoms that constitute water) but nobromines. Therefore the rate-limiting step occurs beforethe bromine enters the reaction. The current interpretationis that the rate-limiting step is proton removal by a wa-ter molecule from the conjugate acid of acetone to formthe enol, CH3C(OH) CH2, as an intermediate that subse-quently reacts rapidly with bromine.

Another classic example is the nitration of benzene andother reactive aromatic hydrocarbons, where the rate de-pends on the concentration of nitric acid but is indepen-dent of the concentration of the aromatic. The current in-terpretation is that the rate-limiting step is formation ofnitronium ion, NO+

2 , which then reacts rapidly with thearomatic.

The mutarotation of glucose is the interconversion ofits α (93) and β (94) anomers, whose rate can be followedby the change of optical activity. The open-chain aldehyde(95) is an intermediate. The reaction is acid-catalyzed, butthe rate is also found to increase with increasing bufferconcentration, even at constant pH. Therefore the rate hasthe following form, where HA is the buffer acid:

v = kobs[glucose]

= {kH[H+] + kHA[HA]}[glucose]. (39)

Such behavior is called general acid catalysis since eachacid is capable of serving as a catalyst. There must be onemolecule of acid, either H+ or buffer HA, present in thetransition state, along with a molecule of the substrate. Thecurrent interpretation is that the transition state for reactionof 94 with HA has structure 96, with H+ detached fromA−. This example shows how the kinetic order tells ushow many of each kind of atom are in the transition statebut it does not itself tell us how those atoms are arranged.

B. Solvent Effects

Although it is not possible to determine the kinetic order insolvent, the fact that polar solvents can have large effectson reaction rates means that solvent effects can be used todiagnose whether a reaction creates or destroys ions. Forexample, reactions of trialkylsulfonium ions RS(CH3)+2with hydroxide show different solvent effects, depend-ing on R. For R = CH3 the rate decreases 2 × 104-fold onchanging from ethanol to water, whereas for R = (CH3)3Cthe decrease is only 3-fold. The large effect in the formercase is consistent with a transition state [HOδ− · · · CH3 · · ·S(CH3)δ+2 ]‡, where ions are being destroyed. According toFig. 12, this is a case where the more polar solvent, water,stabilizes the reactants and renders them less reactive. Thesmall effect in the latter case is consistent with a transi-tion state [(CH3)3Cδ+ · · · S(CH3)δ+2 ]‡ that is still an ionbecause hydroxide is not yet involved.

There are empirical measures more suitable than dielec-tric constant for assessing quantitatively the polarity of asolvent. Even these do not always account for the specificinteractions whereby a solvent stabilizes ions or dipoles.Table VI lists some relative rate constants for nucleophilicsubstitution of Cl− on CH3I (by the DNAN mechanism).There is no correlation with dielectric constant. The lowerreactivity in the first three solvents is because they areprotic, with OHδ+ or NHδ+ groups that approach close tochloride anion and stabilize it greatly (“hydrogen bond-ing”). To make the chloride react, it must be stripped ofits solvation. In contrast, the last two solvents are apro-tic, with their δ+ buried in the center of the molecule,where it cannot stabilize anions as well. Consequently the




TABLE VI Relative Rate Constants for Reactionof Cl− with CH3I

Solvent ε k/kCH3OH

CH3OH 32.6 1

HC( O)NH2 109.5 12.5

HC( O)NHCH3 165.5 45.3

HC( O)N(CH3)2 37 1.2 × 106

CH3C( O)N(CH3)2 37.8 7.4 × 106

chloride is present as a “naked” nucleophile, of enhancedreactivity.

C. Temperature Dependence of Reactivity

Table II showed how rates can depend strongly on temper-ature. Since �G = �H − T �S the free energy of activa-tion in Eq. (23) can be separated into entropy and enthalpycomponents, leading to the logarithmic form

ln

(k

T

)

= ln

(kB

h

)

+ �S‡

R− �H ‡

R

1

T . (40)

Therefore a plot of ln(k /T ) versus 1/T has a slopeequal to −�H ‡ /R and an intercept equal to ln(kB /h) +�S‡ /R, or �S‡ /R + 23.76. Thus both �H ‡ and �S‡ canbe evaluated experimentally.

These quantities are often called activation parame-ters. The enthalpy contribution generally comes from theenergy required to break or reorganize the bonds. Theentropy contribution generally comes from the need toorganize the atoms into the precise arrangement of thetransition state. Thus these parameters provide informa-tion about the nature of the transition state.

Table VII lists activation parameters for the racem-ization of some sulfoxides, RS( O)Ar (97, Ar = C6H4

CH3-p). The sulfur is a tetrahedral stereocenter, owingto its four substituent groups (R, O, Ar, and lone pair).Racemization usually occurs by distorting the sulfur toan achiral transition state 98, with R, O, and Ar coplanar.For entry 1 there is a substantial enthalpy of activation[Eq. (11)] because the sulfur must be distorted from tetra-hedral bond angles to trigonal. There is no entropy con-tribution because reactant and transition state are equallywell organized. For entry 2 racemization proceeds instead

TABLE VII Activation Parameters for Racemization ofSulfoxides, RS( O)C6H4CH3-p

Entry R ∆H‡ (kcal/mole) ∆S‡ (cal/mole K)

1 C6H5 39 0

2 C6H5CH2 43 24

3 CH2 CHCH2 22 −9

by fragmenting the molecule into C6H5CH2 and ArSOradicals. There is a substantial positive entropy becausethe transition state is disorganized, owing to the creationof two particles from one. There is also a higher enthalpythan in entry 1 because a carbon–sulfur bond must bebroken. For entry 3 racemization proceeds via transitionstate 99 and reversible conversion to an achiral interme-diate 100. There is a substantial negative entropy becausethe transition state is highly ordered, with the position ofthe CH2 CHCH2 group restricted. There is also a lowerenthalpy than in entry 2 because formation of a carbon–oxygen bond compensates for breaking the carbon–sulfurbond.

D. Isotope Effects

According to the Heisenberg uncertainty principle, it isimpossible to determine exactly both the position and mo-mentum of a particle. Therefore a hydrogen atom cannotbe motionless at the distance r0

CH, where the energy ofa C H bond (Fig. 16) is minimum, since then its po-sition would be known, and also its kinetic energy andhence its momentum would be known to be zero. Insteadit must have nonzero kinetic and potential energy, the sumof which is its zero-point energy. According to quantummechanics, that energy is given as follows, where kF is theratio F/(r0

CH − rCH) of restoring force to distortion and µ

is the reduced mass, which is approximately equal to mH

or mD:

E0 = 1

4π

(kF

µ

)1/2

. (41)

FIGURE 16 Energy to dissociate a C H or C D bond, each withits zero-point energy.




Because mD > mH the zero-point energy is greater for aC H bond than for a C D, as shown in Fig. 16, and thebond-dissociation energy is less for C H than for C D.Then, according to Fig. 12, a C H bond will break faster.For typical values of kF, the rate constant ratio, kH/kD, is∼7 at 25◦C.

This result can be used to determine whether a C Hbond is broken in the rate-limiting step. The comparisoncan be either intermolecular, between deuterated and un-deuterated reactants, or intramolecular, by analyzing deu-terium content in products from a partially deuterated reac-tant. For example, part of the evidence for transition stateslike 63 is the observation that CH3CHPhCH2Br reactswith NaOC2H5 7.5 times as fast as does CH3CDPhCH2Br.In contrast, the reaction of 1-naphthol-4-sulfonate-2-d(101-d) with o-methoxybenzenediazonium ion (102,Ar = o-CH3OC6H4) to form 103 proceeds at the samerate as that of the undeuterated 101. Therefore the C Hor C D bond is not broken in the rate-limiting step, whichis inferred to be the formation of 104 as an intermediatethat loses H or D in a subsequent step.

The applicability of isotope effects is still wider. Theycan be seen even when a C H or C D bond is not bro-ken but when the carbon bearing H or D undergoes re-hybridization, although the effect is considerably smaller.They can be seen when reaction is carried out in D2Orather than water. They can be detected with other iso-topes, such as of carbon, nitrogen, or oxygen. Again theeffects are smaller.

E. Regiochemical and Stereochemical Studies

It is a truism that the products of a reaction must be knownin order to have any success at proposing a mechanism.Moreover, careful attention to product structures can pro-vide important information about mechanism.

A classic puzzle was the mechanism of addition of HBrto alkenes such as propene. Sometimes the product was2-bromopropane and sometimes it was 1-bromopropane.Eventually the role of the reaction conditions was rec-ognized. The former arises by electrophilic addition viatransition state 65, leading to the carbocation intermediate

(CH3)2CH+ stabilized by six hyperconjugative resonanceforms. The alternative regiochemistry would proceed viathe intermediate CH3CH2CH+

2 , with only three such forms(from delocalization of electrons in two C H bonds andone C C). However, in the presence of peroxides, bromineatoms are generated and a free-radical chain process in-trudes [Eq. (42) followed by Eq. (43)]:

Br· + CH3CH CH2 → CH3CH· CH2Br, (42)

CH3CH· CH2Br +HBr → Br·+ CH3CH2CH2Br. (43)

The regiochemistry here is due to the six hypercon-jugative resonance forms in the intermediate CH3CH·

CH2Br, which is thus more stable than the alternative,CH3CHBrCH·

2, which has only three hyperconjugativeforms (from delocalization of electrons in one C H bond,one C C, and one C Br).

Many nucleophilic substitutions proceed with inver-sion of configuration at the carbon undergoing substi-tution. One of the first demonstrations of this behaviorwas with (−)-2-octanol (105, R = C6H13), which couldbe converted with acetic anhydride to (−)-2-octyl acetate(106). Alternatively, 105 could be activated by conversionto its tosylate (107, Ts = SO2C6H4CH3-p). Reaction withpotassium acetate then formed (+ )-2-octyl acetate (106′),the enantiomer of 106. Conversion of levorotatory to dex-trorotatory did not occur in the reactions with acetic anhy-dride (105 → 106) or tosyl chloride (105 → 107), whichreact at oxygen, not carbon. Therefore in the reaction of107, the nucleophile (CH3CO−

2 ) and the leaving group(TsO−) must be on opposite faces of the carbon undergo-ing substitution. Transition state 108 shows this geometry(with the dotted C C bond behind the plane of the pageand with the other two dotted lines symbolizing partialbonds). It should be noted that this conclusion and thenext were reached without knowing the absolute configu-rations of the molecules.

In contrast, (2R,3R)-3-phenyl-2-butyl tosylate (109,Ts = SO2C6H4CH3-p) reacts in acetic acid to form




(2R,3R)-3-phenyl-2-butyl acetate (110), with retentionof configuration. Even more remarkable is that (2R,3S)-3-phenyl-2-butyl tosylate (111) reacts in acetic acid toform equal amounts of (2R,3S)- and (2S,3R)-3-phenyl-2-butyl acetate (112 + 112′). Again the reaction proceedswith retention of configuration, but here the product isracemic. These are reaction conditions that would fa-vor as rate-limiting step the formation of a carbocation,PhCH(CH3)CH+CH3, rather than attack by acetate, aswith 107. However, that carbocation would be the samefrom either 109 or 111, which should then have giventhe same products. Instead, to avoid so unstable an inter-mediate, the pi electrons of the phenyl group serve as aninternal nucleophile. As in 108, the configuration at C2(the carbon originally attached to oxygen) is inverted asphenyl substitutes for tosylate. The resulting intermediateis a phenonium ion, 113 (from 109) or 114 (from 111).When this then reacts with acetic acid as nucleophile, theconfiguration at C2 is again inverted. Two successive in-versions amount to an overall retention of configuration, asobserved from both 109 and 111. However, 114 is achiral,with a plane of symmetry passing through the six car-bons of the benzene ring. Therefore the product must beracemic. This phenomenon of an internal nucleophile act-ing to avoid formation of an unstable carbocation is oftencalled neighboring-group participation.

F. Labeling and Crossover Experiments

Labeling experiments are a means to tag a portion of amolecule and follow it through the reaction. The label maybe an isotope or it may be a chemical substituent, whichopens the risk of changing the mechanism but which isoften easier to synthesize and to analyze.

Isotopic labeling shows that ester hydrolysis generallyproceeds with cleavage of the acyl–oxygen bond, not thealkyl–oxygen bond. For example, hydrolysis of n-pentyl

acetate in alkaline H218O produces acetate containing the

18O and n-pentanol without any 18O:

CH3C( O)OC5H11 + 18OH−

→ CH3C( O)18O− + C5H11OH. (44)

This experiment could also be performed withCH3C( O)18OC5H11 to verify that the 18O Cpentyl

bond does not cleave. However, it is easier to “label” thatposition with ordinary 16O and run the reaction in H2

18O.A rare exception to this behavior is with the highly hind-ered methyl 2,4,6-tri-tert-butylbenzoate, where hydro-lysis in H2

18O produces CH318OH by alkyl–oxygen

cleavage.Reaction of aromatic halides with NaNH2 in liquid am-

monia produces the corresponding aromatic amine. How-ever, the conversion of o-iodoanisole (115) to m-anisidine(116) shows that this is not simply a substitution of NH2

for I. Instead it proceeds by proton removal to create 117,which undergoes elimination to a benzyne (118) that pref-erentially adds NH−

2 at the meta position to produce 119.The methoxy group is a label to make the rearrangementclear (and to stabilize the anion in 119), and the mecha-nism was further documented through 14C labeling.

Another example is the rearrangement of 120 to 121.The deuterium labeling shows that the reaction does notproceed simply by opening the four-membered ring at theleft. Instead it proceeds by opening the vertical bonds inthe middle to form 122, followed by rearrangement asindicated (to 121′, identical with 121).

Crossover experiments are a form of double-labelingexperiment designed to distinguish between intramolecu-lar and intermolecular mechanisms. For example, methyltransfer from oxygen to carbon in o-(p-CH3C6H4SO2)CH−C6H4SO2OCH3 to form o-(p-CH3C6H4SO2)CH(CH3)C6H4SO−

3 was shown to be intermolecular byusing a mixture of this anion (d0) plus (p-CD3C6H4SO2)




CH−C6H4SO2OCD3 (d6) and observing that a statisticalmixture of d0, d3, and d6 products was formed. The in-tramolecular methyl transfer would require transition state123 but with a 180◦ CH · · · CH3 · · · O angle, to be con-sistent with inversion of configuration at the methyl, andthis geometry is impossible. This reasoning is known asthe endocyclic restriction test.

G. Reactive Intermediates

In multistep reactions there are reaction intermediatesbetween the transition states. Intermediates are ordinarychemical species at local minima in the energy, but theyare not very stable and they are so reactive that they donot persist. Therefore they are at such low concentrationsand of such short lifetimes that it is often impossible todetect them in the reaction mixture. Evidence for inter-mediates sometimes comes from the kinetic behavior. Forexample, an enol intermediate was inferred from the ob-servation that the rate of bromination of acetone [Eq. (38)]is independent of [Br2]. Another example is the absenceof a kinetic isotope effect in the reaction of 101-d with102, showing that the C H or C D bond is not broken inthe rate-limiting step, which is the formation of 104 as anintermediate.

A simple kinetic model for a reaction with one interme-diate is Eq. (45). The implied differential equations canbe simplified with the steady-state approximation. If theconcentration of the intermediate B does not build up ap-preciably, then its time derivative can be neglected relativeto other variations, as expressed in Eq. (46). This algebraicequation is readily solved to give Eq. (47). This then leadsto a simpler differential equation for [A] that has the formof Eq. (35) with kobs given by Eq. (48):

Ak1⇀↽k−1

Bk2→ C, (45)

d[B]

dt= k1[A] − k−1[B] − k2[B] ∼ 0, (46)

[B] = k1

k−1 + k2[A], (47)

kobs = k1k2

k−1 + k2. (48)

If k−1 � k2, this reduces to k1 and the first step is rate-limiting. If k−1 � k2, this reduces to (k1/k−1)k2 and the

second step is rate-limiting. The magnitude of k1 is irrele-vant to the question of which step is rate-limiting, but ifk1 � k−1 + k2, then the concentration of intermediate Bmust be very small.

In some cases k−1 or k2 in Eqs. (45)–(48) is not a rateconstant but a rate coefficient, depending on the concen-tration of some other species. If so, it may be possible tovary that concentration so that the first step is rate-limitingunder some conditions and the second step is rate-limitingunder others. The change of rate-limiting step then appearsas a characteristic dependence of rate on that concentra-tion. For example, Fig. 17 shows the pH–rate profile forthe reaction of acetone with excess hydroxylamine to pro-duce acetoxime, (CH3)2C NOH. The falloff at low pH isdue simply to the conversion of hydroxylamine to unreac-tive HONH+

3 . The falloff at high pH is due to a change ofrate-limiting step from formation of (CH3)2C(OH)NHOHas intermediate at low pH to acid-catalyzed dehydrationof that intermediate at high pH, where this step becomesslow.

The existence of intermediates can also be inferred fromother kinds of evidence. The retention of configuration inreaction of 109 and the racemization in reaction of 111 areevidence for intermediate phenonium ions (113 and 114).The rearrangement in the reaction of 115 with NaNH2

suggests a benzyne intermediate (118). The variable re-giochemistry in the addition of HBr to alkenes was takenas evidence for both the carbocation (CH3)2CH+ and theradical CH3CH· CH2Br, depending on conditions.

Trapping experiments are often used to test for interme-diates. If the chemical reactivity of a presumed interme-diate can be predicted, then addition of a suitable reactantmay intercept the intermediate and convert it to a distinc-tive product. A further control experiment is necessary toverify that the disappearance of substrate is not accelerated

FIGURE 17 The pH–rate profile for reaction of acetone withhydroxylamine.




by the addition, in order to rule out direct reaction be-tween additive and substrate. For example, the hydrolysisof HCCl3 in base to HCO−

2 (plus CO) is thought to proceedvia CCl−3 and the carbene CCl2. Indeed, if the hydrolysisis performed in D2O and unreacted chloroform reisolated,it is found to be largely DCCl3, from trapping of CCl−3 .Moreover, adding PhS− does not increase the rate of disap-pearance of HCCl3, but the product is HC(SPh)3 instead,from trapping of CCl2. Another example is the solvolysisof Ph2CHCl or Ph2CHBr, both of which are thought toproceed via Ph2CH+ as a common intermediate. Addedsodium azide (NaN3) does not increase the rate (exceptthrough its effect on solvent polarity) but does change theproduct from Ph2CHOH to Ph2CHN3.

A variant on trapping experiments is to test substancesthat would inhibit a free radical reaction. For example,certain phenols, like BHA and BHT in food products, canscavenge free radicals. If the reaction does not take its nor-mal course, then it is not necessary to detect a distinctiveproduct in order to conclude that free radical intermediatesare implicated.

A convincing procedure is to observe the intermedi-ate. This is particularly informative since it permits theintermediate to be characterized spectroscopically and itsproperties elucidated. Usually the intermediate is too un-stable to be observed under the reaction conditions. Onepossibility is to modify the substrate so as to stabilizethe intermediate and increase its lifetime. Alternatively,it may be possible to modify the conditions so that theintermediate persists. For example, treatment of triphe-nylmethanol with sulfuric acid converts it to the carbo-cation Ph3C+, stabilized by resonance delocalization ofthe positive charge into the three aromatic rings. Lessstable carbocations can be prepared in “superacid” so-lutions of HSbF6, where there is no nucleophile availableto react, and where they can be characterized by NMRspectroscopy. Similarly, many intermediates can be pre-served by eliminating other species that they might reactwith. Thus carbanions can be observed under conditionsthat scrupulously exclude water and other electrophiles.For example, the basicity of enolates and related anionscan be measured by titration in dimethyl sulfoxide, andTable VIII lists some acidity constants of their conjugateacids. Preserving free radicals and carbenes (R2C) is morechallenging since they react with themselves, but the rem-edy is to immobilize them in a solid matrix so that theycannot encounter another reactive partner. Reactive inter-mediates that rearrange by themselves, such as strainedstructures (40 with n = 6–8, 41, 84) and some carbenesand carbocations, can often be generated at low temper-ature, perhaps photochemically. Similarly, electronicallyexcited states can be generated photochemically and theirbehavior studied.

TABLE VIII Acidity Constants of SomeCarbonyl, Nitro, Cyano, and Sulfone Com-pounds in Dimethyl Sulfoxide

Acid pK a

CH3COCH2COOC2H5 14.2

CH2(COOC2H5)2 16.4

CH3NO2 17.2

PhCOCH2Ph 17.65

PhCH2CN 21.9

CH3COCH2SO2Ph 22.1

PhCH2COOC2H5 22.6

(PhCH2)2SO2 23.9

PhCOCH3 24.7

Cyclopentanone 25.8

CH3COCH3 26.5

[(CH3)2CH]2CO 28.2

A variant is independent synthesis of a presumed in-termediate, which offers the further opportunity to testwhether it is converted to product (or, even better, to theidentical mixture of products) under the reaction con-ditions. For example, it is possible to prepare salts ofNO+

2 and to show that these react with aromatic hydro-carbons to form the same nitration products as with nitricacid.

Another possibility is that the presumption of interme-diacy must be rejected. For example, the Clemmensen re-duction of ketones R2C O to hydrocarbons R2CH2 withzinc metal had been thought to proceed via the alcoholR2CHOH. However, this cannot be an intermediate be-cause it is found to be inert to the reaction conditions.

V. ILLUSTRATIVE EXAMPLES

A. Computation and Structure

The location of the hydrogen in a hydrogen bond is anintriguing question of molecular structure that can now beaddressed confidently by high-level computations. Usu-ally a hydrogen is bonded covalently to one oxygen andhydrogen-bonded to the other. If the two oxygens areof equal or nearly equal basicity, the hydrogen may bebonded to either and jump quickly from one to the other[Eq. (49)]. In contrast, for the monoanion of phthalic acid(124), recent quantum mechanical calculations indicatea symmetric structure, with the hydrogen centered be-tween the hydrogens and bonded equally to both. Thisconclusion is quite sensitive to the oxygen–oxygen dis-tance dO O, which must be short. With a highly accuratemethod known as density functional theory dO O for the




gas-phase ion was computed to be 2.38 A. It does notchange if the ion is embedded in a continuum whose di-electric constant is equal to that of chloroform or water.An alternative approach to solvation is to simulate themotions of the ion and of discrete solvent molecules, asgoverned by their forces of interaction. Interaction withchloroform molecules does not change dO O, but if the ionalso interacts with K+, it becomes asymmetric. In waterdO O increases to 2.46 A and the structure again becomesasymmetric because of the entropy associated with hy-drogen bonding by water to the ion. These results are inagreement with some experimental data.

O H · · · O ⇀↽ O · · · H O. (49)

B. Pericyclic Reactions

The Woodward–Hoffmann rules are a triumph of the ap-plication of molecular orbital theory to reactivity. Theyare concerned with pericyclic reactions, reactions whereelectrons reorganize around a cycle of nuclei, with simul-taneous bond breaking and bond making, as in 67. Thereare three classes of pericyclic reactions: (1) electrocyclicreactions, where a new single bond is made across theends of a pi system, as well as the reverse of such a reac-tion, (2) sigmatropic rearrangements, where a group mi-grates across a pi system, and (3) cycloaddition reactions,where one pi system adds across another, as well as thereverse, a cycloreversion reaction. The full treatment ofthe Woodward–Hoffmann rules involves consideration ofthe symmetry of the molecular orbitals, but there is a sim-pler approach that that focuses on aromaticity of transitionstates.

The electrocyclizations of 1,3,5-hexatriene (125) to1,3-cyclohexadiene (126) and of 1,3-butadiene (127) tocyclobutene (128) offer a revealing comparison. Fromthe initially planar triene (125), rotation about the out-ermost pi bonds allows overlap between the p atomic or-bitals on carbons 1 and 6, to form 129. If those carbonsalso rehybridize toward sp3, the transition state (130) isreached. Eventually those carbons form a sigma bond andthe product (126) results. Similarly, the planar butadiene(127) rotates and rehybridizes to create transition state131, which proceeds to form the sigma bond of the prod-uct (128). Unlike the acyclic reactants or the products,where the sigma bonds do not overlap with the pi or-bitals, the transition states may be characterized by a cy-

cle of atomic orbitals, each overlapping with two neigh-bors. Electrons are delocalized over those cycles sincethey are in bonds that are forming and breaking. In 130there are six delocalized electrons, and in 131 there arefour. Therefore transition state 130 is aromatic and 131is antiaromatic. These reflect stabilization or destabiliza-tion universally associated with cyclic systems containing4n + 2 (n = 0, 1, 2, . . .) or 4n (n = 1, 2, . . .) delocalizedelectrons. Accordingly, 130 represents a low-energy tran-sition state, corresponding to a fast reaction, whereas 131is of high energy and corresponds to a slow reaction. Theconclusion that 125 can undergo this reaction whereas 127cannot is a distinction not accessible from resonance the-ory. It does follow from molecular orbital theory but it re-quires a generalization to transition states of the concept ofaromaticity.

There are two alternatives whereby 127 can react. Toaccount for these it is necessary to generalize aromaticityeven further, to excited states and to a different topol-ogy. According to molecular orbital calculations, the ex-cited electronic states of cyclic systems containing 4ndelocalized electrons are stabilized (relative to acycliccomparisons), whereas those with 4n + 2 are destabi-lized. Thus aromaticity and antiaromaticity reverse in anexcited state. For example, the excited electronic stateof transition state 131 becomes stabilized, and 127 canbe transformed rapidly into 128 under photochemicalconditions.

The alternative topology arises by rotating the outer-most pi bonds of 127 in the same direction. This is calledconrotatory motion, in distinction to the previous disro-tatory motion, where the bonds rotated in opposite direc-tions. This now leads to transition state 132, where the toplobe of the rehybridizing orbital on carbon 1 overlaps withthe bottom lobe of the orbital on carbon 4. It would seem asthough this is an antibonding interaction because a nodalsurface is created between these atomic orbitals. However,according to molecular orbital calculations, this leads toa stabilization. Indeed, this is a general result for cyclicsystems containing 4n delocalized electrons but subject to




the requirement for overlap between the positive lobe ofone orbital and the negative lobe of another. Such an ar-rangement of orbitals has the topology of a Mobius strip,obtained from a rectangular ribbon by twisting one endby 180◦ before fastening it to the other end. The normaltopology, often called Huckel in this context, always haspositive lobes overlapping positive (or negative overlap-ping negative). According to molecular orbital calcula-tions, stabilization and aromaticity can be associated withall Mobius cycles containing 4n delocalized electrons, anddestabilization and antiaromaticity with 4n + 2.

Thus, thermally activated electrocyclizations, or the re-verse reactions, occur in disrotatory fashion when thereare 4n + 2 pi electrons, and conrotatory when thereare 4n. It also follows that these conclusions must bereversed for photochemical reactions. These are verygeneral conclusions, applicable also to molecules withheteratoms (133), to anions (134), and to the ring open-ing of 135 to 136. However, the topology becomesapparent only with labels. For example, disrotatory clo-sure of trans,cis,trans-2,4,6-octatriene (137) producescis-5,6-dimethyl-1,3-cyclohexadiene (138), and conrota-tory opening of cis-3,4-dimethylcyclobutene (139) pro-duces cis,trans-1,3-butadiene (140), even though the transand trans,trans stereoisomers of the products would befavored by stability.

In cycloadditions one pi system adds across another,as in the Diels–Alder reaction (141) of butadiene plusethylene to produce cyclohexene (or a wide variety ofsubstituted versions), or the dimerization of ethene (142)to cyclobutane. These proceed via transitions states 143and 144. The former is aromatic, with six electrons de-localized over the cycle of atomic orbitals. The latter is

antiaromatic, with four electrons delocalized. Thereforethe former reaction is quite facile, but the latter does notoccur thermally. It does occur photochemically, with ap-propriate substituents to facilitate creation of the excitedelectronic state. Alternatively, a cycloaddition with 4n de-localized electrons can occur if the orbital overlaps canachieve a Mobius topology. This requires addition acrossopposite faces of a pi systems. Such an addition is calledantarafacial, in contrast to the usual suprafacial, across asingle face of a pi system. Usually this latter is stericallythe only possibility. However, an alkene can add acrossa ketene R2C C O via transition state 145, with over-lap between the positive lobe of an alkene carbon and thenegative lobe of the carbonyl carbon.

Sigmatropic rearrangements are the most bewilderingof pericyclic reactions because there is always a react-ing pair of electrons from a sigma bond. This feature canbe avoided in the other cases by focusing on the elec-trocyclization or cycloaddition direction, even when thereverse reaction is under consideration. Examples are the1,2 alkyl shift of a carbocation (146), the 1,5 alkyl shiftof a diene (147), the 3,3 shift of two three-atom frag-ments across each other (148), and the 1,3 alkyl shift of analkene (149). The transition states for the first two (150,151) have 4n + 2 (two or six) electrons delocalized over acycle of atoms. Similarly, the 3,3 shift has six delocalizedelectrons in its transition state. All of these are facile reac-tions, with low activation energies, because they proceedvia aromatic transition states. The 1,2 alkyl (or hydride)shift (146) is a remarkably fast process that is responsi-ble for many rearrangements that carbocations undergo.In contrast, the 1,3 alkyl shift (149) has four delocalizedelectrons in its transition state. The suprafacial processtherefore involves a high-energy, antiaromatic transitionstate and is slow. However, if the geometry permits an an-tarafacial process, with migration of the three-atom com-ponent across opposite faces of the single atom, then thereis one overlap between a positive lobe and a negative lobe.This is a Mobius topology and thus an aromatic transitionstate (152). A remarkable example of the success of theWoodward–Hoffmann rules is the rearrangement of 153to 154, with what would otherwise have been a surpris-ing change of the relative stereochemistry of methyl andacetoxy.




These rules are summarized in Table IX. To judgewhether a reaction can occur via a low-energy aromatictransition state, all that is necessary is to find the cycleof overlapping atomic orbitals in that transition state, toascertain whether all the overlaps can be between lobes ofthe same sign or whether there must be one positive-to-negative overlap, and then to count the number of electronsdelocalized over that cycle.

C. Organometallic Catalysis

Olefin metathesis is the interchange of fragments on dou-ble bonds, as in Eq. (50). The reaction is catalyzed by var-ious compounds of tungsten, molybdenum, rhenium, orruthenium. The mechanism had been thought to involvesome sort of complex (155) between two alkenes and ametal (with additional, unspecified ligands, and whose dorbitals may permit a transition state of Mobius topology).However, when the reaction was carried out with a mixtureof 2,2′-divinylbiphenyl and 2,2′-divinylbiphenyl-d4 (156),the product, even at short reaction times, was phenanthrene(157) plus a statistical mixture of ethene-d0 , -d2, and -d4. Ifan intermediate like 155 were formed from either of thesereactants, it could not cleave to ethene-d2. Therefore it wasproposed that the catalytically active form of the metal isa metallocarbene, RCH M, which reacts with an alkeneR′CH CHR′′ to form both 158 and 159 as intermediates.These can cleave to RCH CHR′ and RCH CHR′′ andalso regenerate catalytic metallocarbenes R′CH M andR′′CH M.

TABLE IX General Rules for PericyclicReactions

Conditions 4n + 2 electrons 4n electrons

Thermal Huckel Mobius

Photochemical Mobius Huckel

CH3CH CHC2H5 ⇀↽ CH3CH CHCH3

+ C2H5CH CHC2H5. (50)

D. Ester Hydrolysis

In alkali, 4-(acetoxymethyl)imidazole (160) is hydrolyzed∼105 times as fast as benzyl acetate, CH3COOCH2Ph.The customary mechanism of ester hydrolysis, viaaddition of hydroxide to the carbonyl carbon (69),should not show such a rate disparity. Therefore adifferent mechanism, not available to benzyl acetate,was proposed, proceeding via intermediate 161. As ev-idence, when the hydrolysis was carried out in alkalineH2

18O, the isotope was found in the alcohol product,4-(hydroxymethyl)imidazole, whereas ester hydrolysisordinarily transfers the isotope to the acetate [Eq. (44)].Moreover, when the NH was replaced by NCH3, no rateenhancement was observed.

E. Proton Exchange

Amides, RCONHR′, will exchange their NH hydrogenswith water. The kinetics can be followed with isotopes orwith NMR techniques that distinguish NH protons fromOH. The reaction is catalyzed by both H+ and OH−. Themechanism of the base-catalyzed reaction is simply re-moval of the NH, to create the amide’s conjugate base as anintermediate that next abstracts a different H from water.The mechanism of the acid-catalyzed reaction was eluci-dated with various RCONH2 (162). Advantage was takenof the partial double-bond character (43, R′ = H), whichmaintains H in two different sites (labeled HE and HZ in162), each with its own NMR signal. For electron-donatingR the exchange was found to occur simply by protonation




on nitrogen, to produce RC( O)NH+3 . In this intermediate

there is no longer any partial double-bond character to im-pede C N rotation. Then when one of the three hydrogensis returned to solvent, one possibility is that the originalHE may remain in the amide but be transferred to the HZ

site, to produce 163. For electron-donating R the exchangewas found to occur by a more circuitous mechanism. Pro-tonation on oxygen forms RC( OH) NH+

2 , which canlose either of the two NH to produce two stereoisomericRC( OH) NH as intermediates. Reversal of these twosteps then regenerates the amide, but with one of its NHgroups exchanged. However, those NH exchange onlywith solvent, without exchanging with each other. Thusthe two mechanisms could be distinguished by comparingintramolecular exchange with intermolecular, since onlyvia RC( O)NH+

3 is there opportunity for intramolecularexchange.

F. Decarboxylation

α,β-Unsaturated carboxylic acids, RCH CHCOOH, de-carboxylate to alkenes, RCH CH2, plus CO2. Yet(CH3)3CCH CHCOOH does not show this reactivity.This is a clue that a γ hydrogen is required. The acceptedmechanism involves acid-catalyzed isomerization of theα,β-unsaturated acid (164) to the β,γ -unsaturated acid(165), which can undergo a concerted decarboxylation.Three further pieces of evidence are that α,β-unsaturatedacids do isomerize to the β,γ -unsaturated acid faster thanthey decarboxylate, β,γ -unsaturated acids do undergo de-carboxylation with isomerization (similar to the concerteddecarboxylation of β-keto acids, RCOCH2COOH), anddeuterium is incorporated into the γ position when thereaction is carried out with acid OD.

G. α-Haloketones

The Favorskii rearrangement is the conversion of anα-halo ketone, RR′CX COR′′, in alkali to a car-boxylic acid, RR′R′′C COOH (or ester). Reaction of2-chlorocyclohexanone-2-14C (166, ∗ = 14C) producescyclopentanecarboxylic acid (167) with the label dis-tributed equally between Cα and Cβ. This is evidencefor a symmetric intermediate, the cyclopropanone 168,which can open in two ways. Similarly, reaction of 2-halo-2-methyl-3-pentanone or 2-halo-4-methyl-3-pentanone

(169 or 169′, X = Cl or Br) with sodium hydroxide gavethe same 4:1 mixture of 2,2- and 2,3-dimethylbutanoicacids (170, 170′) from any of the four different reactants.This is evidence for reaction via trimethylcyclopropanone(171) as intermediate. Moreover, when that intermedi-ate was synthesized independently [from dimethylketene,(CH3)2C C O, and diazoethane, CH3CH N+ N−], itproduced that same 4:1 mixture when treated with NaOH.

O

Cl

O-

Cl

O

O-OH

OOH

OOH

O O

X X

H

-H+* -Cl-

*H+

*

OH-

*

*

*

166 168

+

169'

167-α-14C

169

or

167-β-14C

OH-

170171 170'

OH-

+

O OOH

OOH

The reaction of 172 in acetate buffer to form 173 showssome similarity to the Favorskii rearrangement. It had beenthought to occur by elimination of HCl simultaneous withring opening to form 174 as a presumed intermediate that isthen hydrolyzed further. However, when 174 was preparedindependently and subjected to the buffer conditions, itwas unreactive. Indeed, the hydrogen that is removed inthis mechanism is much less acidic than the adjacent oneindicated in 172′, analogous to 166. If, instead, that moreacidic proton is removed, the resulting enolate anion 175can lose Cl−. The resulting 176 is isomeric to a cyclo-propanone, like 168 or 171, but with an open zwitterionicstructure to avoid angle strain. Addition of hydroxide fol-lowed by protonation produces 177. This next loses HCland opens the central ring to form a tautomer of the final173. Support for this mechanism was obtained by labelingthe CCl2 carbon and showing that it acquires H.

ClCl

H

Cl OH

ClCl

H

ClCl

Cl

O-O-O

Cl

O-HOOO

HCl

H

H

O

172 174

buffer

173

175172'

OH-

176

-Cl- +

H+

OH-

-HCl

-H+

O O

O

177

O

-HCl

?




VI. QUANTITATIVE RELATIONSHIPS

A. Brønsted Relationship

Much insight has come from recognizing quantitative reg-ularities among the phenomena of organic chemistry. Thisis simply a highly developed method for reasoning by anal-ogy. It is especially powerful since it permits the extensionof understanding from one class of reactions to another.

The earliest case was the treatment of reactions subjectto general acid or general base catalysis [Eq. (51), adaptedfrom Eq. (39)]. The rate constant kHA reflects the speedwith which acid HA can act as a proton donor, and kB

reflects the speed with which base B can act as a protonacceptor. There ought to be a relationship between thekinetic power of an acid or base and its thermodynamicacidity or basicity: kHA ought to increase as HA becomes astronger acid, and kB ought to increase with the basicity ofB. A quantitative expression of this relationship is Eq. (52),where c is a constant and where the basicity constant of Bhas been replaced by the acidity constant of its conjugateacid BH+:

kobs = kH[H+] + kHA[HA]

or kobs = kOH[OH−] + kB[B−], (51)

log10 kHA = −αpK HAa + c

or log10 kB = βpK BH+a + c . (52)

Therefore if kHA or kB is measured for a series of acids orbases and its logarithms plotted against −pK HA

a or pK BH+a ,

a straight line should result with slope α or β. Such anequation is known as the Brønsted relationship.

For example, the acid hydrolysis of ethyl vinyl ether,C2H5OCH CH2, proceeds by rate-limiting proton trans-fer to produce C2H5O+ CH CH3, which then reactsrapidly with water and eventually cleaves to C2H5OH andCH3CH O. Each acid has its own rate constant kHA forproton transfer; a plot of log10 kHA versus −pKa of thatacid is shown in Fig. 18. The slope α is 0.68 and the in-tercept c (at pKa = 0) is 0.38.

B. Linear Free Energy Relationships

Figure 18 is an example of a linear free energy relation-ship. The horizontal axis is related to the free energy ofacid dissociation via Eq. (14) and the vertical axis is re-lated to the free energy of activation via Eq. (23). A modelfor such behavior can be obtained by constructing the re-action coordinate as a composite of the dissociation of thereactant O H bond and the formation of the product C Hbond, as indicated in Fig. 19. The transition state may betaken as the point where the two curves cross, when bondformation begins to compensate for the bond breaking.

FIGURE 18 Brønsted plot for the hydrolysis of ethyl vinylether. Acids RCOOH, from left to right, are R = NCCH2, ClCH2,CH3OCH2, H, HOCH2, CH3, and CH3CH2.

The activation energy is indicated as �G‡ and the freeenergy of the reaction as �G◦. What happens if the prod-uct is modified so as to become less stable, as indicatedby the dashed curve? The activation energy is now �G‡′

(distinguished with a prime), higher than �G‡, and thefree energy of reaction is �G◦′, less negative than �G◦.However, the change of activation energy, �G‡′ − �G‡,is less than �G◦′ − �G◦, as can be seen from the verticaldistances between the two sets of arrow tips. This resultcan be written in the following form, where α is between0 and 1:

�G‡′ − �G‡ = α(�G◦′ − �G◦). (53)

It can be shown that Eq. (52) is an example of this equation.The slope α is a selectivity parameter. If α is small,

then all reactions proceed at nearly the same rate, without

FIGURE 19 Variation of energy along a reaction coordinate con-structed from cleaving the reactant bond and forming the productbond. Solid curve is for the standard product, the dashed curvefor the product denoted with a prime.




selection for thermodynamic favorability. If α is near 1,then the thermodynamic stabilities fully influence the acti-vation energy. Thus α is a quantitative measure of product-development control, or the extent to which the transitionstate partakes of the stability of the product. In a very ap-proximate sense, α represents a measure of the extent towhich the transition state resembles the product. For ex-ample, in the above hydrolysis of ethyl vinyl ether, the αof 0.68 can be identified loosely with the partial negativecharge δ− acquired by A in the transition state (178).

C2H5O CH

CH2

H A

178

δ+ δ-‡

If instead the product becomes more and more stable,the product curve in Fig. 19 would drop, and its inter-section with the reactant curve would also drop in en-ergy. Thus the activation energy is lowered, opposite tothe change above, where the product became less stable.Moreover, as the product curve drops, its intersection withthe reactant curve moves back (to the left) along the re-action coordinate. Thus the transition state resembles thereactant more and more. There is less and less product-development control, and α diminishes. In the extremecase of a very stable product the reaction becomes fast butunselective.

Some of these conclusions can be seen in the reac-tions of some hydrocarbons with radical species. The rel-ative rate constants of hydrogen abstraction [Eq. (54)]by chlorine atoms, bromine atoms, trichloromethyl radi-cals, and under conditions of N -bromosuccinimide (NBS)bromination are listed in Table X. Also included are theC H bond-dissociation energies. For each radical the ratestend to increase as the BDE decreases. This is product-development control, arising because the weaker the bond,the faster it is cleaved. Each of these is also a linearfree energy relationship, although the linearity is not per-fect because phenyls and methyls have different steric re-

TABLE X Rate Constants of C H Abstraction (per H, Relative to Toluene) by Cl·,Br·, Cl3C·, and with N-Bromosuccinimide (NBS)

R k /ktol(Cl) k /ktol(Br) k /ktol(Cl3C) k /ktol(NBS) BDE(R H)

CH3CH2 0.77 1.6 × 10−5 98

(CH3)2CH 3.3 3.4 × 10−3 95

(CH3)3C 4.6 0.30 92

PhCH2 =1 =1 =1 =1 90

PhCH(CH3) 2.5 17 50 21 87

Ph2CH 2.0 9.6 50 10 85

PhC(CH3)2 5.6 37 260 45 85

Ph2C(CH3) 42 650 84

Ph3C 7.3 18 160 83

quirements. The chlorination reaction is rather unselec-tive, consistent with a very stable product, owing to thestrength of the H Cl bond that is formed (BDE = 103kcal/mole, compared to 87 kcal/mole for H Br). The otherreactions are more selective, especially with Cl3C·. Fur-thermore, the selectivity with NBS is the same, withinexperimental error, as with Br·. It would be very fortu-itous for the N -bromosuccinimidyl radical to have thesame selectivity as Br·. Therefore it was concluded thatthe reactive radical in NBS brominations is Br· and notN -bromosuccinimidyl.

R H + X· → R· + H X. (54)

C. Hammett Equation

There is no requirement that the �G‡ and �G◦ in Eq. (53)must refer to rates and equilibria of the same reaction. Thefollowing Hammett equation is an extension to the com-parison of substituent effects in two different reactions:

log10 kX = ρσX + c. (55)

Here the substituent constant σX is defined by the fol-lowing equation from the measured acid-dissociation con-stants of substituted and unsubstituted benzoic acids, andkX and kH are rate constants for substituted and unsubsti-tuted reactants:

σX = log10

(

K XC6H4COOHa

/

K C6H5COOHa

)

= pK C6H5COOHa − pK XC6H4COOH

a . (56)

Table XI is a short list of substituent constants. In-creasingly positive values correspond to greater electron-withdrawing power, which stabilizes the negative chargeof the carboxylate anion and thus increases the acidity. Inpractice, rate constants kX for a series of reactants withdifferent substituents X are plotted against σX to obtain astraight or nearly straight line whose slope is the reaction




TABLE XI Substituent Constants Derivedfrom pKa of Benzoic Acids, XC6H4COOH

X σX X σX

p-MeO −0.28 p-Br 0.23

p-Me −0.17 m-Br 0.39

m-Me −0.06 p-NC 0.70

H = 0.00 m-O2N 0.71

m-MeO 0.11 p-O2N 0.78

constant ρ and whose intercept is c (which equals log10 kH

in an ideal case).Figure 20 shows such a plot for the reaction of substi-

tuted benzonitriles with HS−:

XC6H4CN + HS− → XC6H4C( S)NH2. (57)

‡

179

N

HS

X

δ-

δ-

The slope is 2.1. The positive value reflects the fact thatelectron-withdrawing substituents accelerate the reaction.This is consistent with a transition state 179 that carriesa negative charge. This transition state is not the sameas a carboxylate anion, but it is analogous, in that anysubstituent that stabilizes the carboxylate also stabilizesthe transition state. If the slope had been negative, as insome reactions, it would mean that electron-donating sub-stituents accelerate the reaction, consistent with the devel-opment of a positive charge in the transition state.

FIGURE 20 Hammett plot for reaction of XC6H4CN with HS−.Substituents X from left to right are p-CH3O, p -CH3, H, p -I,p -Br, p -Cl, and m -Br.

For many reactions it is possible to recognize an analogyeither to acid dissociation of substituted benzoic acids or toother model reactions. Among these models are aciditiesof phenols, XC6H4OH, or aliphatic acids, XCH2COOH,rates of solvolysis of XC6H4C(CH3)2Cl, and even the abil-ity of a solvent to stabilize the ground electronic stateof a dyestuff. Then the effects of substituents on bothequilibrium constants and rate constants can be correlatedquantitatively with the substituent effects in those modelreactions. No analogy is perfect, but one may provide anunderstanding of a new reaction in terms of a more familiarone.


CHEMICAL KINETICS, EXPERIMENTATION • CHEMICAL

THERMODYNAMICS • HYDROGEN BOND • KINETICS

(CHEMISTRY) • ORGANIC CHEMICAL SYSTEMS, THEORY

• ORGANIC CHEMISTRY, SYNTHESIS • ORGANOMETAL-LIC CHEMISTRY • PERIODIC TABLE (CHEMISTRY) • PHO-TOCHEMISTRY, MOLECULAR • POTENTIAL ENERGY SUR-FACES • QUANTUM CHEMISTRY • STEREOCHEMISTRY

BIBLIOGRAPHY

BERNASCONI, C. F. (ED.). (1986). “INVESTIGATION OF RATES AND

MECHANISMS OF REACTIONS,” 4TH ED., WILEY, NEW YORK.CAREY, F. A., AND SUNDBERG, R. J. (2000/1990). “ADVANCED OR-

GANIC CHEMISTRY,” PART A, 4TH ED. (2000), PART B, 3RD ED. (1990),KLUWER/PLENUM, NEW YORK.

CARROLL, F. A. (1998). “PERSPECTIVES ON STRUCTURE AND MECHA-NISM IN ORGANIC CHEMISTRY,” BROOKS/COLE, PACIfiC GROVE, CA.

LOWRY, T. H., AND RICHARDSON, K. S. (1987). “MECHANISM AND THE-ORY IN ORGANIC CHEMISTRY,” 3RD ED., HARPER & ROW, NEW YORK.

MARCH, J. (1992). “ADVANCED ORGANIC CHEMISTRY: REACTIONS,MECHANISMS, AND STRUCTURE,” 4TH ED., WILEY, NEW YORK.

MASKILL, H. (1985). “THE PHYSICAL BASIS OF ORGANIC CHEMISTRY,”OXFORD UNIVERSITY PRESS, NEW YORK.

MILLER, B. (1998). “ADVANCED ORGANIC CHEMISTRY: REACTIONS AND

MECHANISMS,” PRENTICE HALL, UPPER SADDLE RIVER, NJ.MULLER, P. (ED.). (1994). “GLOSSARY OF TERMS USED IN PHYS-

ICAL ORGANIC CHEMISTRY,” Pure Appl. Chem. 66, 1077–1184(HTTP://WWW.CHEM.QMW.AC.UK/IUPAC/GTPOC/).

PROSS, A. (1995). “THEORETICAL AND PHYSICAL PRINCIPLES OF OR-GANIC REACTIVITY,” WILEY, NEW YORK.

STOWELL, J. C. (1994). “INTERMEDIATE ORGANIC CHEMISTRY,” WILEY,NEW YORK.

TIDWELL, T. T., RAPPOPORT, Z., AND PERRIN, C. L. (EDS.). (1997).“PHYSICAL ORGANIC CHEMISTRY FOR THE 21ST CENTURY (A SYM-POSIUM IN PRINT).” Pure Appl. Chem. 69, 210–292 (HTTP://WWW.IUPAC.ORG/PUBLICATIONS/PAC/SPECIAL/0297/INDEX.HTML).

WILLIAMS, A. (2000). “CONCERTED ORGANIC AND BIO-ORGANIC

MECHANISMS,” CRC PRESS, BOCA RATON, FL.

http://www.chem.qmw.ac.uk/IUPAC/GTPOC/

http://www.iupac.org/PUBLICATIONS/PAC/SPECIAL/0297/INDEX.HTML

http://www.iupac.org/PUBLICATIONS/PAC/SPECIAL/0297/INDEX.HTML).

P1: GRB Final Pages Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology EN016B-738 July 31, 2001 14:0

StereochemistryErnest L. ElielUniversity of North Carolina

I. IntroductionII. History. Optical ActivityIII. StereoisomerismIV. Stereoisomerism of Alkenes and CyclanesV. Properties of Stereoisomers

VI. SignificanceVII. Separation (Resolution) and Racemization of

Enantiomers

VIII. Structure, Configuration, NotationIX. Determination of ConfigurationX. Chirality in Absence of Chiral CentersXI. ConformationXII. Cycloalkanes and their ConformationsXIII. Chiroptical Properties. Enantiomeric PurityXIV. Prochirality

GLOSSARY

Angle strain Excess potential energy of a moleculecaused by deformation of an angle from the normal,e.g., of a C–C–C angle from the tetrahedral. Also called“Baeyer strain.”

Anti or antiperiplanar Conformation of a segmentX–C–C–Y in which the torsion angle (which see) is180◦ or near 180◦. X and Y are said to be anti or an-tiperiplanar (to each other).

Cahn–Ingold–Prelog (C-I-P) Descriptors DescriptorsR and S (and others) used to describe the spatial ar-rangement (configuration) of ligands at a chiral centeror other chiral element (chiral axis, chiral plane).

Chiral center An atom to which a C-I-P descriptor canbe attached, usually a group IV (tetrahedral) or group V(pyramidal) atom. Reflection of the molecule must re-verse the descriptor.

Chirality Handedness; the property of molecules ormacroscopic objects (such as crystals) of not being su-perposable with their mirror images. Substances may

be called chiral if the constituent molecules are chiral,even if the substance is racemic.

Configuration The spatial arrangement of atoms bywhich stereoisomers are distinct, but disregardingfacile rotation about single bonds.

Conformation The (differing) spatial array of atoms inmolecules of given constitution and configuration pro-duced by facile rotation about single bonds.

Conformers Stable conformations (located at energyminima). Isomers which differ by virtue of facile ro-tation about single bonds, usually readily intercon-vertible.

Constitution (of a molecule) The nature of its constituentatoms and their connectivity.

Diastereomers (diastereoisomers) Stereoisomers thatare not mirror images of each other.

Eclipsed conformation Conformation of a nonlinear ar-ray of four atoms X–C–C–Y in which the torsion angle(which see) is zero or near zero (also “synperiplanar”).X and Y are said to be eclipsed.

Enantiomer(ic) excess In a partially or completely

79

P1: GRB Final Pages


80 Stereochemistry

resolved substance, the excess of one enantiomer overthe other as a percentage of the total.

Enantiomers Stereoisomers that are mirror images ofeach other.

Enantiomorphs Macroscopic objects (e.g., crystals)whose mirror images are not superposable with theoriginal entity.

Factorization The analysis, where possible, of chiralityin terms of individual chiral elements, such as chiralcenters, chiral axes, and chiral planes.

Gauche Conformation of a segment X–C–C–Y in whichthe torsion angle (which see) is ±60◦ or near ±60◦

(also “synclinal”). X and Y are said to be gauche (toeach other).

Kinetic resolution Separation of enantiomers by virtueof their unequal rates of reaction with a nonracemicchiral reagent.

Optical activity The property of chiral assemblies ofmolecules or chiral crystals of rotating the plane ofpolarized light.

Optical purity The ratio of the observed specific rotationof a substance to the specific rotation of the enantiomer-ically and chemically pure substance.

Optical rotation Rotation of the plane of polarized lightproduced by the presence of chiral molecules or chiralcrystals in the light path.

Racemic mixture (racemate) A composite of equalamounts of opposite enantiomers.

Racemization Conversion of individual enantiomers intoa racemic mixture.

Resolution Separation of enantiomers from a racemicmixture.

Staggered conformation Conformation of a nonlineararray of four atoms (A–B–C–D) in which the torsionangle (which see) is near 60◦. Ligands A and D are saidto be gauche or synclinal to each other.

Stereochemistry Chemistry in three dimensions; topo-graphical aspects of chemistry.

Stereogenic center An atom where exchange of two ofits ligands changes the configuration. Such an atom isusually, but not necessarily, a chiral center.

Stereoisomers Isomers of the same constitution but dif-fering in spatial arrangement of their constitutingatoms.

Torsion angle In a nonlinear array of four atoms A–B–C–D, the angle between the plane containing A, B, andC and the plane containing B, C, and D.

Torsional strain Excess potential energy of a moleculecaused by incomplete staggering of pertinent bonds(e.g., deviation of the torsion angle in H–C–C–H fromthe normal 60◦). Also called “Pitzer strain.”

van der Waals or compression strain Excess potentialenergy of a molecule caused by approach of two or

more of its constituting atoms within the repulsiveregime of the van der Waals potential. Also called“nonbonded (repulsive) interaction.”

I. INTRODUCTION

This article deals with the stereochemistry of organic com-pounds, although many of the general principles also applyto organometallic and inorganic compounds.

The term “stereochemistry” is derived from the Greek“stereos” meaning solid—it refers to chemistry in threedimensions. Since nearly all organic molecules are threedimensional (with the exception of some olefins and aro-matics to be discussed later), stereochemistry cannot beconsidered a branch of chemistry. Rather it is an aspect ofall chemistry, or, to put it differently, a point of view whichhas become increasingly important and, as will be shown,essential for the understanding of chemical structure andfunction.

II. HISTORY. OPTICAL ACTIVITY

Historically, stereochemical thinking developed from theobservations of J. B. Biot in 1812–1815 that quartz crystalsas well as solutions of certain organic substances rotatethe plane of polarized light; this phenomenon is called“optical rotation.” When the beam of light coming to-ward the observer rotates the plane to the right (clock-wise) we call the rotation positive (+); when the rota-tion is to the left or counterclockwise, it is negative (−).The angle of rotation α is proportional to the concentra-tion c (conventionally expressed in grams per milliliter,i.e., density, for solids and pure liquids and gases) andthe thickness of the layer—usually the length of the cellin which the liquid or solution is contained (convention-ally expressed in decimeters): α = [α] · l · c (“Biot’s law”).The proportionality constant [α] is called “specific rota-tion”: [α] = α/ l (dm) · c(g/cm3). For solutions, where theconcentration c′ is conventionally expressed in g/100 ml,[α] = 100α/ l (dm) · c′ (g/100 ml). The constant [α] (con-ventionally reported without units) is used for the char-acterization of “optically active” substances (i.e., whichdisplay optical rotation). It varies with temperature andwith the wavelength of the light used in the observation,which need to be indicated (as subscripts and superscripts,respectively): thus, [α]t

λ, where t is the temperature in◦C and λ is the wavelength in nanometers (nm). Unfor-tunately, since most substances are prone to solvationand intermolecular association in solution or in the liq-uid state, and since such association varies with concen-tration and solvent, these items must also be recorded. A

P1: GRB Final Pages


Stereochemistry 81

typical specific rotation might thus read [α]20D + 57.3 ± 0.2

(95% EtOH, c = 2.3), denoting measurement at 20◦C atthe sodium D line (589 nm) in 95% ethanol at a con-centration of 2.3 g/100 ml. (Monochromatic light ofthis wavelength is easily generated in a sodium vaporlamp and has thus classically been used for polarimetricmeasurements.)

In 1822 Sir John Herschel found that mirror-image crys-tals of quartz (discovered by R. J. Hauy in 1801 and called“enantiomorphs”) rotate the plane of polarization in oppo-site directions. This provided the first correlation of opticalrotation with enantiomorphism, in this case of crystals. In1848 Louis Pasteur achieved separation of the enantiomor-phous crystals he detected in the sodium–ammonium saltof the (optically inactive) paratartaric acid (today calledracemic tartaric acid, see below) and thereby obtained twodifferent substances, one of which rotated polarized lightto the right, the other to the left, even in aqueous solu-tion. Pasteur concluded that the enantiomorphous crys-tals were made up of molecules that themselves differedas object and (reflection) image on the molecular scale.Such mirror-image molecules are called “enantiomers”and the separation which Pasteur accomplished is called“resolution.” Since the difference between enantiomersresembles the difference between a right and a left hand(which are also mirror images of each other, but other-wise essentially identical in their dimensions), we callsuch molecules “chiral” (Greek “cheir” = hand) and wecall the property of certain molecules to display enan-tiomerism “chirality” (terms coined by Lord Kelvin in1893). While every molecule has a mirror image, chiral-ity exists only if the image is nonsuperposable with theoriginal molecule, just as a right shoe is not superposablewith a left one. (In contrast, socks worn on right and leftfeet are superposable; they are “achiral.”)

III. STEREOISOMERISM

The understanding of molecular structure was not wellenough advanced in 1848 for Pasteur to explain enan-tiomerism (chirality) in terms of atomic arrangement. Thebasis for that was laid only a decade or so later when,in separate publications, A. S. Couper, F. A. Kekule, J.Loschmidt, and A. Crum Brown illuminated the struc-ture of molecules in terms of the connectivity betweentheir constituent atoms. Then, in 1874, J. H. van’t Hoffin the Netherlands and J. A. Le Bel in France simulta-neously proposed the structural basis for chirality: Whenfour different atoms or groups ( jointly called “ligands”)are attached tetrahedrally to a given carbon atom, twomirror-image arrangements are possible (Fig. 1) corre-sponding to the two enantiomers. The enantiomers are said

FIGURE 1 Tetrahedral molecule Cabcd (the carbon atom at thecenter of the tetrahedron is conventionally not shown). [Reprintedwith permission from Eliel, E. L., and Wilen, S. H. (1994). “Stere-ochemistry of Organic Compounds,” Wiley, New York.]

to differ in “configuration.” Thus they explained not onlythe stereoisomerism of such simple molecules as CHF-BrI or CH3CHOHCO2H, but also that of more complexmolecules such as tartaric acid (Fig. 2). The two tartaricacids, A and B in Fig. 2 (separated by Pasteur as sodium–ammonium salts), are enantiomers. Pasteur’s starting saltcame from a 1:1 mixture of the two, called “racemic tar-taric acid.” A “racemate” is a mixture of equal quantitiesof corresponding enantiomers. Since the numerical valuesfor the optical rotations of the two enantiomers (negative,or −, for one; positive, or +, for the other) are equal andopposite, the racemate displays no optical activity.

In later work Pasteur discovered a fourth species of tar-taric acid [if we count the enantiomers A and B and theracemate (A + B) as three distinct species]. He was notable to explain the nature of this optically inactive iso-mer, but the Le Bel–van’t Hoff theory led to assignmentof the structure C (Fig. 2) to this stereoisomer. It has thesame connectivity (constitution) as A and B but differsin configuration. The reason for its lack of optical activ-ity is that it is superposable with its mirror image D; i.e.,it is achiral. It is a stereoisomer of A and B, but not anenantiomer. Such stereoisomers that are not mirror im-ages of each other are called “diastereomers”; thus C is adiastereomer of A and of B (and vice versa), whereas Ais an enantiomer of B (and vice versa). Compound C iscalled “meso-tartaric acid,” the prefix “meso” indicatingthat it is the achiral member in a set of diastereomers thatalso contains chiral members.

IV. STEREOISOMERISM OF ALKENESAND CYCLANES

So far it would appear that stereoisomerism is dependenton three-dimensional structure, but van’t Hoff recognizedthat stereoisomers can also exist in two dimensions, asin cis- and trans-disubstituted ethylenes (Fig. 3). The sub-stituents may be on the same side (cis) or on opposite sides(trans) at the two ends of the (planar) double bond.

P1: GRB Final Pages


82 Stereochemistry

FIGURE 2 The tartaric acids. [Reprinted with permission from Eliel, E. L., and Wilen, S. H. (1994). “Stereochemistryof Organic Compounds,” Wiley, New York.]

The classical cis/trans nomenclature for alkenes workswell for 1,2-disubstitued ethenes (A, B), but with tri- (C)and tetrasubstituted species it is not unequivocal. Thus inthe propenoic acid (C), Br is trans to CO2H but cis to Cl.For an unequivocal description, substituents at the sameterminus are ordered by the Cahn–Ingold–Prelog system(see below; for the present purpose it suffices to recognizethat substituents of higher atomic number have priorityover those of lower atomic number). Descriptors in thiscurrently used system, are Z (for the German zusammen)if the higher priority ligands are on the same side of thedouble-bond system and E (for entgegen) when they areon opposite sides. Thus C in Fig. 3 is (Z )-2-chloro-3-bromopropenoic acid.

Cis–trans isomerism is also found in cyclanes, which,for the purpose of counting stereoisomers, may be consid-

FIGURE 3 The E, Z nomenclature for cis and trans substitutedalkenes.

FIGURE 4 Stereoisomers of 1,2- and 1,3-disubstituted cyclobutanes.

ered planar (but see below). Figure 4 shows the cis–transisomerism of 1,2- (A–C) and 1,3- (D, E) dichlorocyclobu-tanes. The situation in the 1,2 isomers is similar to thatin the tartaric acids: there are two enantiomers (A, B;trans) plus an (achiral) meso isomer (C; cis), which isa diastereomer of A and B. The situation in the 1,3 iso-mers (D, cis; E, trans) is different. While D and E arediastereomers, neither of them is chiral (each one is su-perposable with its own mirror image). Carbons 1 and3 are not chiral centers since there is no chirality in themolecule, and yet, changing their relative position (cis ortrans) gives rise to (dia)stereoisomers. Carbons 1 and 3are therefore called “stereogenic.” All chiral centers arestereogenic, but, as seen in this case, not all stereogeniccenters are chiral centers. The E/Z system is not used forcycloalkanes.

V. PROPERTIES OF STEREOISOMERS

Enantiomers, though not superposable, resemble eachother very closely (as do right and left hands). Thedistances (both bonded and nonbonded) between cor-responding constituent atoms are the same, and thus

P1: GRB Final Pages


Stereochemistry 83

enantiomeric species have been called “isometric.” In theabsence of a second source of chirality their physical andchemical properties are identical. This is true of melt-ing and boiling points, density, refractive index, a wholehost of spectral properties (UV, IR, NMR, mass spectrum,etc.), thermodynamic properties, chromatographic behav-ior, and reactivity with achiral reagents (thus, enantiomericesters are hydrolyzed in the presence of achiral acids orbases at the same rate.) A macroscopic analogy is that ofright and left feet (chiral) which fit equally into either oneof a pair of socks (achiral).

However, when the chemical or physical agents arethemselves chiral, this equivalence breaks down, just asleft and right feet do not fit equally well into a right shoe.Examples are esterification of racemic α-methyl benzylalcohol, C6H5CHOHCH3, with the 2,4-dichlorophenylether of (+)-lactic acid, Cl2C6H3OCH(CH3)CO2H [the(−) enantiomer reacts faster], irradiation of racemic α-azidopropanoic acid dimethylamide, CH3CH(N3)CON(CH3)2, with left-circularly polarized light (chiral; see be-low), which leads to faster photochemical destruction ofthe (+) enantiomer and chromatography on chiral station-ary phases.

The situation is otherwise with diastereomers, which arenot isometric and therefore do not have identical physicalor chemical properties even in the absence of externalchirality. Thus, meso-tartaric acid (C in Fig. 2) melts at140◦C, whereas the diastereomeric (+) and (−) acids Aand B both melt at 170◦C (as explained above, A andB have the same melting point even though their opticalrotations are opposite).

VI. SIGNIFICANCE

Optical rotation is useful to characterize enantiomers ofchiral substances and has played an important historicalrole in the development of stereochemistry. However, thereal significance of stereoisomerism lies in the concept offit or misfit, epitomized by the fit of a right glove with aright hand and the corresponding misfit of a left glove. In1858, Pasteur discovered that, when a solution of racemictartaric acid is inoculated with the mold Penicilliumglaucum, the (+) acid is metabolized (and thereby de-stroyed) by the microorganism and the (−) acid remainsbehind. This selectivity is due to the organism’s enzymesbeing able to interact with the naturally occurring (+)enantiomer, but not with the other (−). A further exampleis the hydrolysis of the acetyl derivative of a racemicα-amino acid, RCH(NHCOCH3)CO2H, promoted by theenzyme hog kidney acylase. Only the acyl derivative ofthe (naturally occurring) L-amino acid (see below fornotation) is hydrolyzed to L-RCH(NH2)CO2H, whichcan then be readily separated from the unhydrolyzed

D-RCH(NHCOCH3)CO2H. The latter can separately behydrolyzed (with aqueous acid) to D-RCH(NH2)CO2H;both pure D and L acids can thus be prepared from the race-mate. This process is called “kinetic resolution” (“kinetic”because it depends on reaction rate; one enantiomer reactsrelatively rapidly, the other much more slowly or not atall). The hand-and-glove analogy applies here: One enan-tiomer fits into the active site of the acylase enzyme andits hydrolysis is thus promoted; the other enantiomer doesnot fit, just as a right glove (substrate) fits on a right hand(enzyme analog), whereas a left glove does not fit. Consid-ering the matter more generally, the right-hand/right-glovecombination is diastereomeric to the right-hand/left-glovecombination; diastereomers have different free energiesand so, presumably, have the transition states leading totheir formation. The effectiveness of the kinetic separa-tion (resolution) depends on the difference in activationenergies (and the associated difference in reaction rate):The greater the difference, the better the resolution.Enzymes tend to excel here because their relativelycomplex topography usually leads to a large difference inactivation energy between enantiomeric substrates (fit vs.misfit in the transition state) and hence a high degree ofdiscrimination.

A similar discrimination is seen in pharmaceuticals,whose action depends on their interaction with enzymes(if they are enzyme inhibitors) or other receptors. By wayof example, only the (−) enantiomer of α-methyldopa(α-methyl-3,4-dihydroxyphenylalanine) is an antihyper-tensive since it is enzymatically converted in the bodyto the pharmacologically active α-methylnorepinephrine.The (+) enantiomer is inactive. We call the active enan-tiomer “eutomer” and the inactive one “distomer.” Sincemost drugs have some side effects and therefore shouldbe used at the minimal effective dose, the pharmaceuti-cal industry is much interested in specifically preparingthe eutomer, free of the ineffectual and potentially detri-mental distomer (which constitutes 50% of the racemate).Other areas where enantiomers differ in function are inthe agriculture and flavor industries. Thus in the case ofthe agrochemical paclobutrazol, the (+) enantiomer is afungicide, whereas the (−) enantiomer is a plant growthregulator. Concerning flavor, the (−) enantiomer of car-vone (2-methyl-5-isopropylidenecyclohexanone) has theodor of spearmint, whereas its (+) enantiomer has the odorof caraway.

VII. SEPARATION (RESOLUTION) ANDRACEMIZATION OF ENANTIOMERS

Since enantiomers are so similar to each other, it is not sur-prising that their separation requires special methodology,which can be summarized as follows:

P1: GRB Final Pages


84 Stereochemistry

1. Separation by crystallization (Pasteur’s first method,1848)

2. Separation by formation and separation ofdiastereomers (Pasteur’s second method, 1853):(a) separation by crystallization, (b) separation bychromatography

3. Asymmetric transformation (a) of diastereomers,(b) of enantiomers

4. Kinetic resolution: (a) chemical, (b) enzymatic(Pasteur, 1858)

5. Separation by chromatography on chiral stationaryphases

6. Enantioselective synthesis7. Synthesis from enantiomerically pure precursors

(sometimes called enantiospecific synthesis)8. Miscellaneous methods

Pasteur’s method of manually separating enantiomor-phous crystals of enantiomers is obviously not practical;it is also not general. Racemic mixtures can lead to threedifferent kinds of crystals: conglomerates, racemic com-pounds, and racemic solid solutions. Compounds (wherethe unit cell, the smallest unit of a crystal, contains equalnumbers of enantiomeric molecules) are unsuitable forPasteurian resolution, as are racemic solid solutions. Onlywhen the enantiomers crystallize in discrete crystals (i.e.,as a macroscopic mixture called a “conglomerate”) canthe two types of crystals be separated even in principle.But only ca. 10% of racemic mixtures (ca. 20% in thecase of salts) crystallize as conglomerates. When they do,separation can be achieved by a modification of Pasteur’stechnique (called the “method of entrainment”) involv-ing alternate seeding with one enantiomer, separating theadditional crystalline material formed, replenishing thesolution with racemate, then seeding with the oppositeenantiomer, thereby inducing crystallization of the secondenantiomer. Large quantities of enantiomers, for exam-ple, of glutamic acid, have been separated in this mannercommercially.

Pasteur’s second method, formation of diastereomers byreaction of a racemate with an optically active auxiliaryor “adjuvant” (“resolving agent”), is much more common.Common resolving agents are naturally occurring chiralacids [such as (−)-malic acid, HO2CCHOHCH2CO2H]for chiral bases, and naturally occurring chiral bases, suchas (−)-quinine, for chiral acids. The salts formed are of-ten crystalline and can be separated by fractional crys-tallization. After separation, the resolved acid or base isliberated by treatment of the salt with mineral acid orbase, respectively. Alternatively, covalent diastereomersmay be formed [e.g., by esterification of a racemic acidwith (−)-menthol (2-isopropyl-5-methylcyclohexanol)]and separated by some type of chromatography on an

achiral stationary phase. After separation the chiral aux-iliary is removed (e.g., by hydrolysis in the case of anester).

To understand the third method, “asymmetric transfor-mation,” we must first take up “racemization,” i.e., theconversion of one of the enantiomers into a racemate. Thisapparently counterproductive process is actually useful:In resolution the undesired enantiomer is produced as anequimolar by-product. Racemization of that enantiomerallows one to start the resolution process over. Actually,racemization involves converting one enantiomer into theopposite one, but since enantiomers have the same freeenergy, the equilibrium constant between them is unity,i.e., the product of equilibration is the racemate. However,racemization, to be feasible, requires a chemical pathway.For example, a chiral ketone, RR′CHCOR′′ may be racem-ized by base via the resonance-stabilized achiral enolateanion RR′C−COR′′ ⇔ RR′C CR′′O−.

If such equilibration occurs concomitant with reso-lution, it is sometimes possible to convert the race-mate entirely into one of the enantiomers. This process,called “crystallization-induced asymmetric transforma-tion,” may be observed during crystallization of diastere-omers when the stereoisomers to be resolved can simul-taneously be equilibrated. An example is phenylglycine,C6H5CH(NH2)CO2H, required as the (−) isomer in man-ufacture of the antibiotic ampicillin. Equilibration of theenantiomers is effected by adding benzaldehyde to theracemic material (resulting in reversible formation of aSchiff base which is readily racemized) and precipitationof the desired (−) acid as its salt with (+)-tartaric acid.The (+) isomer is concomitantly reconverted to the race-mate; in the end, nearly the entire amino acid crystallizesas the (+)-tartrate of the (−) acid.

Kinetic resolution has already been discussed. Purelychemical approaches (exemplified by the resolution of chi-ral allylic alcohols, e.g., C6H11CHOHCH CHCH3, withSharpless’ reagent, which contains isopropyl tartrate asthe chiral constituent) are currently rare; enzymatic meth-ods (e.g., hydrolysis of esters of chiral alcohols catalyzedby lipase enzymes) are more common since enzymes arefrequently highly selective for one enantiomer over theother and the effectiveness of kinetic resolution dependson the degree of selectivity.

While ordinary chromatography does not separateenantiomers (though it can lead to separation of diastere-omers), enantiomers can be separated by chromatographyemploying a chiral stationary phase enriched in a singleenantiomer. In that case, the interactions between the twoenantiomers of the analyte and the chiral stationary phaseare diastereomeric in nature and therefore often differ instrength, the stronger interaction leading to longer reten-tion time.

P1: GRB Final Pages


Stereochemistry 85

Asymmetric (or enantioselective) synthesis is a targetedmethod for obtaining individual enantiomers from achiralprecursors. Ordinarily, the introduction of a chiral center(or other chiral element) in the course of synthesis leadsto equal formation of the two enantiomers, i.e., to a race-mate. Selectivity can, however, be achieved by using achiral (enantiomeric) reagent or catalyst, in which casethe transition states leading to the two enantiomeric prod-ucts are diastereomeric, or by attaching a chiral auxiliary(see above) to the starting material so that the products arediastereomers rather than enantiomers, in which case theirfree energies and the activation energies for their forma-tion also differ and one isomer is formed in preference overthe other. At the end of the reaction, the chiral auxiliaryis chemically removed. By way of example: Reduction ofpyruvic acid, CH3COCO2H, to lactic acid with an achiralreagent such as sodium borohydride gives racemic lacticacid, CH3CHOHCO2H, since approach of hydride fromeither face of the keto-carbonyl group is equally likely.But when reduction (by a reducing coenzyme) is carriedout in the presence of the chiral enzyme lactic acid dehy-drogenase, only (+)-lactic acid is formed.

A distinct process, sometimes called “enantiospecificsynthesis,” is one in which an enantiomerically pure start-ing material (natural or man-made) is converted by stan-dard reactions into an enantiomerically pure product.

Among other methods are diffusion through chiralmembranes and partition methods involving chiralsolvents.

VIII. STRUCTURE, CONFIGURATION,NOTATION

By “structure” of a molecule we understand the totalityof the nature and array of its atoms. This comprises theidentity and connectivity of these atoms (“constitution”)and their arrangement in space (“configuration,” “con-formation”; see below). Structure may be determined byX-ray, electron, or neutron diffraction of a crystal of thesubstance in question.

Constitution can usually be inferred from elementalanalysis and chemical degradation or by various spectro-scopic methods, such as nuclear magnetic resonance. Con-stitutional isomers, such as butane and 2-methylpropane,have the same elemental composition but differ in con-nectivity. In contrast, the two enantiomers of lactic acid,CH3CHOHCO2H, identical in constitution, differ in thespatial disposition of the ligands at C(2): They are said todiffer in “configuration” and may be called configurationalisomers. While configuration is a property of the moleculeas a whole, it is convenient to “factorize” it into elementsof chirality, notably the “center of chirality” [e.g., at C(2)

FIGURE 5 Ordering substituents (A > B > C > D) according tothe Cahn–Ingold–Prelog chirality rule.

in lactic acid]. This “factorization” allows us to specifyconfiguration by assigning a configurational “descriptor”to each chiral center or other chiral element.

The configurational descriptors now universally usedare “R” (for rectus, right, in Latin) and “S” (for sinister,left) (Cahn et al., 1966); they are sometimes called “CIPdescriptors” after their proponents. To assign a descrip-tor to a chiral center, assume that its four ligands, a, b, c,and d, can be ordered by some convention: a > b > c > d.The molecule is then viewed from the side away from thelowest ranked ligand (d) such that the other three ligands(a, b, c) lie in a plane (Fig. 5). If a–b–c then describe aclockwise array, the descriptor is R; if the array is coun-terclockwise, the descriptor is S. To establish priorities ina real molecule one orders the ligands by atomic number,thus in CHFClBr, Br > Cl > F > H. For lactic acid the pri-ority is O > C > H, but no immediate decision is reachedfor CH3 versus CO2H. Here one goes out to the next atomaway from the chiral center: O for CO2H and H for CH3

and since O > H, CO2H has priority over CH3. Wherestill no decision is reached, one goes out one tier more,thus –CH2CH2OH has priority over –CH2CH2CH2OH.Once a decision is reached, the process stops. In the out-ward path, one always gives preference to the atom ofhigher priority; thus –NHCl has priority over –N(CH3)2:Cl > C overrides C > H. All ligands on the atom reachedmust be probed, thus –CO2H > –CHO (two O’s over oneO). When the lack of a decision is caused by a doublybonded ligand (e.g., CH O vs. CH2OH as in glyceralde-hyde, HOCH2CHOHCH O) the double bond is replacedby two single bonds with the ligands “complemented” ateither end; thus –CH O is considered as O–CH–(OC) andthus has priority over CH2OH. An absent ligand (as thelone pair in N: or :O:) is considered to have atomic num-ber zero. When chirality is due merely to the presenceof an isotope, as in C6H5CHDOH, the ligand of higheratomic weight is given priority: O > C > D > H. For morecomplicated cases, the reader is referred to standard texts.

There are two exceptions to the current use of CIPdescriptors, α-amino acids and sugars, where an oldernomenclature is often used. Before considering this point,a discussion of projection formulas is required. Sincemolecules are three dimensional but paper is planar,

P1: GRB Final Pages


86 Stereochemistry

FIGURE 6 The D and L nomenclature for α-amino acids and aldohexoses. [Reprinted with permission from Eliel, E. L.,and Wilen, S. H. (1994). “Stereochemistry of Organic Compounds,” Wiley, New York.]

several conventions have been developed to representmolecules (or rather their three-dimensional models) intwo dimensions. In one of them, the molecule is writ-ten with as many atoms as possible in the plane of thepaper and with additional atoms being connected with aheavy line (—) when the attached ligand is in front of theplane and with a dashed line (- - -) when it is behind. Inanother, the so-called “Fischer projection” (after its origi-nator Emil Fischer) the tetrahedron is oriented so that twoof the groups (top and bottom) are pointed away from theobserver, with the other two (sideways groups) pointingtoward the observer, and the molecule is projected in thisfashion. Figure 6 shows Fischer projections of α-aminoacids and of monosaccharides (aldoses) with the furtherproviso that the most oxidized group is to be placed atthe top and the NH2 or OH and H ligands on the side.In α-amino acids the symbol D is used when the NH2

group is on the right, L if it is on the left; it turns out thatall naturally occurring α-amino acids are L. In the caseof monosaccharides, the chiral center furthest away fromthe aldehyde or ketone function determines the descriptorand the symbol used is D when the OH is on the right inthe Fischer projection formula, L if it is on the left, in-dependent of the configuration of any of the other chiralcenters.

The Fischer projection formulas shown above, whileuseful for the assignment of the descriptors, do not corre-spond to the actual shape of most molecules. As explainedbelow under the topic of conformation, most molecules ex-ist in “staggered” (Fig. 7) rather than the “eclipsed” con-formations implied in Fischer projections. A more realis-tic representation of, say, (R,R)-tartaric acid is shown inFig. 7, which, in addition to the unrealistic three-dimen-

FIGURE 7 Fischer, sawhorse, and Newman representations of (2R, 3R )-tartaric acid.

sional formula corresponding to the Fischer projec-tion, displays the staggered conformation in a three-dimensional, so-called “saw-horse” formula and itsprojection (seen from one end of the molecule) in a so-called Newman projection. [This staggered representationis generated from the eclipsed one by rotation about theC(2)–C(3) bond.]

IX. DETERMINATION OF CONFIGURATION

Since configuration is an integral part of structure, the de-termination of the architecture of any molecule, naturallyoccurring or synthetic, is not complete until its configura-tion is known. For example, the fit of a drug with its recep-tor or of an inhibitor with an enzyme cannot be understood(or modeled, or rationally improved) absent informationon its configuration.

Configuration may be relative or absolute. To say thata right hand fits a right glove is to make a statement ofthe relative configuration of the two. Even a small childmay be able to make this correlation. But to recognizethat a picture of a glove is that of a right glove in anabsolute sense is more difficult. The same applies to thedetermination of the absolute configuration (or sense ofchirality) of a molecule.

There are many ways of determining relative configu-ration (Eliel and Wilen, 1994). The most straightforwardone, where accessible, is an X-ray crystal structure; sincean X-ray (or neutron or electron) diffraction picture leadsto the positions of the constituting atoms in space, theirrelative orientation (i.e., configuration) can be determined.Relative configuration thus correlates one chiral center

P1: GRB Final Pages


Stereochemistry 87

with another. The two chiral centers need not be in thesame molecule; as will be shown later, configurational cor-relations between two similar molecules can sometimesbe based on comparison of their optical rotations, opti-cal rotatory dispersion, or circular dichroism spectra (seebelow). It is also possible to tie two chiral centers, oneof known absolute configuration, the other unknown, to-gether chemically by either ionic or covalent bonds anddetermine their relative configuration by X-ray diffractionor other means. Since the absolute configuration of one ofthe chiral centers is known, that of the other can then be de-duced. If the compound containing that center can then beseparated by an appropriate chemical reaction, its opticalrotation can be measured and thus the necessary correla-tion between optical rotation ( + or − ) and configuration(S or R) is established. [There is no general relation what-ever between + and − (experimental quantities) and Rand S (descriptors).]

Determination of the absolute configuration (R or S)of an isolated species is more difficult. Enantiomers areindistinguishable in their physical behavior in scalar mea-surements (i.e., in measurements not involving absoluteorientation in space); ordinary X-ray diffraction is of thistype. However J. M. Bijvoet, in the Netherlands, found in1951 that this impediment can be circumvented by em-ploying X-rays of a wavelength close to the absorptionedge of one of the constituent atoms in the molecule to beexamined. This specific absorption (usually by a relativelyheavy atom, such as sulfur or bromine introduced in thespecies to be examined by chemical transformation if nec-essary) leads to a phase shift of the wavefront diffractedby this particular atom. This phase shift causes a pair ofspots in the normally centrosymmetric diffraction pattern(so-called “Bijvoet pairs”) to become unequal in inten-sity; from the relative intensity of these spots the absoluteconfiguration of the compound under investigation can be

FIGURE 8 Chiral compounds lacking chiral centers.

inferred. Thanks to the availability of powerful computers,it has also become increasingly feasible to derive absoluteconfiguration from optical rotation or circular dichroismspectra (see below) by theoretical computation. In somecases, absolute configuration can also be established byexamining the crystal habit (macroscopic dimensions) ofa crystal in the presence of certain impurities (Addadiet al., 1986).

Once the absolute configurations of a few chiralmolecules are known, those of others can be establishedby correlation.

X. CHIRALITY IN ABSENCEOF CHIRAL CENTERS

Although Le Bel’s and van’t Hoff’s understanding of chi-rality rested on the concept of tetrahedral carbon or, moregenerally, of what is now called a chiral center, chiralityis not dependent on the existence of chiral atoms. Anymolecule that is not superposable with its mirror imageis chiral. An example, shown in Fig. 8E, is twistane.A secondary criterion for chirality is the absence of aplane of symmetry or a point of inversion. However, chi-ral molecules may contain simple (proper) axes of sym-metry; twistane, in fact, has three mutually perpendiculartwofold symmetry axes. One class of chiral moleculesalready foreseen by van’t Hoff (though obtained as in-dividual enantiomers only much later) are appropriatelysubstituted allenes, as shown in Fig. 8A. The orbitals areso disposed that the two double bonds are perpendicularto each other, and so two mutually different substituentsat the two termini will give rise to chirality (in contrast tothe cis–trans isomerism of alkenes, Fig. 3).

Related chiral molecules are appropriately substitutedspiranes (Fig. 8B) and alkylidenecycloalkanes (Fig. 8C).These molecules are said to possess chiral axes (along the

P1: GRB Final Pages


88 Stereochemistry

FIGURE 9 Conformers of meso-tartaric acid and of 1,2-dibromoethane.

line of the double bonds or bisecting the rings). Also inthis category are the biphenyls, to be discussed later. Yetanother common type of chirality due to “chiral planes”is seen in the benzenechromium complex in Fig. 8F. (Thenormal symmetry plane of the benzene rings is abolishedby the out-of-plane coordinated chromium atom.)

XI. CONFORMATION

van’t Hoff’s counting of stereoisomers was based onthe assumption that rotation about single bonds was“free,” otherwise there should be at least three isomersof meso-tartaric acid shown in Fig. 9A–C two of which(Fig. 9B, C) would be chiral and enantiomeric. (This ison the assumption that the substituents are staggered—see below—otherwise many more isomers could exist.)In 1932, However, S.-I. Mizushima discovered by vibra-tional spectroscopy that there are, in fact, two isomers of1,2-dibromoethane (Fig. 9D, E; vibrational spectroscopycannot distinguish between enantiomers). It was estab-lished later that the bromine substituents are “staggered”rather than “eclipsed,” meaning that the torsion angle Br–C–C–Br is 60◦ (Fig. 9D) or 180◦ (Fig. 9E) rather than 0◦.The structures in Fig. 9D, E are said to differ in “confor-mation” (rotation about single bonds).

Why are there differing isomers for 1,2-dibromoethaneeven though they cannot be isolated? The answer to this

FIGURE 10 Eclipsed and staggered conformations of ethane. Since the hydrogen atoms on the front carbon obscurethose in the rear in the eclipsed conformation, the torsion angle is offset by a few degrees in the Newman formula.[Reprinted with permission from Eliel, E., Allinger, N. L., Morrison, G. A., and Angyal, S. J. (1981). “ConformationalAnalysis,” American Chemical Society, Washington, DC. Copyright 1981 American Chemical Society.]

question was provided by K. S. Pitzer in 1936 when hediscovered that ethane itself (Fig. 10) exists in staggeredconformation and that the three possible staggered confor-mations are separated by energy barriers of 12.1 kJ/mole(corresponding to the eclipsed conformation as the en-ergy maximum). Such barriers are high enough to allowdetection of the individual conformers by vibrational spec-troscopy but far too low to allow chemical separation.What one sees in chemical behavior (and also in physi-cal measurements involving “slow” time scales, such asmeasurement of dipole moments or of electron diffractionpatterns), is an average of the contributing stable con-formations (called “conformers”) produced by their rapidinterconversion.

The staggered conformations of butane (in Newmanprojections) are shown in Fig. 11 and resemble those in1,2-dibromoethane (Fig. 9). There are three; in two ofthem (the enantiomeric gauche conformers in Fig. 11A,C the terminal methyl groups are close enough togetherto give rise to van der Waals repulsion. Thus theseconformers are less stable (by 4 kJ/mole) than the anticonformer (Fig. 11B), in which the methyl groups areremote from each other. [The relative instability of thegauche (Fig. 9D) relative to the anti confomer (Fig. 9E)is even greater in BrCH2CH2Br because of additionaldipole repulsion in Fig. 9D and its enantiomer.] Theconformations of Fig. 11A, C in which the torsion angleω[C(1)–C(2)–C(3)–C(4)] is 60◦ are called “gauche”

P1: GRB Final Pages


Stereochemistry 89

FIGURE 11 Stable conformers of butane.

(French for “skew”). In Fig. 11B the torsion angle is180◦; this conformation is called “anti.” Torsion anglesoften deviate from the ideal values (60◦ or 180◦) for stag-gered conformations; thus it may be desirable to specifythe exact torsion angle ω when known (e.g., from X-raystructure determination). When C(1)–C(2)–C(3)–C(4)describe a right-handed helical turn, ω is positive; fora left-handed turn, it is negative. As an alternative, asystem of semiquantitative conformational descriptorsmore detailed than gauche and anti has been developed byKlyne and Prelog (1960) and is described in the originalreference and in Eliel and Wilen (1994).

In straight-chain hydrocarbons larger than butane, ro-tation about each single bond is possible, giving rise toa large number of conformations; this situation exists es-pecially in linear polymers, where it was studied early byP. Flory. Even when conformations in which the chain coilsupon itself in such a way as to generate excessive van derWaals (steric) repulsion are excluded, the number of low-energy conformers will be quite large and a Monte Carloapproach may have to be used to find the family of pop-ulated (low-energy) conformers. (Conformers which lie12 kJ/mole or more above the lowest energy conformer arepopulated to the extent less than 1% of the total and maybe neglected for most purposes.) By way of an example, inlinear polyethylene, only a minor fraction of the moleculeswill be in the most stable zigzag (all-anti) conformation.

These considerations are important in the conforma-tion of proteins (natural polymers). When polypeptides aresynthesized—in the laboratory and perhaps also in vivo—they are first formed as linear strings (so-called “randomcoil” conformations, similar to those of a polyethylene).But in polypeptides and proteins additional considerationscome into play, notably hydrogen bonding between aminoacid residues and hydrophobic forces generated by the re-luctance of hydrocarbon side chains to be in contact withthe common water solvent. Additional interactions be-tween nonadjacent amino acids may come about becauseof oxidation of cysteine to cystine residues (2 –SH →–S–S–). These various interactions lead to a folding of thechain into so-called secondary structures, which include ahelical (α-helix) conformation (Pauling et al., 1951) sta-bilized mainly by hydrogen bonding between nonadjacentbut close amino acids in the polymer chain, and a doubled-

up conformation called a β or pleated sheet stabilized byhydrogen bonding between rather more distant membersof the chain folded onto each other. “Tertiary structure” ofproteins comprises the combination of α-helices, pleatedsheets, and some random-coil areas, which gives rise totheir three-dimensional shape.

The distinction between configuration and conforma-tion is usually based on whether the interconversion ofthe pertinent stereoisomers is slow or fast. Since a bar-rier of 84 kJ/mole between two species corresponds toan interconversion rate at 25◦C of 1.3 × 10−2 sec−1, i.e.,a half-life of 1 min, making isolation of the individ-ual species quite difficult, one might say that the divi-sion between configuration and conformation comes atbarriers of about 84 kJ/mole. However, such a precisedistinction is problematic. At lower temperatures, inter-conversion rates decrease and isomers that differ onlyin conformation (cf. Fig. 9) may become isolable. Also,the technique of observation matters. Infrared and Ramanspectroscopy are “very fast” and thus the vibrational spec-tra of the conformational isomers of 1,2-dibromoethaneare distinct. Nuclear magnetic resonance (NMR) is in-termediate, and there are numerous instances where anaveraged NMR spectrum is seen at one temperature butspectra for the individual conformers emerge at lowertemperatures.

An interesting example of the fluidity of the delin-eation between configuration and conformation is seenin the biphenyls (Fig. 12). In biphenyl itself, rotation isfast; thus a 3,3′,5,5′-tetrasubstituted biphenyl (Fig. 12A)cannot be resolved into enantiomers, even though con-formations in which the two rings are not coplanar arechiral. However, as soon as sizable substituents are intro-duced at positions 2, 2′, 6, and 6′ (Fig. 12B, X = Y) thecompounds become resolvable; they display axial chiral-ity. When X and Y (in Fig. 12B) are different and otherthan F or CH3O, the enantiomers are stable. When oneof the four substituents is H, however, the compoundsare resolvable but usually racemize readily either at roomtemperature or above by rotation about the Ar–Ar bond.Biphenyls with only two ortho substituents are generallynot resolvable unless the substituents are very bulky, asin 1,1′-dinaphthyl (Fig. 12C, Z = H). (The enantiomeric2,2′-dihydroxybinaphthyls, Fig. 12C, Z = OH, have foundmanifold uses, e.g., as parts of chiral reagents and chiralcatalysts.)

XII. CYCLOALKANES AND THEIRCONFORMATIONS

Before considering conformation in cyclic molecules(which is more complex since rotations about individual

P1: GRB Final Pages


90 Stereochemistry

FIGURE 12 Stereoisomerism in biphenyls.

bonds are not independent of each other, unlike in alkanechains) we need to consider the topic of “strain.” AlreadyA. von Baeyer in 1885 realized that formation of smallrings, such as cyclopropane or cyclobutane, required de-formation of the normal tetrahedral or near-tetrahedralbond angle of 109◦28′. Thus in cyclopropane the inter-nuclear bond angle is 60◦, i.e., it deviates 49◦28′ from thenormal and this causes angle strain, which in turn destabi-lizes the cyclopropane molecule. If one takes the contribu-tion of a CH2 group to the heat of combustion as 658.7 kJper group [this is the difference in heat of combustion be-tween two large homologous alkanes CH3(CH2)nCH3 andCH3(CH2)n +1CH3], the calculated heat of combustion ofcyclopropane is 3 × 658.7 = 1976 kJ/mole, whereas theexperimental value is 2091 kJ/mole; the difference of115 kJ/mole is a measure of the strain in cyclopropane.Corresponding values are, for cyclobutane, 110 kJ/mole;for cyclopentane, 26.0 kJ/mole; and for cyclohexane,0.5 kJ/mole. The low value for cyclohexane is at firstsight surprising; Baeyer thought that cyclohexane wasplanar, with bond angles of 120◦, and should thereforeshow strain of 120◦ − 109◦28′ or 10◦32′, though thisstrain would be due to enlargement rather than diminu-tion of the bond angle. There are two ways of accountingfor this discrepancy. First, while strain increases againfor the so-called “medium rings” (C7, 26.2 kJ/mole; C8,40.5 kJ/mole; C9, 52.7 kJ/mole; C10, 51.8 kJ/mole; C11,47.3 kJ/mole), it diminishes thereafter for the “large rings,”e.g., to 8.0 kJ/mole for C14. This is due to the fact that ringsother than cyclopropane are actually not planar and thereare different sources of strain in these rings (and actu-ally even in cyclopropane). One source is strain due toeclipsing of bonds (“torsional strain”), as explained abovefor ethane. In planar cyclobutane and cyclopentane, thisstrain (due to four or five pairs of eclipsed hydrogen atoms,respectively) is large enough to cause these species to benonplanar, even though this increases angle strain. (A non-planar polygon has smaller angles than a planar one.) Inthe larger cycloalkanes, however, where (if they were pla-nar) the angles would be expanded beyond the tetrahedral,puckering actually diminishes not only the torsional oreclipsing strain (see the discussion on ethane above) butalso the angle strain. In fact, much of the strain in mediumrings is due to nonbonded atoms getting too close to each

other; this causes so-called “nonbonded” or van der Waalsstrain (steric repulsion).

Cyclohexane, which is virtually strain-free, is a spe-cial case. In 1890 (only 5 years after Baeyer proposedhis strain hypothesis) H. Sachse realized that C6H12 isnot planar, but can be constructed from tetrahedral car-bon atoms, either in the shape of a chair (Fig. 13A) orthat of a boat (Fig. 13B). Today we know that, because ofsteric repulsion between the hydrogen atoms at C(1) andC(4) pointing inside plus eclipsing strain at C(2, 3) andC(5, 6), the shape in Fig. 13B is actually deformed to a“twist-boat” form (Fig. 13C) and that the chair (Fig. 13A)is the most stable conformer. But it took some 60 yearsafter Sachse for the physical and chemical consequencesof the chair shape of cyclohexane to be recognized, by K.Pitzer, O. Hassel, and D. H. R. Barton. Chemically speak-ing, axial substituents are more hindered (crowded) thanequatorial ones and therefore generally less stable, andreact more slowly (e.g., in the esterification of acids andalcohol and the hydrolysis of the corresponding esters).Also, the bimolecular elimination reaction (e.g., of H2Oin cyclohexanols or HX in cyclohexyl halides and tolue-nesulfonates) proceeds more readily when the substituent(OH or X) is axial than when it is equatorial. Barton sawthese consequences (and others) of cyclohexane confor-mation (Eliel et al., 1965) in the rigid cyclohexane systemsof steroids and terpenes. Thus in 3-cholestanol (Fig. 14)the equatorial or β isomer is more stable than the axial one(designated α, meaning that the substituent is on the sideopposite to the angular methyl groups, whereas β impliesthat it is on the same side). Also, the β isomer is more eas-ily esterified than the α, but elimination of water to give acholestene is more facile for the axial α isomer.

In monocyclic cyclohexanes the situation is more com-plex since the barrier to interconversion of the ring is only

FIGURE 13 Conformations of cyclohexane.

P1: GRB Final Pages


Stereochemistry 91

FIGURE 14 Conformation of 3β-cholestanol.

about 42 kJ/mole and interconversion of the two conform-ers is extremely rapid at room temperature. Thus chloro-cyclohexane (Fig. 15, X = Cl) exists in rapid equilibriumbetween axial and equatorial conformers which differ infree energy by only about 25 kJ/mole, corresponding to74% of the equatorial and 26% of the axial isomer at 25◦C.As expected, in the infrared spectrum there are two C−Clstretch frequencies, but in the laboratory, chlorocyclohex-ane, even though a mixture of two conformers, appears as ahomogeneous substance with the average properties (suchas chemical shifts in NMR) of the two conformers. Whenone cools the substance to ca. −60◦C (the exact tempera-ture required depends on the operational frequency of theNMR instrument), however, two different NMR spectrabegin to emerge, and at −150◦C the equatorial conformerhas actually been crystallized from trideuteriovinyl chlo-ride solution, with concomitant enrichment of the axialisomer in solution.

Equilibria for a large number of monosubstituted cyclo-hexanes have been determined and tabulated; they weremostly determined by low-temperature 13C NMR spec-troscopy (Eliel and Wilen, 1994).

The conformations of piperidine (azacyclohexane) andtetrahydropyran (oxacyclohexane) are qualitatively simi-lar to those of cyclohexane. (Some quantitative differencesare seen, for example, in the equatorial preferences ofsome substituents resulting from dipolar interactions withthe ring hetero atom and from the fact that C–N and C–Odistances are shorter than C–C in cyclohexane.) Thesering systems are important, being found in alkaloids andhexose sugars, respectively.

Because of torsional (eclipsing) strain, cyclobutane andcyclopentane are not planar. Cyclobutane is wing-shaped;cyclopentane oscillates among a number of low-energy

FIGURE 15 Conformational inversion of substituted cyclohex-ane. [Reprinted with permission from Eliel, E. L., and Wilen,S. H. (1994). “Stereochemistry of Organic Compounds,” Wiley,New York.]

conformations of which the envelope (four carbon atomsin a plane, one out of plane) and the half-chair (three ad-jacent carbons in a plane, the fourth above that plane andthe fifth below) are the most symmetrical. The barrier be-tween these conformations is very low; thus their rapidinterconversion, which involves up-and-down motions ofsuccessive adjacent carbon atoms, has the appearance ofa bulge moving around the rings; this process has there-fore been named “pseudorotation.” Higher cycloalkanesfrom C(7) on display families of conformations whichare separated by barriers similar to those in cyclohexane,but within a given family there may be several individualmembers interconverted by pseudorotation. This subjectis discussed in detail in Eliel and Wilen (1994).

XIII. CHIROPTICAL PROPERTIES.ENANTIOMERIC PURITY

By “chiroptical properties” are meant optical propertiesthat differ between enantiomers and can be used to char-acterize them. They comprise optical rotation, optical ro-tatory dispersion (ORD), and circular dichroism (CD).

Optical rotation has already been discussed. Becauseof its critical dependence on solvent (including cosol-vents, such as ethanol in chloroform), temperature, con-centration, and the potential presence of impurities, es-pecially chiral impurities, in the sample, experimentaldetermination of [α] requires considerable care and manyof the values given in the literature cannot be trusted. Thisis unfortunate since it is often desirable to determine the“enantiomeric purity” of a sample to see whether the de-sired enantiomer is obtained free of the other. (For exam-ple, in pharmaceutical chemistry one wants to obtain thepure eutomer free of the distomer; see above.) Since in allbut a few cases optical rotation is proportional to the frac-tion of the major enantiomer in the total substance, onemight expect enantiomeric or optical purity to be equal to100[αobs]/[αmax]%, where [αmax] is the (presumed known)specific rotation of an enantiomerically pure sample. How-ever, this is true only if both values have been accuratelydetermined under exactly the same conditions (solvent,temperature, etc.).

Because of this difficulty, other, more reliable methodsof determining enantiomeric purity (now no longer called“optical purity”) have been developed. Basically these de-pend on converting the enantiomers into diastereomers,either by covalent chemical bonding or by complexingin some fashion, with another, usually enantiomericallypure chiral auxiliary. (For some methods the auxiliaryneed not even be enantiomerically pure.) Once the enan-tiomers have been converted into diastereomers, their ratiocan be determined by NMR or chromatographic methods,

P1: GRB Final Pages


92 Stereochemistry

including NMR in a chiral solvent or with a chiral com-plexing agent, or by chromatography of unmodified enan-tiomers on a chiral stationary phase. The methods justdescribed provide the mole fractions of each individualenantiomer, which may be expressed as an enantiomerratio (n+/n− or n−/n+). However, because optical pu-rity had been used in the past, it is more common touse an expression for “enantiomeric excess” (e.e.): e.e. =|(n+ − n−)|/(n+ + n−). Since this represents the molefraction of the major enantiomer diminished by that of theminor one, it is equal to the above-defined optical purity.

Although there is no direct relation between sign ofrotation and configuration, it may now become possibleto infer configuration from optical rotation data (Kondra,Wipf, and Beratan, 1998). The method is based on van’tHoff’s “optical superposition” rule, which says that in amolecule containing several chiral centers, the total molarrotation ( = 100α/MW ) is the sum of the contributionsof each chiral center. As originally proposed the rule hadseveral shortcomings, including (1) that it would requiremodel compounds of known absolute configuration to de-termine the contribution of a given chiral center and (2)it does not hold when the centers are close to each otherand thus influence each other’s contributions. This secondlimitation means that for closely connected chiral centers,it is necessary to determine the contribution of a segmentcontaining all of these centers. The first problem is beingattacked by performing a priori computations of the mo-lar rotation contribution of individual chiral centers or ap-propriate groupings thereof (Kondra, Wipf, and Beratan,1998). This is becoming possible due to the advances inquantum chemical calculations (e.g., by density functionalmethods) and the increasing power of computers to handlecomputationally demanding problems.

Other chiroptical techniques to infer configuration areoptical rotatory dispersion (ORD) and circular dichroism(CD). ORD relates to the change in optical rotation withthe wavelength of the light employed in the measure-ment. Normally the absolute value of rotation increasesas wavelength becomes shorter; observation at shorterwavelengths is thus a convenient way to increase rota-tion (and thereby the accuracy of measuring it) when αD

is small. However, as the wavelength approaches that of aUV absorption band (e.g., of C O in a ketone), its abso-lute value (whether positive or negative) suddenly dropsprecipitously, passes through zero near the UV absorp-tion maximum, reverses sign rapidly approaching anotherextremum (of opposite sign to the first), and then gradu-ally declines. This phenomenon of rapid change at the UVmaximum is called the “Cotton effect” after the French sci-entist who discovered it. Similarity in Cotton effects of re-lated compounds, one of known and one of unknown con-figuration, can sometimes be used to assign the configura-tion of the unknown. However the use of ORD has largely

been superseded by the simpler to interpret CD (Nakanishiet al., 2000). While it might appear that plane-polarizedlight is achiral, it may actually be considered to be a su-perposition of right- and left-circularly polarized light inwhich the sense of polarization changes as a right-handedor left-handed helix along the direction of propagation ofthe light beam. If the right- and left-circularly polarizedbeams proceed at the same speed, the result is light polar-ized in an unchanging plane, but if one of the two generat-ing beams moves faster than the other, the plane of polar-ization will keep turning as the light propagates, i.e., therewill be optical rotation. Since the speed of light dependson the refractive index of the medium it traverses, thepolarization is thus caused by unequal refractive indicesfor right- and left-circularly polarized light. There are de-vices that can produce right- and left-circularly polarizedlight beams separate from each other. Using such beams,it is found that not only the refractive indices, but alsothe absorption coefficients for the two beams differ. Thephenomenon resulting from this difference in absorptionis called “circular dichroism” and manifests itself in whatlooks similar to an absorption curve in the UV, except thatit is signed. (In fact its maximum or minimum occurs atthe wavelength of the UV maximum.) Comparison of CDabsorption spectra can be used to infer configuration sim-ilarly as was the case for ORD; however, CD spectra arebetter resolved than ORD spectra since there is less bandoverlap resulting from multiple UV absorption maxima.

CD is also very useful in throwing light on confor-mation. Thus the (weak) CD absorption spectrum of arandom-coil polypeptide chain is essentially a superposi-tion of the spectra of the individual constituting aminoacids. However, when secondary structure comes intoplay, as in an α helix of β-pleated sheet, a large and char-acteristic increase in CD absorption is observed, which, inturn, allows one to infer the nature of the secondary struc-ture, if any. CD is used to infer not only protein but alsonucleic acid and polysaccharide conformation (Fasman,1996).

XIV. PROCHIRALITY

The phenomenon of prochirality (Mislow and Raban,1967) is important both in NMR spectroscopy and inenzyme chemistry. An atomic center (e.g., a tetrahedralcarbon atom) in a molecule is considered “prochiral” ifreplacing one of two identical ligands at the center by adifferent one not previously attached to that center pro-duces a chiral center. Thus the carbon atom in CH2ClBris prochiral since hypothetical replacement of one of thehydrogen atoms by deuterium yields CHDClBr, whichis chiral. The apparently identical (or “homomorphous,”from Greek “homos,” same, and “morphe,” form) hy-drogen atoms in CH2ClBr are in fact distinct; they are

P1: GRB Final Pages


Stereochemistry 93

called “heterotopic” (from Greek “heteros,” different, and“topos,” place). In contrast, the carbon atom in CH3Cl isnot prochiral since replacement of H by, say, D wouldproduce CH2DCl, which remains achiral.

In the former case (CH2ClBr) replacement of one orother of the two hydrogen atoms gives rise to enantiomericproducts. The hydrogen atoms are therefore called “enan-tiotopic.” In the latter case, replacement gives the samecompound and the hydrogens in CH3Cl are called “ho-motopic.” In a molecule such as CH2BrCHOHCO2H,replacement of one of the terminal hydrogens by, say,chlorine would give one or other of the diastereomers ofCHBrClCHOHCO2H; in this case the terminal hydrogensare said to be diastereotopic. These definitions of homo-topic, enantiotopic, and diastereotopic ligands also pro-vide a means for their recognition: Replacement of oneof two or more homotopic ligands by a different ligandgives identical products, analogous replacement of enan-tiotopic ligands gives enantiomeric products, and such re-placement of diastereotopic ligands gives diastereomericproducts. There is also a symmetry criterion which maybe applied to the appropriate molecules above: Homotopicligands in a molecule are interchanged by operation of bothsimple symmetry axes and symmetry planes; enantiotopicligands are interchanged by operation of a symmetry planebut not by operation of a simple symmetry axis, and di-astereotopic ligands are interchanged neither by symmetryaxes nor by symmetry planes.

Diastereotopic ligands (e.g., protons or C-13 atoms)generally display distinct signals in NMR spectra, but ho-motopic and enantiotopic ligands have coincident (iden-tical) signals (except possibly in the case of enantiotopicligands, in a chiral solvent, or in the presence of a chi-ral complexing agent) since NMR is an achiral technique.Both enantiotopic and diastereotopic ligands may be dis-tinguished by enzymes (which are chiral). Thus in citricacid, HO2CCH2C(OH)(CO2H)CH2CO2H, all four methy-lene hydrogen atoms are distinguished by enzymes in thecitric acid cycle (to demonstrate this distinction, they mustbe individually labeled as deuterium atoms). On the otherhand, the CH2 groups are pairwise identical in NMR (e.g.,C-13) but the geminal hydrogen atoms in each are di-astereotopic and provide an (AB)2 system in the protonNMR spectrum. Further details may be found in Eliel(1982) and Eliel and Wilen (1994).


BIOPOLYMERS • ENZYME MECHANISMS • NUCLEAR

MAGNETIC RESONANCE • ORGANIC CHEMICAL SYSTEMS,

THEORY • ORGANIC CHEMISTRY, SYNTHESIS • PHYSICAL

ORGANIC CHEMISTRY • PROTEIN STRUCTURE • PROTEIN

SYNTHESIS • RHEOLOGY OF POLYMERIC LIQUIDS

BIBLIOGRAPHY

Addadi, L., Berkovitch-Yellin, Z., Weissbuch, I., Lahav, M., andLeiserowitz, L. (1986). “A link between macroscopic phenomena andmolecular chirality: Crystals as probes for the direct assignment ofabsolute configuration of chiral molecules. In “Topics in Stereochem-istry,” Vol. 16, pp. 1–85, Wiley, New York.

Cahn, R. S., Ingold, C., and Prelog, V. (1966). “Specification of molec-ular chirality,” Angew. Chem. Int. Ed. Engl. 5, 385–415 .

Eliel, E. L. (1982). “Prostereoisomerism (prochirality).”In “Top-ics in Current Chemistry,” Vol. 105, pp. 1–76, Springer-Verlag,Heidelberg.

Eliel, E. L., and Wilen, S. H. (1994). “Stereochemistry of Organic Com-pounds,” Wiley, New York.

Eliel, E. L., Allinger, N. L., Angyal, S. J., and Morrison, G. A. (1965).“Conformational Analysis,” Wiley, New York [reprinted (1981),American Chemical Society, Washington, DC.]

Fasman, G. D. (ed.). (1996). “Circular Dichroism and the Conforma-tional Analysis of Biomolecules,” Plenum Press, New York.

Gawley, R. E., and Aube, J. (1996). “Principles of Asymmetric Synthe-sis,” Pergamon Press, Oxford.

Hegstrom, R. A., and Kondepudi, D. K. (1990). “The handedness of theuniverse,” Sci. Am. 262(January), 108–115.

Jacques, J., Collet, A., and Wilen, S. H. (1981). “Enantiomers, Race-mates and Resolutions,” Wiley, New York.

Juaristi, E. (ed.). (1995). “Conformational Behavior of Six-MemberedRings,” VCH, New York.

Kagan, H. B., and Fiaud, J. C. (1988). “Kinetic resolution.” In “Topicsin Stereochemistry,” Vol. 18, pp. 249–330, Wiley, New York.

Klyne, W., and Prelog, V. (1960). “Description of stereochemical rela-tionships across single bonds,” Experientia 16, 521–523.

Kondru, R. K., Wipf, P., and Beratan, D. N. (1998). “Atomic contribu-tions to the optical rotation angle as a quantitative probe of molecularchirality,” Science 282, 2247–2250; id. (1998). Theory-assisted de-termination of absolute stereochemistry for complex natural productsvid computation of molecular rotation angle, J. Am. Chem. Soc. 120,2204–2205.

Mislow, K., and Raban, M. (1967). “Stereoisomeric relationships ofgroups in molecules.” In “Topics in Stereochemistry,” Vol. 1, pp. 1–38, Wiley, New York.

Nakanishi, K., Berova, N., and Woody, R. W. (eds.). (2000). “Cir-cular Dichroism: Principles and applications,” Wiley-VCH, NewYork.

Pauling, L., Corey, R. B., and Branson, H. R. (1951). “The structure ofproteins: Two hydrogen bonded helical configurations of the polypep-tide chain,” Proc. Natl. Acad. Sci. U.S.A. 37, 205–211.

Prelog, V., and Helmchen, G. (1982). “Basic principles of the CIP systemand proposals for a revision,” Angew. Chem. Int. Ed. Engl. 21, 567–583.

Ramsay, O. B. (1981). “Stereochemistry,” Heyden & Son, Philadelphia.Sih, C. J., and Wu, S.-H. (1989). “Resolution of enantiomers via bio-

catalysis.” In “Topics in Stereochemistry,” Vol. 19, pp. 63–125, Wiley,New York.

ebooksclub.org encyclopedia of physical science and technology 3e organic chemistry

Documents