modelingthechemistryofcomplex petroleum mixtures · modelingthechemistryofcomplex petroleummixtures...

Modeling the Chemistry of ComplexPetroleum MixturesRichard J. QuannMobil Technology Company, Paulsboro, New Jersey

Determining the complete molecular composition of petroleum and its refined products is notfeasible with current analytical techniques because of the astronomical number of molecularcomponents. Modeling the composition and behavior of such complex mixtures in refineryprocesses has accordingly evolved along a simplifying concept called lumping. Lumping reducesthe complexity of the problem to a manageable form by grouping the entire set of molecularcomponents into a handful of lumps. This traditional approach does not have a molecular basisand therefore excludes important aspects of process chemistry and molecular propertyfundamentals from the model's formulation. A new approach called structure-oriented lumpinghas been developed to model the composition and chemistry of complex mixtures at a molecularlevel. The central concept is to represent an individual molecule or a set of closely related isomers

as a mathematical construct of certain specific and repeating structural groups. A complexmixture such as petroleum can then be represented as thousands of distinct molecularcomponents, each having a mathematical identity. This enables the automated construction oflarge complex reaction networks with tens of thousands of specific reactions for simulating thechemistry of complex mixtures. Further, the method provides a convenient framework forincorporating molecular physical property correlations, existing group contribution methods,molecular thermodynamic properties, and the structure-activity relationships of chemical kineticsin the development of models. Environ Health Perspect 106(Suppl 6):1441-1448 (1998).http.//ehpnet1.niehs.nih.gov/docs/1998/Suppl-6/144 1-1448quann/abstract.html

Key words: complex mixtures, structure-oriented lumping, petroleum, mathematical modeling,structure-activity relationships, chemical kinetics, molecular thermodynamics

Many chemical reaction systems found inindustry and nature are complex in termsof the large number of molecular compo-nents and chemical reactions. Examplesinclude petroleum refining, biologic sys-tems, and combustion processes. The num-ber of components can exceed 104 and thenumber of chemical reactions can be anorder of magnitude greater. Developingmodels of such large reacting systems is dif-ficult because of the scale and the lack offundamental information. Analytical tech-niques cannot identify and measure all theindividual molecular components when somany are present a mixture. It is alsoimpractical to study the mechanisms andkinetics of all possible chemical reactions.

This paper is based on a presentation at theConference on Current Issues on Chemical Mixturesheld 1 1-1 3 August 1997 in Fort Collins, Colorado.Manuscript received at EHP 17 February 1998;accepted 30 July 1998.

Address correspondence to R.J. Quann, MobilTechnology Company, Paulsboro, NJ 08066.Telephone: (609) 224-3251. Fax: (609) 224-3843.E-mail: richard2 [email protected]

Abbreviations used: LFER, linear free-energy rela-tionship; SOL, structure-oriented lumping.

Further, how could this information bemanaged, organized, and formulated into amodel simulating the reactions of so manycomponents? Accordingly, models devel-oped for complex reacting systems haveused a technique called lumping (1,2) tosimplify the representation of compositionand chemistry.

Lumping partitions the entire molecularpopulation into a small number of lumps(approximately 10) determined by the lim-itations in measuring mixture composition,the similar chemistry and physical proper-ties of molecular components within alump, and the model's intended purpose.This approach reduces complexity andeliminates the need for a detailed molecularcharacterization, albeit at the expense of itsusefulness. Using petroleum refining as anexample, the model shown in Figure 1 wasdeveloped in the 1970s to predict gasolineyield for the fluid catalytic cracking processthat converts heavy gas oil containing high-molecular-weight hydrocarbons to the lowermolecular weight components of gasoline.It represents the complex composition ofgas oil as eight lumps based on the crudeanalytical capabilities of that time-broad

molecular categories and boiling range.Although this model can predict gasolineyield from these lumps, it is not structuredto predict gasoline composition or theimpact of composition on the required qual-ity specifications. In addition, the gasolineyield is not reliably predicted because thelumping scheme cannot represent otherimportant variations in gas oil composi-tion and because the process chemistry isinadequately represented.

Models developed for refining processesmust have the capability of predicting notjust accurate product yields but also theproduct's molecular composition, theboiling point distribution, certain mixturephysical properties and quality specifica-tions that depend on composition, andthe plant's operating conditions. A molecu-lar approach, if possible, could satisfy theserequirements by incorporating moredetailed petroleum characterization, processchemistry, and molecular property corre-lations. Recently developed analyticalmethods are capable of probing someaspects of petroleum's molecular structureand composition, and there is an exten-sive, but not complete, knowledge base ofthe chemistry of refining processes fromlaboratory studies of selected hydrocar-bons. A recently developed approachcalled structure-oriented lumping (SOL)attempts to model the complex compo-sition and chemistry of petroleum mix-tures at the molecular level (3). Thispaper reviews the concepts of SOL andthe strategy for developing molecular-based models of reaction systems wherethe number of molecular components isvery large and their molecular structurecannot be completely determined.

Organizing theComposition of PetroleumThe basic concept of the SOL approach isthat any hydrocarbon molecule can beconstructed from a set of different struc-tural features or increments. A structuralincrement is a specific combination or con-figuration of C, H, S, N, and 0 atoms,often occurring in different molecules. Theset of 22 increments is shown in Figure 2.The increments consist of three types ofaromatic rings (A6, A4, A2), six types ofnaphthenic rings (N6-NI), a methylene-CH2- group (R), bridging between rings(A-A), hydrogen deficiency (H), hetero-atom structures containing S, N, and 0,and the degree of branching (br) and ring

Environmental Health Perspectives * Vol 106, Supplement 6 * December 998 1441

R.J. QUANN

$C__I < | 650°F T

i,0gwBg \ \ ~~~~~~~~~carbonson

* m 2 B 9 ~~~~~650°F+ aromatic||65°Fparaffins Gasoline carbon

430 -650OF 430-650°F|,, . c-O co \ \naphthenes aliphatic

4¢< m~ ~~w\ V \ aromatics

1 X~ 1><\\ \ I I I / 430-650OF|~~~# ~~430-650OF aoai

and coke

Figure 1. Modeling the fluid catalytic cracking process using a simplified representation of petroleum's complexcomposition.

A6 A4 A2 N6 N5 N4 N3 N2 NI R br me H A_A S RS AN NN RN 0 RO O=

899 O U 3 9 ) > ~c< C12 SH N H NH2 ON 0

H H HH H H H HH

H H H

tA_A

A6 A6Niigi.g

Figure 2. Structural increments and their stoichiometry for constructing molecules. Modified from Quann andJaffe (3).

substitutions (me). Each increment has aspecific C, H, S, N, and 0 stoichiometryand thus molecular weight. The termincrement is applicable because most can-not exist independently but must occur asan incremental part of a molecule.

The SOL method mathematicallyorganizes this set of increments into avector, with each element of the vector cor-responding to one of the increments. Amolecule is represented as a 22-elementnumerical vector, with each moleculedistinguished by different numbers andtypes of increments, as shown in Figure 3.

For example, a vector with a one in thefirst element or position (A6) and zero forall others represents a benzene molecule.Naphthalene is constructed with A6 andA4 increments, having one for the first twoelements of the vector and zero for all oth-ers. Other examples are shown in Figure 3.Molecular elemental stoichiometry andmolecular weight are easily computed fromthe contributions of the increment'selemental stoichiometry.

Any molecule found in petroleum canbe represented by this method, altough notnecessarily uniquely. Many molecules may

have identical vector representation. Theincrements count structures present in themolecule. Increment spatial orientation isnot represented by the structure vector.Molecules having the same structure vec-tor but whose increments are arranged dif-ferently are called structural isomers.Examples of structural isomer sets thathave identical vector representations butwith clearly different spatial arrangementof the increments are shown in Figure 4.The br and me increments do not con-tribute to the molecule's stoichiometry butare used for a limited description of cer-tain isomeric features: the number ofbranches on a paraffin or alkyl chain andthe number of methyl substituents onrings, respectively. Again, the method asshown here does not assign specific loca-tions for these branches or ring substitu-tions, only the total number for themolecule. For example, there are only 21possible structure vector representationsfor all possible C43 paraffins, yet there areover a thousand-billion isomeric permuta-tions. Not distinguishing structural iso-mers is a practical consequence ofanalytical limitations, but it also has otherimplications, as we will discuss.A complex mixture in SOL is represented

as a set of vectors with each vector corre-sponding to a molecule or an ensemble ofstructural isomers. The composition of themixture is expressed as the weight or molepercent of each vector. It is helpful to visual-ize the mixture using the concepts of molec-ular class and homologous series. Somecommon molecular classes found in petro-leum or refinery products are shown (with-out alkyl substituents) in Figure 5. They areall described by the SOL method and havedifferent combinations of structural incre-ments. A homologous series is composed ofmolecules from the same molecular class,varying only in R, br, and me because of thepresence of alkyl substituents. The numberfor R is the total number of carbons for aparaffin or olefin or on the alkyl substituentsof a ring compound. The simplest exampleof a homologous series in petroleum is thenormal paraffins (br = me = 0) extendingfrom methane (R= 1) to over 60 carbonatoms. Other homologous series are morecomplicated in that br and me may varywith R to represent some average branch-ing and substitution for the set of isomersat each carbon number of each molecularclass. A complex petroleum mixture inSOL is thus composed of some 5000 to10,000 vectors-one for each molecule orisomer ensemble-and organized into

Environmental Health Perspectives * Vol 106, Supplement 6 * December 1998

Increment StoichiometryC: 6 4 2 6 5 4 3 2 1 1 0 0 0 0 -1 0 -1 -1 0 -1 0 0H: 6 2 0 12 10 6 4 2 0 2 0 0 2 -2 -2 0 -1 -1 1 -2 0 -2S: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0N: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 00: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1

C - carbon, H - hydrogen, S - sulfur, N - nitrogen, 0 - oxygen

Structural Assembly:

A6 N4

1442

MODEUNG THE CHEMISTRY OF COMPLEX PETROLEUM MIXTURES

A6 A4 A2 N6 N5 N4 N3 N2 Ni R br me H A_A S2,3,4-Trimethylpentane

Benzene

0Naphthalene

00Phenanthrene

Pyrene

Carbazol

Benzofuran

Fluoranone

Adamantane

Cholesterol

WaterH20

RS AN NN RN 0 RO 0=

0 0 0 0 0 0 0 0 0 8 3 0 1 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 00

1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

2 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0

1 0 0 0 0 1 0 0 0 0 0 0 -1 0 0 0 0 0 0 1 0 0

2 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 1

0 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 1 0 2 1 0 0 10 2 2 -1 0 0 0 0 0 0 0 1 0

0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0

Figure 3. Examples of the structure-oriented lumping vector representation of molecular structure. Modified from Quann and Jaffe (3).

Figure 4. Examples of isomers having the same structure-oriented lumping vector representation.

about 1 50 molecular classes and their including isomers, of the low-boilinghomologous series. (< 450°K), gasoline-range material in

Describing molecular structure at this petroleum or products. However, thelevel is consistent with our ability to analyze numerous possible increment spatial orien-petroleum. Modern gas chromatography tations result in far too many componentstechniques routinely identify and quantify (isomers) for analytical techniques tomost of the several hundred components, identify in specific molecular structures that

boil above 450K. The only measurementis the number of molecules contributing toa molecular mass. By combining chro-matographic separations with mass spec-trometry (4), the most likely molecularclass and the total carbon number (R) ofalkyl groups are inferred at each molecularmass, providing the homologous seriesdistribution of each molecular class.

The structure of the alkyl groups andall their possible permutations cannot bedetermined at each mass, with the possibleexception of n-alkyls. An average or arche-type structural arrangement of alkylgroups can be inferred from nuclear mag-netic resonance measurements of the mix-ture but this does not necessarily imply acommon arrangement. Similarly, the spa-tial arrangement of ring structures canhave many permutations. The biologic ori-gin of petroleum would suggest that thecata configuration found in terpenes andsteroids is the likely isomeric form of themultiring structure (5). Hence, the SOLdescription of molecular structure is consis-tent with the limitations of current analyti-cal methods. The SOL approach assigns amolecular structure to the large number(approximately 5000) of distinguishable


0

ternQ..T .T.T

0

jI

uJQTg

H

_]mT

Figure 5. Common molecular classes found in petroleum.

masses or isomer sets in petroleum, lumpingonly the isomers at the same carbon numberin the same molecular class. Additionaldescriptors in the SOL method could, ofcourse, be incorporated to locate and sizealkyl substitutions and branching on non-

cyclic molecules should more detailed ana-

lytical information become available. Ifspatial orientation is crucial to the modelingof chemical processes, as in biochemicalsystems, more sophisticated methods ofmolecular descriptions could be devised.

Describing the Chemistryof Refinery ProcessesWhy describe molecules mathematicallywith structural increments and why theseparticular increments? Although the SOLmethod provides a basis for organizing themolecular components of petroleum, its

main purpose is to provide a convenientframework for developing a mathematical,or algorithmic, description of complexprocess chemistry. A characteristic featureof the catalytic and thermal chemistry ofrefining processes is that certain specificmolecular rearrangements occur repeatedlyon many different types of molecules.These particular increments correspond tothose structural entities that are rearrangedduring reactions.

Using a limited set of structural groupsto describe thousands of componentsenables the use of a limited set of reactionrules to establish the complex reactionnetworks involving tens of thousands ofreactions. Each reaction rule consists of a

reactant selection rule and a product gener-ation rule. The reactant selection rule firstidentifies which molecules in the system can

undergo a certain structural rearrangementthat characterizes a particular type ofchemical reaction. Logical constructs are

applied to the vectors to determine whichcomponents have the increment(s)required for the reaction. The product gen-eration rule is used to convert each reac-

tant's vector to a corresponding productvector. Examples of reaction rules for aro-matic ring saturation, ring opening, anddealkylation are given in Figure 6. Forexample, the aromatic saturation ruleshown in Figure 6 determines if a moleculehas the A4 ring required for the reaction,then converts this ring to an N4 naph-thenic ring through a simple mathematicaloperation. The reactant vector's A4 ele-ment is decreased by one and the N4 ele-ment is increased by one to create theproduct vector. Molecules without an A4


RJ. GUANN

00

Il.

gJrj 0 1JI.I.J.T.T

.li

S

rs FeAlO

H

H

tJ

H

H

07S3 1

in,us-id&

1 444

MODEUNG THE CHEMISTRY OF COMPLEX PETROLEUM MIXTURES

Aromatic saturation A6 A4 A2 N6 N5 N4 N3 N2 Ni R br me H A_A S RS AN NN RN 0 RO 0=

1 i 0 0 0 9 1 2 O O O 0 O O O O O O1 L O O O 0 9 1 2 0 0 0 0 0 0 0 0 0 0

2H2o

Reactant selection rule: A4 1Product generation rule: A4*- A4 - 1

N4-N4+ 1

Naphthenic ring opening A6 A4 A2 N6 N5 N4 N3 N2 Ni R br me H A_A S RS AN NN RN 0 RO 0=

, 1 1 0 0 0 0 0 0 9 1 2 0 0 0 0 0 0 0 0

H2 1 1 0 0 0 0 0 |13 2 2 2 O O 0 0 O0

Reactant selection rule: N4 2'1Product generation rule: N4*- N4 - 1

R*- R + 4br- br + 1

Dealkylation A6 A4 A2 N6 N5 N4 N3 N2 Ni R br me H A_A S RS AN NN RN 0 RO 0=

an

* d 1 1 0 0 0 0 0 0 0 13 2 2 0 0 0 0 0 0 0 0 0 0

H2 1 1 0 0 0 0 0 0 0 3 0 2 0 0 0 0 0 0 0 0 0 00 0 0 0 0 0 0 0 0 10 2 0 1 0 0 0 0 0 0 0 00

Reactant selection rule: (A6 2 1) A (R > 3)Product generation rule:

First product: R*- 1 + rings

Second product: R4- R - 1 + rings

Figure 6. Formulating chemical reaction rules using the structure-oriented lumping representation for molecular structure. Modified from Quann and Jaffe (3).

ring are not selected for this reaction.Although hydrogen is also a reactant in thisand many other reactions, it is not neces-sary to explicitly include it in the formula-tion of the rule. Information on thehydrogen stoichiometry is automaticallyobtained from the difference in the H con-tent of reactant and product molecules andreadily computed from the H content ofthe increments.

Some reaction rules, particularly thoseinvolving transformations of alkyl groupssuch as dealkylation or cracking, are obvi-ously dependent on what is assumed forstructure of the alkyl groups. Only experi-mentation with petroleum mixtures canreveal how these rules should be formu-lated, as it must apply to isomers having avariety of substituent structures.

Applying a reaction rule to all molecules(or vectors) in the mixture generates areaction class for the specific chemicaltransformation represented by the rule. Areaction class may have thousands of reac-tants and their corresponding products.Each reactant has the necessary increment

for the rule. Figure 7 illustrates how reactionrules generate numerous reactant-productpairs (only a limited number are shown) fora mixture for two reaction dasses.

Additional complexity arises from thefact that a molecule may satisfy the reac-tant selection criteria of more than onerule, i.e. a reactant may have parallel reac-tion pathways leading to different prod-ucts. For example, the partially saturatedmultiring aromatic molecule in Figure 6may undergo further saturation, ringopening, dealkylation, or even cracking ofthe alkyl group. Application of all rules toall vectors generates the entire complexreaction network. Application of severalrules in Figure 8 to just one componentand several of its offspring illustrates thegeneration of only a small piece of thecomplex network. Catalytic refiningprocesses have 20 to 40 reaction classes orrules, resulting in over 50,000 distinctchemical transformations for the model ofthe process chemistry. Computer pro-grams using sorting procedures automaticallyuse the network to construct a reactor

model's differential rate equations foreach component.

Modeling the Kinetics ofComplex Reaction SystemsConstructing complex reaction networksfor thousands of components and theirtens of thousands of reactions createsanother challenge. Each reaction requireskinetic parameters, including rate con-stants, activation energies and adsorptionconstants on catalytic sites, and chemicalthermodynamic properties. Obviously, it isnot feasible to conduct experimentalstudies of 50,000 isolated reactions. TheSOL concept of reaction classes and theconcept of structure-activity relationshipsprovide a basis for determining the largenumber of kinetic parameters.

The first approximation for estimatingreaction rate parameters is to assume thatall reactions in a class have the same kineticparameters because they undergo the sameintramolecular transformation. The numberof parameters required for the entirenetwork of reactions is reduced from 50,000


R.J. QUANN

Figure 7. Application of reaction rules to complex mixtures containing many components. (A) A4 ring saturation:reactant selection rule A4>0; product generation rule A4=A4-1, N4=N4+1. (B) N4 ring cracking: reactantselection rule N4>0; product generation rule N4=N4-1, R=R+4.

to 20 to 40 unique rate constants andactivation energies. This generally provesinsufficient-within each reaction class themolecular structure of the reactant or prod-uct can influence the rate of reaction.Therefore, structure-activity relationshipsare required for each reaction class.

The structure-activity relationship inkinetics is not a new concept. It beganwith Hammet (6), who organized familiesof reactions to study the influence of

substituents on the rates of homogeneousreactions. The basic premise of thestructure-activity relationship is that therate constants for a family of moleculesundergoing the same reaction can becorrelated to a measurable or calculablemolecular property. One form of thisrelationship is:

ln ki = a+ b (Rhi)

where ki is the rate constant for moleculei, a and b are constants to be determinedexperimentally, and RIi is defined as thereactivity index for molecule i. This equa-

tion can be derived from transition-statetheory and is also known as a linear free-energy relationship (LFER). The RIi is thecalculable molecular property that cor-

relates with the free energies of the transi-tion states of the reactants. The correctchoice of RI reflects the controlling mech-anism of the reaction family, such as car-

benium ion stability in cracking reactions,free radical stability in thermal reactions,or electronic properties of aromatic ringsin hydrogenation reactions (7). Structure-adsorption relationships have also beendeveloped for adsorption and poisoningon catalyst sites (8). In principle, the a

and b parameters and the appropriate RIcan be determined from a limited amount

of experimental work. The LFER can thenbe used to calculate the rate constant forany reaction in a large class of reactions.

The use of structure-activity relationshipspresumes knowledge of exact molecularstructures. As discussed previously indetail, this is not the case for most of thecomponents in petroleum or for the SOLrepresentation. The structure-activity rela-tionship must be developed for the isomersets that correspond to the individual SOLcomponents. This information can only beobtained from detailed analytical examina-tion of products from kinetic studies on

petroleum mixtures.

Calculating the Propertiesof MixturesRefinery products must meet certain qualityspecifications for combustion properties,fluid characteristics, or polluting poten-tial. They include boiling point distribu-tion, octane number, cetane number,pour point, aniline point, smoke point,cloud point, pour point, refractive index,viscosity, viscosity index, vapor pressure,

surface tension, and sulfur and nitrogencontent. In addition, thermodynamicproperties are required for process engi-neering calculations and for determiningthe extent of reversible chemical reactions.Each mixture property is determined bythe collective properties of the mixture'smolecular components. Unfortunately,individual molecular properties exist onlyfor a limited number of compounds andthere is little information on the collectiveproperties of the closely related isomersdefined as an SOL component. A mixtureproperty can be calculated if a simulated


A

49-~ 2H 0--

C - M2H ,-C9'-2H2_ -

&6_ em12 o

B

@~ H2 o H2 3 -

aizH2

H2_a_-

4 > H2-~

H' H2

_ _UH2a gH2

-

2H2_~

060"-. 09,Y.

1 446

oer --2H2_~ cor

cor -2H2_~ C&C

MODELING THE CHEMISTRY OF COMPLEX PETROLEUM MIXTURES

CH4+

+

.1+ ^~

+ -

+CH4

+ CH4

Paraffin isomerizationand cracking rules

C H4+ ALJ

+ A^

+ A

+ CH4

Figure 8. Developing complex reaction networks through application of reaction rules.

or measured composition is available andif a structure-property correlation existsto estimate the contributions of the indi-vidual molecules or isomer sets in themixture. There are different approaches to

obtaining structure-property correlations.One approach is to develop a new

correlation for a molecular property usingthe molecular class and homologous seriesconcepts as shown in Figure 9 for boilingpoint, specific gravity, and viscosity. Thecorrelation relies on the property's smoothvariation with carbon number for thehomologous series of each molecular class.The only requirement is the availability ofan extensive database on individual molecu-lar properties. Fisher (9) surveyed variousalgebraic forms to represent physical proper-ties of n-paraffins as a function of carbonnumber. Correlations for other molecularclasses and their homologous series can be

developed as a deviation from the n-paraffinseries. With the exception of the n-paraffinseries, the homologous series property corre-

lation in SOL must reflect the isomer group

nature of each SOL component.Another approach is to develop or

utilize existing group contribution tech-niques, particularly for thermodynamicproperties. Examples include Bensongroups (10) for calculating molecular heatcapacity, enthalpy and entropy, and uni-versal functional activity coefficient groups

(11) for liquid phase activity coefficients.Properties of molecules are computed as a

sum of the properties of the groups thatcompose the molecules. These techniquesgenerally have group definitions that differfrom SOL and must be mapped to theSOL increment representation.A third method for obtaining molecular

properties is to use existing empirical

relationships between molecular properties.For example, critical properties for equationsof state, including critical temperature,critical pressure, and the acentric factor,can be estimated from the boiling point,molecular weight, and specific gravity ofthe molecular components. Many of therefinery-product-quality properties can

also be obtained from industry-standardcorrelations that use these more common

physical properties.

SummaryModels of refinery processes must havethe capability of predicting the yields,composition, and properties of various prod-ucts. Lumping schemes with too simplistic a

representation of petroleum's compositionand refining chemistry have insufficientscope and reliability. A molecular-basedapproach to modeling is required to


I~~~~~~~I-~0 --

Hydrogenation and ring opening rules--

-iQ + CH4

_06- +-

>)6- +

+

+1 ,'

Alkyl group cracking rules

1447

R.J. QUANN

1400 Alkyl benzenes 1.4- Alkyl benzenes 1000- Alkyl benzenes14001_-O o1200 - 01.2 100

800 -1.0 10

600-0.8 _ _

200 0.6 ____________________0.11000 -

1400 Naphthalenes 14hthalenes Naphthalenes

1200 112 Dnithaenes 100|

600 -10-1400 -0.8200- U

10

0 0.6 0 0.11400 - cm 134 O 10041200 -Mononaphthenes ~., Mononaphthenes ~ 100 Mononaphthenes

gu 1000seres QE 800 - C 1.00 600 -10Z 400 -0.8

0 0.6 0.11400 -Dinaphthenes 1.4 -Dinaphthenes 1000 -Dinaphthenes

1200-12- C 01000 1.2M800 1.-1600

400 -0.8 --- - - - - -

200 06I10 .6-0.1-

10 20 30 40 50 4 10 20 30 40 50 4 10 20 3 40 50

Carbon no.

Figure 9. Correlating molecular properties using molecular class and homologous series concepts. Modified from Quann and Jaffe (3). (A) Boiling point, 760 mm, of homolo-gous series. (B) Specific gravity of homologous series. (C) Absolute viscosity at 100 and 2001F.

capture the necessary details of mixturecomposition, chemistry, and properties. Amolecular-based model of complex mix-tures can be developed if the followingcriteria are satisfied. First, analyticaltechniques must provide some level of

detail on the molecular composition andmolecular structure of the mixture. Sec-ond, a method must be devised to repre-sent and manipulate the molecularstructure and chemistry of a large numberof components. Third, sufficient data on

molecular properties must be available todevelop computational methods to estimatemolecular properties. Finally, structure-activity relationships are required toestimate the kinetic parameters for a largenumber of reactions.

REFERENCES AND NOTES

1. Jacob SM, Gross B, Voltz SE, Weekman VW Jr. A lumpingand reaction scheme for catalytic cracking. AICHE J22(4):701-713 (1976).

2. Wei J, Kuo JCW. A lumping analysis in monomolecularreaction systems. IEC Fundam 8:114 (1969).

3. Quann RJ, Jaffe SB. Building useful models of complex reac-tion systems in petroleum refining. Chem Eng Sci51:1615-1635 (1996).

4. Altgeltt KH, Boduszynski MM. Composition and Analysis ofHeavy Petroleum Fractions. New York:Marcel Dekker,1994.

5. Tissot BP, Welte DH. Petroleum Formation andOccurrence. New York:Springer-Verlag, 1994.

6. Hammet LP. The effect of structure upon the reactions oforganic compounds. J Am Chem Soc 59:96-103 (1937).

7. Neurock M, Klein MT. Linear free energy relationships inkinetic analysis. Polycyclic Aromat Comp 3:231-246 (1993).

8. Ho TC, Katritzky AR, Cato SJ. Effect of nitrogen com-pounds on cracking catalyst. IEC Res 31:1589-1597 (1992).

9. Fisher CH. Equations correlate n-alkane physical propertieswith chain length. Chem Eng September 20 (1982).

10. Benson SW. Thermochemical Kinetics. New York:JohnWiley & Sons, 1976.

11. Fredenslund A, Jones RL, Prausnitz JM. Group-contributionestimation of activity coefficients in non-ideal liquid mix-tures. AICHE J 21:1086-1098 (1975).

1448 Environmental Health Perspectives * Vol 106, Supplement 6 * December 1998

modelingthechemistryofcomplex petroleum mixtures · modelingthechemistryofcomplex petroleummixtures...

Documents