regularities ii{ the infrared absorption spectra of silicate minerals · ·...

REGULARITIES Ii{ THE INFRARED ABSORPTIONSPECTRA OF SILICATE MINERALS

Pnrr.rp J. LeuNon, Standard Oil Company (Indiana), Whi,l,i,ng, India.na

Assrnlcr

In order to augment the published infrared data on silicate minerals so that any rela-tionship between spectra and structures might be revealed, the infrared absorption spectraof eighteen silicates have been measured from 2 to 15 microns. These spectra were ob-tained from thin films of powder having particle diameters less than 2 microns. Study ofall available silicate spectra shows that the infrared spectrum can be used to characterizethe type o{ silicon-oxygen group. The groups that have been considered are the isolatedSiOr tetrahedron, SigOg ring, single chain, double chain, layer, framework, and silica.The region of strongest absorption tends to shift toward shorter wave length as the ratioof silicon to oxygen increases.

fNrnopucrrow

Although the absorption of infrared radiation by minerals was firststudied a half century ago (1,2),it was not unti l recent years that thecommercial availability of accurate, automatic-recording infraredspectrometers enabled mineralogists to take advantage of the possibilitiesafiorded by this exceedingly useful tool. The present work concerns thecontributions of infrared spectroscopy to the knowledge of the struc-tures of silicate minerals.

The atoms of a solid are constantly vibrating about their equilibriumpositions with frequencies of 1012 to 101a cycles per second. Becauseinfrared or heat rays have frequencies in this range, it is possible todetermine some of these vibration frequencies by either of two methods.fn the absorption method, infrared radiation is passed through a solid;if the frequency of the radiation coincides exactly with a vibration fre-quency, the solid absorbs some or all of the infrared radiation of thisfrequency. The reflection method is based on the principle that thereflecting power of a solid increases rapidly in the vicinity of a strongvibration frequencyl the reflection maxima show the approximatepositions of strong vibration frequencies.

The selective absorption of infrared radiation by silicate mineralshas been known since the early 1900's, when Coblentz (1, 2) obtainedinfrared absorption spectra of cleavage sheets or polished sections ofmany silicates. The slices he used were about 0.1 mm. thick; they wereopaque to infrared radiation having wave lengths longer than 9 microns.Coblentz also examined the reflecting power of polished surfaces ofminerals and found that the silicates had reflection maxima in theregion of 9 to 12 microns. He made the important observation that "thevarious silicates have quite different reflection spectra."

764

INFRARED ABSORPTION SPECTRA OF SILICATE MINERALS 765

In the 1930's Schaefer et al. (3) and Matossi and Kriiger (4) publishedthe infrared reflection spectra of some silicate minerals. Matossi andBronder (5) examined the infrared absorption, in the range 2 to 8microns, of plates cut from silicate crystals. They assigned the absorp-tion bands observed in this region to combination and overtone frequen-cies derived from the longer-wave length fundamental vibrations, whoseapproximate positions were known from reflection spectra. As in thepioneering work of Coblentz, the samples they examined were opaque towave lengths longer than 9 microns, and the positions of the strongfundamental-vibration frequencies in this region could not be obtainedby absorption measurements.

An outstanding advance in experimental technique resulted from thework of Pfund (6) and of Barnes et al. (7), who showed that the infraredabsorption spectrum of a solid could be obtained by using a thin filmof powder, provided that the powder particles had diameters smallerthan the wave length of the infrared radiation. When the particle sizewas larger than the wave length, the infrared rays were refracted andemerged from the layer in random directions giving an apparent opacity;when the particles were smaller than the wave length, they did not actas refracting centers. The greatest exponent of the powder techniquehas been I-ecomte (8), who used it to obtain the infrared spectra qf manyinorganic compounds containing polyatomic anions such as POa, SOr,and HCOa.

The powder technique has only recently been applied to the study ofthe silicates. Keller and Pickett obtained infrared spectra of some clays(9) and silica minerals (10). A report by Adler et al. (ll), under the aegisof the American Petroleum fnstitute, consists of a collection of infraredstudies of clay minerals made at several laboratories. In many of thepublished powder-film spectra the absorption bands are poorly definedbecause of large-particle scattering. Hunt et ol. (12), who obtainedspectra of several silicates, used only particles having diameters lessthan 5 microns, with the result that their silicate spectra show well-defined absorption bands.

In principle, it is possible to calculate theoretically the vibration fre-quencies for a solid; however, except for very simple structures, formi-dable mathematical difficulties are encountered. fn such instances it isconvenient to adopt an empirical approach that has been useful in thedetermination of the structures of organic compounds. This approachconsists of comparing the infrared spectra of a series of compounds, eachcontaining the same group of atoms. Often such comparisons show thatcertain groups of atoms have characteristic vibration frequencies givingrise to absorption bands which persist in different compounds.

PHILIP J. LAUNER

d d. i dH A I sj A i^a rl r ',. I"S C s ? B qH 2 e e. ?E a s d s ?d

..:

.! .e .f; -.e .i .e .e g_q lFEs '= . 'E i ;HEV, | + F, ?, h 'E E ; i Z: I F ; i > , i l E L a =d | = E i { vF t r ! rn U

il.

a

a

Fl

cj

766

Ha

Fz

h

6i

r'lva

F


; i i a4 v ' .= 5 a

4 A

F

zobo

(J

B

z

E

bo

Fl

=

I s 6 ^iiEiB dB Ed de 6 . H p g d d 4 ^ y s f , C t ) e*Z Z ; 5 a | n Y e S ? ! / > \ , e E M

cE*Boo .EAFEe !. ; A ! * ; t = H

i l €€€E 6' 3. ro : . d o . : i i

d

a

F.l

isNs.s

U

Fl

F

768 PHILIP T. LAUNER

The purposes of the present work are to present the infrared absorptionspectra of eighteen silicate minerals, to compare these spectra with thosewhich have appeared in the literature, and to point out certain similari-ties in the spectra of minerals having the same type of silicon-oxygengroup.

Exppnrlrowrer,

The minerals used in this study are described in Table 1. They wereselected because their absorption spectra were needed to complementpublished infrared data on silicates.

The infrared absorption spectra were obtained using thin films ofpowder having particle diameters less than 2 microns. To isolate particlesof this size, about 2 g. of the silicate, ground to pass a 300-mesh screen,was shaken with 100 ml. of C.P. isopropyl alcohol in a glass-stoppered,100-ml. graduated cylinder. The suspension was then allowed to stand16 hours or longer. At the end of this time an upper layer of the suspen-sion was drawn ofi to include only particles finer than 2 microns, thedepth of this layer being calculated from Stokes'law. A serviceable form(15) of Stokes'law for this calculation is:

c -D'sbFpz)(TX60)

18a\ lQa

where S:distance in cm. fallen by particle in I minutes,D:diameter of particle in microns,g:acceleration of gravity (980 cm./sec.2),

p1:density of the particle (S./ml.),pg:density of the liquid (g./ml.),

4:absolute viscosity of the liquid (poises).

Isopropyl alcohol rather than water was chosen as a sedimenting liquidbecause of the work of Webb (15), who showed that during the sedimen-tation of silica-alumina cracking catalysts flocculation occurs in waterbut not in isopropyl alcohol.

The isopropyl alcohol suspension withdrawn was evaporated to 15ml. and centrifuged. Most of the supernatant alcohol was drawn off, thepowder and remaining alcohol were stirred, and a few drops of the result-ing slurry were transferred with a 1-ml. hypodermic syringe to a weighedsodium chloride plate. When the alcohol evaporated, a thin film of pow-der was left on the plate, which was again weighed. The weight of thefilm divided by the area of the plate was recorded on the spectrum inmgf cm.z

The spectra were obtained on a Beckman 1R2 infrared spectrometerequipped with a sodium chloride prism. The wave-length scale was

tm

EO

H . ozF 4 0

20

0


calibrated against known absorption maxima of liquid toluene and ofatmospheric water vapor and carbon dioxide. The slits were automati-cally adjusted by means of a slit drive. The 10, or reference, curve wasobtained by recording the radiant power transmitted by a sodiumchloride plate having the same dimensions as the plate used to supportthe sample. The 1, or sample, curve was obtained by recording the radiantpower transmitted by the powder-coated plate.* The repeatabil ity of arecorded curve obtained using the -slit drive was about * O.3/6 of therecorder chart width. Wave lengths were accurate to *0.02 micron. Atypical recording is shown in Fig. 1. Percent-transmission curves wereobtained from these recordings by measuring the ratio I/Ioin a point-by-point manner. Measurements of percent transmission (1/1sX100)were then transcribed to a chart having a linear wave-length scale.

20 30 40 i0 90 r00 i lo t2Q

\ \AVE LENCTH IN MICRONS

l'rc. 1. Automatically recorded infrared curves of anthophyllite. 10 curve shows radiantpower transmitted by a sodium chloride plate.Values on lowercurves give sample thick-nesses in mg,/cm.2.

The spectra of the eighteen minerals studied are shown in Figs. 2-5,and the positions of the absorption bands are listed in Table 2. Eachspectrum has a transmission maximum at or near 8 microns and trans-mission minima, or absorption bands, in the range of 8 to 15 microns.Scattering by refraction is responsible for the percent-transmissiondecrease at the lower end of the wave-iength scale. Many of the spectrashow a weak band at 6.1 microns due to a small amount of entrainedwater. The 6.l-micron bands are stronger in the spectra of attapulgiteand sepiolite, two silicates which contain water of crystallization. Theearly researches of Coblentz (l) showed that both water of crysta.lliza-tion and liquid water absorb strongly at 6.1 microns.

x In order to obtain reproducible intensities, it is important to compare the.Io and the.I sodium chloride plates for equal infrared transmission before each sample preparation.

Plates with non-parallel faces cause a prisrnatic deviation of the infrared beam with a

subsequent decrease in the amount of energy reaching the thermocouple, The naturalcleavage planes of sodium chloride, lightly polished with jeweler's rouge, make goodparailel surfaees,

770 PHILIP J. LAUNER

6 . 9 t O r l

60

.lo

to

o8 ! a . 5 6 7 e 9 | o l l 1 2 t 3 1 4 1 6

WAVE LEI{GTH I1{ NICRoi{S

Frc.2. Infrared absorption spectra of forsterite, B-CazSiOr, willemite, phenacite,

ano zlrcon.

2

2IOO r

2

2

zoao=ozEF

F2UJoGlrJc

i-Q.l

,.61

\ \ A f\, /ssp-Co2SiOa

).3

I \

,A \ vPHENACITE

too

ao

00

.lo

20

o

too

00

60

Q

20

o

6 ro0o9 s o=o? .oGF

40F28 z oElrJo - 0

roo

80

60

Q

20

o

ro0

to

€o

.lO

?o

o

INFRARED ABSORPTION SPECTRA OF SILICATE MINERALS 77I

o 2

\ o.9

\ \

L KYANTTE t

t5 t4

tI

\\

IBEN ITOITE

t3 t4

ll trl

| l t a

8 1 4 5 6 7 E e r o i l n | l nwAvE LEI{GTH til rtCRof{S

Frc.3. Infrared absorption spectra of kyanite, benitoite, rhodonite, and aegirite.

t l

o.5

/

\ I IV

RHODONITE

1u.D

/

/1 l l7 R l

(AFT€R HYORO(

t t lr t l. IODONITEHLORIC ACIO TREATMENT

\ ^AIJ

a\ o !

ft

// AEGIRITE


t3 t4

| l l a

.fo

co

o

WAVE LENGTH IN MICRO{S

Frc. 4. Infrared absorption spectra of enstatite, clinoenstatite, jadeite, and spodumene.

40

zo

o

100

80

60

tlo

20

o

z t @o9 e o=o?eoE, Q

F2

U e oEtrJA O

too

80

60

4

m

o

too

80

ENSTATITE ( PA )

i zO.2

\ 6 \

v Ensrairre inc)

o.2

\ / \ nd, \

\ v V" Vt

/!,cr-ngqnsrarye

\\0 .1

Iu.c

/

JADEITE

2 l 4 5 a t 8 9 r o i l t e t ! ! l l 5

zIo.2=ozEF

F2IJocUJG

2

2

2


\'o.2

,6, \

I

\ ORTHOCLASE

2 ! 4 6 6 7 8 9 l O l l l 8 1 3 1 4 1 5

CO

.l(,

go

o

roo

to

@

.lo

go

IYAVE LENGTH If{ TIOROIIS

Frc. 5. Infrared absorption spectra of anthophyllite, hornblende, attapulgite, sepiolite,and orthoclase.

0.3

I

HORNELENDE

T.lnr,r 2, PosrrroNs ol fNrnannn AssonprroN BeNos

Mineral Wave Length in Microns

Forsteri te 10.00 l l . l710 .45 11 .89

d-cazSior 9 .7 10 .04 11 .1r 0 . 7 1 1 . 5

l t .78

Wil lemite 10.25 11.10r0.74 tr.52

Phenac i te 9 .3 10 .25 11 .10 12 .3 13 .45 14 .401 0 . 6 6 1 2 . 9 0 1 3 . 6 210 .95 13 .94

Zircon 9.7 11.05

Kyan i re 9 .5 10 .3 11 .10 13 .62 14 .659 .92 10 . 58 11 .35

Benitoite 9.66 10.75. 13.13

Rhodon i te 8 .95 9 .18 10 .28 l t . l7 13 .889.47 10 .519 . 7 5 1 0 . 8 5

Enstati te 8.86 9.35 10.28 11.09 13.18 14.449.90 10 .62 11 .55 13 .42

13. 75

cl inoenstat i te 8.35 9.07 10.08 11.12 12.50 13.57 14.638.68 9 .29 t0 .62 11 .65 13 .80

9 .88

Aegir i te 9.00 10.35 11.62 13.629 . 5 59 . 8 4

Jade i te 8 .85 9 .35 10 .70 11 .65 13 .409 . 9 7

Spodumene 8 .65 9 .16 10 .80 11 .609 . 7 0

Anthophyl l i te 8.86 9.02 10.21 11.11 12.80 13.00 14.039. 11 t0 .97 13 .25 14.909.46 13.649.79b

Hornblende 8.84 9.01 10.14 13.25 14.529.43 10 .49

10.87

Attapulgite 8.39 9.68 10.128.87 10 .95

Sepiol i te 8.38 9.64 10.138.92 10 .93

Orthoclase 8 . 85 9 .66 12 .90 13 .7 5

" Possibly has a shoulder at 10.87 microns.b Band center uncertain, because talc, a possible contaminant, has its strongest band

at 9.8 microns.

INFRARED ABSORPTION SPECTRA OF SILICATE MINERALS 7 7 5

The spectrum of rhodonite, in Fig. 3, shows absorption bands due toa carbonate at 5.58, 7.0, and 11.51 microns. The positions of these bandsstrongly suggest that the carbonate is rhodochrosite, MnCOa, reportedto have absorption maxima at 5.58, 7.0, and 11.50 microns (12). As aconfirmatory test for carbonate, an infrared spectrum was obtained ofrhodonite which had been treated with hydrochloric acid. In the spec-trum of acid-treated rhodonite the carbonate bands are weaker and thesilicate bands are stronger.

Both the infrared spectrum, shown in Fig. 5, and an rc-ray powderphotograph reveal the presence of a-quartz and calcite in the sepiolitesample used for this study. Seven infrared bands can be accounted forby a-qu.artz (8.6,9.2, 12.53, 12.83 microns) and calcite (6.95, 11.40,14.03 microns).

An absorption spectrum of hornblende published by Hunt et, al. (I2)has bands at 9.8 and 14.95 microns and is unlike the spectrum shown inFig. 5 for the hornblende sample used in the present study and authenti-cated by r-ray difiraction.

Drscussrox ol RESULTS

The silicon-oxygen groups to be considered in this report are arrangedaccording to the o-ray-determined structural classification of W. L.Bragg:

I. Isolated SiOr tetrahedronII. Si:Oe ring

III. Single chainIV. Double chainV. Layer

uli: Ill;"*"'oIn order to ascertain whether infrared absorption can be used to char-

acterize the various types of silicon-oxygen groups which occur inminerals, the experimental results are summarized and correlated withpublished infrared data by means of the bar charts shown in Figs. 6-12.In these charts, strong infrared bands are depicted by tall bars andweak bands by short bars. Some published spectra were not includedbecause the identity of the silicate was in doubt or because the bandswere poorly defined. If the absorption spectrum of a silicate was notavailable but its reflection spectrum was known, the reflection spectrum(labelled R) was used in the bar charts. The reflection maxima serve toindicate the approximate wave lengths at which absorption bands areexpected to occur. Since the reflection method gives only the strongvibration frequencies, the reflection spectrum will usually show lessdetail than the absorption spectrum.

I

I

I ,lI

I

I

I

I

tlI tl I T

RI

I T l l I I

R

oLtvtNE ilzt

FORSTERITE

p -CosSioa

ALMANDITE (i l,ret

PYROPE ( I I I

SPESSARTITE (II)

GROSSULARITE (II '

uvARovlTE { i l ,

ANDRADITE ( I t I

WILLEMITE

PHENACITE

ZlRooftl

TOPAZ (2,3)

KYANITE

olroNoRoDrTE (4,

8 9 r O | | t 2 1 3 t 4 1 5WAVE LENGTH IN MICRONS

Frc. 6. Infrared absorption maxima of silicates containing the isolated SiOr tetra-hedron.

"R" indicates the reflection spectrum. Spectra from the literature are identified byreference numbers.

ill ,llltl II I

R ,l - M I S SI N G -


8 9 l O l r t 2 1 3 1 4 1 5YVAVE LENGTH IN MICRONS

Frc. 7. Infrared absorption maxima of benitoite and other silicatesthat may contain the SirOg ring.

BENITOITE

RHOOON ITE

WOLLASTONITE (I2}

PECTOLIT€ ( I I

ENSTATITE

CLINOEITISTATITE

AUGITE ( I2 }

AEGIRITE

JADEITE

SPOOUMENE

ANTHOPHYLLITE

TREMOLITE (12}

AOTINOLITE (12}

HORNBLENOE

ATTAPULGITE

SEPIOLITE

8 9 r O i l t 2 1 3 1 4 1 5WAVE LENGTH IN MICRONS

Frc. 8. Infrared absorption maxima of single-chain silicates.

I 9 r O i l l e 1 3 1 4 1 5WAVE LENGTH IN MICRONS

Fro. 9. Infrared absorption maxima of doubie-chain silicates.

t, ,l l l I I

l l t, ,l I l l I

l, t,t l I I

T I I

I tl I

I ,rl I I t l

,l t, I I

I II, I I

I I [ , I I

l l

I


I s olated. S i.O + T elr ahedr on

Infrared data on orthosilicates are collected in Fig. 6. All of theseorthosilicates have two, three, or four strong bands in the range of 10 to12 microns. Since these compounds embrace a variety of positive ionsand crystal structures, it is reasonable to assign these bands to vibrationsof the independent SiOa group, which all orthosilicates have in common.

T l l I I

I I I I

I T T

I

I I T

t l tl I I I

I

,l t l t I

I I I

9 r o l t t 2 1 3 1 4 1 5

WAVE LENGTH IN MICRONS

Frc. 10. Infrared absorption maxima of layer silicates.

The absorption spectra of olivine and forsterite are very nearly identi-cal. Synthetic B-Ca2SiOr has a spectrum similar to the spectra of olivineand forsterite. A weak band at 11.9 microns in the olivine and forsteritespectra corresponds to a sharp, much stronger band at 11.78 microns inthe 0-CazSiO4 spectrum.

The spectra of the garnet minerals (almandite to andradite) were ob-tained by Adler (11), who found that the absorption bands of thegarnets shift to longer wave lengths as the lattice constant increases.

Phenacite, Be2SiO4, and willemite, Zr'zSiO4, the two major silicates

K A O L I N I T E ( i l , 1 2 1

0 t c K t T E ( i l 1

HALLOYSITE (I I }

MoNTMoR|LLoNtTE ( i l ,12'

N O N T R O N I T E ( I I )

H E C T O R I T E ( I I )

P Y R O P H Y L L I T E { I I )

TALC ( i l , r2)

S E R P E N T I N E I I I I

M U S C O V I T E i l r , 1 2 )

B I O T I T E ( l l , l ? )

INFRARED ABSORPTION SPECTRA OF SILICATE MINER-ALS

of the phenacite group, have strikingly similar absorption bands at10.25,10.7, and 11.1 microns, despite a sevenfold difference in the atomicweights of the cations. Phenacite has strong absorption bands between

ORTHOCLAS€

MtcRocLtNE ( t2 )

ALBITE t IE)

LEUCITE I ' )

I {EPHELITE I I2 I

CANCRINITE 14)

sooALtrE (4)

scaP0LtTE (4)

NATROLTTE il,4)

9 l o | r t 2 1 3 1 4 1 5WAVE LENGTH IN T' ICRONS

Fro. 11, Infrared absorption maxima of framework silicates

4-QUARTZ t to , t l , te l

4 -CRISTOBALITE ( ro , l r )

0PAL ( to i i l ,12)

FUSEo SIL ICA ( to , l l )

I 9 l O i l 1 2 1 3 1 4 1 5WAVE LENGTH IN MICRONS

Frc. 12.Infrared absorption maxima of silica minerals.

12 and 15 microns-a region where most of the other orthosilicates donot absorb. It is reasonable to attribute these bands, and possibly the9.3-micron band, to Be-O vibrations. By the same reasoning, thekyanite 13.6- and 14.6-micron bands are thought to be due to Al-Ovibrations.

I

I I I

tl I

R

t l I

RI I

RI

RT

R

T I

I T I I

I t

I I

I T


The spectrum of zircon shows less structure than the spectra of theother orthosilicates. There is some arbitrariness in selecting 9.7 micronsas the center of the shorter wave-length band.

The reflection maxima of topaz, AlzSiO+Fz, and chondrodite,2 MgzSiOr.Mg(F,OH)2, fall in the region of the spectrum where theother orthosilicates have their characteristic absorption maxima.

SizOs Ring

Spectra of benitoite, rhodonite, wollastonite, and pectolite are com-pared in Fig. 7. Benitoite, in which the silicate ion is known to consistof three tetrahedra linked corner to corner in a ring to form SirOs (16),has strong bands at 9.66, 10.75, and 13.13 microns. The band at 13.13microns occurs in a region where most silicates do not have strong absorp-tion bands.

Structure determinations of rhodonite, MnSiOe, and wollastonite,CaSiOs, are not available; however, these silicates are thought to con-tain SisOg rings. Benitoite's single bands at 9.66 and 10.75 microns ap-pear as triplets in rhodonite and wollastonite. If rhodonite and wollaston-ite have Si3Oe rings, it is surprising that they Iack the strong band inthe 13-micron region shown by benitoite. Although these data neitherconfirm nor disprove the presence of SigOg rings in rhodonite and wollas-tonite, the marked similarity of the spectra of these two minerals leavesno doubt that rhodonite and wollastonite contain the same type of silicateion.

Pectolite, NaHCaz(SiOa)r, which is reported (17) to be isomorphouswith wollastonite and possibly to have an analogous structure, hasstrong reflection maxima at 9.4 and 10.8 microns. These maxima cor-respond to the centers of the wollastonite trifurcated absorption regions.To confirm the structural similarity of pectolite with wollastonite, anabsorption spectrum of pectolite is needed.

Single Choin

Infrared spectra of six pyroxenes are compared in Fig. 8. With theexception of aegirite, all of these single-chain silicates have three strongbands between 9 and 12 microns.

The spectra of the MgSiO3 polymorphs, enstatite and clinoenstatite,have the same over-all appearance but show small difierences in thewave lengths of two of the strong absorption bands and numerous differ-ences in the weak absorption bands.

Augite, CaMg(SiO)r, has three strong absorption bands Iike those ofenstatite and clinoenstatite, but its spectrum shows fewer weak bands.

The resemblance of the 11.6-micron bands of aegirite, jadeite, and

INFRARED ABSORPT:ION SPECTRA OF SILICATE MINERALS 78I

spodurnene suggests that in all three minerals these bands are due to theiame kind of vibration. Jadeite, NaAl(SiOa)2, and spodumene,LiAI(SiOrr, have very similar absorption spectra. In this pair, thesubstitution of Na+ for Li+ causes a bathochromic shift of four absorptionbands.

Double Chain

The double-chain silicates, whose spectra are depicted in Fig. 9,

have two strong absorption bands near 10 microns, two or more weak

bands between 8.4 and 9.5 microns, and another weak band near 10.9microns. The four amphiboles have a band near 13 microns.

Attapulgite, a clay with a double-chain structure (18), has a spectrumwhich Iacks the 13-micron band of the amphiboles. The examination of

a I.4 mg./cm.2 layer (Fig. 5) confirmed the absence of a 13-micron bandbut did reveal bands at 12.53 and 12.83 microns characteristic of a-quartz, an impurity which was not detected from an r-ray powderphotograph of this sample.

The five absorption bands of sepiolite coincide in positions (within a

few hundredths of a micron) and relative intensities with the bands

of attapulgite. The infrared spectrum of sepiolite supports the double-

chain silicate structure proposed for sepiolite by Longchambon (19).

Layer

Absorption data on minerals containing silicon-oxygen sheets are

collected in Fig. 10. Despite the complexity of these minerals, they all

have one or two strong absorption bands between 9.4 and 10.4 microns'It is reasonable to assign this strong absorption to a vibration character-istic of a silicon-oxygen layer, the only grouping of atoms which all of

these minerals have in common.Another interesting regularity is that the substitution of magnesium

for aluminum reduces the number of absorption bands and causes the

strong silicon-oxygen band to undergo a bathochromic shift. These

effects are illustrated by going from pyrophyllite, AI2Si4O10(OH)2, to

ta lc , MgrSiaOro(OH)z; and f rom muscovi te, KAl2AlSi3Oro(OH)2, to

biotite, KMg3AlSiaOro(OH)r. In the series (20) of related clays: mont-

morillonite, nontronite, and hectorite, these effects are apparent in

going from the spectra of montmorillonite and nontronite to the spec-

trum of hectorite, the aluminum-free magnesium end-member of the

series.Serpentine, originally described as a double-chain structure, has

been classed by Warren (21) with the Iayer structures. The dissimilarity

of the spectrum of serpentine and the spectra of the double-chain silicates

supports this change.


Frameworh

A comparison of infrared data on framework structures, given in Fig.11, shows that the sil icon-oxygen framework, l ike the layer, tends toabsorb strongly near 10 microns. All of the absorption spectra and someof the reflection spectra show weak bands in the region of 13 to 15microns.

The difierence between the infrared absorption spectra of two crystalforms of KAISi3O8, orthoclase and microcline, shows again that theinfrared spectrum of a mineral depends not only upon chemical consti-tution but also upon the way in which the atoms are arranged in thecrystal structure.

Nephelite, once classed with the orthosilicates, now is known to havea tridymite-like framework (17). The similarity of the nephelite spec-trum to the spectra in Fig. 11, along with the absence in nephelite of asecond strong band near 11 microns (cf. Fig. 6), is consistent with aframework structure.

Silica

fnfrared absorption maximafused sil ica are shown in Fig. 12.

of a-quartz, a-cristobalite, opal, andThese silica minerals have in common a

S]NGLE CHAIN

DOUBLE CHAIN

FRAMEWORK

WAVE LENGTH IN MICRONSFrc. 13. WaveJength ranges of strong infrared absorption bands

of silicon-oxygen groups.


strong 9.2-micron absorption band and weaker bands between 8.2and 8.6 microns and between 12.5 and 12.8 microns.

CoNcrusrow

A chart showing the wavelength ranges of the strong infrared absorp-tion bands of silicon-oxygen groups is given in Fig. 13, where the regionin which strong bands occur is shown by a horizontal line to the left ofeach group. The silicon-oxygen groups here are arranged according toan increasing ratio of sil icon to oxygen. As the Si:O ratio progressesfrom 0.25 in the isolated SiOa tetrahedron to 0.50 in sil ica, the region ofstrong absorption shifts to shorter wave lengths, and the waveJengthranges tend to narrow.

One of the best features of the infrared spectrum is that absorptionor lack of absorption at certain wave lengths often can be correlatedwith the type of silicon-oxygen group in the mineral. Small displace-ments of the vibration frequencies characteristic of a silicon-oxygengroup must arise mainly from the fact that these vibrations are not com-pletely independent of the rest of the crystal. It is to be hoped that suchfrequency shifts ultimately can be interpreted to give information aboutthe structure adjacent to the silicon-oxygen group.

AcrNowlBnGMENTS

The author is indebted to L. J. Schmauch of this Laboratory for thedesign and construction of the automatic slit drive for the infraredspectrometer; and to W. A. Kimball, also of this Laboratory, for carry-ing out the *-ray diffraction studies. He wishes to thank the followingpersons for generously supplying specimens: Robert K. Wyant, ChicagoNatural History Museum; Wilfrid R. Foster, Champion Spark PlugCompany; K. K. Kelley, Pacific Experiment Station, Bureau of Mines;George Switzer and C. Lewis Gazin, United States National Museum.

RnlntntcBs

1. Corr-rNrz, Wrr,r.r.q,u W., Investigations of Infra-Red Spectra, Parts III and IV,Washington, D. C.: Carnegi.e Institution oJ Wa.sltington (1906).

2. CormNtz, Wrr-r,reu W., Supplementary Investigations of Infra-Red Spectra, PartsV, VI, and VII, Washington, D. C.: Carnegie Institittion oJ Washington (19O8).

3. Scnenrnn, C., Merossr, F., .tNo Wrnrz, K., Das uitrarote Reflexionsspektrum vonSilikaten: Zei,t. f. Physih.,89, 21W233 (1934).

4, Mlrossr, Fnawn, ,rNo Krtictn, HeNs, Das ultrarote Reflexionsspektrum von Sili-katen: Ze'it. J. Physi,k, 99, l-23 (1936).

5. Mlrossr, F., .tm BnoNonn, O , Das ultrarote Absorptionsspektrum einiger Silikate:Zei t . f . Physik, r l r ,1-17 (1938).

6. PruNn, A. H., The optical properties of metallic and crystalline powders: J. OfliealSoc. Am.,23, 375-378 (1933).

7U PHILIP J, LAUNER

7. Ber.nes, R. Bowrrxo, Bnerterx, R. Ronrnr, nlto FrnnstoNn, R. S., Powder filters

for the infrared: Phys. Rer.,47,792 (1935).

8. Lncourn, Jnem, Applications analytiques des spectres infrarouges de poudres aux

compos6s min6raux et d I'eau de cristallisation: Analytica Chimica Acta, 2, 727-733(1e48).

9. Knr.r.nn, W. D., .tuo Prcrntt, E. E., The absorption of infrared rariiation by clay

minerals: Am. f our. Sci., 248, 26+-27 3 (1950).

10. Knr,r,rn, W. D., nuo Prcxlrr, E. E. , Absorption of infrared radiation by powdered

silica minerals: Am. Mineral.,34, 855-868 (1949).

11. Aor,an, H,tNs H, BRAy, Err,rs E., Srrvrns, Nnr,sox P., Huwt, JonN M., Knlmn,

W. D., Prcrnrt, E. E., eNo Knnn, Plur, F., Infrared spectra of reference clay minerals:

American Petroleum Inslitute Project 49, Prelimi'nary Report No. 8, Columbia Uni-

versity, New York (1o50).

12. HuNr, JonxM., Wrsnano, Manv P., em Botlu,mr, LlwnnNcn C., Infrared absorption

spectra of minerals and other inorganic compounds: Anal,. Che.m., 22t 1478-1497( 1esO).

13. Krnc, E. G., Heats of formation of crystalline calcium orthosilicate, tricalcium silicate

and z inc or thosi l icate: J. Am. Chem.Soc. ,73,65G658 (1951).

14. Fosrnn, Wrrlnro R., High-temperature f-ray difiraction study of the polymorphism

of MgSiOr: f . Am. Ceram. 50c.,34,255-259 (1951).

15. Wr'ss,Gr,rNNM.,Particlesizedistributionof crackingcatalysts:'PetroleumProcessing,2,497-502 (t947).

16. Btlcc, W. H., mll Bnlcc, W. L., The Crystalline State, Vol. I, pp. 131-136, London,

G. Bell and Sons Ltd. (1949).

17. Gnumn, Jorrm W., Progress in silicate structures: Am. M'ineral.,33r 679-691 (1948).

18. Bnenr,v, W. F., The structural scheme of attapulgite: Am. Minerol'., 25, 405-410(1940).

19. Lonccrr,q.unoN, HrNRr, La s6piolite d'Ampaidrandava, Madagascar: Compt. rend,.,

200, 949-951 (1935).

20. Knnn, Paur- F., .q.No H.lurr.toN, P. K., Glossary of Clay Mineral Names: American

Pel,roleum InsLitute Projecl 49, Pretriminory Report No. It Columbia University, New

York (1949).

21. WannnN, B. E., eNo HnnrNc, K. W., The random structure of chrysotile asbestosl

Phys. Ret:., 59, 925 (1941).

Manuscribt recei,t:ed. Jan. 11, 1952.

regularities ii{ the infrared absorption spectra of silicate minerals · ·...

Documents