Identification of Polymorphic Forms of Crystals by Infrared Spectroscopy

Identification of Polymorphic Forms of Crystals by Infrared Spectroscopy DAVID

N. KENDALL

Calco Chemical Division, American Cyanamid Co., Bound Brook, .V. J. niques for study of polymorphism are discussed. The value of the infrared method is illustrated by results obtained on such pigments as the Para Reds and copper phthalocyanine blues. Results on the titanium dioxides illustrate the limitation of this infrared technique when applied to certain inorganic crystals. From application of the method knowledge can be obtained of how the molecular or molecularionic species are arranged and oriented in the crystal lattices of polymorphic substances by empirical interpretation of infrared absorption spectra.

While studying the relation between the spectra and structure of pigments, it was found that infrared techniques can be conveniently and successfully applied to distinguish between different crystalline forms of substances of the same chemical composition. Infrared spectroscopy has been applied to the identification and determination of polymorphic forms of both organic and inorganic crystals. The method complements and supplements the x-ray, microscopical, and electron diffraction techniques. The advantages and disadvantages of these tech-

T

HE use of x-ray (SO), light microscopical(16), electron micro-

and titanium dioxides. Interpretation of these spectra has been made by means of empirical correlations based on both published [ I ) and unpublished observations.

scopical, and electron diffraction ( 7 )techniques in the identification of polymorphic forms of crystals is known. It has now been discovered that infrared techniques can also be applied conveniently and successfully to distinguish between different crystalline forms of substances of the same chemical composition. This new infrared method can in certain cases detect smaller quantities of polymorphs than previously used techniques. It is rapid, employs small samples, and is not affected by the difficulties of crystal morphology characteristic of x-ray and microscopical methods. Permanent records in the form of infrared spectrograms are obtained which can be quickly referred to a t any time in the future when the occasion demands. The infrared spectroscopical technique described herein was developed primarily for analyzing polymorphic pigments and dyes, but the same procedures can be applied to other organic and inorganic crystals. The method involves obtaining the near infrared spectra of the pure polymorphs and then using these spectra to identify qualitatively and determine quantitatively the polymorphic forms that are present in any given sample of crystals. Earlier workers have used infrared experimentally to show differences in the polymorphic forms of calcium carbonate (IO, 16, ZI), diamond (22, 16,27, 29, S I ) , and quartz (28). On the theoretical side papers have been published on the selection rules for the Raman and infrared spectra of molecular (6) and ionic (3) crystals, on the general theory (2,9,11) of the vibrational spectra of crystals, and on the polarized infrared spectra ( l a ) of inorganic salts. To date, infrared spectroscopy has been used only to show differences between polymorphic forms of inorganic crystals. The present work extends the infrared techniques to the identification and determination of polymorphic forms of both organic and inorganic crystals. The sharpest spectra of ground particles of a crystalline material will be obtained, as first pointed out by Pfund (18,19),when the crystals under study are of a size less than the wave length of the infrared radiation, because of scattering. The spectra reported on herein were obtained on particles of diameters less than 5 microns, with the exception of the brookite spectrum. The absorption curves illustrated in this paper are shown for the region from 5000 to 650 em.-' Spectral frequencies of the bands unique to each polymorph of a given crystal are shown under Results and Discussion. Absorption spectra are reported for polymorphs of the para toners, copper phthalocyanine blues,

EXPERIMENTAL

All spectra reported were obtained with a Perkin-Elmer Model 12B infrared spectrometer. The instrument was calibrated using known .absorption bands of water vapor, carbon dioxide,,'and ammonia.

Figure 1. Proposed-Structural Formula of Para Red

.

I

FREQ. IN Figure 2.

CM."

Polymorphism of Para Red

A . Blueshade B . Yellow shade Samples run as Nujol mulls

382

873

V O L U M E 25, NO. 3, M A R C H 1 9 5 3 The crystalline powders were mulled in Nujol or perfluorokerosene on a hard opal glass plate with a glass muller. Several mullings were made in the preparation of each sample, the number depending on the ease of dispersion. Csually two grindings were sufficient. With t,he exception of brookite, the particle diameters of all samples examined were less than 5 microns before the mulling process began. The niulled sample was placed on a flat 073

383 rocal centimeters. The energy of the source is, of course, superimposed upon the radiation transmitted by the sample in each spectrum. While complete spectra from 5000 to 650 cm.-1 were obtained for each substance, as will be observed in some plots, only segments of such spectra are shown for reasons of space conservation. Breaks in the spectral curves in certain figures indicate that less significant regions of the spectrum have been omitted. On the “intensity” ordinate, spectrograms have been arbitrarily displaced vertically in order to present comparisons of polymorph spectra over the same spectral intervals. The dashed and full frequency verticals were for the purpose of oral presentation of this paper. In general, the full lines represent the absorption frequencies where the differences between two given polymorphic forms are most readily observed. Most samples were run in the solid state dispersed in Nujol or perfluorokerosene. Several solution spectra are presented, however, for reasons given later.

,1350

031

I

FREQ. IN CM.-’ Figure 3.


A . Blueshade B . Yellow shade

Samples run as Nujol mulls

rock salt plate and spread out into a uniform film by applying manual pressure to another flat plate atop the sample. The plates and sample were then held in position in a simple cell comprising flat metal fore- and backplates held together by three threaded pins, two a t one extremity, one a t the other, Threaded knurled knobs allowed the maintenance of the proper amount of pressure. Visual observation of the sample through the plates quickly indicated whether a satisfactory dispersed film had been obtained.

Figure 5 .


A . Blueshade B . Yellow shade

Samples run ae Nujol mulls

I


’


A . Blueshade B. Yellow ahade R . Radiation


Infrared absorption spectra were then run and the resulting spectra measured and interpreted. Over each of eight segments of the spectrum covering the range from 5000 to 650 cm.-l a constant slit width was employed, and these slit widths were the same for every spectrum. Philpotts et al. (go) have emphasized recently the importance of constant slit width to infrared spectroscopical investigations. RESULTS AND DISCUSSION

Figures 1 to 19 show the infrared absorption spectra of polymorphic forms of Para Red, copper phthalocyanine blue, and titanium dioxide as well as the structural formulas of these substances. The graphs show intensity versus frequency in recip-



A . Blueshade B. Yellow ahade Samples run as Nujol mulls

Para Reds. Figure 1 shows the structural formula of Para Red proposed on the basis of its chemical constitution and properties, but most particularly on the basis of the present infrared study, Para Red pigment is prepared (53) from diazotized p-nitroaniline, which yields when treated with sodium hydroxide a stable isodiazotate, 02NCsH,N20Na. Upon acidification with hydrochloric acid and coupling with 2-naphthol, Para Red results.

ANALYTICAL CHEMISTRY

384 Control of the appropriate variables in the process gives rise to the blue shade or the yellow shade of the pigment. The coupling in either case takes place in the I-position. Figures 2 through 6 show the infrared absorption spectra of the blue and yellow shades of Para Red as Nujol mulls. The spectral regions of greatest interest in the5000 to 650 em.-' range are presented. Inspection of the figures mentioned reveals that the blue shade and the yellow shade of Para Red have different solid state absorption spectra. The blue shade crystals show bands a t 1179, 1024, 1015, and 852 em.-' which the yellow shade crystals lack. The latter absorb a t 1620, 1572, 1000, 946, 844, 754, and 741 em.-' while the blue shade crystals do not. Figures 7 and 8 present several segments from the spectra of the blue and yellow shades of Para Red in 0.5% by weight chloroform solution together with the corresponding spectral regions of chloroform itself for comparison. It is evident that a sufficient concentration of each shade of Para Red dissolves in chloroform t o make a valid comparison of their spectra. The presence of the 845-cm.-' band-e.g., in the spectra of the blue and yellow shades (curves B and C, Figure 8)-and its absence in the chloroform spectrum (curve A ) prove that the Para Reds have dissolved in chloroform in sufficient concentration to reveal their presence in their absorption spectra. Several other bands characteristic of the Para Reds will be observed in the 1075 to 1250 em.-' region of curves B and C in Figure 8. The absence of these bands in curve A , the chloroform spectrum, will be noted. The complete rock salt spectrum of the blue shade in chloroform solution is identical to that of the yellow shade in the same solvent. This experimental evidence proves the two shades are identical in chemical composition. The fact that the infrared absorption spectrum of the blue shade is different from that of the yellow shade in the solid state shows the two shades to be polymorphic forms.

by Nyswander (16) in his work on calcite and aragonite. The infrared technique also is applicable to those crystals in which ionic, covalent, van der Waals, and other types of chemical binding are all involved in the spatial arrangements of the ions and molecules in the unit cell and the crystal lattice. Figure 4 shows that no free or bonded hydroxyl absorptions are found in the spectrum of either the blue or yellow shade Para Red crystals. Figure 7 gives the absorption spectra of the blue shade (curve B ) and yellow shade (curve C) of Para Red in chloroand -OH region toform solution in the high frequency -NHgether with the spectrum of chloroform (curve A ) in the same region. Here also, no free or bonded hydroxyl bands are found for either shade. It is, therefore, likely that extremely strong hydrogen bonding exists between the hydroxyl hydrogen and the diazo nitrogen nearest the nitro-substituted benzene ring in both shades of Para Red in the crystalline state as well as in solution in chloroform. This spectroscopical finding is compatible with the known virtual alkali insolubility of the Para Reds. That the N . . .H-0 bonding is to the diazo nitrogen nearest the nitrosubstituted benzene ring is very certain on the grounds of ring strain, bond distances, and steric considerations (17).

\

874

Figure 8.


A . Chloroform B . Blueshade C. Yellow shade B and C i n chloroform solution

FREQ. IN Figure 7.

CM.-'


A . Chloroform B. Blueshade C. Yellow shade B and C in chloroform solution

Polymorphism is particularly common in the case of organic dyes and pigments, as first shown by extensive x-ray power diffraction studies begun about 1933 in Germany (SO). Polymorphism of organic pigments has also been observed by electron microscopy and electron diffraction (7). The experimental evidence cited above for the two shades of Para Red shows that substances of the same chemical composition which have their molecules or molecular species oriented differently spatially to form crystals of different space lattices can be readily identified and distinguished from each other by means of their different infrared spectra. This finding applies also to crystals made up structurally of ions, as was first reported

The possibility of a very weak band a t 3030 em.-' was noted in a perfluorokerosene mull spectrum of Para Red blue shade. The presence or absence of such a band was ambiguous, however, and the spectra of the Para Reds should be reinvestigated in the -"-and -OH region with the higher dispersion afforded by lithium fluoride. This 3030-em. -1 absorption, however, if it does exist, would appear to be of too low frequency for a bonded It is probably an aromatic C-H stretching frequency. -OH. For the Para Reds in the solid state or in solution, a bonded hydroxyl absorption is entirely missing, or owing to the great strength of the hydrogen bonding it is shifted into and obscured by C-H absorptions. The absence of both free and bonded -OH absorptions in a substance containing an -OH grouping is unusual but not unknown. Its absence was observed by the author in these laboratories in 1946 in the Nujol mull spectrum of l-hydroxyanthraquinone (6, 14). Here also the reason for the absence of any hydroxyl absorption was found to be the presence of extremely strong hydrogen bonding, which in this case, was an 0 .H-0 bond. The possibility of a ketonic grouping with the hydroxyl hydrogen shifting to the nitrogen in the molecular structure of Para. Red both as i t exists in solution or in either of its dimorphic forms is ruled out by the absence of any ketonic carbonyl or -NHabsorptions in any of the Para Red spectra. The ketonic

V O L U M E 2 5 , NO. 3, M A R C H 1 9 5 3 carbonyl absence is evident from examination of Figure 9, which shows the C=O stretching region for the blue shade ( A ) and the yellow shade ( B ) . For ready comparison, the radiation curve for the same region is shown as curve R. The 1620-cm.-l band observed in the yellow shade spectrum is too strong for a para .uhstituted phenyl ring absorption and is believed to represent n type of quasi-carbonyl absorption. Khen the hydrogen of a 6-OH is strongly bonded t o a diazo nitrogen linked to the same n: omntic ring, the carbon-oyygen hond ncquire. considerable

385 33, than the blue shade crystals, which gave 31 absorptions. This probably means that the yellow shade has a lesser degree of crystalline order, more types of intermolecular linkages, and hence more absorption bands. The blue shade is then probably the thermodynamically stable one, that is a t room temperature. Frequencies of the absorption bands for the two polymorphic forms of Para Red as Kujol mulls in the rock salt region are given in Table I.

Table T.

h’ujol Mull Spectra of Blue and Yellow Shade Para Reds in the Rock Salt Region

I

(Frequency in em.->)

Ih 1596 m 1590 m 1313 m 1330 s 1295 W 1268 m 1230 m 1203 s

Blue Shade 1179 w 1015 w 98.5 u s 1x66 w 1133 m 981 m 1140 u. 963 n. 873 u. 1104s 862 m 1094 w 832 m 1037 w 1024 U‘ 838 s

808 793 787 768 748 685 682

Yellow Shade 1620m 1 1 6 8 ~ 1000 w 984 m 1572m 1 1 5 5 m 1516s 1 1 4 0 m 978 w 1332s 1111 m 961 v w 1 2 9 4 ~1107s 946 m 1261 m 1 0 9 2 m 874 w 861 s 1228 m 1039 u. 844 m 1208s 1 0 0 8 w

836 s 810 u791 w 766 w 754 m 748 m 741 w 688 w fim ”-- w

m , medium; s , stiong: w, weak.

Figure 9.


A . Blueshade B . Yellow shade R. Radiation


douhle bond character. In the present case, the 1620-crn-’ absorption of this hybrid carbon-oxygen is below the normal frequency of about 1680 em.-’ for a true carbonyl, yet well nbove the normal 1235-cm.-l absorption for a true single bond C-0 of a +--OH. The packing of the Para Red moleculesin the t h e shade crystal is such as not to allow the occurrence of this 1620-em.--1 absorption observed in the yellow shade. The -NHabsorption absence is evident from inspection of Figure 4. The structural formula proposed for Para Red given in Figure 1 is based on the spectroscopical evidence and reasoning given above. Considerations of bond angles and distances ( 1 7 ) show that the Para Reds are trans isomers with respect to the azo linbagr. While Figures 2 through 4 point out the infrared spectral differences between the blue and yellow shades of Para Red, Figures 5 and 6 emphasize the spectral similarities obsrrved when two polymorphs are examined in the infrared. These similarities are to be expected, since the two forms are but different spatial expressions of the same chemical composition. I n general, polymorphic forms of organic crystals with their preponderance of nonpolar intermolecular linkages will give rise to a greater nuinher of infrared spectral differences than inorganic crystals with their preponderance of ionic linkages. Evidence that the infrared spectral differences between crystals of the blue and yellow shades of Para Red were no accident of chance is shown by the fact that three samples of this pigment of shade unknown to the author were successfully identified from their infrared spectra as to polymorphic form. From empirical correlations of the data in Figures 2 through 6 Imed on published material ( 1 ) and studies in these laboratories, i t is probable that the molecules in the crystals of Para Red !-dlow shade are oriented in such a way as to give the 2-naphthalene ring more freedom to vibrate than in the Para Red blue shade crystals. I n the blue shade crystal, on the other hand, the benzene ring containing the hydroxyl grouping has more freedom to vibrate than in the yellow shade crystal. The yellow shade crystals were found to give more infrared absorption bands,

A band is classified as weak if its absorbance is less than 10% of the absorbance of the strongest band, as medium if between 10 and 60%, and as strong if greater than 60%. Copper Phthalocyanines. Figure 10 shows the structural formula of copper phthalocyanine. Linstead et al. ( 4 ) found the phthalocyanine molecule is composed of four isoindole units joined together by four extracyclic nitrogens to form a compound having a strainless 16-membered central ring with one metal atom in thr ctantrr i n the case of the metallic derivatives of phthnlocyanine.

Figure 10. Structural Formula of Copper Phthalocyanine

Robertson (23, 24) and Robertson and Xoodward (26) i n careful studies investigated the complete crystal structure of phthalocyanine and its metallic derivatives. Their results confirmed the earlier work ( d ) , yielded bond lengths, and showed the metal atom to be located in the center of the molecule and that it did not alter substantially the form of the phthalocyanine portion. The structure of copper phthalocyanine must be regarded as one continuous conjugated system. Robertson had only one form of copper phthalocyanine available to him-namely, the alpha form. This was prepared by sublimation as large needle crystals totally unsuited for pigment use but of definite value for x-ray diffraction study. He reported no data on any other form. Susich (50) showed by x-ray diffraction that copper phthalocyanine exists in two different crystal forms. Here the Robertson form is called “alpha,” because historically this crystal form was the first to be identified, and the first commercial pigment form is called “beta ”


386 Figures 11 through 14 show the infrared absorption spectra of two polymorphic forms of copper phthalocyanine. The spectra from 4000 to 650 cm.-' are presented for the alpha form, curve A , and the beta form, curve B , as Nujol mulls. Inspection of these figures reveals that the alpha and beta forms of copper phthalocyanine yield different spectra which are readily distinguishable one from the other. Here again the techniques of infrared spectroscopy enable one to identify and distinguish between polymorphic forms of organic crystals.

.

794

1288

1101

730 I

1190

I

V

FREQ. IN C M - '

10'91

Figure 13. Polymorphism of Copper Phthalocyanines A . Alpha form B. Beta form


XB

FREQ. IN CM."

Figure 11. Polymorphism of Copper Phthalocyanines A . Alpha form B . Beta form


I

\ 899

1011

' I

1003

I

1015 '

998 83 J

940

I

900

I

I R E 0 IN CM -I

within their respective crystal planes. The molecules in the alpha form present a "regular" array, those in the beta form a "staggered" array. Because the rock salt infrared spectra of the two forms are different, the spectral differences must arise from the different effects that the intermolecular forces present in one form have on the intramolecular forces of that form as compared with the interaction of similar forces in the other form. That the effect of intermolecular forces on intramolecular forces ail1 be different in the alpha and beta forms of copper phthalocyanine is evident upon examination of Figure 15. Empirical interpretation of the spectral differences between the two forms reveals that probably the ortho disubstituted benzene rings and the C--T groupings are freer to vibrate in the alpha form of copper phthalocyanin~than in the beta, while the C-H groupings are freer to vibrate in the beta form than in the alpha form. Infrared spectra in the rock salt region were run on copper phthalocyanines crystallized in xylene and in aniline. Each of ~ characteristic of these spectra showed a strong 7 3 0 - ~ m . -band the alpha form and also a weak 720-em. -1 band characteristic of the beta form. It, therefore, became of interest to work out a quantitative method of analysis for the determination of alpha and beta form copper phthalocyanines in mixtures of the two forms.

Figure 12. Polymorphism of Copper Phthalocyanines \:587

A . Alpha form B . Beta form Samples run as Nujol mulls

The alpha form, spectrum A , has 10 bands not found in the spectrum of the beta form. These are the 730-, 879-, 956-, 980-, 1003-, 1101-, 1173-, 1606-, 3152-, and 3210-cm.-' bands. The beta form, spectrum B , has three bands not observed in the spectrum of the alpha form. These are the 720-, 865-, and 3115cm.-l absorptions, The most striking infrared spectral difference between the two polymorphic forms is the strong 730-cm.-l band in the alpha form as opposed t o the also strong 720-cm.-l band in the beta form (as seen in Figure 11). The other readily observable differences will be found upon inspection of Figures 11 through 14. The difference between the crystal structure of the alpha and a suggested structure for the beta form (3.2)when a projection is taken parallel to the crystal plane of the respective polymorphic forms is shown in Figure 15. The two crystal forms differ only in the arrangement of the copper phthalocyanine molecules

3880

I

A

FREQ. IN CM. -1

Figure 14. Polymorphism of Copper Phthalocyanines A . Alpha form B . Beta form Samples run as Nujol mulls

V O L U M E 2.5, NO. 3, M A R C H 1 9 5 3

Alpha

387

Beta

Figure 15. -4tomic and 5lolecular Structure of Copper Phthalocyanine 735 7 2 0

1050

I

cause of the obvious difficulties in reproducing uniform dispersions of samples in Nujol, and because of the scattering in the region the absorbance measurements are made. The qualitative and quantitative method described above for the identification and determination of polymorphic forms of copper phthalocyanine has been employed successfully on many preparations. The method has proved to be reliable, rapid, and convenient. The results have shown that not only can polymorphs of organic crystals be identified and distinguished by means of infrared spectroscopy, but they can also be quantitatively analyzed as to percentage of each polymorph present when samples containing more than one polymorphic form are encountered. The polymorphic forms present in a series of six chlorinated copper phthalocyanines xere also identified and determined by the procedure given in this work. Reference to Figure 16, which gives the most characteristic absorption region for a typical sample of this series, shows bands a t 735,760,909, and 1050 cm.-' S o n e of the unchlorinated forms of copper phthalocyanine has an) of these four bands. ahchlorine was found in all six samples by means of chemical analysis and the i35- and 760-cm.-' bands are empirically of correct frequency for C-CI absorptions, it is certain the six samples contain chlorine. -4s the four absorptions which are unique to the chlorinated samples are also absorptions not found in polychlorinated copper phthalocyanine, the spectrum of TT hich in the same region is shown in Figure 17, it is probable that all si.; samples are monochloro copper phthalocyanines.

1050

FREQ. IN CM.-'

v -

Figure 16. Polymorphism of Copper Phthalocyanines A . Alpha form B . Beta form C. Monochlorinated


A solid state technique was a necessity, as polymorphs give identical solution state spectra. A quantitative analysis was, therefore, worked out employing dispersion of the samples in Nujol. Samples of pure alpha form and pure beta form were used as standards. T h a t these standards were pure was ascertained by noting the absence of the strongest beta absorption in the alpha form, and the absence of the strongest alpha absorption in the beta form. Spectra of the alpha and beta form standards and of known mixes of 98% alpha and 2% beta, and 95% alpha and 5 % beta, were run as Nujol mulls in the region from 700 to 745 cm.-l The procedure for the preparation of each sample was to use the same quantity of Sujol, the same quantity of pigment, and the same number of mullings, 5 . Mull thickness-Le., cell thicknesswas then adjusted until the absorptipn maxima of the Nujol 2850-em. -l doublet and the nearby N U J Otransmission ~ maximum a t 2700 cm.-' were the same. The known mixes were ground together thoroughly in a mortar and additional mixing occurred in the mulling process. The well-known "base-line" technique (8) was employed for the quantitative analyses of percentage alpha and beta forms present in a sample. The absorbance measurements were made a t 730 cm.-' for alpha and 720 cm.-l for beta, once adjustment of the cell length had been made to give equivalent thickness for known mix and unknown, as explained above. The copper phthalocyanine crystallized in xylene was found to contain 95% alpha form and 5% beta, while that Crystallized in aniline was found to contain 96% alpha form and 470 beta. No great accuracy is claimed for this quantitative method be-

9[

760

I

+

> t cn

z

w

c

z

I

FREQ. IN CM.-'

Figure 17. Polymorphism of Copper Phthalocyanines A . Polychlorinated B . rMonochlorinated

Samples run a s Nujol mulls

To illustrate the results obtained on the series of six chlorinated copper phthalocyanines, Sample 1 containing 6.4% chlorine was found by the infrared method described in this paper to be monochloro copper phthalocyanine of beta form. I t contained no alpha form crystals. Sample 2, which was crystallized into the alpha form by exposure of the beta form to boiling quinoline, containing 5.9% chlorine, was found to contain 62% monochloro alpha form and 38% monochloro beta form. Similarly, for the four remaining samples of this chlorinated copper phthalocyanine series i t was possible to identify the polymorphic forms of the pigment present and to determine the percentage of each form that was present in the sample using the infrared technique described. The absorption bands observed for five of the chlorinated copper phthalocyanine pigments in the principal regions of interest are shown in Table 11. The complete rock salt spectrum at least must be considered in identifying the polymorphic form or forms present in organic, inorganic, or mineral crystals. Experience in these laboratories has shown that both the sodium chloride and potassium bromide prism infrared regions are sometimes necessary to an unamhigu-


388 Table 11. Infrared Absorptions of Chlorinated and Unchlorinated Copper Phthalocyanines Alpha=

Beta0

1

.....

720 S

720 Si 726 Si

.....

......

2 720

Sz

3

4

5

720 SI

720 Si 726 S i

720 SI 726 Si

..

....

. . . . 786'$1 73b'& .. 735 735 735 735 735 760 760 760 760 760 909 909 ... 909 909 909 1050 1050 1050 1050 ... TO50 a .41pha and beta are unchlorinated. the other samples are chlorinated. .4U absorptions are in om.-1 S = s t r o n i , SI > €32. The criterion for claseification of a band as "strong" is the same as used in Table I. 730 S

... ...

... ... ,,,

under the influence of the force fields of other titanium and oxygen ions in other Ti06 octahedra. The observed absorptions, then, result from the directive influence of the covalent binding between octahedra, One would, therefore, expect a close similarity between the infrared absorption spectra of the anatase, rutile, and brookite forms of titanium dioxide. This great similaritv is actually observed.

..,

.

ous determination of the polymorphic modifications present in crvstalline materials. Titanium Dioxides. The use of infrared spectroscopical techniques to differentiate between polymorphic forms of inorganic ciystals is not new (IO,16). The power of the infrared method, however, has not been sufficiently appreciated or applied. It is only in recent years that any intensive application has begun to be made of the techniques of infrared spectroscopv to the study of inorganic compounds and minerals. Recently Keller and Pickett ( I S ) reported an extensive survey of the infrared spectra of clay minerals. Hunt, Wisherd, and Bonham's ( I O ) survey of the 2- to 16-micron spectra of 64 minerals, polymineralic sediments, and inorganic chemicals is also of recent origin. These workers show the spectra of polymorphic forms of calcium carbonate. The identification of polymorphic forms of inorganic pigments is illustrated in this paper using 'titanium dioxide as an example. The results obtained on this pigment illustrate the limitation of this infrared technique when applied to certain inorganic crystals. Titanium dioxide exists in three polymorphic forms: anatase, rutile, and brookite. The first two forms are well known through their widespread use as commercial white pigments of unexcelled hiding power. The third form, brookite, is a naturally occurring mineral of no known commercial importance. Figures 18 and 19 show certain regions of the rock salt infrared absorption spectra for anatase, rutile, and brookite. Inspection of these figures readily reveals the spectral differences among the three polymorphic forms of titania are slight. The infrared technique described in this work, then, would not be a very satisfactory m-ay to identify and determine quantitatively the polymorphic f o r m in titanium dioxide pigments. The wellknown x-ray technique is a much more satisfactory method. The limitation of the infrared technique illustrated by the titanium dioxides as an example applies to those inorganic substances whose cations and anions are both highly ionic in character. That this limitation is not serious is realized when one considers that very many pigments are organic in nature and that certain inorganic pigments are highly covalent. The fundamental structural differences among anatase, rutile, and brookite are verv slight indeed. Each polymorphic form consists of Ti06 octahedra. It is only in the different manner that these octahedra are bound together in three-dimensional space that serves to distinguish the structures of the three polymorphic forms. Because these Ti06 octahedra are shared in different ways, however, the Ti-0 interatomic distance varies as between anatase and rutile, and probably also brookite. As the force constants of bonds which are one factor in determining the vibrational frequencies of polyatomic molecules are closely related to interatomic distances, it follows that the force constants and hence vibrational frequencies will differ as between the polymorphic forms of titania. Because of the known approximately 50% ionic-50% covalent character of the binding in crystalline anatase, rutile, and brookite, it is not possible to interpret their observed infrared absorptions in a simple manner in terms of their crystal structures. Each ion in each Ti06 octahedra come?

COMPARISON OF TECHNIQUES

Three techniques are currently used for the study of polymorphism: the x-ray diffraction, the light microscopical, and the electron microscopical methods. The x-ray technique for the study of polymorphism has iwen of most value up to the present. Polymorphism, according to Susich (SO), was observed by x-ray diffraction a t least as early as 1933 when i t was found that aniline-azo-2-naphthol, known as Sudan Orange R, existed in four different modification3. The practical significance of this finding was not appreciated because this dyestuff is used for the coloring of fats and male3 in the state of a solid solution, The existence of polymorphic forms is. however, practically important for pigments and dyes applied in the solid state for coloration of paints, lacquers, plastics, gum, paper, and other materials. Here the different phwical properties of the vaiiouq forms result in difference in tinctorial qualitieq.

,874

\

Ya

NUJOL

\'

C

I

FREO. IN CM. -1

Figure 18. Polymorphism of Titanium Dioxide A. B r o o k i t e B. R u t i l e C. A n a t a s e S a m p l e s run as Nujol mulls

The x-ray technique for studying polymorphism has certain advantages. I t requires only small samples, from 10 to 20 mg. It does not require the comparatively large-size crystals neccssary to successful application of the light microscopical technique. It is simple, as only powder patterns obtained with the use of flat cameras are required and from 10 to 12 powder patterns can he taken on one 9 X 12 cm. x-ray film. I t is rapid, a 10- to 20minute exposure time being usually sufficient. More recrnt technique employs monochromatic radiation and Geiger counters rather than filtered radiation and film. The x-ray method can also be used for the study of polymorphic forms that are not entirely free of impurities, because the diffraction pattern is not sufficiently sensitive to reveal foreign substances. It is often the best technique for studying the polymorphism of inorganic suhstances. The limitations of the x-ray method are several. It is ordinarily not suitable for detecting the presence of a polymorphic form unless such is present in a sample in greater than 5 % by weight concentration. The x-ray technique can detect less than 5 % of anatase in rutile, however, an exception to the above state-

V O L U M E 25, NO. 3, M A R C H 1 9 5 3 ment. In those cases where the crystallites of a substance are too small or where the crystal lattice is too distorted, the power pattern of a substance may show lines which are either diffused or too few in number to give exact characteristics. In such instances the x-ray method is not applicable. Furthermore, difficulty is encountered when one polymorphic form is in a nearly perfect state of crystallizat,ion while another is in a poorly crystallized state. The extensive use of the x-ray method is dependent upon the economical availability of a large collection of comparative data. To date published x-ray data of the joint American-British committee contain few organic pigments and dyestuffs among them. The collection is continually being extended. however.

b

389 For inorganic polymorphs whose absorptions are few and n-eak the x-ray method of studying polymorphism is superior. SUMMARY Polymorphic forms of both organic and inorganic crystals can be identified and in certain cases quantitatively determined by means of infrared spectroscopical techniques. Knowledge can he obtained of how the molecular or molecularionic species are arranged and oriented in the crystal lattices of various polymorphic substances by empirical interpretation of their infrared absorption spectra. Infrared studies of polymorphic pigments have proved uselul in relating their crystal form and structure to pigment properties.

997

ACKNOWLEDGMENT The author is grateful to R. H. Kienle and G. L. Royer for their encouragement and support of this work, to J. P. Leineweber and F. R. Tarantino for supplying samples n-ith their histories, and to 0. E. Sundberg and associates for the chlorine determinations. -4sample of Arkansas brookite was generously provided by Roy Dahlstrom and C. H. Moore, Titanium Division, National Lead Co., Inc. LITERATURE CITED

FREQ. IN CM. -1

Figure 19. Polymorphism of Titanium Dioxide A. Brookite B. Rutile

C. Anatase


Light microscopy has been used to determine among other physical properties the form and polymorphism of crystals. The examination can be completed quickly using a small amount of material. If crystals greater than about 1 micron in average diameter are available, the light microscopical technique is satisactory. The method is not a t all applicable to pigmentq, however, because the particle size of these materials is below the range of the light microscope. The crystals being studied must also be free of the usual difficulties associated n i t h optical properties such as morphology, tm-inning, and variations in degree of crystal perfection. Moreover, a relatively large minimum concentration of a given polymorphic form is necessary in a mixture of polynioiphs to ensure qualitative detection. The electron microscopical method of investigating the polymorphism of crystalline substances has the advantage of being able to handle crystals of a size ordinarily encountered in commercial pigments. It is subject to the optical property difficulties of light microscopy and cannot detect a polymorphic form present in small concentration. I t is also sometimes a disadvantage when the electron beam changes one polymorphic modification into another, although such can be an aid to the identification of the polymorphic form present. The infrared technique proposed in this paper for the identification and determination of polymorphic forms of crystals has advantages over methods previously used. It is rapid, qualitative detection of a polymorphic form present to the extent of only 1% by nreight can be made, and quantitative determinations of the various forms present in a mixture can be accomplished. One is not troubled by the difficulties of crystal morphology. The small-sized particles encountered in commercial pigments can be handled successfully and information can be obtained about how the molecules are oriented in the crystals studied.

(1) Colthup, S . B., J . Opt. Soc. Amer., 40, 397 (1950). (2) Couture, Jean, Rea. Sci., 85, 86 (1947). (3) Couture, Lucienne, J. Chem. Phys., 15, 153 (1947). (4) Dent, C. E., Linstead, R. P., and Lowe, A. R., J . Chem. Soe.,

1934, 1036.

( 5 ) Flett, XI,,J . Chem. Soc., 1948, 9. (6) Halford, R. S.,J . Chem. Phys., 14, 8 (1946). (7) Hamm, F. A , and Van Korman, E., J . A p p l . Phys., 19, 1097

(1948).

(8) Heigl, J. J., Bell, XI. F., and White, J. U., Aii.4~.CHEM.,19,

293 (1947). (9) Hornig, D. F., J . Chem. Phys., 16, 1063 (1948). (10) Hunt, J. M.,Wisherd, XT. P., and Bonham, L. C., ~ ~ N A CHEM., L.

22, 1478 (1950). (11) Kastler, Alfred, Compt, rend., 227, 1024 (1948). (12) Keller, W. D.. and Halford. R. S..J . Chem. Phzls.. 17. 26 (1949).

Keller, W.D., and Pickett, E. E., Am. J . Sci.,-248, 264 (1950). Kendall, D. N., “Application of Infrared Spectroscopy to the Molecular Structure of Anthraquinone Derivatives,” presented before Symposium on Molecular Structure and Spectroscopy, Ohio State University, June 1949. Kofler, Bdelheid, “Micro Methods for Identification of Organic Substances and Their Mixtures,” Volumes 1 and 2, Innsbruck, Austria, Sovitas, 1937. Kyswander, R. E., P h y s . Rev.,28,291 (1910). Pauling, L. C., “The Nature of the Chemical Bond,” Ithaca, h’. Y . , Cornel1 University Press, 1940. Pfund, A. H., J . Opt. SOC.Amer., 23,375 (1933). Ibid., 24, 143 (1934). Philpotts, -4. R., Thain, W., and Smith, P. G., ANAL.CHEM.,23, 268 (1951). Rawlins, F. I. G., and Rideal, E. K., Proc. Roy. Soc., A116, 140 (1927). Reinkober, Otto, Ann. P h y s i k , 34, 343 (1912). Robertson, J . XI., J. Chem. Soc., 1935, 615. I b i d . , 1936, 1195. Robertson, J. RI., and Woodward, I., Ibid., 1937, 219. Robertson, Robert, Fox, J. J., and Martin, A. E., P h d . T r a n s . Roy. Soc. ( L o n d o n ) , A232, 463 (1934). Robertson, Robert, Fox, J. J., and Martin, -4.E., Proc. R o y . Soc. ( L o n d o n ) , A157, 579 (1936). Saksena, B. D., Proc. I n d i a n Acad. Sci., 12A, 93 (1940). Simon, F., .Vuture, 125, 855 (1930). Susich, George, ANAL.CHEM.,22,425 (1950). Sutherland, G. B. B. hl., and Willis, H. A,, Trans. Faraday SOC., 41, 289 (1945). Tarantino, F. R., Stubbs, D. H., Cooke, T. F., and Melsheinier, L. 4.,Am. I n k M a k e r , 29, 35 (1951). %*hitmore, F. C., “Organic Chemistry,” New York, C. Van Nostrand Co., 1937.

RECEIVEDfor review July 19, 1952. Bccepted December 22, 1952. Presented before the Division of Physical and Inorganic Chemistry a t the l l Y t h Meeting of the AMERICAN CHEMICILSOCIETY,Boston, Mass.

Identification of Polymorphic Forms of Crystals by Infrared Spectroscopy

Recommend Documents