090
P. F. Krause, J. E. Katon, and R. W. Mason
The Journal of Physical Chemistry, Vol. 82, No. 6, 1978
The Polarized Infrared Spectra and Structure of Crystalline Bromoacetic Acid? P. F. Krause, Department of Chemlstry, Universlty of Central Arkansas, Conway, Arkansas 72032
J. E. Katon,* and R. W. Mason$ Department of Chemlstty, Miaml Unlverslty, Oxford, Ohio 45056 (Received September 23, 1977) Publlcation costs assisted by the U.S. Alr Force
The polarized infrared spectra of oriented polycrystalline films of one polymorph of bromoacetic acid have been recorded. The observed factor group splittings are consistent with a hydrogen-bonded polymeric structure. The enhanced resolution afforded by the infrared polarization studies along with appropriate fador group analyses has allowed approximation of both site and factor group symmetries. These are discussed in relation to recent x-ray crystallographic data.
Introduction I t is firmly established that carboxylic acids in the solid state may form more than one hydrogen-bonded structure, since a number of crystalline acids have been studied by x-ray diffraction. These compounds often exist in polymorphic modifications, however, many of which have not been studied structurally. Recent studies on a variety of compounds have shown that structures may be predicted through the polarized infrared spectra of oriented polycrystalline f i l m ~ . l - ~ Single-crystal and oriented polycrystal polarized midinfrared spectra for a-chloroacetic acid6are consistent with a hydrogen-bonded polymer structure or the hydrogenbonded tetramer structure more recently found by x-ray diffraction studiesS6 Polarized mid-infrared spectra of oriented polycrystalline films of y-chloroacetic acid7 are consistent with a polymer structure, while 0-chloroacetic acid single-crystal polarized infrared spectra are consistent with a hydrogen-bonded dimer7in agreement with recent x-ray studiesS8Vibrational analyses on polycrystalline but nonoriented samples of both bromoaceticgand iodoaceticlO acids have supported dimeric structures. The vibrational spectra of dimeric carboxylic acids are complicated by the fact that the fundamental modes occur as phase-related pairs. The monomeric units are weakly coupled and the selection rules derived from group-theoretical arguments are not rigorous. In an attempt to solidly establish the solid-state structure of bromoacetic acid via vibrational spectroscopy, the infrared spectrum of crystalline bromoacetic acid has been reinvestigated using the polarized infrared technique. As the roomtemperature and low-temperature spectra of solid but nonoriented bromoacetic acid are similar to the melt, the structure of the polymorph of bromoacetic acid under investigation was thought to remain unchanged. For this reason, the dimeric structure was chosen as a starting point for the analysis of the polarized infrared data. Three polymorphs of bromoacetic acid have been reportedll of which two are stable only at elevated pressures. However, a recent x-ray diffraction study12 reported crystal-structure data for two polymorphs of bromoacetic acid. These were prepared by evaporation from different solvents. The results of the vibrational study of crystalline bromoacetic acid presented in this 'Supported in part by the U. S. Air Force under contract No. F33615-77-C-5013. *Present Address: Olin Corporation, Lake Charles, La. 70605.
study will be discussed in relationship to the x-ray diffraction results. Since the crystal structures of two polymorphs are known, this investigation further serves to indicate the usefulness of recording polarized infrared spectra of polycrystalline compounds in terms of the structural arguments that can be forwarded.
Experimental Section Reagent grade bromoacetic acid purchased from Matheson Coleman and Bell was purified by vacuum distillation. The fraction used in all experiments was collected at 74 "C under a vacuum of less than 1Torr. The samples were stored in vacuo and redistilled frequently. Oriented polycrystals of organic liquids have been grown by simply cooling the sample between alkali halide plates.13J4 In the present cases, oriented polycrystalline films of bromoacetic acid were prepared by warming the sample to just above its melting point (48 "C) and then pressing the sample between KBr windows. The sample was then placed in a desiccator where crystallization was allowed to occur. A qualitative indication as to the degree of orientation was provided by observing the degree of light extinction produced by the sample between crossed Polaroids. If the sample was sufficiently oriented, the cell was transferred to a conventional liquid-nitrogen Dewar or to the sample compartment of a Cryogenic Technology, Inc., Model 20 Cryostat. The sample was allowed to cool to the desired temperature before recording of the spectra. Spectra were recorded on a Perkin-Elmer Model 180 infrared spectrophotometer. The estimated accuracy of the frequencies is at worst il cm-l for measurements on different samples. For different polarizer settings for a particular sample, the precision, which is more significant than the accuracy, is estimated to be at least i0.2 cm-l. Results Typical mid-infrared spectra of oriented polycrystalline bromoacetic acid at 77 K are shown in the two traces in Figure 1. The differences between the top and bottom spectra are due to a 90" rotation of the wire-grid polarizer. The sample chosen shows excellent orientation and no vibrational modes yielded maximum or minimum intensities at polarizer settings other than 90° apart. The corresponding infrared data are given in Table I. Comparison of spectra at temperatures as low as 20 K with room-temperature spectra show little variation. Changes in the spectra that were observed are those due to typical band sharpening upon cooling and also the observation of
QQ22-3654/78/2Q82-Q6QQ~Q7 .QQ/Q. @ 1978 American Chemical
Society
Infrared Spectra and Structure of Bromoacetic Acid
The Journal of Physical Chemistry, Vol. 82, No. 6, 1978 691
100
m
it4
!IM x)
0 4
1w
80
,I '%
20
0
Figure 1. Polarized infrared spectrum of crystalline bromoacetic acid. Top trace: 6' Table I).
a few additional bands at lower temperatures. These new bands were observed only when the polarized infrared technique was utilized and are readily attributed to resolution of factor group components. It is important to note that the spectra recorded a t a given temperature for all crystal films was independent of sample preparation. This indicates that the same polymorph is being investigated in all cases, Except for frequency shifts with the OH vibrational modes as the temperature was lowered, the experimentally observed frequencies were also essentially temperature independent. Three crystallographic modifications of bromoacetic acid were proposed in 196111 with no information reported concerning their molecular or crystallographic symmetry. Leiserowitz and vor der Bruck12have recently reported two crystallographic forms of bromoacetic acid as indicated by x-ray diffraction studies. Form I was grown by slow evaporation from carbon tetrachloride while form I1 was prepared by slow evaporation from methylcyclohexane. The x-ray diffraction results are summarized in Table 11. It is of real interest to determine whether our crystal-film structure is the same as one of those reported. Crystals of bromoacetic acid were grown from both carbon tetrachloride and methylcyclohexane and mid-infrared survey spectra (Nujol mulls) of these samples were compared with mid-infrared survey spectra of crystal films of bromoacetic acid. Differences are noted in the spectra of the samples recrystallized from the two different solvents but they are fairly minor; striking differences are observed between these spectra and spectra obtained from crystal films. Figure 2 shows a portion of the mid-infrared spectrum (1400-600 cm-l) comparing bromoacetic acid recrystallized from carbon tetrachloride (upper trace) with a nonpolarized crystal-film spectrum of bromoacetic acid (lower trace). Both spectra were recorded at room temperature. The two spectra show many differences. Three of these differences are as follows: (1)A very strong band at about 1200 cm-' exists in the crystal-film spectrum, which has been assigned to the CH2 wag. This mode is apparently
+ 90'
spectrum. Bottom trace: 6 spectrum (see text and
Figure 2. Upper trace: Nujol-mull infrared spectrum (1400-600cm-') of bromoacetic acid recrystallized from CCI4 recorded at room temperature. Lower trace: mid-infrared spectrum (1400-600 cm-') of a crystal film of bromoacetic acid recorded at room temperature.
greatly shifted in the Nujol-mull spectrum of bromoacetic acid recrystallized from CCl,, since it is not clearly visible. (2) The OH out-of-plane deformation at about 900 cm-l in the Nujol-mull spectrum is shifted to about 870 cm-l in the crystal-film spectrum. (3) The CBr stretching mode shows a shift from about 630 cm-l in the Nujol-mull spectrum to 729 cm-l in the crystal-film spectrum. These observations are certainly convincing evidence that the oriented polycrystals grown from the melt between alkali halide windows are of a different structural modification
692
The Journal of Physical Chemlstry, Vol. 82, No. 6, 1978
TABLE I: Infrared Spectral Datau for Crystalline Bromoacetic Acid Oriented crystal Nonoriented crystal 77 K 298 Kb 77 KC e c l d e + 90C~d Assign 3213.4 2900-3200 3223.8 3212.8 3141.9 -3153
1
P. F. Krause, J. E. Katon, and R. W. Mason
TABLE 11: X-Ray Diffraction Results for Two Polymorphs of Bromoacetic Acida I1 (Dimer) (from I (Dimer) (from CCl,) methylcyclohexane) Monoclinic Orthorhombic Monomer site C, Dimer site Ci Conformation angle 9"
Monbmer site C, Dimer site Ci Conformation angle 22O
a L. Leiserowitz and D. vor der Bruck, Cryst. Struck Commun., 4, 647 (1975).
3009 2954 2580 248 1
1814 1720
1385
1385.7
I
6 (CH,)
(scissors)
1380.6
1296
1197
1200.3 1150.8
1151
1153.3
871
886.2
1150*3}6(cH,) (twist) 1141.1 1141.2 930.0
915.3 913.2
900
6 (CH,) (rock)
902.3 895.7
729
730.1
656 539
668.4 541.9
887.4
895.1 874.8 14C-C)
669.1
670.9}6
538.7
538.5 S ( C 0 , ) (wag)
(CO,)(scissors)
410 All frequencies in cm-'. Relative polarizer angle,
a
Reference 8.
This work.
than the polymorphs reported by Leiserowitz and vor der Bruck.12 This apparent third modification will be indicated as form 111. The relatively high frequency (729 cm-l) for the carbon-bromine stretching vibration is consistent
with a conformational angle of about Oo in other carbonyl compounds.15 This can be compared with the conformational angles of 9' for form I and 22O for form 11. The form I11 polymorph can be obtained from the bromoacetic acid recrystallized from CC14by simply melting the sample, pressing the melt between alkali halide windows, and allowing the sample to cool. The conclusion that we are studying a third crystalline modification, whose structure is unknown, makes the analysis of the polarized spectra somewhat more complex and speculative. It should be noted that form I11 has been prepared by cooling a liquid film between two alkali halide windows under pressure. This polymorph is presumably metastable at ambient conditions and if disturbed by scratching or scraping will convert to one of the more stable polymorphs. Such behavior has been noticed previously with chloroacetic acid. The y polymorph is formed when liquid chloroacetic acid is cooled between alkali halide windows and is readily converted to the a polymorph upon mechanical di~turbance.~ It cannot, therefore, be removed from the windows and studied in other ways.
Discussion The analysis of the polarized infrared spectra of bromoacetic acid begins with the following pertinent observations. (1) For any molecular vibrational mode of the cyclic dimer, only two bands are observed in the crystal. It appears that there are two components for all bands. (2) There are a number of relatively strong infrared bands observed in the polarized spectra that appear at frequency separations greater than one would expect for factor group componentsbut which are not resolved in the nonpolarized spectra. The C-0 stretching region illustrates this behavior (see Table I). The single medium-strong band at 1296 cm-l in the crystal-film spectrum of a nonoriented sample of bromoacetic acid yields bands at 1316.0 cm-l (strong) and 1296.5 cm-l at one polarizer setting in the polarized spectra. With a 90° rotation of the polarizer, bands at 1315.1 cm-l and 1294.4 cm-l (strong) are observed. The bottom trace of Figure 3 illustrates the former behavior while the top trace shows the latter. The CH2 scissoring mode (1385.7 cm-l in the nonpolarized spectrum) also illustrates this behavior. At the setting of the polarizer shown in the top trace of Figure 3, two bands at 1397.4 cm-l and 1380.6 cm-l are observed. With a 90° rotation of the polarizer (bottom trace) two other bands with frequencies 1392.8 cm-l and 1380.3 cm-l are observed. A previous analysis of polycrystalline, nonoriented bromoacetic acidgindicated that there were bands at 1385 and 1296 cm-' in the infrared (non-Raman observable) and bands at 1395 and 1306 cm-l in the Raman spectrum (non-infrared observable). This indicates the presence of a center of symmetry and, by implication, a hydrogenbonded dimer structure. The very large splitting5 observed in the polarized spectra, 20 cm-l for the C-0 stretch and
Infrared Spectra and Structure of Bromoacetic Acid
1400 l
1200 1
1
I
i
I
I
The Journal of Physical Chemistry, Vol. 82, No. 6, 1978 693
Fmac k w
SITE G#up
+v
Cs
F
m GAOUP Dai
1000
KL KfES YIELD THIEE INFWRP ACTIVE FAexlR OROUP c[EpoBTTs
100
Flgure 4. Correlations between possible site and factor groups for bromoacetic acid.
I
Flgure 3. A portion (1400-1000 cm-’) of the polarized infrared spectrum of crystalline bromoacetic acid at 77 K. The upper trace is of the 6 90’ spectrum while the bottom trace is of the 6 spectrum.
+
12 cm-I for the CH2scissors, suggest that the bands at 1392 and 1306 cm-l correspond to the Raman bands observed in the dimer structure. This indicates that the polymorph under investigation does not have a center of symmetry and is, therefore, a hydrogen-bonded polymer. Other portions of the polarized spectra show similar behavior, notably the OH stretching region. In the infrared spectrum of a nonoriented sample of crystalline bromoacetic acid, a strong, broad, unresolved band (3200-2900 cm-l) was observed a t room temperature. Nonoriented sample spectra have clearly indicated two bands a t 3223.8 and -3153 cm-l at 77 K, however. The enhanced resolution afforded by the polarized infrared technique yields further splittings; 3212.8 and 3132.7 cm-l at one polarizer setting and 3213.4 and 3141.9 cm-l a t a polarizer setting rotated by 90°. The most plausible explanation follows the same reasoning as employed above; that is, the bands a t 3213 and 3135 cm-l are assigned to the infrared-active and Raman-active “dimeric” OH stretching modes which are now both infrared active because of the structural change, while the smaller splittings observed are due to
the resolution of two factor group components for each band. In terms of gross structural features of the particular polymorph investigated, the data seem most consistent with monomer molecules which are hydrogen-bonded in a polymeric or tetrameric framework. Analogous structures to these have been reported for chloroacetic acid (y and a,respe~tively).~~’ Some aspects of the crystal structure can be derived from the polarized infrared results and relationships with the molecular structure of this particular polymorph can be postulated. As stated above, each fundamental vibrational mode apparently splits into four infrared-active factor group components. These occur in pairs. One pair corresponds to the Raman-active modes of the cyclic dimer and the other pair corresponds to the infrared-active modes of the cyclic dimer. Clearly, the polymer structure still shows the strong coupling of adjacent molecules through the hydrogen bonding since these pairs are split by a fairly large amount (10-20 cm-l). The two components of each pair are split by much smaller amounts (
