MSU RETURNING MATERIgg§z Place in book drop to LIBRARIES remove this checkout from —__ your record. FINES will be charged if book is returned after the date stamped be10w. M “IE- 96 . a. 'vfri t“ u‘ "' " .-" 1"" ‘-" 5 is; A STUDY OF THE VIBRATIONAL SPECTRA OF POLYCRYSTALLINE CARBON TETRACHLORIDE, CARBON TETRAFLUORIDE, AND THE CHLOROFLUOROMETHANES BY Donald Dexter Fontaine A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1983 ABSTRACT A STUDY OF THE VIBRATIONAL SPECTRA OF POLYCRYSTALLINE CARBON TETRACHLORIDE, CARBON TETRAFLUORIDE AND THE CHLOROFLUOROMETHANES by Donald D. Fontaine The far-infrared spectrum of solid CC14, the Raman spectrum of solid CF4 and both the far-infrared and Raman spectra of the solid chlorofluoromethanes have been obtained. These molecules have few interaction types, are chemically similar, and are expected to have weak intermolecular forces in solids. It was of interest to determine the extent to which a simple non-bonded interatomic potential could reproduce some of the structural and spectroscoPic properties due to these forces. An interatomic potential parameter set which is consistent with the observed structure and available lattice vibrational frequencies of CF4 has been obtained. In the course of this work the reported CF4 structure was found to be in error and the corrected structure is given here. Fontaine, Donald D. None of the currently available interatomic potential parameter sets were found to fit the structure of phase II CCl4 solid as closely as those tested with CF4. This is perhaps not surprising given the unusually large unit cell population in the CCl4 structure. Those potential parameter sets which best fit the CCl4 phase II structure are given. The far-infrared spectrum of CCl4 phase II obtained here is the first spectrosc0pic data which requires, for its interpretation, a unit cell population approaching that determined crystallographically. In order to test the transferability of the carbon, chlorine and fluorine interatomic potential parameters -—obtained for CF4 and CCl4-—the crystal symmetries of the chlorofluoromethanes had to be inferred, either from that of molecules of similar symmetry, or from an interpretation of the spectroscopic data. In the case of CFZCl2 the spectroscopic data indicate a C4h factor group and a C2 site group for the crystal. Unfortunately the computer program used for the lattice dynamics calculations was limited to crystals of orthorhombic or lower symmetries, and this limited the computation of lattice frequencies to the structures of molecules analogous to CF2C12; the CHZBrz, CH2C12, SpectroscoPic data obtained for CFBCl and CFCl3 did not limit the possible crystal symmetries to a number which and CF4 strucures. would allow meaningful calculations to be made. TABLE OF CONTENTS CHAPTER ONE INTRODUCTION CHAPTER TWO EXPERIMENTAL A) Infrared Spectroscopy B) Raman Spectroscopy CHAPTER THREE INTERFEROMETRY A) Optical System B) Data Aquisition and Instrument Control C) Data Processing D) Detector Comparison: Bolometer versus DTGS CHAPTER FOUR CARBON TETRACHLORIDE A) Structures and Spectroscopic Predicitons B) Previous Spectrosc0pic Results C) Spectrosc0pic Results D) Packing Analysis of CCl4 Phase II 1) Description of Calculations 2) Description of Potential Parameter Sets used in the Packing Analysis 3) Results and Discussion CHAPTER FIVE CARBON TETRAFLUORIDE A) Crystal Structure of the Low Temperature Solid B) Lattice Dynamics C) Spectrosc0pic Predictions and Results ii Page number 1 5 ll 11 18 22 30 34 34 39 40 47 47 54 59 65 65 75 78 CHAPTER SIX CFCl CFZCl CF Cl 3 2 3 A) Previous Spectroscopic Results-CF2C12 B) The Far Infrared and Raman Spectra of Solid CF Cl 2 2 C) Factor Group Analysis of CF2C12 D) Structural Analogues E) Packing Analysis of the Structural Analogues F) Spectrosc0pic Results--CFC13 G) Factor Group Analysis--CFC13 H) Spectroscopic Results-—CF3C1 I) Conclusion APPENDIX REFERENCES iii 86 87 87 95 100 102 109 114 119 125 126 132 Table Table Table Table Table Table Table “Table Table Table 10 LIST OF TABLES Observed lattice frequencies of CCl in Solid Phase II. 4 The observed frequencies of the v2(e) band of CCl4 in solid phase II. Test for Convergence of Lattice Sums for Carbon Tetrachloride Using Potential Parameter Set II. Change in the CC1 Phase II Lattice Parameters a, b, c, B Necessary to Achieve Minimum Static Lattice Energy at Various Lengths of the Lattice Sum (Potential Parameter Set II). The Changes in Molecular Position and Orientation of Each of the gour CCl Molecules in the Asymmetric Unit of Solid Phase II, Necessary to Achieve Minimum Static Lattice Energy, Computed at Various Lengths of the Lattice Sum (Potential Parameter Set II [59]). Test for Convergence of the Lattice Energy Sums for the CC1 Phase II Structure (Potential Parameter Sgt II [69] ), with a Partial Charge of -0.3 Electrons on the Chlorine Atoms. . . 0°K Estimation of AHsub of CC14. Potential Sets Used in the Packing Analysis of Solid Phase II in CC14. The Changes in Lattice Constants a, b, c, B Necessary to Achieve Minimum Static Lattice Energy Computed with Ten Different Potential Parameter Sets. (Direct Lattice Summation Limit 5.0 A.) The Changes in Molecular Position and Orientation of Each of the gour CCl Molecules in the Asymmetric Unit of Solid Phase II Necessary to Achieve Minimum Static Lattice Energy, Computed with Three Different Potential Parameter Sets Each with and without the Inclusion of Partial Charges on the Chlorine Atoms. (Direct Lattice Summation Limit 5.0 A.) iv Page 43 46 50 52 53 55 56 60-61 63 Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 11 13 14 15 16 17 18 19 21 22 23 24 25 26 27 28 29 Fractional Atomic Coordinates in Bol'shutkin et al. Structure 32 . Fractional Atomic Coordinates in Sataty, et al. Structure 28 . Fractional Atomic Coordinates, This Work . Some Atomic Distances of CF4, Phase II. Calculated versus Observed Diffraction Patterns. 67 . Potential Parameter Set I CF Lattice Frequencies Calculated with Po ential I. Atomic Coordinates of the Molecular Frame, and the Orientation Matrix. Potential Parameters: Set II. CF 11‘} Lattice Frequencies Calculated With Potential Changes to Obtain Equilibrium Structure. Correlation Table for CF in Phase II. 4 Observed Raman Frequency Shifts for CF Phase II. 4 in Solid The Observed Raman Frequency Shifts and Infrared Absorptions of Solid CFZClZ' Frequencies of the Fundamental Vibrations of_l CF2C12 in the Gas Phase and in the Solid (cm ). Site Group-Factor Group Combinations Consistent With the Lattice Spectra of CF2C12. Group Theoretical Predictions for Several Site Group/Factor Group Combinations. Potentials Used in CF Cl2 Packing and Lattice Dynamics Calculations. Structure and Lattice Frequencies Obtained with the Given Factor Group and Site Group. 66 67 68 68 69 70 74 74 76 77 79 80 82 90-91 94 99 101 104 105 Table 30 The Observed Raman Frequency Shifts of CC13F Solid at 15°K. 112 Table 31 Site Group-Factor Group Combinations CFZCl2 118 Table 32 Raman Active Fundamentals of Polycrystalline CF Cl at (20-25 K): Positions, Assignments, an Chlorine IsotOpic Splitting. 122 vi Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 10 11 12 13 14 15 16 17 18 19 LIST OF FIGURES Page number Optical diagram of interferometer Far-infrared Spectrum CF2C12 with and without CSI filter Aliased CF2C12 spectrum Unapodized transform of a sinusoid Triangularly apodized sinusoid transform Split cosine bell apodized sinusoid transform Apodized versus Unapodized CF3C1 solid spectra Average ADC value versus Averaging time (bolometer vs. DTGS) Correlation diagram phase II CCl4 Correlation diagram phase III CC14 Far-infrared spectrum CCl4 phase II (13 K) Temperature dependence of CCl4 UR-active lattice modes The v2 e internal vibration in the infrared Static Lattice energy versus b-fractional coordinate CF4 Raman spectrum CF4 (20-25 K) Raman spectrum CF2C12 (20 K) Infrared spectrum v4, v5 regions CF2C12 (.14 K) Far-infrared spectrum CFZClz-lattice regions The temperature dependence of the IR-active lattice modes CFZCl2 vii 12 19 20 25 26 27 28 33 36 38 41 44 46 71 83 88 92 96 97 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 20 21 22 23 24 25 26 A-1 A-2 A-3 A-4 A-5 A-6 Raman spectrum CFCl3 solid-internal region Raman spectrum CFCl3 solid-lattice region Far-infrared spectrum CFCl region 3 solid-lattice The temperature dependence of the IR-active lattice modes CFCl Raman spectrum CF3 3 Cl solid Far-infrared Spectrum CF3C1 solid The temperature dependence of lattice modes CCl3 The 6-12 potential potential The 6-12 potential potential The 6-12 potential potential The 6-12 potential potential The 6-12 potential potential The 6-12 potential potential and and and and and and viii fitted fitted fitted fitted fitted fitted the IR-active exp-6 exp—6 exp-6 exp-6 exp-6 exp-6 110 111 115 116 120 123 124 126 127 128 129 130 131 CHAPTER 1 INTRODUCTION The use of infrared and Raman spectroscopy to establish or confirm molecular structure has a long and fruitful history. With the develOpment of the appropriate group theoretical procedures in the late 1940's [1,2], extension to include interpretation of the vibrational spectra of molecular solids was made possible. A recent description of these procedures has been given by Rousseau, Bauman, and Porto [3]. In favorable cases these methods allow the prediction of crystal symmetries [4-8]. More typically the symmetry is known and is used to predict spectroscopic activity [9-11]. Lattice dynamical and ab initio calculations have been undertaken in order to quantify the forces acting in molecular crystals [12-15]. The purpose of lattice dynamic calculations has been to determine the adequacy of simple potentials between non-bonded atoms for the prediction of the lattice vibrational spectra, as well as to test the transferability of a given potential among the members of a homologous series of compounds. The system chosen for study in the present work is CF(n)C1(4—n)' n=0,...,4. These molecules have few interaction types and the intermolecular forces in the solids are expected to be weak, so this system seemed amenable to calculation with a simple potential. However, the crystal structures necessary for these calculations have not been reported for the compounds with n=l,2,3. In these cases the crystal symmetry must first be inferred, either from that of molecules of similar symmetry, or from an interpretation of the spectroscopic data. For a selected crystal symmetry, an equilibrium structure is obtained by packing analysis [16] under the constraints imposed by the Symmetry relationships. This packed structure is then used in a lattice dynamics computation of lattice frequencies. A comparison of the observed frequencies to those calculated determines the appropriateness of both the structure and the potential parameter set used in the calculation. Packing analysis of CF4 and CC14, for which crystal structures are known, was used to determine the interatomic potential parameter sets used in the calculations involving the chlorofluoromethanes. Three solid phases of CCl are known to exist in 4 equilibrium with the solid's vapor pressure at low temperatures [24,25]. Two higher-pressure phases of CCl4 are also known [26,27]. The present study is concerned only with the lowest temperature phase-—the so-called phase II of CCl4 [24]. The low temperature phase of carbon tetrachloride has been studied in the mid-infrared [17,18], far-infrared [19,20], and by Raman spectroscopy [21-23]. The far infrared spectrum of this phase obtained in this study has shown a larger number of lattice modes than had been previously reported [19,20]. This is the first spectro- sc0pic information which requires, for its interpretation, a number of molecules in the unit cell approaching that determined crystallographically [24]. Previous results were interpretable with a "reduced" unit cell containing only four molecules [19]. The far—infrared [28], mid-infrared [29,30], and Raman [29] spectra of the low temperature solid phase of carbon tetrafluoride, phase II, have been investigated previously. The far-infrared spectra [28] suggested a reassignment of the previously determined space group symmetry [31,32]. Molecular packing analysis, lattice dynamic calculations, interatomic distances and calculated diffraction patterns indicate that this reassigned structure [28] is incorrect, and that the correct structure results from an alternative correction to that given by Sataty et al. [28] for the error present in the atomic coordinates of reference 32. The conclusion [28] that the space group symmetry given in reference 32 should be reassigned is unchanged. The Raman and mid-infrared spectra of solid CFCl3 have been reported by other workers [33,34]. The present work includes the first far-infrared results obtained for this compound. Crystalline CF2C12 and CF3C1 have not been previously investigated by either far-infrared or Raman spectrosc0py. CHAPTER 2 EXPERIMENTAL The dichlorodifluoromethane and the carbon tetra- fluoride were Chemically Pure grade obtained from the Matheson Company, with minimum stated purities of 99 mole percent. The trichlorofluoromethane and the chlorotri- fluoromethane were obtained from PCR Research Chemicals with stated purities in the range 97 to 99 percent. The carbon tetrachloride used was Mallinckrodt spectrophoto- metric grade. Mass spectra of each compound were obtained using electron energies of 20 and 70 eV and a photoionization energy of 21.2 eV; none gave an indication of impurities. Nor was there spectroscopic evidence for impurities. A. Infrared Spectrosc0py The liquid samples of CCl4 and CC13F were frozen, pumped and thawed three times before being deposited from the vapor. The gases, CF2C12, CF3C1 and CF4, were deposited directly from the vapor without further treatment onto a polished copper plate attached to the cold tip of an Air Products Model CS-202 or CSA-202 cryogenic refrigerator. Thermal contact between the cold tip and the copper plate was made with indium foil. The copper cold plate made a 45 degree angle to the optical axis of the interferometer. The far infrared radiation, after passing through the sample, is reflected by the copper plate and, after an additional pass through the sample, is collected by a liquid-helium cooled Infrared Laboratories composite Ge-bolometer located 90 degrees to the optical axis of the interferometer. Temperatures were measured with a gold (0.07% iron) versus chromel thermocouple, or with a hydrogen vapor bulb thermometer. Since the temperature sensors were attached to the cold tip of the refrigerator, the tempera- ture of the sample on the cold plate was probably slightly higher than indicated by the sensors. The temperature of the samples was controlled with an Air Products Model 3610 Temperature Controller which employed a platinum resistor temperature sensor. The measured temperature is believed accurate within a few degrees, with temperature control better than one degree. Samples were deposited at the lowest temperature that the refrigerator could achieve, 10-15 K. The temperature of the copper plate was then slowly increased until the sample began to sublime to colder regions in the cryostat. The cell is not separately enclosed and the refrigerator vacuum shroud was pumped during this annealing process. Inadequately annealed samples displayed broader and less intense absorptions than those shown by well annealed samples. Solid carbon tetrachloride was annealed at 110-145 K. At the higher annealing temperatures (130-145 K) the sample sublimed, forcing short annealing times, on the order of fifteen to thirty minutes. The sublimation precluded the observation of the higher temperature solid phases of carbon tetrachloride. Use of the lower annealing temperatures (110-120 K) reduced sublimation but required 12 to 24 hours for proper annealing. Carbon tetrachloride is a weak absorber and requires samples thicknesses of one to two millimeters to give reasonable spectra. The trichlorofluoromethane samples were annealed at 130 K for thirty minutes. At this temperature there was little sublimation although this annealing temperature is a much higher proportion of its melting point (162 K) in comparison to CCl4 (melting point 250 K). The CC13F samples were about 0.2 mm thick. Samples of dichlorodifluoromethane (melting point 115 K) were annealed at 105 K for fifteen minutes. Some sublimation was noted at this temperature. These samples were about 0.1 mm thick. Chlorotrifluoromethane (melting point 93 K) was annealed at 79 K for fifteen minutes. This sample was also about 0.1 mm thick. Solid carbon tetrafluoride (melting point 89 K) was annealed at 55 K. In order to obtain far infrared spectra of this non-polar molecule of quality similar to those of the chlorofluoromethanes, samples about one millimeter thick were required. After annealing, the samples were recooled to 10 to 15 K in about twenty minutes by turning off the temperature controller. Slow cooling under temperature control was also used. Identical spectra were obtained at both cooling rates. Vacuum shroud windows were dimpled 1/8, 1/16, or 1/32 inch thick polyethylene, each of which gave a characteristic fringe pattern to the background. To adequately ratio out these fringes it was necessary to take background spectra without having moved the refrigerator. The far infrared spectra were obtained with a Beckman-RIIC FS-720 Fourier Spectrometer with FS-820 stepper motor mirror drive, and an Infrared Laboratories Ge bolometer. Data acquisition and mirror stepper motor movement were controlled by a PDP-8/M or -8/I computer. To obtain typical lattice mode interferograms the 100 gauge (25 micron) Mylar beamsplitter and a Beckman number two far-infrared filter were employed. The first beam— 1 splitter fringe minimum is at 120 cm- and the filter passes frequencies less than 120 cm-1. This allows the use of a 40 micron step size, with its corresponding "folding frequency" of 125 cm-1. A typical double-sided scan contained 715 or 1001 steps, giving resolution of 0.7 or 0.5 cm-l, respectively. Interferograms were cosine bell weighted, zero-filled to 4096 points and transformed. This gives the transformed points a spacing of 0.061035 cm-l. In this work the reported far-infrared frequencies of well-resolved bands are believed to have an accuracy of 0.5 cm—l. For overlapped and broad bands the accuracy of the reported frequencies is between one and two wavenumbers. A 25 gauge (6 micron) beamsplitter was used to observe v4 and v5 of CC12F2, and 02 of CC14. The internal modes of CC12F2 were observed with 1334 15 micron steps, giving a 333 cm.1 folding frequency and one cm"l resolution. No filters were used in these experiments. B. Raman Spectroscopy The Raman spectra of CF4, CF3C1, and CFZCl2 were obtained with an enclosed vapor deposition spray-on cell described elsewhere [13]. The Raman spectra of CCl4 and CFCl were obtained with the open cell described in the 3 previous section; glass was substituted for the polyethy- lene windows. The Raman scattering was observed at 90° to the incident excitation with both cells. Use of the enclosed Raman cell allowed higher annealing temperatures than in the infrared cell or, if desired, sample liquefaction. Such liquefaction was used to obtain glassy samples, while vapor-deposited samples were powdery. Spectra of both sample morphologies were obtained for CF4 and CF2C12. In neither case did the different morphologies exhibit significant spectroscopic differences. Only the powdery form of the other compounds was investigated. Temperature control of the Raman cell for annealing purposes was achieved by the use of a germanium or platinum resistor temperature sensor in a circuit which controlled the current to a zener diode functioning as a heater. All Raman spectra were obtained at the lowest possible sample block temperature, 20 to 25 degrees. This temperature is higher than that achieved with the infrared cell because of the absence of radiation shielding and the larger sample cell mass in the Raman cryostat. The Raman spectrometer consisted of a Jarrell-Ash model 25-100 double Czerny-Turner monochromator with a thermoelectrically-cooled RCA C31034A photomultipler tube detector. The 514.5 nm line of a Spectra-Physics model 164 argon ion laser was used as the excitation source for all the Raman experiments. The resolution of the Raman spectra varied from 0.3 cm-1 for strong fundamentals, to 1 cm.1 for weak bands. The spectrometer was calibrated with the argon ion emission lines present in the laser beam [35]. In this work the accuracy of the Raman bands is believed to be 1 cm-1 except in the case of broad or weak bands, where the accuracy may be only 3 cm-l. 10 CHAPTER 3 INTERFEROMETRY [36-38] A. Optical System The Optical system of a typical Fourier spectrometer, shown in Fig. l, is that of a Michelson interferometer. The spectra range of the interferometer is determined by the source, the beamsplitter, the moving mirror drive, and the detector. The source used in the RIIC-Beckman F5720 is a 200 watt mercury vapor lamp with a dimpled quartz envelope. It is a broad band source with output from the far-infrared to the ultraviolet. Beamsplitters are polyethylene terepthalate (Mylar, Melinex) sheets of varying thickness. The useful spectral range of these beamsplitters is determined by their thickness. The thinnest beamsplitter, (25 gauge, 6 microns) is useful over the widest range, 1 to 600 cm-1, with efficiency much reduced from 25 cm‘ at the extrema. Thicker beamsplitters are more efficient at lower frequenices, but they "black out" certain spectral regions due to the interferences caused by internally reflected radiation. The moving mirror is on a stepper-motor driven micrometer which has a minimum step size of 2.5 microns and a maximum total excursion of 5 cm. 11 SOURCE l BEAMSPLITTER MOVABLE )— MmROR COMPENSATOR C: ‘ a ‘—" ' 9 DETECTOR 1.) L j FIXED MIRROR Figure l. Fourier spectrometer 12 This results in a maximum possible alias-free spectral range of 1000 cm"1 and a maximum possible resolution of 0.2 cm-l, respectively. The interferometric experiment consists of recording the variation in intensity at the detector of the trans- mitted far infrared radiation as a function of moving mirror displacement. The resulting interferogram arises from the variation in optical retardation experienced by radiation traversing the arm of the interferometer containing the moving mirror, relative to that of the radiation traversing the fixed mirror arm of the interferometer. The relationship of the observed interferogram to the desired spectrum can be demonstrated by considering the intensity due to the superposition of two monochromatic waves traveling in the same direction, but with a relative phase shift. This intensity is given by [39]: I=I+I 1 +2i/IlI2 cosy (1) 2 where y is the phase difference between the two waves. Considering 50:50 transmission:reflection by the beamsplit- ter, and I1 and 12 to be the radiation from the two arms of the interferometer we have, I1 = 12 = 15/4 (2) The factor of four arises because the source intensity, 13 IS, is beamsplit twice before it passes to the detector. Substituting the above into Equation (1) we obtain Id = IS(l-+c05'y)/2 , (3) where Id is the intensity reaching the detector. The phase difference between the radiation from each arm of the interferometer is a function of the wavenumber of the monochromatic radiation and the displacement of the moving mirror relative to the fixed mirror, and is given by |( Hz. (4) v = 2n; mono xfixed-xmoving The factor of 2 at the end of the previous equation, is a result of the arrangement of the two arms of the inter- ferometer. The radiation travels the distance from the beamsplitter to the moving mirror twice, once to the mirror and once after reflection from the mirror. This gives the optical path difference as twice the physical displacement. Defining the optical path difference 62H )l2. (5) Xfixed-xmoving and substituting for the phase difference in Eq. (3) we obtain I = IS[1+ cos (2n 1m(moon/2 . (6) d As the moving mirror passes the points corresponding to optical path differences of 0, 1/2, 31/2, 2x,..., successive maxima of intensity, IS, and minima of intensity, l4 zero, appear at the detector. Considering the intensity due to polychromatic radiation as the limit of the sum over all the wavenumbers and noting that negative wavenumbers have zero intensity gives Id(6) = s +321I 15(3) cos(21r'\76)d§)_ . (7) —m Dispersion in the beamsplitter and imperfect optical alignment contribute sine components to the interferogram. Including these into the above expression we obtain Id(5) = __s_2__+_22[_ 1.0018(3) exp (21TiU-5)dU . (8) Incorporation of the constant offset, the first term on the right in Equation (8), and the scale factor in front of the integral into the interferogram gives 0 w _ _ _ Id(6) = J 15(0) exp (21rivé)dv . (9) Recognition that this is one-half of a Fourier transform pair gives, co 15(5) = J 1:3(6) exp (-21Ti-\-)-O)d(5 . (10) —m This is the desired spectrum, IS(U), as a function of the observed interferogram, Ié(6). Since we sample the interferogram at discrete points, the discrete Fourier transform is used to acutally obtain the spectrum. This transform is given by [40]: 15 1(5.) = lnElIUS )exp (~2wi3.6 ) , (11) J nk=0 k 3 k with 5k = k)4 :5 {"3 1 —1 (D '7 4.1 .5 {h J - 'l 0‘ 'I'rFI'I'I‘TIPI'I'I'T'I'I'I'I'I'I'I'I'I‘I'I'I'I‘ ° 50 7O 90 110 130 150 170 190 210 2.30 250 270 290 310 3.30 350 Wovenumber (cm-1) Figure 2. Far-infrared spectra of solid CF C1 2 2 with and without Cs I filter. 19 CF2c12 folding frequency 166.667 cm”, Wavelength (microns) “V N §8 § 8 1 It] I 1 1 l ... L- >\ .4: (I) C - _ (D 4...: .5 O I I I I I I I l I I I l I I I l I l I I I l I l I I I I I I I 10 .30 50 70 90 1 10 130 150 170 Wovenumber (cm-1) Figure 3. Far-infrared spectra of solid CF2C12 with Cs I filter and aliased absorptions. 20 instructions has been described elsewhere [42]. The data acquisition section of the interface inverts the output of the FS-200 amplifier demodulator with an Analog Devices AD521K instrumentation amplifier. The output of the instrumentation amplifier is monitored by an Analog Devices ADC-QU 12 bit A to D converter. Three functions are provided by the interface: 1) the initiation of a conversion, 2) the reading of the digital value present at the accumulator input, and 3) the activation of the skip line upon the completion of a conversion. These functions are provided by the PDP8 assembly language instructions (D836, IOPl), (D837, IOP4) and (D837, IOPl), respectively. The circuitry necessary to implement these instructions has also be described elsewhere [42]. Functions which must be performed manually include setting the F8-200 amplifier gain, time constant and offset. One must also manually position the mirror at approximately zero path difference to begin the scan. There are two PDP8 operating programs for the interferometer, FORRUN and FTIR. They differ only in that FTIR allows multiple scans while FORRUN does not. Either program calculates the parameters necessary to define a scan from user-supplied values. The output of either program is a file containing the interferogram or interferograms. The programs require five user-defined inputs: the desired resolution, the maximum frequency of interest, the settling time, the integration time, and the output data file 21 specification. The desired resolution is used to calculate the total mirror excursion and the number of five micron (optical path difference) steps needed to cover that distance. The maximum frequency of interest is used to determine the stepsize and the number of five micron increments needed to make one step. The stepsize obtained is that which gives the minimum folding frequency greater than the maximum frequency of interest. The settling time is the waiting period after a mirror step has been taken before data conversion is begun. The integration time is the period during which the analog to digital converter is read and averaged. As currently written, the program reads the A to D every five milliseconds. This relatively long time is due to averaging and storing values as floating point numbers. The five milliseconds is, however, a small fraction of the shortest instrument time constant of 125 milliseconds [42]. A non-zero filled, non—apodized 512 point Fourier transform called SHORT has been written for judging sample quality. Larger transforms and final graphics are performed on the RSX-llM system available in the Chemistry Computer Graphics Facility. C. Data Processing Zero-filling has been chosen as the interpolation method because of its simplicity. One appends zero to the data file until the file length has been extended by a factor of 2m, m==0,l,2,..., where m is called the number 22 of zero fillings. After zero filling the discrete spectrum is defined at the (2m)(n) frequencies — = :1 vi 2mYn>m3 mnemoo pmuwcommcc may no EH0mmcmua .v munmfim ATkBIQQdB: a". 0.. co. oo on on oPphbbb-PFh—bnhpL—r-pP—Lb-ppPLpL—yhlPhpphh-h—thban-P—o V A LI I ‘ n d111<4qdq—fi-q-‘duu-‘dqfi-«qd-uu—chuud-uq—d-d-u-d15 0 8w 0:. 8' on on Oh gt... Scio n—lx 2. OWE 8...? 02 $650-$203: QOEmE .m..OI>> ._.OZ 25 .v mwsmwm ca mm ccomoa .uom mumc Umuwpomm “Madonnauu map mo Enemmcmna A723 269... ON— 0: 8p 8 00 Oh oFPbbrhpphbth-pph.pL—h-pphpnppr-bhprLFb-bn—p—nbb— 4.. Pill I ON— qfidfidddififid«dq-and‘u—dI—«dddfidW-ddaq-ddqd—4]_dddifiddfi o: 8.. am an nplx Oh 8.3—... Can? ”2.5231... mafia-.3! QOEMQ m30I>> HOZ on o .m mwsmflm 26 .v munmwm Ga mm ccmmmq .oaflm mumo coungwo3 dawn mcflmoo uflamm can «0 Enemmcmna A723 6689: ON— 0: cop co 8 on o .- b p p p - F. p h — — L F n - b b p F — u r h n — b]_ (F h —L p h p h b p b p — h N LP F — hLi — r— O p 1] I p N J] I N n m. n C 1] l t n H d d‘ a: .- 1 u q u — q 4 d d — u u a q — u q u — _ d a u d — — d d .1 — A d d A — d u d n d a q A 4 n ON— 0: 00— cm on Oh A753 98:53:: 3 .. 85.3 205.22... mE Sou OZEMQS. njmm mZ_mOO ._.3n_m .m whomflm 27 O 20 4O 60 80 100 120 I40 5 I I I I I I4 L I I I I I I I I I I I I I I I I I J I - + 4 _ CI . CI 5 w% — a . [MN (I) w c +- E 2 i 1 *— l 1 —‘W P I O I ' I ' I ' I ' I I I ' I ‘ I j I ' I ' I r I ' I ' I ' 0 20 4O 60 80 100 720 140 Wovenumber (cm—1) Figure 7. Upper curve: Ratio of unapodized sample to unapodized background. Lower curve: Ratio of apodized sample to apodized background. 28 b: intensity. One therefore expects the ratios to be very noisy in spectral regions where there is little intensity in the background. The regions 0-20 cm”1 and 110-133 cm-1 have very little intensity in either the background or the sample spectrum, and it is in those regions that the noise is greatest. Apodization reduces the spurious oscillations in the ratio but it does not change the wavelength dependence of those oscillations. Split cosine bell apodization is used for all spectra, because it gives low height side- lobes. The Fourier transformation routine uses two real arrays, one for the cosine and one for the sine parts of the transform. These are referred to as the real and imaginary parts respectively, and must be dimensioned a power of two. On entry, the imaginary part contains zeros and the real part contains the double-sided weighted and zero-filled interferograms. The interferogram is loaded into the real part of the array as follows. The interfero- gram is divided at the maximum data point, the first half is shifted to the end of the array, the second half is shifted to the beginning of the array and the middle is filled with zeros to make the desired power of two. If the point where the optical path difference between the two arms of the interferometer is zero, called ZPD, fell on a sampled point, the above arrangement would put all the spectral information into the cosine part of the transform. Since one cannot ensure that ZPD is sampled, the sine part 29 of the transform contains some spectral information. By recording [(real part)2-+(imaginary)2] it becomes unnecessary to locate ZPD. The above splitting and shifting of the interferogram is not required. The zeros need only be appended to either end of the interferogram. The result is to shift more of the spectrum into the imaginary part of the transformed interferogram. The above splitting and shifting was done to ease the transition from double- sided to single-sided cosine transforms, if this would become desirable. Single-sided transforms are not planned, but the following would be necessary for this change: 1) change the instrument control program to step the one-sided data file; 2) modify the data handling program to accurately calculate the position of ZPD, and hence phases for the data file; 3) modify the transforming routine to use the phase information; 4) move the interferometer fixed mirror so that ZPD falls near an end of the moving mirror travel. If this were implemented, the theoretical resolution would 1 1 be increased from 0.2 cm- to 0.1 cm- . D. Detector Comparison: Bolometer versus Deuterated Triglycine Sulfate The relative sensitivities of the DTGS and Ge bolometer detectors may be roughly determined from the analog-to- digital conversion (ADC) values of the signal from the instrument under identical conditions, after accounting for 30 the differences in gain used for each detector. The comparison was made using the 25 gauge beamsplitter, no sample or filters, and a time constant setting of 0.5 sec (its minimum). The bolometer value of 2330 ADC units was obtained with the bolometer preamp gain at 200 (its minimum), the FS-200 amp-demodulator gain setting at 10 (also its minimum), and the output zero shifted down 3 volts. The DTGS value of 1750 ADC units was obtained with its preamp gain at 600, the F8-200 amp-demodulator at 25, and no zero shift. A change in the F8-200 gain from 10 to 25 was determined to correspond to a 6-fold increase in gain. A zero shift of 3 volts down corresponds to a loss of 1230 ADC units for a 0-10 V 12 bit A to D converter. That the bolometer cuts off frequencies higher than 375 cm-1 while the DTGS remains sensitive to frequencies throughout the mid-IR results in an intensity increase of about a factor of 2 reaching the DTGS detector. These factors may be combined to show that the bolometer is roughly 70 times more sensitive than the DTGS detector. The standard deviation of the measured value for the bolometer at this level is about 7 ADC units. The standard deviation of the measured value for the DTGS detector is about 40 ADC units. Thus the improvement in signal-to-noise with the use of the bolometer is about 400. Although this is less than the factor of 1000 one would expect from a ratio of the noise equivalent powers of the 31 two detectors, the bolometer is clearly the far more sensitive detector. Figure 8, the integration time (averaging time) versus the average ADC value for each of the detectors, shows that in neither system is the noise reduced by increasing averaging times. That the standard deviations of the measurements increase with measurement time is probably a consequence of DTGS preamplifier instabilities and source instabilities. 32 BOLOM 2375 ETER 2350 2325 ADC Value F£F4F49491EH 2300 L I ‘ I ‘ I ' I ’ I ' I ' l ' I ' I ' 7_'_| 0.0 1.0 2.0 3.0 4.0 5.0 Averaging Time (seconds) 2275 DTGS 1875 1850 __ 1825 -' 3 O 1800 0 3 1775 0 1750 1725 ' 1700 g 1675 53 1650 ' § E: O 1625 ADC Value 3.1 1600 I I I I I l I I I l I I I I I l ITIfi—l 00 L0 20 30 4L) 50 Averaging Time (seconds) Pique 8. DeuKXDr.mm Hmmm.mmai .4. .4, mmhh.am Hmmm.mmal oasn.mm momm.mmal omvh.mm momm.mmal voom.mm oamm.mmal aavm.mm bmmN.HmHI mmoa.nm mowm.mval Aooav mmnm.ohai ooa mmnm.onal mmmm.mm mmnm.o>al mmmm.mm vmnm.o>al mmam.mm nonm.onal mamm.mm mmvm.ohai mam>.mm mmav.o>al Hmmm.mm mmmh.mmai mmmm.hm avam.mmal ovmm.hm mmva.mvai uHEHA AmHOE\be mo monocm unmoumm mowuuoq wam.ma mmom meme mmhm mem mmam vmma mmoa bow mvv «Hm mma hm womv mmhm Hmmm mmam vmmH mmoa hon ovv «Hm hm mufimm mo Esm ooeuuoa uuouflc mo mocomuo>aoo new name mEDm mofluumq uneven omvm em. «mad mm. wmva em. HNHH me. man vv. mhm mm. Nmm vm. omm mm. «ea em. mm ma. makes mo .02 AHIcou How puma .m mqmda 49 w>u0m 0 08¢me O OPHNW O 0>Hmm O .88.888u8 8 8 8888.878 8 o <888.88u8 an <888.8~u8 8 m 888 88.8 8.888 8888.8 8.888 8888.8 8.888 8888.8 II. 8.88 888 88.8 8.888 8888.8 8.888 8888.8 8.888 8888.8 .1. 8.8 888 88.8 8.888 8888.8 8.888 8888.8 8.888 8888.8 (I. 8.8 888 88.8 8.88 8888.8 8.88 8888.8 8.88 8888.8 .1) 8.8 88 88.8 8.88 8888.8 8.88 8888.8 8.88 8888.8 .1: 8.8 88 88.8 8.88 8888.8 8.88 8888.8 8.88 8888.8 :1. 8.8 18888 88.8 18888 8888.8 .8888 8888.8 i888. 8888.8 88. 8.8 888 88.8 888 8888.8 888 8888.8 888 8888.8 88. 8.8 88 88.8 8.88 8888.8 8.88 8888.8 8.88 8888.8 88. 8.8 888 88.8 8.88 8888.8 8.88 8888.8 8.88 8888.8 88. 8.8 888 88.8 8.88 8888.8 8.88 8888.8 8.88 8888.8 88. 8.8 88 88.8 8.88 8888.8 8.88 8888.8 8.88 8888.8 88. 8.8 88688 188008888 88588 a 88588 8 88588 < 88288 88688 0:08 0860 5:08 0800 ocoa 0860 0:08 0800 008uumq 0080008 MG ml 0< D... DC DI ”Q ”I . . 888 x 88 8808 888 x I08 300 888 x In8 when 888 x 88 308. 8308880688 800.888 .888 00w umuoEmumm 8c8uc0uomv 85m 0080008 ecu mo mnumcmq m50880> um amumcm 0080008 08000m E9E8a82 0>08£0< 00 >88m00002 m .0 .n .o muouosmumm 0080008 HH ounce 8800 on» G8 000020 .8 mqmme of the variation in the lattice parameters provided by the longer limit sums. The results of a test of the effect of the length of the direct lattice summation limit upon the equilibrium orientation and position of the four independent CCl4 molecules in an asymmetric unit is given in Table 5. The choice of 5.0.A as the summation limit again gives over 90% of the variation given by the longer length summation limit. The large number of parameters to be varied, 24, gave calculation times so long as to prohibit inclusion of the reciprocal sum in this test. The contribution to the energy of the reciprocal space Coulobmic sum when a partial charge of -0.3 electrons is given to each chlorine atom and +1.2 electrons to the carbon atom is seen in Table 6 to be about 28 kJ /mole. Neglect of the Coulombic reciprocal space sum for this structure would overestimate the static lattice energy by this amount. The contribution of the Coulombic term to the static lattice energy can be estimated only when all terms are included and summed to convergence. For CCl4 in solid phase II and potential set II with a charge of -0.3 electrons on the chlorines the contribution of the Coulombic terms to the static lattice energy is estimated to be 179.9713-l70.8739==9.0974 kJ/mole, or about 5% of the static lattice energy calculated without charges. 51 .pm8m800mwc5 m8 00m8m mmxmu cofiunuou mac» 308:3 uaonu 088M 0:» no c08u0088oa .008w800omco m8 cowuuancduu man» no co8u0muwoo .mm 00cmumumu :8 van» m30880u pang uwuumesxmm 030 G8 muaauodoa 8800 0:» mo mcwuunaoZd vmm.N mwm.o mmm..~ hom.o who.m mHm.o hOH.¢ mawco m.o ham.N th.o Newfm mom.o mmo.N MHm.o VHo.v Hmooc m.m Hmbom hmmoo vmm.H VNm.o ovm.H mhv.o me.m Nfimno o.m vmv.o moooo mbw.o Noo.o Hhv.o Noo.o MH0.0 VOOoo v.v . C s: 8.8 :8 C s: C .5 .5 8.888 «o 8.888 8-8800 80 8.880 «.8800 no 8-8800 8880 88 8.88800 8858.8 mmmE mo uo mmme mmmE NO NO mmoE mmmE no mo mmmE mmmE no No name cowumasdm umucm0 030 Mo 8088000 0008800 083 Mo 8082.800 kucmu 083 mo umucmu 8008800 083 no 8008800 008093 cmnoucu on» we :oaoucu may uo cmsounu 0:» mo nmaounu on» no 000880 muxo cowunacnua maxm coauMHmcmua m8xu cowumamcmue m8xm caduudmumua 5... usona 8 cu £528 8 cm usonm 8 ca 88688 8 u ceauuuom 0 88080688088 0 cOHumuom 0 880800038 .8_mo_ HH umm umuwemumm Hawucuuomv 55m 008uumq on» NO mnumcmq mooaum> 00 00950800 .xmumcm 0080008 uwumum Essacwz 0>0820¢ ou mummmmuwz .HH madam vHHOm mo 88uwcn 08uumEEhm¢ may :8 M0890waoz 8800 850m 0:» «o z0wm mo sawumuc08uo 0cm cOHuwmom umaoomaoz c8 mmmcnnu 029 .m 3&8 52 TABLE 6. Test for Convergence of the Lattice Energy Sums for the CCl4 Phase II Structure (Potential Parameter Set II [69]), with a Partial Charge of -0.3 Electrons on the Chlorine Atoms. Direct Lattice Reciprocal Lattice Static Lattice Summation Limit Summation Unit Energy (A) (A-l) (kJ/mol) 4.4 0.19 -l74.0023 5.5 0.24 -194.2392 6.6 0.29 -l78.0300 7.7 0.34 -l79.7779 9.9 0.44 -179.9684 11.0 0.49 -l79.9708 12.1 0.54 -179.9712 13.2 0.59 -179.9713 14.3 0.64 -l79.9713 4.4 -— -175.9812 5.0 -— -206.7162 5.5 - -220.1129 6 6 - -207.7690 7.7 - -208.9874 8.8 —- -208.9960 919 —— -208.9940 20.0 - -208.9940 53 The calculated static lattice energy is related to the heat of sublimation as follows: E = -(AHO K + 3 sub E zero point) ° 0 K sub based on available [25,60,611 An estimate of AH thermochemical data is 74 kJ/mole. The estimate is outlined o . =‘1 z . . in Table 7. It is assumed that Ezero point /§j_hlvl. where Vi is the observed frequency of the ith external optical mode of CCl4 in solid phase II. If one takes 1 36 cm- for the average frequency of the 95 external optical modes, one obtains E =183 kJ/mole. zero point Thus BOSE-75 kJ/mole and this value, along with the shifts in molecular orientation and position, is used to judge the acceptibility of various different potential parameter 89125. 2. Description of Potential Parameter Sets Used on the Packing Analysis The parameter sets tested in this work are listed in Table 8. Potential I is that of Reynolds, Kjems, and White as reported by Bates and Busing [57]. 'It was obtained for use in dynamical calculations on several chlorinated benzenes. Potential II is Bates and Busing's potential A, which was obtained by a fit to the structural and thermodynamic prOperties of hexachlorobenzene. Bates and Busing compare their potential to two others: potential A of Bonadeo and D'Alessio [54], our potential X, and that 54 0 TABLE 7. Estimation of AH0 K of CCl sub 4 225.7 + c (solid II)dT = 3682 110 cal/mol 03.74] 3.5 P + AH(II to Ib) = 4631.120 [25] 250. 53 + p(solid Ib)dT = 704110 [25] 225. 7 + AH(Ib to liquid) = 2562 16 [25] 374. 69 + Cp (liquid)dT = 4000 i150 a) 250. 53 AH(liquid to vapor) = 7630 1200 a) 374. 69 + 0p (ideal gas)dT = 5460:1200 [84] 3. 5 = 17.7 10.6 kcal/mol = 74.1 82.5 kJ/mol a) Handbook of Chemistry and Physics, 53rd edition, Robert C. Weast, ed., The Chemical Rubber Co., Cleveland, Ohio. 55 <..U 0 8| 8 82558882 8803 08 88808888 88 8858.8 8 8.6.0880 8.8.8.88. . u 83> 0805\nd. 5808 8803: 8 888 8058.8- .8 :8 0800080800 8080:0000 8000.: 880.8: 008.81 08N.8| 0m8.ml 0mv.m| 888.8: 000.8: 800.8: M8 hm0.m 0N0.m >m0.m 000.m m0>.m 080.0 000.0 000.m 000.0 0 00.0 808.m mm.m 880.8 8m.m 8m.m 000.0 8m.m m0.m U 80:80 000000 0000888 000000 0000080 0000088 0000088 0000808 000000 0000088 m 0800 0000 8808 mmm0 000m8 08008 0880 80mm mm00 d 0000.: 088.: 888.0: mm0.0| 08N.0I 000.0: 000.0: 000.0: 000.0: Ow 000.8 008.0 000.m ~80.m 000.8 080.0 000.m 080.m 800.m 00 800.m 808.0 N0.m 880.8 mm0.m 000.0 m0.m 0m.m «0.0 o 8010 000M8m 0088N0 0000mm 000M008 0088mm 000808 000000 000000 008mmh m mmm0 080m 080m 00mm NO0N 080m: 000m 0088 0888 m 8N8.| >~8.I 0N8.0I 0m8.0| 000.0: 00m.0| 00m.0| 00m.0| 00m.0| Mw 088.0 m00.m 000.0 80m.m ~00.m ~00.m ~00.m ~00.m ~00.m 0 M80.~ 000.m mm.m 880.8 00.m 00.0 00.m 00.0 00.m U 010 00mmmm 000800 008088 000800 00008m 00008m 00008m 00008m 00008m m 00088 0008 0888 N008 88mm 0000 000m 00mm 00mm d xH HHH> HH> H>. .> >8 HHH HH H .8800 :8 88 00050 08800 00 m8mx8mc< 028x000 0:0 28 0003 0000 8080:0000 .m mamflfi 56 of Reynolds,Kjems, and White, our potential I. They find that these two potentials do not give as good a fit to the observed hexachlorobenzene structure as their potential A. They also find that for the case of optimization with partial charges on the atoms, their potential B which places a charge of -0.106 e on the chlorines, much improves the fit to the observed structure although contributing only 13% to the total energy. Potential II is one for argon suggested by Hill [62]. This potential has been used by Gavezzotti [55] to calculate the lattice energies of some chlorinated benzenes. Potentials IV, V and X are potentials B, C and A of Bonadeo and D'Alessio [54]. They were obtained from a least squares fit to the heat of sublimation and lattice frequencies of p-dichlorobenzene. These authors find that their potential A (our potential X) gave the best fit of the three potentials to the observed properties of p-dichlorobenzene. Several Lennard-Jones 6—12 atom-atom potentials have been suggested for CCl4 in the liquid and plastic solid phases [56]. Because the packing programs available to us require potentials of the exp-6 type, these literature potentials had to be fitted to the appropriate mathematical form before they could be tested. Potentials VI, VII and VIII were obtained by non- linear least squares fits, over three different ranges of the radius, of the exp-6 parameters to potential I of McDonald, Bounds and Klein [56]. Potential VI was obtained 57 with the fit between 3.0 and 4.954A for all three interaction types. Potential VII was fit between 3.5 and 4.95.4 for all three interaction types. The C-C interaction of potential VIII was fit between 3.2 and 4.95.8, and the C-Cl and Cl-Cl interactions were fit between 3.3 and 4.95 A. Potential IX is a fit to potential II of McDonald et a1. [56], with the C-C interaction fit between 4.6 and 6.55.A, the C—Cl interaction fit between 4.0 and 6.55.A, and the Cl-Cl interaction fit from 3.5 to 6.9514. Comparisons of McDonald's Lennard-Jones type potential I to the fitted Buckingham potential VI for the three interaction types are shown in Figures 1, 2, and 3 of appendix one. Compari- sOns of McDonald's L-J type potential II to the fitted Buckingham potential IX for the three interaction types are shown in Figures 4, 5, and 6 of appendix one. McDonald, Bounds and Klein [56] conclude, on the basis of molecular dynamics calculations on liquid and cubic Ia CC14, that their less anisotopic potential II-—our potential IX—-is the preferred potential. They also find that the inclusion of a partial charge of -0.3 electrons on the chlorines has a negligible effect on the structure of the liquid. Righini and Klein [63] used the McDonald potentials to study the lattice dynamics of the high pressure solid phase III of CC14. They find that the McDonald potential II (our potential IX) gives the best agreement with the observed Raman-active lattice frequencies of that phase. They also find that the inclusion of partial charges of the 58 above magnitude contributes only approximately 7% of the binding energy and has no significant effect on the calculated lattice spectrum. 3. Results and Discussion A comparison of the results of the packing analysis calculation using the ten different sets of interatomic parameters is given in Table 9. None of the ten potential parameter sets accurately reporduces the estimate of the static lattice energy of phase II CCl4. The calculated static energies vary from 2 to greater than 4 times the -75 kJ/mole estimate delineated in Table 7. Potentials I, IV, V and X have the deepest Cl-Cl potential well depths and give the poorer results. Potential IX, the fit to the large carbon model (potential II) of McDonald, gives significantly larger shifts in the molecule's positions and orientations than the other potential sets. In particular the use of potentials VI, VII, and VIII, the simulation of the more anisotropic potential I of McDonald, give a better fit to the structure of CCl4 in phase II than the use of potential IX, the simulation of the less anistropic potential II of McDonald. In general, the shifts in molecular positions and orientations are surprisingly uniform for most of the potentials tested. The uniformity and magnitude of the shifts suggest that no simple potential of exp-6 form can adequately model the forces which give rise to such a 59 .mmmm uxw: may :0 mm muocuoom m 000- 808i 008i 008i 808: 080- 000i 008- 008- 0801 m m- N- 8- m- m- 8- m- N- 8- 8- 808 A00 080.0 000.0 880.0 008.0 080.0 000.0 080.0 000.0 000.0 008.0 co8umuou mmmnm>¢ A¢ A.8.o.m 88588 888882580 wo8uumq uowu8ov .m8m8ucmuom ucm8w888o awe mnu How wmnwcm wo8uumq 08umum 5:58:82 w>m8£o¢ ou mumwmmomz mco8umuom can mco8uw8mcmne nmum8so8mo on» 80 D8CD 08Humafixmd gnu c8 mm8oom8oz 8800 ucmpcwmmpCH 8508 may Hm>o mmmnm>d one 000- 808- 808i 808- 008i 0001 000- 008| 0081 080: mm A882\8xv m8- N- m- m- 8- NH- 8- 8- m- m- 808 0.8 00.0 00.0 80.0 80.0 08.8 80.0 00.0 00.0 08.0 808 000.0 000.0I 080.0I 880.01 080.0: 080.0I 880.0I 080.0: 080.0u 880.0: Sumoov< 800.0 000.0 000.0 000.0 000.0 808.0 008.0 808.0 000.0 000.0 8.08 000.0 008.0 00.0 80.0 00.0 00.0 08.0 88.0 00.0 00.0 88n< 000.0 808.0 088.0 800.0 808.0 088.0 080.0 080.0 088.0 080.0 8888 8 0 x8 888> 88> 8> > >8 888 88 8 8888:8808 8.8.0.0 u8E88 co8umEEom m08uumq pomu8ov .wuwm Hmumenmm 8m8ucmuom pcw8m888o owe £883 omusmeoo woumcm w08uumq 08umum 5588c8z m>m8£o¢ on 8800mmomz 0 .o .n .m mucmumcoo mo8uumq c8 mwmcmsu was .0 mqm¢8 60 .8800 0:0 80 000800000880 000 000808000 :8 00800880> 0:0 808 888 0:0 0800050800 0080008 :8 00800880> 0:0 808 8888 003 08888 008006500 028 .00:0080>000 000 80 008008080000 0903083 0008 0803 0008008850 .8800 08003 800 00> 0:0 808 0808 008080800 0008 050 8808000 00: 080 x 808000000 005000m8 61 large unit cell containing such a large number of molecules in the asymmetric unit. It would, however, be necessary to Optimize the potential parameters to fit this particular phase of CCl4 in order to demonstrate this failure. The calculated shifts in positions and orientations of the four independent CCl4 molecules in the asymmetric unit in the phase II solid, for various values of the Cl partial charge, are shown in Table 10. The fit to the phase II structure is very much improved if partial charges of the magnitude suggested by McDonald et al. [56] are added to the potential. This is similar to the case of hexachloro- benzene, where inclusion of partial charges had a small effect on the energy but provided much improved fit to the observed structure. Although none of the potential parameter sets tested gives both the experimental static lattice energy and the observed structure, the parameter sets of Bates and Busing (potential II), McDonald's potential I (Potential VI, VII and VIII), and somewhat surprisingly the potential obtained [55] from Hill's [62] data for argon represent a signifi- cant improvement of the potentials having lower energy minimums, such as potentials IV, V and X. The less anisotr0pic potential of McDonald (our potential IX), although giving reasonable fits to the less-ordered phases of CC14, fails badly when transferred to CCl4 phase II. This indicates that it may be very difficult to otain 62 .0 08808 0008 008.0 880.0 000.0 008.0 008.0 888.0 000.0 000.0 880.0 000.0 008.0 008.0 000.0 800.0 800.8 800.8 000.8 800.0 008.8 800.8 800.0 000.0 800.0 880.0 000.0 800.0 000.0 000.0 800.8 000.8 080.8 000.0 000.8 000.8 008.0 000.0 808.0 080.0 088.0 088.0 000.0 008.0 800.0 008.0 000.0 800.0 000.0 800.0 000.8 080.0 080.0 800.0 888.0 000.0 080.0 808.0 0.0I 0 0.0! 0.0I 0 0.0I 0.0I 88>. 8H> 888 888 888 88 88 808.0 800.0 800.8 800.0 080.8 088.0 080.0 080.0 0 88 ~00 ~00 ~00 ~00 880 A<0 880 A08000 00 880000002 88 00000 08800 80 80800 0880085800 000 08 008000802 8800 8000 000 80 0000 80 00800000880 000 00808000 808000802 08 0000000 009 .08 00009 63 one simple potential which can give the structural differences represented by different phases. In general, potentials II, III and those fit to McDonald's potential I give the best fit to the structure of CCl4 phase II and the fit of all three of these potentials is very much improved if partial charges of a relatively large magnitude are included, even though the effect on the calculated energy is small. These are the potentials chosen for use in calculations with the chlorofluoromethanes. O’\ .5 CHAPTER 5 CARBON TETRAFLUORIDE A. The Crystal Structure of the Low Temperature Solid Phase of Carbon Tetrafluoride Two crystal structures have been proposed for the low temperature solid phase of CF4, sometimes called phase II or the a phase. Both of these suggested structures are based on the analysis of X-ray powder pattern diffraction data [31,32]. Sataty, Ron and Herbstein [28], hereafter designated SRH, obtained the far infrared lattice spectrum of solid phase II CF4. In order to obtain a crystal symmetry consistent with the observed lattice spectrum they gave an alternative space group assignment to the structure of Bol'shutkin et al. {32]. The results of molecular pack- ing analysis [16], analysis of the diffraction data [64]. lattice dynamics calculations [65], and a computation of the interatomic distances [66], all indicate that the SRH structure is incorrect. The error in the SRH structure is shown to result from a mistaken correction of an error present in the list of atomic coordinates of Bol'shutkin et al. [32]. The corrected structure in this work is shown to result from an alternative correction of 65 the error present in the work of Bol'shutkin et al. [32]. The Bol'shutkin structure is P21/c No. 14 with four molecules per unit cell on C1 sites, with a==8.435, b=4.320, c=8.369 A, and B=1l9.40. The reported fractional atomic coordinates of the basis CF4 molecule are listed in Table 11. These fractional coordinates are TABLE 11. Fractional Atomic Coordinates in Bol'shutkin et al. Structure [32]. Atom x y 2 C 0.250 0.323 0.250 F(l) 0.392 0.500 0.286 F(Z) 0.289 0.500 0.394 F(3) 0.108 0.147 0.214 F(4) 0.211 0.147 0.106 incorrect, as is verified by a calculation of the intratomic F-F distances. Inspection of the x and z coordinates suggests the presence of a two-fold axis parallel to b, located at x==0.25, z==0.25, containing the carbon atom. This two-fold axis interchanges F(l) with F(3), and F(2) with F(4). Interchanging the y coordinate of F(2) with that of F(3) or interchanging the y coordinate of F(l) with that of F(4) gives a molecule that is tetrahedral within the experimental error and which is consistent with a two-fold axis parallel to b. The SRH structure is C2/c No. 15 with four molecules per unit cell on C2 sites with a =8.435, b==4.320, 66 c==8.478.&, and B==120°42'. This structure was derived from that of Bol'shutkin et a1. by interchanging the y coordinate of F(2) with that of F(3), and recognizing that the systematic extinctions are consistent with a higher symmetry space group. SRH also make the change in the choice of the lattice vectors given by the transformation a I 0 0 a' b = 0 1 0 b' c l 0 1 c' where a', b', and c' refer to the lattice vectors of Bol'shutkin et al. The fractional atomic coordinates in the SRH structure are listed in Table 12. To obtain the TABLE 12. Fractional Atomic Coordinates in Sataty, et a1. Structure [28]. Atom x y 2 C 0 0.323 0.250 F(l) -0.106 0.500 0.286 F(2) 0.105 0.147 0.394 F(3) 0.106 0.500 0.214 F(4) -0.105 0.147 0.106 fractional atomic coordinates in one monoclinic basis from the fractional atomic coordinates in the other monoclinic basis, it is convenient to first transform the coordinates into an orthonormal basis set and then from these into the other monoclinic basis set. The structure used in this work is given by inter- changing in the Bol'shutkin structure the y coordinate 67 of F(2) with that of F(3) as was done by Sataty et al. The fractional atomic coordinates of this structure are given in Table 13; the lattice vectors are identical to those of Sataty et al. TABLE 13. Fractional Atomic Coordinates, [This Work]. Atom x y 2 C 0 0.323 0.250 F(l) -0.106 0.147 0.286 F(2) 0.105 0.500 0.394 F(3) 0.106 0.147 0.214 F(4) -0.105 0.500 0.106 A comparison of some atomic distances calculated from the structure of SRH and this work is shown in Table 14. TABLE 14. Some Atomic Distances of CF4, Phase II. ---- Sataty Structure ---- ------- This Work ------- Atoms Distance (A) Atoms Distance (A) F(2)-Carbon 1.323 F(2)-Carbon 1.325 -F(3) 2.160 -F(3) 2.160 -F(1) 2.161 -F(l) 2.161 -F(4) 2.165 -F(4) 2.165 -F(4)' 2.349 -F(l)' 3.020 -F(1)' 3.020 -2F(4)' 3.065 -F(3)' 3.076 -F(3)' 3.076 -F(1)' 3.109 -F(2)' 3.101 -F(1)' 3.109 Note that the non-bonded F(2)--F(4)' distance in the SRH structure is unreasonably short. Low angle reflections of the powder X-ray diffraction patterns calculated from the two structures are compared in Table 15. Average Cu Ka radiation, and isotropic 68 TABLE 15. Calculated versus Observed Diffraction Patterns Experimental SRH This Work Bol'shutkin [32] 20 I 20 I 23.87 100.0 100.0 100 23.84 100.0 23.97 17.8 33.8 24.22 41.5 76.6 57.2 24.14 62.3 24.42 39.9 73.7 90 9 24.38 84.6 24.54 26.0 48.0 ' 29.52 5.0 6.7 29.72 3.2 85.7 64.1 29.64 63.2 38.62 32.3 98.7 73.8 38.62 65.4 38.79 7.5 2.4 45 2 38.88 48.7 58.1 ' 38.88 36.0 41.82 2.5 4 6 3.4 41.80 2.7 42.42 9.6 6.2 14 9 42.46 9.7 42.66 7.4 13.7 ’ 43.71 2.0 12.8 9.6 43.70 7.5 47.28 4.3 0.6 48.91 2.9 9.2 11 3 49.02 5.2 5.9 ° 48.98 12.3 49.42 5.2 5.5 4.1 49.36 4.5 49.62 3.8 10.7 8.0 49.58 7.6 49.93 2.3 14.1 50.07 2.2 0.1 50.31 8.2 15.1 11.3 50.32 7.9 53.75 0.0 3.6 2.7 53.62 2.5 54.08 3.1 0.5 69 temperature factors equal to 2.0 for all atoms were used to obtain the calculated diffraction patterns. In particular, the relative intensities calculated for 20 =29.72, 47.28 and 53.75° indicate that our structure gives the correct structure. That the intensities calculated for this work are generally too high may be attributed to the choice of isotropic thermal parameters for the atoms. Molecular packing analysis of the CF4 phase II SRH structure was undertaken both as an independent method of determining the correctness of that structure, and as a means of testing the interatomic potential parameters available for F-F interactions. The form of the potential was A.. _ _ 1 __ V(rij) - gig... Bij eXP( Cijrij) with the values of the parameters listed in Table 16. These will hereafter be referred to as potential set I. TABLE 16. Potential Parameter Set I [67] A B C do so kJ/mol A6 kJ/mol A'1 A kJ/mol F-F 541.8 154800. 4.26 3.113 -0.325 C-F 1135. 232700. 3.93 3.455 -o.372 c-c 2377. 350000. 3.60 3.882 -0.396 The position of the well minimum, do’ and the well depth, 80' are also included in this table. The F-F parameters were obtained [67] from thermodynamic data for neon 70 [62]. Static Lattice Energy (kJ per mole) Figure 14. CF4 Sataty et al structure "J -5.E .1 -10-'_3 E -15 '1 1 -20; SRH '1 A: Thh wont -25111lIlIl—rlilfirIrTI1j 0.0 0.5 1.0 Fractional Coordinate of carbon along 6 Static lattice energy versus fractional coordinate along the 2-fold axis parallel to b. 71 As these interatomic potential parameters do not result from a fit to the observed structure or thermochemical properties of fluorocarbons one might expect the results of the calculation of such properties for CF4 to be qualitative. Only two degrees of freedom remain for the structure if one assumes a rigid molecule and fixes the lattice constants and the space group. These correspond to translation along the two—fold axis parallel to b, and to rotation about that two-fold axis. Figure 14 shows static lattice energy versus the position of the CF4 center of mass along the two-fold axis parallel to b, with rotation about b held to zero. The Sataty structure has a y fractional coordinate of 0.323 which gives a static lattice energy of -15.2 kJ/mol. This is far from a minimum, either that at y==0.172 or that at y==0.672. The structures at these two minima are identical, as is suggested by the symmetry of the lattice energy diagram. A translation of the origin by 0.5 analog either a or b interchanges the two structures. The structure used in this work is identical to that obtained with the SRH basis CF4 molecule if that molecule is given a carbon y fractional coordinate of 0.177 (0.677). That this structures lies very much nearer to the static lattice minimum shown in Figure 14 suggests that it is the more reasonable structure. 72 Lattice dynamics calculations [65] were performed using potential I applied to the SRH structure and that of this work. The primitive cell used in the lattice dynamics calculations was obtained from the cell of Sataty et al. by choosing the b axis as the vector from the origin to 1/2, 1/2, 0 (the face center). It has a=8.435, b=4.738, c=8.478A,01=ll6.9°, B=120.7° and y=27.12°. The atomic coordinates in the molecular frame, and the orientation matrix (that matrix which transforms the coordinates from the molecular frame to an orthogonal crystallographic frame) for both structures are listed in Table 18. The two structures yield infrared-active modes of nearly identical frequencies, but the six Raman-active modes are predicted to have widely differing frequencies. These are listed in Table 17. The infrared frequencies calculated from both structures are lower than those observed; the structure suggested by Sataty et al. predicts less reasonable Raman-active frequencies. Although the Raman-active lattice modes have not been observed experimentally, two of the calculated frequencies are imaginary. This result suggests that the SRH structure is far from equilibrium as was also indicated by the packing analysis. In summary, all considerations indicate that the Sataty structure is in error and that the alternative structure presented in this work is the correct one. 73 mwooo.on mmhmw.o owwam.o mnvmm.ol hmmoo.o bmmoc.o oo.o momam.o| mmmmm.o xwuumz coflumucofluo one vow.o| ¢m>.o1 vmh.ou www.01 «mh.c umb.on Avvm ¢m>.o «$5.01 vo>.o ¢w>.o vmh.o vmh.o Amvm vo>.o| vm>.o vmh.o emh.o1 voh.o1 qwn.o Amvm voh.o wob.o vwb.on em>.o vou.o: vo>.ou AHVm o.o o.o o.o o.o o.o o.o U IIAWII Ilqdll IIMNII IImWII [Infill x IIIIIIIIII xuoz mane IIIIIIIII 1111:: onsuosuum mumumm 1111:: .xwuumz cofluoucowuo on» cam .oEmum Hmaoooaoz on» no woumcficuooo anoufi .ma mamda we sea on as om mm mm as 50 am mm mm Hm sumcflmmefl mm mm ha mumcflmmafl Hm mm x003 mate musuonnum saw is can 60>nmmno x003 mane 11111111 o>auo¢ cmfimm IIIIIIII 1111111111111111 0>Huo< concumcH HI. mhfiuufluum mMm mm mm ED mN .H Hafiucwuom spa: cwumHDOHMU mowocmovonm oofluuoq who .hH mamfla 74 B. Lattice Dynamics The goal of lattice dynamics calculations performed with the CF structure was to obtain non-bonded fluorine— 4 fluorine potential parameters which gave reasonable static lattice energies, and which predicted lattice frequencies close to those observed. The static lattice energy is related to the heat of sublimation of 0 K and the zero point energy as follows, _ 0 K -Estatic - AHsub+Ezero point The static lattice energy calculated using potential I 10 K was -20.2 kJ/mol. This is too low, given that -AHsub is equal to 17.6 kJ/mol and an estimate of the zero point energy is 1.8 kJ/mol [68]. The lattice frequencies calculated using potential I are also too low. For these reasons it was felt necessary to obtain new potential parameters. The procedure initially adopted was to fix the position of the well minimum, re, and to fix the well depth, V(ro), at the values given by potential I; for each selected value of the parameter C, values for the other parameters, A and B could then be calculated. Given the potential V(r) = -A/r-+B exp(-Cr) , one obtains 75 ro7CV(ro) A = (6-roC) and B = 6 r0 V(ro)-+A 6 r0 exp(-Cro) The choice of values of C between 3.8 and 6.2 A.1 gave little improvement in the calculated lattice frequencies. Potential parameters which increased the equilibrium position and decreased the well depth were then chosen and the lattice frequencies and energy were calculated for each set chosen. The potential parameters obtained by this trial and error procedure are listed in Table 19. TABLE 19. Potential Parameters: Set II . A B C do so kJ/molA6 kJ/mol A.1 A kJ/mol F-F 559.7 154800. 4.00 3.493 -0.176 C-F 1153. 232700. 3.80 3.673 -0.268 C-C 2377. 350000. 3.60 3.882 -0.396 The static lattice energy obtained with this potential was -13.9 kJ/mol. This about 80% of the experimental 101< -AHsub . If a charge of -0.2 electrons [69,70] is placed at the fluorine positions and 0.8 electrons is placed at the carbon positions one obtains a static lattice energy of -l8.4 kJ/mol. The lattice frequencies obtained with each of these potentials are listed in Table 20. Potential II was also tested in the packing analysis program. If the lattice parameters and rotation about the two-fold axis were not allowed to vary, potential II gave a 0.013.A shift of the 76 TABLE 20. CF4 Lattice Frequencies Calculated With Potential II. ---------------- a. Infrared-Active Modes ----------------- Potential II Potential II-tQ‘I Observed (40 K) 47 cm-1 47 51 57 56 57 68 63 66 ------------------ b. Raman-Active Modes ---—---------—---- Potential II Potential II-tQ‘z 31 cm'1 28 31 29 52 51 57 54 63 60 75 76 “Partial charge on the fluorine equal to 0.2 electrons. 77 CF4 center of mass in the direction parallel to b before reaching the minimum energy position. Potential II with the above-mentioned partial charges gave a 0.0417A shift parallel to b to the minimum. If the lattice parameters and orientation about the two-fold axis, as well as the position along the two-fold axis, are allowed to vary, one obtains the result presented in Table 21. Although these variations are larger than the experimental uncertainties, Potential II gives an equilibrium structure somewhat nearer to that determined experimentally. The fit of the observed frequencies and of the static lattice energy are much better for potential II, and it was deemed adequate for further packing and lattice dynamics calculations. C. Spectroscopic Predictions and Results The Raman spectrum of crystalline phase II CF4 was investigated. In a previous study of this phase, by Fournier et al. [29] a Toronto arc was used as the excitation source, and no lattice modes were observed. The present laser Raman instrumentation provided somewhat better resolution of the factor group components of the fundamentals, but again no lattice modes were observed. Spectroscopic symmetry predictions based on the group theoretical analysis [3] are presented in Table 22. This analysis is based on the crystal symmetry given by Sataty, Ron, and Herbstein [28]. The observed internal 78 1.1 .n 0» Hmoamumm mnxm toomuo30 map 05008 cooumuomo A<:::::4Bg (2v3.2v4) F2 48 5Au (v1,202,v3,v4) (V3IV4) / ' F2 Bu (2v3,2v4) -------------------- b. External Modes --------—----------- Molecule Site Factor Td C2 C2h 2A (2 mixed) 4B (4 mixed) X >< g 2______4B\2A (mixed, acoustic) 4B (2 mixed, 2 acoustic) Raman active: 6 IR active: 3 80 mode Raman frequency shifts from the present and the earlier work [29] are presented in Table 23, and the Raman spectrum obtained here is shown in Figure 15. The two sets of assignments are identical except for the components of v3. The highest frequency component of v3 was previously assigned [29] as a longitudinal optical (LO) component of a lower-frequency, factor-group split, transverse optical (TO) mode, analogous to the assignment in the plastic and liquid phase of CF [71,72]. However, 4 phase II has been shown since to have a centrosymmetric unit cell, and this precludes the possibility of LO-TO splitting [73]. Other previous interpretations of Raman scattering at 90° from similar polycrystalline samples in terms of LO-TO splitting have been shown to be in error [74]. The definitive evidence for such LO-TO splitting would be marked changes in the intensity and splittings of such bands in the Raman spectrum obtained from single crystals as a function of the scattering angle. Such scattering- angle-dependent spectra would also be expected with polycrystalline samples. Although scattering-angle- dependence has not been investigated because the crystal structure of phase II CF4 is reasonably well established, an alternative interpretation would seem to be preferred. One possibility might be to suggest Fermi resonance between v3 and 2v4, which has been used to explain the gas phase Raman spectrum in the 1200-1300 cm-1 region, as well as 81 TABLE 23. Observed Raman Frequency Shifts for CF4 in Solid Phase II. This Work(25 K) Reference 28 (70 K) . -1 . -l A351gnment Av (cm ) Ass1gnment Av (cm ) 436.1 435.8 v2e 439.3 V3 438.5 628.0 629.6 V4f2 630.0 v4f2 631.2 2v2 868.1 2v2 867.7 vlal 906.1 vlal 907.2 1227.4 TO 1230.1 v3f2 1241.5 v3f2 TO 1243.7 1321.6 LO 1321.0 82 . :05 on: 93868 05 Sam mumdLm ohm mmflocwovoum . OH omv who paaom mo gmug .3 953m mm; was .0 S m mg vww mm 5.. .<. 3,. VNN. gum. hm. 8m. [IT 83 the liquid-argon solution spectrum in this region [75]. The band positions in the infrared spectrum of CF4 in liquid Argon solution are similar to those of the Raman spectrum and are also given this assignment [76]. Such Fermi resonance is permitted by symmetry since the representation of the f2 Xfé direct product which gives the symmetry species of the 2v4 combination contains the f2 irreducible representation of the v band. If there 3 were Fermi resonance between 2V4 and v3 in the solid one would expect the 2v4 and v3 components to appear to be shifted by approximately equal but opposite amounts from their unperturbed frequencies. The unperturbed 2v4 can be estimated as twice the frequency of v4 in the solid. Two times the frequency of v4 in the solid gives 1256 cm—1. If this were the higher of the two unperturbed frequencies Fermi resonance would shift it even higher. In this case the highest frequency band at 1321 cm.1 would be associated with 204 and one would estimate its Fermi resonance shift as 65 cm-l. One would then estimate the unperturbed 1 frequencies of the 1227 cm- and 1242 cm"1 bands as l and 1307 cm-1. These are higher than the 1292 cm- 1256 cm"1 unperturbed 2v4 frequency contrary to our initial assumption. If there is Fermi resonance between v3 and 2v4 the higher frequency component must be due to v3. This results in the prediction of either 1270 cm-1 or 1285 cm-1 for the higher frequency (1321 cm-1) band depending on whether the 1242 cm.1 or the 1227 cm-1 band 84 is used to estimate the Fermi resonance shift. The best explanation seems to be to describe the appearance of the three highest frequency bands as the three group components of v3. The appearance of the third factor group component of the v3 band is not required. As previously mentioned in the case of CCl the number of 4 factor group components predicted by the group theoretical method may very well exceed the number of observed components. Indeed this is even the case in the present spectrum where only two of the three predicted components of the v4(f2) band appear. 85 CHAPTER 6 CF3C1, CF2C12 AND CFC1 3 The far infrared and Raman spectra of the solid chlorofluoromethanes have been obtained in order to test the transferability of the chlorine and fluorine interatomic potential parameters-—obtained for CF4 and CCl4-to these chemically similar compounds. Since the crystal structures of these compounds are unknown, the crystal symmetry must first be inferred, either from that of molecules of similar symmetry, or from an interpretation of the spectroscopic data. For a selected crystal symmetry, an equilibrium structure is obtained by packing analysis under the constraints composed by the symmetry relationships. This structure is then used in a lattice dynamics computation of lattice frequencies. A comparison of the observed frequencies to those calculated determines the appropriateness of both the structure and the potential parameter set used in the calculation. This procedure has been carried farthest for CF2C12, and this chapter begins with the discussion of this compound. 86 A. Previous Spectroscopic Results More than 100 vibrational bands, including funda- mentals, overtones, combinations, and "hot" bands of the three isotopic species CF235C12, CF235C137C1, and CF237C12 have been identified in the infrared spectrum of the gas in the range 1600-400 cm"1 [77]. An analysis of the band contours, (and differences of combination and hot bands) was used to obtain an almost complete assignment of the positions of the 9 fundamentals for all of the isotopic species. The gas phase infrared spectrum of a chlorine-35 enriched sample was obtained in the 923 cm-1 1 (v8) and 1161 cm- (v6) regions in order to distinguish the overlapped hot bands and isotopic bands [78]. CFZCl2 has also been included in Urey-Bradley force field calculations of a series of halogenated methanes [79]. The Raman spectrum of the gas has been measured only with non-laser sources, and under low resolution [80-81]. The present work is the only spectroscopic investigation of the solid of which the author is aware. B. The Far Infrared and Raman Spectrum of Solid CF2C12 Seven of the nine internal modes and six lattice modes are observed in the Raman spectrum of solid CF2C12, which is shown in Figure 16. Not shown are the internal modes v8, which appear as a weak singlet at 903.4 cm-l, and v6, which is recorded as a very weak, overlapped triplet 87 SCH. on: Eflfloxw 05 Eoum 333m mum 83:96on . Cw omv NHU NB pHHOm mo 58mmlqmfimm .3 chug: .o «9.3. .9»: .00“ gm an mum awe W? .05» a; n) .0 0923 am [llwlwm m8 m8 com mom 80. NB. mJN mm 4443521. womb 88 between 1142-1152 cm-l. The symmetry species of the fundamentals are labeled so as to be consistent with reference 79. The multiplets within a band have been assigned as due to either chlorine isotopes or factor group splitting, according to the relative intensities of the multiplet component, and the size of the splitting between the components. The natural abundance of chlorine isotopes gives the ratio 100:65:11 for the possible combinations of isotopic masses of chlorine in CF2C12. If the relative intensities of the multiplets within a band are in this proportion they are assigned as due to chlorine isotopes. The remaining multiplet components are assigned as due to factor group splittings. The assignments and frequencies of the Raman spectrum are given in Table 24. Since v7 and v9 are overlapped the assignment of a particular band in the 430-440 cm-1 range to one or the other fundamental is uncertain. Among the possibilities is the assignment of 35 C1 - isotopic components for each, with overlap of the 2 35 37 component of v and the C1 Cl-component of v7. The 9 two fundamentals observed in the far infrared are shown in Figure 17. These are the v4a1 and vsa2 normal modes, and their frequencies are also listed in Table 24. The 05a2 mode is gas-phase infrared-inactive and its appearance here is due to crystal symmetry. The v4a1 mode is a triplet with spacing consistent with its assignment as due to chlorine isotopes. The bands do not have the 89 TABLE 24. The Observed Raman Frequency Shifts and Infrared Absorptions of Solid CF2C12 Raman Relative Peak Infrared Assignment Av (cm’l) Height v (cm'l) CF237C12 259.0 9 259.1 CF237C135C1 261.6 54 262.0 Vkal CF235C12 264.2 100 264.5 factor group 266.1 37 -— CF237C12 452.1 9 V331 CF237C135C1 456.2 54 CF235C12 459.2 100 vzal 664.4 shoulder 664.4 100 factor group 1063.3 100 ”1&1 factor group 1065.9 28 factor group 319.5 100 319.9 V532 factor group 322.0 71 321.6 sh? Ugbz 903.4 875.8 V3+V9 878.9 1142 vw v5b1 1148 vw 1142 vw 431.8 5 ng2 434.8 40 437.3 sh Table 24 Continues. 90 Table 24 Continues. Raman Relative Peak Infrared Assignment Av (cm’l) Height v (cm’l) 439.4 100 V7b1 441.5 55 Lattice Modes Raman Infrared 20°K 14°K 33.4 cm'1 28.0 cm’1 42.8 33.3 44.6 39.4 48.0 47.7 52.9 56.8 91 0012152 LOWER 010908 01090C UPPER Figure 17. The far-infrared spectrum of solid_$ v4 al and v5 a2 regions (14 K) (cm 240 250 260 270 280 290 300 310 320 330 111111111111111111111 llllllJlllLL'llllllllllllllllllllllll'LllLlllllllllllllllllLlLlllALllllL >‘ 1— l.- 1) -.-4 U) C 0) 4J c 1 #4 C1) > "4 p — f5 .— ,4 (D D’. 0 III!‘IIIIIIITIIIIIITTIIIIIIII'YIIIIIIIYTIIIIIIIYIIIIWIII(III![IIIIIIIYIIIIIIIIIITIIIIT1TIITII 240 250 260 270 280 290 300. 310 320 330 WAVENUMBER FZCl2 1n the 100:65:11 intensity ratio expected for the chlorine isotopes. This is probably due to the higher frequency component "bottoming out" due to the sample thickness. The frequencies of the fundamentals in the gas and in the solid are compared in Table 25. The 2.6 cm.1 splitting of vl observed in the solid is in contrast to the infrared spectrum of the gas, where no isotopic splitting is observed, and this supports its assignment as a factor group splitting. The absence of isotopic splitting in the vla1 mode is consistent with the calculated potential energy distribution [77,79], which designates this mode as primarily the symmetric C-F stretch. The v2 band is a poorly resolved doublet; although the small displacement of the weaker component to lower frequency would be consistent with isotopic splitting, neither the relative intensity nor the splitting can be accurately determined and factor group splitting cannot be discounted. The multiplets of the v3 and v4 bands have intensity distributions and splittings consistent with their assignment as due to chlorine isot0pes, with the exception of the high frequency shoulder of the v4 band at 266.1 cm-l. This feature is assigned as a factor group component of the v4 band. The relative intensity of the two components indicates that the v5 doublet is due to factor group splitting. The va1 mode is a very weak and overlapped triplet. No attempt is made to assign this structure, or the overlapped v7 and v9, components. The v8 mode appears 93 .>~u>«uuoanuu H180 o.mom can o.ooo .H.coa couoaauHoo nuoao> conuauuonca am>+ n) nu“; oucocOnou «Sham one .nucan uon can :OADICanEou coo3uon mucou0uuwv on» EOuu oouofiuumuo .vHOAm vouch ouco~a> «chance a vows: vaucaou~ouu n.hao ..noo o.omo o.moo o.oso as..ns H.o ~.m H.n «.onv m.nnv o.om. on. o.am. o.vno o; no n.n m.n cocoaauoao o.n ov.oao . on.-o o~.n~o n~o ..moo o5 ~n o.a o.a . an.onv o.oo. o.~ov cocoaauoso n.5no 1 ..ono m.a.v so an m.a ~.H H.~ H.~ o.oo~a o.ooaa o.ooaa soaa woos soda Noon so an o.o o.o a...nn am.oan os.o~n o.on coo.aum Nan m.on o.-n m: as H.a ~.H s.a m.~ ao.om~ ou.om~ oo.ao~ H.om~ o.~u~ m..o~ do” o.am~ o.ao~ ~..o~ H.oo~ «a as n.~ n.~ o.~ m.~ o.~ o.~ a.o a~.ooc an.vo. m.oo. on. a.~mq ~.om. «.oo. no as a.v ~.v uoamoauo>o H.v o.n so.ooo m.moo o.ooo soc c.4oo ..ooo ~> as m.“ n.a o.~ so.~oaa ..~oad o.~oo~ oooa n.nooa o.mooa Ho Ho v.o o.o o.~ a N N N a anon mo ”condomn so Homo to 1H. 660 cuaom .moaau oaaom vouauuan adadz a 150. cauom on» so can among ado onu ca «HUNmU no unaduouna> HaucoEdocam on» no nowunosooum .mN m4m49 A 944 as a singlet, weaker than the neighboring v3-+v9 combination band with which it is probably in Fermi resonance. The spectroscopic observations of fundamentals later used in the factor group analysis are these: that the a modes may have two factor group components in the Raman, and that the a2 mode has at least one factor group component in the infrared and two factor group components in the Raman spectrum. The far infrared spectrum of CF2C12 in the lattice region is shown in Figure 18; the frequencies of the four observed modes are listed together with the six Raman active lattice mode frequencies in Table 24. It should be noted that at least two of the infrared lattice frequencies are non-coincident with Raman lattice frequencies. The temperature dependence of the frequencies of the infrared-active lattice modes has also been obtained and a summary of these spectra is given in Figure 19. The atypical increase in frequency with increasing temperature shown by the two lowest frequency lattice modes may indicate an unusual lengthening of one or more of the lattice vectors with increasing temperature. C. Factor Group Analysis of CF2C12 The combination of the crystallographic site or sites chosen for the molecule and the factor group chosen for the unit cell determines the number and spectroscopic activity of the crystal normal modes. These numbers and activities 95 .25 .30 35 40) 45> 50I 551 60 lllllllllllllllll‘lllIt'll!lllllllllJllllllllllllllllllllllljllllllll Ission Relative Transm 'l'l'l'l'l'l'l'Trl'l‘l’I'l'l'lrlrl'l'l'['l‘i‘l‘Ijl'l'l'l‘l‘l'l'l'l‘l' 25 50 35 40 45 5O 55 60 Wovenumber (cm-1) Figure 18. The far-infrared spectrum of solid (Zli‘ZCl2 in the lattice region. (14 K) 96 wavenumber Figure 19 . 5O 45 4O 35 .30 25 CF2012 10 2O .30 4O 50 60 70 80 90 100 11114—11llll'llLlllJllll‘lllJJlllllJll .J . “000 0000vo O - icon) 0051vo o :- _ OA 0 0 o - _ O o O O O O; .. .000 O .. O -00 .. ._ A .— .. O — .. 00d) _ -DOD .- ‘3‘l‘l'lli‘i'l'lrl‘l‘lri'l‘l‘l'l'l'lT 10 20 .30 4O 50 6O 7O 80 90 100 temperature (Kelvin) 50 45 40 .35 30 25 The tarperature dependence of the inf rared-active lattice frequencies of solid CIE‘2C12 (14 K) o- 0.1 mm thick A- 0.3 nm thick (blacking out the band at 32.5 cn'l) 97 must be consistent with those observed experimentally. The site group must also be consistent with the molecular point group. Often, however, the number of factor group components predicted exceeds the number observed experimentally. This was the case for both CF and CCl 4 4 which were considered in previous chapters. The absence in experimental spectra of predicted factor group multiplets is usually ascribed to the weakness of the intermolecular coupling from which they stem. The number of factor group multiplets observed is therefore taken as the minimum number acceptable. If, for example, the site group-factor group combination leads to the prediction of a smaller number of factor group components than that observed experimentally, it is rejected. If the predicted number of factor group components exceeds that observed it is rejected only if the predicted number is larger than a specified number greater than that observed. Given these constraints, the site group-factor group combinations determined from diffraction measurements for CF4 would have been among those determined as consistent with the spectrosc0pic results. The large disparity between the number of crystal normal modes predicted and observed for the large, populous unit cell of CCl4 would have precluded the above procedure from predicting the correct site group-factor group combination. The factor group analysis of solid CF2C12 was begun by restricting the site group—factor group combination to 98 those consistent with the lattice spectra. The seven site group-factor group combinations which give predictions of at least four, but less than seven, infrared active lattice modes, some of which are not coincident with Raman active lattice modes are listed in Table 26. Correlation diagrams TABLE 26. Site Group-Factor Group Combinations Consistent With the Lattice Spectra of CF2C12 Number of Infrared Lattice Total Number Modes Not Total Number of Infrared- Coincident With of Raman- Site Factor Active Lattice the Raman Active Active Lattice Group Group Modes Lattice Modes Modes C2 C4h Cs C4h sz Th 4 4 7 C2v 0h 4 4 8 Cs C3h 4 2 8 CS C6h 4 _4 9 C2 C6h 4 4 8 were then drawn in order to obtain the spectroscopic activity and factor group multiplicities of the funda- mentals under those site group-factor group combinations. All but two of the site group-factor group combinations predict the appearance of a singlet for the vsa2 mode in the Raman spectra. A doublet is observed, as predicted by either the combination of the C2 site group with the C4h factor group or the combination of the C2 site group with the C6h factor group. These two site group-factor 99 group combinations exist in six tetragonal and one hexagonal space groups, respectively. The hexagonal space group requires that the CF2C12 molecules be arranged around a six-fold axis, as it is difficult to imagine that such an arrangement could be packed efficiently, the hexagonal space group seems unlikely. Unfortunately, the lattice dynamics program TBON can calculate lattice frequencies only for crystals of orthorhombic or lower symmetry. Lacking the capability of comparing the observed lattice frequencies with computed values as a test of the accuracy of the minimum energy, packed structures, the minimum energy structures consistent with the seven space group symmetries were not computed. Such packing and lattice dynamics calculations were restricted to the structures of molecules of similar geometry which have been determined by x-ray diffraction. D. Structural Analogues Dichloromethane and dibromomethane have geometries similar to that of dichlorodifluoromethane and their crystal structures have been determined from X-ray diffraction measurements [82-83]. A summary of the predictions of spectroscopic activity and factor group multiplicity made from the correlation diagrams drawn for these structures is shown in Table 27. The C2v symmetry of CF2C12 is compatible with the occupied site symmetry of the CF4 structure and this structure has been 100 o m m o H H H H H H H H H H H H H H H H H N H H H N H H H N H H H N H H H N H H mm cwEmm mm cmEmm mon No Humor coo no coo oon No cmo NNNNNNN NNNNNNN N NH 5 o N N II .QMH>o N N I: 3> N N II .mmH>o N N In 3> H N H N H N H N H N In H H N In H H N II N mm :mEom mm coEom N N mucmcomeou.mdono A H0 SUV HOHUMW ©0>H¢mfl0 H 3N m." u.m o aHo "wOHDDMH "m? "m9 No "Nu? "m? Ho "m9 No “v? “n? Hm "N9 "H9 mcoHnocHQEou moouw HODOMM\QDOHO muHm Hmum>ow HON mcowuowcmum HMUHumuomnB msouo .hN mqmdfi 101 included in Table 27. Also included is the site group- factor group combination obtained as that most consistent with the spectroscopic observations. It can be seen that none of the three possible analogues crystallizes in a space group permitting the site group-factor group combination determined to be most consistent with the spectroscopic data. Both the CH2C12 and CHZBr2 structures predict considerably more lattice modes than are observed for CF2C12, but the prediction of too many modes does not rigorously exclude these as possible structures. The CF4 structure predicts too few lattice modes and too few fundamental multiplet components, and it is therefore a much less likely structural analogue. E. Packing Analysis of the Possible Structural Analogues The method used to obtain equilibrium structures begins with a model CF2C12 molecule having tetrahedral valence angles, a C-Cl bond length of 1.77.A, and a C-F bond length of 1.33.A. These values are within the uncertainties of the experimentally obtained values [84]. The carbon atom of the CF2C12 molecule is placed in the position of the carbon in the analogous structure, then the CF2C12 molecule is oriented so that its bonds point as closely as is possible in the directions of the analogous bonds. CF4 is tetrahedral and it was possible to make the directions of the CF Cl2 bonds the same as those of CF4. CH Cl2 has 2 2 its two-fold axis aligned with a crystallographic two—fold 102 axis, and the CF2C12 was oriented so that its F-C-F plane coincided with the H-C-H plane of CHZClZ. After placement and orientation of the CF2C12 molecule(s) in the analogous structure, the parameters in a potential of the exp—6 form are chosen and those structural parameters unconstrained by crystallographic symmetry are allowed to vary until an equilibrium structure is obtained. The same potential is then used to compute the lattice frequencies of the equilibrium structure. Two different parameter sets were used in the packing analysis of CF2C12, potential set IIIb and potential set VIb. The values of the parameters in those potentials are shown in Table 28. The interatomic parameters of set IIIb are those of the CC1 potential III 4 and the CF4 potential II with the Cl-F parameters obtained by the usual combining rules. These give the Van der Waals constant and the pre-exponential constant of a cross interaction as the geometric mean of the two uncrossed interactions. The constant in the exponent of a cross interaction is given as the arithmetic mean of the two uncrossed interactions. The C-C and Cl-Cl parameters of set VIb are those of CCl4 set VI. The F-F parameters are those of CF4 set II and the other parameters in this set are again obtained by using the usual combining rules. Both potentials were used to obtain equilibrium structures for CFZCl in the D%: space group of CH2C12. 2 These structures and lattice frequencies calculated from them are listed in Table 29. The volume per molecule of 103 TABLE 28 . Potentials Used in CF2C12 Packing and Lattice Dynamics Calculations IIIb VIb A 2377 1592 A - kJ/mol A6 c-c B 349900 621700 B —- kJ/mol c 3.60 4.044 c - A‘1 A 3820 3344 c-01 E 596600 1803000 0 3.63 4.044 A 6142 6832 c1-01 B 1018000 5109000 c 3.66 4.044 A 560.0 560.0 F-F B 154800 154800 c 4.00 4.00 A 1855 1956 Cl-F B 397000 889300 c 3.83 4.022 A 1153 944.2 C-F B 232700 310200 c 3.80 4.022 104 .mmscHucoo mN mHnaB OHOE\hx m.NNI v hm.H m.mov om b.0NH om wmv.m Nm0.v mom.OH DHHH .uom ouHm No cmo Homoc OHOE\bx m.mHI m Hm.H m.mmm om mh.mm om vmm.vd Hmm.¢ www.ma QHHH .uOm ouHm Ho cmo ANumNmov oHos\ox N.mN- a mo.N N.mom om om om on.o mom.m HHm.a QH> .uom N SN 0 H u.m U vHQ ANHoNroc OHOE\hx m.mNI v HH.N m.omm om ca om mmm.m vmm.m bmv.v QHHH .uom N SN ouHm HH 0 D ANHonoo moumcm wofluuma aflumum u N AmEo\Emvo AHHoo> Hay» Hovm “ova A490 AH0 may QDH3 pwcwmuno mowocozvmum oofluumn pom ouauosnum .mcouw wuwm .mN mqmflfi 105 NOH as so No No Nm so oN sN HIEU mo mHm oHHH .uoo ouHm No nmo Aomov an mma vHH mm mm wma mm mm vm mm vm mv Nm m.mm NN um ma Hm MH Hm .mmEH .meH HIEU mNH mHm nHHH .uom ooHo Ho cwo ANumNmoo os ms .uoocH N- so oo mo ss om os mo so mo so oN oo sN so mN om oH sm alEU mNH mHs nH> .ooo 0on No ”mo ANHonoV os os 3.008.: N. mm eh o.so m.Ns om os mo mo so om oN m.om mN om NH om m.o m.mm HIEU mNH mHs nHHH .uoo ouHm No ”mo ANHonov m.wm m.Nm o.mv h.bv w.v¢ v.mm m.Nv m.mm «.mm o.mN mo me H.Eo co>uomoo .ooscHucoo mN 0Hnoe 106 the two cells differ from each other by only 4%; their average value is 97.0 A3/CF2C12 molecule. This is very nearly equal to the average of the 131.6 A3/CC14 molecule and 64.2 A3/CF4 molecule values obtained for the CCl4 and CF4 structures. This suggests that CF2012 could pack efficiently into the structure of the CH2C12 analogue. The lattice frequencies of the 0%: structure calculated by the two potentials are broadly similar, although those calculated with potential VIb are systematically higher than those calculated by IIIb. It is also apparent that neither interatomic potential parameter set gives frequencies which match those observed. Both potentials give far too few low frequency infrared-active lattice modes, and both predict several Raman-active external modes at higher frequency than those observed experimentally. The structure obtained through the use of the CHZBr2 analogue gives the poorest results, both in terms of the packing analysis and the lattice dynamics calculations. The volume per CF2C12 molecule obtained in the packing analysis is 110.8 A3/molecule, significantly larger than that obtained with the CH Cl structure. The static 2 2 lattice energy is also significantly higher than that obtained from the use of the CH2C12 structure. There are two possible explanations for the poor packing of CF2C12 into the CHZBr2 structure. First, it may be that the crystal symmetry constraints prevent efficient packing. This seems unlikely, since the point group symmetries of 107 the two molecules are identical. A second possibility is that the packing algorithm has iterated to a structure corresponding to a false minimum. That this is the more likely of the two possible explanations is supported by the appearance of imaginary frequencies in the lattice dynamics calculations. If the packing had resulted in a true equilibrium structure it seems likely that any arbitrary, small displacement would give the typical restoring forces, and real eigenvalues. This persistence of false minima has been previously noted by Williams [70,85]. The packing into the CF4 structure gave a reasonable volume per molecule and real albeit somewhat high lattice frequencies. The packing analysis yields reasonable CF2C12 structures for two of the three crystal symmetries adOpted by analogous molecules, and the lattice dynamics calculations can provide a good test of the quality of the packed structure. Both the lattice dynamics and packing calculations are sufficiently sensitive to provide predicted frequencies and structures which reflect the relatively small changes between potential parameter set IIb and potential parameter set VIb. There are, however, rather large deviations of the calculated frequencies from those observed. Without the necessary crystallographic and thermochemical data it is difficult to determine whether these deviations are due to incorrect structures, or to 108 bad potential parameter sets. If the calculation of lattice frequencies could be completed for tetragonal or hexagonal space groups, it would be of interest to perform the packing analysis with those space groups determined to be most consistent with the observed spectroscopic data. These would provide a more direct determination of the quality of the potential parameters. F. Spectroscopic Results-—CFC13 The infrared [32] and Raman [33] spectra of polycrystalline CFC13 at 77 K have been reported previously. The infrared spectrum from 500-1300 cm.1 of the matrix- isolated molecule has been reported, along with a calculation of the chlorine isotopic splittings [17]. The positions of the six fundamentals were measured and the assignment of the observed bands to their respective symmetries was made in each of the above investigations. The Raman-active lattice modes of polycrystalline CFC13 were also observed. The Raman spectrum of polycrystalline CFC13 at 15 K has been obtained in this laboratory and is shown in Figure 20. The Raman spectrum in the lattice region is reproduced in Figure 21. The measured peak positions and assignments are listed in Table 30. A detailed discussion of the assignments of the fundamentals is given in reference 33. It will not be repeated other than to note the large number of components for all fundamentals 109 . SC.“ ocHH 05.398 05 EOHM BmEm mum moflocmsqoum . 0H m3 maosHU UHHOm mo goommlg .ON mHDmHm .-Eu ..Eu .-Eu TE“. 3... .oNA ova so: men .on a on» own an: cos T...» 7.5 :13 «an a... 3o .3. on. no. 32 2.0. coo. .r l l l t. l l more mMDOE 44213.2. 110 LN’TICE MODES (31:013 .11. 8-. L .1 cm Figure 21. Raman spectrum of lattice region of CFCl3 (13 K). 111 251 TABLE 30. The Observed Raman Frequency Shifts of CC13F Solid at 15°K. Relative Av (cm—l) Av (cm-l) Peak Assignment Height This Work Shurvell (77°K) a1 W 1075.8 1075.8 m Fermi 1071.1 Sh a1 multiplet 1068.2 w sh involving 1066.2 sh V1,2V2 and W 1063.1 1063.1 w v4+v6 1060.6 w sh 35C13 100 535.4 536.7 vs 35c123701 61 532.9 534.1 s a1V2 35c137012 17 530.2 531.0 m 37013 4 527.5 528.2 vw 35C1§ 100 353.6 353.1 vs a1V3 ::c13:012 76 351.4 350.8 s c1 C12 26 348.9 348.2 m 37013 3 346.6 346.0 vw 7 843.0 factor 21 841.4 842.9 m group 71 838.2 839.2 s splitting 38 834.1 833.9 m alv4 with 64 822.5 822.6 3 possible 64 822.5 822.6 S LO-TO 100 816.8 816.9 S splitting Table 30 Continues. 112 Table 30 Continued. Relative Av (cm-l) Av (cm-1) Peak Assignment Height This Work Shurvell (77°K) C1 100 398.5 398.6 vs ev5 isotopic 46 394.4 394.8 m splitting 20 392.5 392.6 w 31 247.2 248.5 m Cl 100 245.5 246.6 vs ev6 isotopic 50 243.5 244.7 m splitting 13 241.0 242.0 w 240.6 w LATTICE MODES Infrared (13°K) Raman (15°K) 25.7 26.4 w 28.3 29.7 33.7 33.5 37.2 40.4 40.0 49.9 49.0 52.0 53.2 55.2 overlapped 59.6 59.7 62.0 66.0 113 and the relatively large splitting of the V4al multiplets. This is the only fundamental that clearly displays at least some factor group splitting. This splitting has previously been assigned as due in part to the longitudinal and transverse optical branches of the v4al mode. Such LO-TO splitting is permitted only for acentric unit cells. The far infrared spectrum of polycrystalline CFC13 at 13 K has been obtained and is shown in Figure 22. Nine prominent absorbances and at least three weak bands are observed between 25 and 66 cm-l, and their frequencies are also listed in Table 30. The infrared- and Raman-active lattice modes appear to be very nearly in coincidence, especially after making allowances for the broadness of the Raman bands. This supports the suggestion that CFC13 crystallizes in a noncentric unit cell. The temperature dependence of the infrared-active lattice modes was investigated and a plot of the band positions versus temperature is given in Figure 23. There are no unusual features that would suggest a phase change in the temperature range studied. G. Factor Group Analysis-—CFC13 Only those factor group-site group combinations which give the coincidence of at least some of the infrared- and Raman-active lattice modes were considered. Restricting the choice of site group-factor group combinations to those which give at least 9, but less than 15, infrared 114 crous 25 30 35 40 45 5O 55 60 65 70 lllllllllllllllllllllllllll'1llllll‘LlllllllllllllllllllllllllLlllllllilLLlllllllllllll|l Ission Relative Transm 'l‘l'l'l'l'l'l‘l‘l‘l‘lrl‘l'l'l'l’l‘l'l‘lW'l‘Fl‘lW‘l‘Fl'l‘l'l'l‘l‘l‘l'l‘l'l'l'l'l'l'l‘ 25 30 35 4o 45 50, 55 60 65 70 Wovenumber (omn1 ) Figure 22. The far-infrared spectrum of solid CC13F in the lattice region 115 CCI3F CLJFO! THROUGH 15 10 2O 30 4O 50 50 7O 80 90 100110120130 85 4.4IIIJIJIJIJIILLIIILIIIIII'IILIIIIIJllllllllllll 65 do n 60—: CDOOO 0° C 60 10QDOOO 00 O O ” o o - 21 o O _ H O O _ 55:00:)000 oo o o C_-_-55 .. o L. L aocoooo 00 o 0 o o L- a) .‘2 O L .0 soqomooooo o 0 L50 E Z 0 o : 3 .. 0 o .. C ‘45-: r45 9 r: c H L o 40-00300000 o o 0 ~40 3 : O O O C: d l- .35': L35 j003000 oo o o o o o _ q r- 30 OCDOO JO qomo O 00 o o [- .. O o - 25 'I’I'I'T'l‘l’l'l'l‘l'l'l‘l'l‘l'l'i'l‘l‘l'l'l'l 25 10 2O .30 4O 5O 60 70 BO 90 100110120130 temperature (Kelvin) Figure 23. The tarperature dependence of the nine Host intense infrared- active lattice nodes of €013 . 116 active lattice modes as well as giving less than 16 Raman active lattice modes, leaves nine possible combinations of site group and factor group. There are 21 space groups which can accommodate these nine combinations of site group and factor group. These are collected in Table 31. Recognition that the v4e fundamental displays some factor group components does not further restrict the number of possible space groups, as all the listed site group factor group combinations give 2, 3 or 4 factor group multiplets for this band. It is interesting to note that, as was the case with CFZClz, neither of the structures of the possible analogues, chloroform and bromoform in this instance give predicted activities consistent with the experimental observations for CC13F. Chloroform crystallizes on CS sites with a D2h factor group [86]. This combination results in the prediction of six infrared-active lattice modes, non-coincident with 12 Raman-active lattice modes. The bromoform structure [87] results in the prediction of two infrared-active lattice modes coincident with four Raman-active lattice modes. The suggested [33] combination of a CS site group with a C2V factor group results in the prediction of six infrared-active lattice modes, again contrary to the spectroscopic observations In summary, the far infrared lattice spectrum of CC13F clearly shows its structure to be different from either that of chloroform or bromoform. The previously-suggested 117 manoum macaw uamuomwwp HN > m >m ooH .soH .ooH a o o ocouoooHo a .ooH .ooH .ooH .soH .ooH o so N >mo mo ucouoooao N ooH .ooH .55H 8 mo No ocouoooao N oaH .NHH N No No ucouoomHo m ooH .ooH .ooH .moH N No Ho o .o a no N mo mo ocouoooHo N a .o .s .o o 9o N ha Ho o .o .m N No Ho H .oz N Ho Ho ocouoooHo N museum moomw aHnHmmom mo HmnEaz IIMNII macho Houomm maouw ouwm coHoooHoHocooH NHono mcoHumcHnEoo osouo uooocmuaoouo oon .Hm mamas 118 structure [33] is also unlikely, although the conclusion that the structure is acentric is supported by the present work. Twenty-one space groups which allow structures consistent with the lattice spectra have been determined. The large number of structures implicit in this number of space groups prohibits further testing of the interatomic potential parameter sets. Such calculations await the X-ray diffraction determination of at least the space group. H. SpectroscoPic Results-—CF3C1 No spectrosc0pic investigations of solid CF3C1 have appeared in the literature. The infrared spectrum of the gas phase is very well characterized [88-94]. The infrared spectrum of matrix isolated CF3C1 is also available [95], as is the gas phase Raman spectrum [96]. These works have been concerned primarily with obtaining molecular geometry parameters and force constants from the analysis of the rotational structure and band contours. Of primary interest here is that estimations of the magnitude of the chlorine isotopic splitting for all six of the fundamentals have been reported. The Raman spectrum of crystalline CF Cl is shown in 3 Figure 24. Chlorine isotopic splitting has been resolved only in the v3al mode. Only the v6 e mode shows splitting 119 EU CH aaHH maHuHoxw map scum mDMHSm mum moHoaoavon .AM ONV HUmmU GHHOm mo Esupommmlamemm .vN oHsmHm Q0\~ _U M\~ wn\~ . ON\~ .vm mom so... omc coo omro _.ss oms 853 .o S we; mm mm no.0. more. om: oo.N_ coho 120 due to crystal effects. The observed frequencies and assignments are given in Table 32. The far infrared lattice spectrum at 13 K has been obtained and is shown in Figure 25. There are four peaks in the range from 35 to 70 cm-l. The temperature dependence of the frequencies of these bands, shown in Figure 26, provides no indication of a phase change over the temperature range studied. The frequencies of the infrared-active lattice modes are listed together with those of the Raman-active bands in Table 32. The low intensity and broadness of the Raman active lattice modes makes the determination of their coincidence or lack of coincidence with the infrared active lattice modes difficult. The absence of a Raman band near 62 cm.1 suggests that at least some of the infrared bands are noncoincident with Raman bands. However, the number of observed Raman lattice bands falls short of the number of observed infrared bands. This can only mean that some Raman bands have not been located, probably because of their low intensity. Whether one of these low-intensity, missing bands could occur at 62 cm.1 is impossible to determine. Given the sparsity of the Raman lattice region, the factor group analysis would entirely depend on the infrared lattice spectrum. An analysis based on such few data would provide little illumination of the structure. It may only be said that the appearance of four infrared 121 m.Hw o.vm m.mv HIEO v.~v Axovav UOHMHMGH s.Nm o.Ho HLao m.oN HmooNuoNo chaos mamas mafiuumq v.Nmm HN .ooH N.H o.oom uooHsoco moo m.ovm oH .mm. H. m.mmm HN mm> .Nal N.o a.ooHH oH 66> mm“ 3 2 Km 5%.... 0.... r... s mow HN Husm .oo. H.H o.sss ooH HcNo .oa. om.A osoH m HoHs .uam A IEoV mcHuuHHmm HHIEOV uanam xowm ucaecmfimmd UH anamH ocHHOHnu coHuwmom w>HumHam mcwuuHHQm OHQODOmH aCHHoHsu pom .muamscmflmmm .maowufimom "AxomNIONV um HUmmU ocHHHmumxuowHom mo mHoucaEmpcam a>Huo< cmeom .Nm mqmda 122 CF3CI 2E5 436’ 2E5 4K) 445 E“) 2E5 6K) GES ‘7C> lllllllll‘llllllllllllllllllllllllllll'llllllllllllllllIll!llllllll‘lll|1|llllllllllllLl Ission r‘”\ Relative Transm WTI'I‘I'l'l'lTl'l‘l'l'l‘l‘l‘l'l‘l‘l‘l‘l‘l'l‘l'I‘IW‘I'I‘I‘T‘i'I'I‘l'I‘I'I‘I'T'I‘I‘I‘ITI‘ 25 3O 35 40 45 5O 55 60 65 70 Wavenumber (cm-1) Figure 25. The far-infrared spectrum of solid CF region (13 K) . 3C1 in the lattice 123 85 60 M 0| wavenumber Us O Q. VI 40 35 Figure 26. CF 30: 1O 20 30 4O 50 60 7O 80 90 L1 1 1441.4 I [J 1 IJ I l I l l 1 I 1 l_l_l_l l l | l l l l I _ :5 C) _ — O _ i ° 0 t . O o _ q C) . - r- . Eb C) E -4 O O . o I - % O O O O P -‘ O O o L- -J O r- - +- 4 t : ED A m A : .4 A O 40 . _‘ ‘A ,— - r j T I I I 11 I I I T—] I IjT V rj—Ij T 1 I I I I I I l I 10 20 30 4O 50 60 7O 80 90 temperature (Kelvin) 65 50 55 50 45 4O 35 The temperature dependence of the infrared-active lattice modes of CF Cl. 0, o - 0.13m 11 - 0.31nn 124 lattice modes indicates that there is more than one molecule in the unit cell. I. Conclusion The vibrational spectra of the polycrystalline chlorofluoromethanes obtained in this work have provided information about the symmetries of these crystals. For CF2C12 this specification of the crystal symmetry is and CF Cl it is very much less nearly unique; for CFC13 3 complete. The results of packing analysis and lattice dynamics calculations with the CF2C12 molecule have indicated that a reasonable range of lattice frequencies can be obtained from structures obtained with nonbonded interatomic potential parameter sets transferred from other molecules. Further tests of the quality of the potential sets were precluded by the inability of the lattice dynamics program to accept structures with Bravais lattice symmetry higher than orthorhombic. 125 APPENDIX CARBON-CARBON O—BUCKINGHAM A-U -O.O-3 -o.1—: o 3 0 -'.I g : 2 42‘: V j E -o.3—: E a 0- : —OA i .005.-[TI[IIIc111ITIITITIIIITYITIIIIIIIIIIIIII 3.0 3.5 4.0 4.5 5.0 RADIUS(A) Figure A-l. The 6-12 potential and fitted exp-6 potential. 126 CARBON-CHLORINE O-BUCKINGHAM A-LJ POTENTIAL (kJ/mol) RADIUS (A) Figure A-2. The 6-12 potential and fitted exp-6 potential. 127 CHLORINE-CHLORINE 'O-BUCKINGHAM A-LJ -m 3 ._J -o.3 -: Q .. O .— 3 5 ‘ -o. — 3 .4 V — é -o.7 — E -o.9 -: 0‘ .1 —1.1 5 _103 I I I I I I I I 1 I I I I l I I I I Ii I I I I I I I I I I I I I I I I I I I 3.0 3.5 4.0 4.5 5.0 RADIUS (A) Figure A-3. The 6-12 potential and fitted exp-6 potential. 128 cunnudummmatadinmwmmu Orb! LARGE CARBON MODEL -00 ? —o.1-E ¢> : ° 7". E : 2 "0-27: " : é-os-E rs l : —OA i -Oo5dIIIIIIIIIIIIIIIIIIIIIIII'FWIIIIjll'IIIII £0 45 50 55 60 RAWMS 00 Figure A-4. The 6-12 potential and fitted exp-6 potential. 129 I I I .o .o 9 N d O I .0 u POTENTIAL (kJ/mol) I I c: o 0‘ # I 9 a I .0 a llllIllllIIlllIIIllIHllIHIIIllllIIHlIllllIllllIllllIllllIllllIllllIHllIJLUI I 9 a: CARBON-CHLORINE O-BUCKINGHAM A-LJ LARGE CARBON MODEL 9 Figure A-5. O IIIIIIIIIIIII'IIII'IIIIIIIIIIIIIIIIIIII 4.5 5.0 5.5 RADIUS (A) The 6-12 potential and fitted exp-6 potential. 130 6.0 CHLORINE-CHLORINE O-BUCKINGHAM A-LJ LARGE CARBON MODEL -0. 3 i -03 L o .4 O .— E 05 I :2 ‘ - V _ g —07 j E -—0.9 — Q' L -1.1 —: -103 IIITIIIII—[IIIIIIIIIlIIII'IIIIIIjII'TIII] 3.0 3.5 4.0 4.5 5.0 RADIUS (A) Figure A-6. The 6-12 potential and fitted exp-6 potential. 131 LIST OF REFERENCES 10. ll. 12. l3. 14. 15. LIST OF REFERENCES D.H. Hornig, J. Chem. Phys. $2, 1063 (1948). H. Winston and R.S. Halford, J. Chem. Phys. 1;, 607 (1949). D.L. Rousseau, R.P. Bauman and S.P.S. Porto, J. Raman §pectrosc. $2, 253 (1981). M.P. Marzocchi and P. Manzelli, J. Chem. Phys. =;, 2630 (1970). M.P. Marzocchi, V. Schettino, and S. Califano, J. Chem. Phys. 3;, 1400 (1966). A. Kimoto and H. Yamada, Bull. Chem. Soc. Japan 3;, 1096 (1968). E. Burgos, E. Halac, and H. Bonadeo, J. Chem. Phys. g3, 1546 (1981). M.P. Marzocchi and S. Dobos, §pectrochim. Acta 0A, 1437 (1974). H.F. Shurvell, S.E. Gransden, J.A. Faniran, and D.W. James, Spectrochim. Acta 32A, 559 (1976). A. Anderson, B. Andrews, and B.H. Torrie, J. M01. Struct. g2, 409 (1982). H.F. Shurvell and D.W. James, J. Raman Spectrosc. éé' 28 (1983). E.B. Halac, E.M. Burgos, H. Bonadeo and E.A. Alessio, Acta Cryst. 5;}, 86 (1977). G.R. Elliott and G.E. Leroi, J. Chem. Phys. 32, 1217 (1973). K. Suzuki and K. Iguchi, J. Chem. Phys. =3, 4594, (1982). G. Taddei, H. Bonadeo, M.P. Marzocchi, and S. Califano, J. Chem. Phys. 2;, 966 (1973). 132 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. D.E. Williams, PCKS/PCK6, QCPE ;;, 373 (1979). C.F. Cook, W.B. Person and L.C. Hall, Spectrochim. Acta 23A, 1425 (1967). S.T. King, J. Chem. Phys. 49, 1321 (1968). A. Anderson, B.H. Torrie and W.S. Tse, Chem. Phys. Lett. 2;, 119 (1979). G.W. Chantry, H.A. Gebbie, B. Lassier and G. Wyllie, Nature 214, 163 (1967). M. Ito, §pectrochim. Acta ;;, 731 (1965). H.F. Shurvell, §pectrochim. Acta 27A, 2375 (1971). R.J.H. Clark and B.K. Hunter, J. Chem. Soc. 1_1, 2999 (1971). cm I.» U1 s S. Cohen, R. Powers and R. Rudman, Acta Cryst. 1670 (1979). J.A. Morrison and E.L. Richards, J. Chem. Thermodynamics g, 1033 (1976). G.J. Piermarini and A.B. Brown, J. Chem. Phys. 2g, 1974 (1973). C.E. Weir, G.J. Piermarini, and 5. Block, J. Chem. Phys; 22, 2089 (1969). Y.A. Sataty, A. Ron and F.H. Herbstein, J. Chem. Phys. 2;, 1094 (1975). R.P. Fournier, R. Savoie, F. Bessette and A. Cabana, J. Chem. Phys. 22, 1159 (1968). F. Bessette and A. Cabana, Spectrochim. Acta 25A, 157 (1969). S.C. Greer and L. Meyer, J. Chem. Phys. 2;, 4583 (1969). D.N. Bol'shutkin, V.M. Gasan, A.I. Prokhvatildov and A.I. Erenburg, Acta Cryst. B28, 3542 (1972). H.F. Shurvell, S.E. Gransden and D.W. James, Can. J. Spectrosc. ;Q, 27 (1975). H.F. Shurvell, J. Raman Spect. ;, 347 (1973). 133 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. H.C. Craig and T.W. Levin, Appl. Spectrosc. =;, 475 (1975). P.R. Griffiths, "Chemical Infrared Fourier Transform Spectroscopy", Wiley-Interscience, New York (1975). R.J. Bell, "Introductory Fourier Transform Spectroscopyf, Academic Press, New York (1972). J. Chamberlain, "Principles of Interferometric Spectroscopy", Wiley- Intersc1ence, New York (1979). M. Born and E. Wolf, "Principles of O tics", Third edition, Pergamon Press, New York (1965) P. Bloomfield, "Fourier Analysis of the Time Series: An Introduction", John WiIey and Sons, New York (1976). C. Giancaspro and M.B. Comisarow, Appl. Spectrosc. g;, 153 (1983). T.M. Niemczyk, Ph.D. Thesis, Michigan State University (1972). M.B. Comisarow and J.D. Melka, Anal. Chem. §_, 2198 (1979). —— J.F. Rabolt and R. Bellar, Appl. Spectrosc. 3_, 132 (1981). _— J. Bertie in "Analytical Applications of FT- IR to Molecular and Biological Systems , J. R. _Durig, ed., NATO Advanced Study Institute, Series C, Vol. 57, p. 25, D. Reidel Pub. Co., Dordrect, Holland (1980). R. Rudman and B. Post, Science 154, 1009 (1969). G. J. Piermarini and A. B. Braun, J. Chem. Phys. §_, 1974 (1973). ‘- B. Post, Acta Cryst. ;;, 349 (1959); K. Nishikawa, K. Tohji and'Y. Murata, J. Chem. Phys. 14, 5817 (1981). _— G. Herzberg, "Infrared and Raman S ectra", D. Van Nostrand, Princeton, New Jersey (1945) Y. Onodera and Y. Toyozawa, J. Phys. Soc. Japan 24 341 (1968). —— R. G. Whitfield and G. E. Leroi, J. Chem. Phys. =g, 2151 (1978) 134 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. I.W. Levin and W.C. Harris, J. Chem. Phys. =;, 2715 (1972). S Abramowitz and J.J. Comeford, Spectrochim. Acta , 1479 (1965). ||N H H. Bonadeo and E. D'Alessio, Chem. Phys. Lett. ;2, 117 (1973). A. Gavezzotti and M. Simonetta, Acta Cryst. A31, 645 (1975). I.R. McDonald, D.G. Bounds, and M.L. Klein, M01. Phys. 32, 521 (1982). J.B. Bates and W.R. Busing, J. Chem. Phys. g2, 2414 (1974). D.E. Williams, Acta Cryst. A28, 629 (1972). D.E. Williams, Acta Cryst. 27, 452 (1971). J.F.G. Hicks, J.G. Hooley and C.C. Stephenson, J. Amer. Chem. Soc. 2g, 1064 (1944). T. Atake and H. Chihara, J. Chem. Thermo. ;, 51 (1971). T.L. Hill, J. Chem. Phys. ;g, 399 (1948). R. Righini and M.L. Klein, Chem. Phys. Lett. g;, 604 (1982). ANAFAC, private communication, D. Cromer. TBON, private communication, G. Taddei and H. Bonadeo. DISTAN, private communication, Allan Zalkin, Lawrence Berkeley Lab., as modified by Don Ward, Michigan State University. A. Gavezzotti and M. Simonetta, Acta Cryst. 31, 645 (1975). T. Shinoda, Mol. Cryst. 5 101 (1979). =’ C.R. Brundle, M.B. Robin, H. Basch, J. Chem. Phys. 2;, 2196 (1970). A. Rosenberg and G. Birnbaum, J. Chem. Phys. 1396 (1968). ll“ 00 ‘ 135 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. M. Gilbert and M. Drifford, J. Chem. Phys. 66, 3205 (1977). ‘— M. Yvinec and R.M. Pick, J. Chem. Phys. ;;, 3440 (1979). G. Turrell, "Infrared and Raman Spectra g; Crystals", Academic Press, New York (1972). E.R. Bernstein and G.R. Meredith, J. Chem. Phys. 9;, 4132 (1972). F.P. Daly, A.G. Hopkins and C.W. Brown, Spectrochim. Acta 30A, 2159 (1974). A.C. Jeannotte II, D. Legler and J. Overend, Spectrochim. Acta 29A, 1915 (1973). S. Giorgianni, A. Gambi, L. Franco and S. Ghersetti, J. Mol. Spectrosc. ii, 389 (1979). M. Morillon-Chapey, A.G. Diallo and J.-C. Deroche, J. Mol. Spectrosc. fig, 424 (1981). L.H. Ngai and R.H. Mann, J. Mol. Spectrosc. =g, 322 (1971). W. Holzer and H. Moser, J. Mol. Spectrosc. g2, 188 (1966). H.H. Claassen, J. Chem. Phys. 2, 50 (1954). T. Kawaguchi, K. Tanaka, T. Takeuchi, and T. Watanabe, Bull. Chem. Soc. Japan gg, 62 (1973). T. Kawaguchi, A. Wakabayashi, M. Matsumoto, T. Takeuchi and T. Watanabe, Bull. Chem. Soc. Japan ég, 57 (1973). D.R. Stull and H. Prophet, JANAF Thermochemical Tables, NSRDS-NBS 37 (1971) and subsequent supple- ments. D.E. Williams, Acta Cryst. A25, 464 (1969). R. Fourme and M. Renaud, C.R. Acad. Sc. Paris, Series B 263, 69 (1969). T. Kawaguchi, K. Takashina, T. Tanaka, and T. Watanabe, Acta Cryst. B28, 967 (1972). L.C. Hoskins and C.J. Lee, J. Chem. Phys. 2:, 4932 (1973). 89. 90. 91. 92. 93. 94. 95. 96. W.G. Golden, D.A. Horner and J. Overend, J. Chem. Phys. 2;, 964 (1978). S. Biedermann, H. Burner, A. Ruoff, C. Alamichel and M. Morillon, J. Mol. Spectrosc. 22, 367 (1975). K. Scanlon, I. Suzuki and J. Overend, J. Chem. Phys. ;5 3735 (1981). A. Ruoff and H. Burger, Spectrochim. Acta 6A, 171 (1980). S. Giorgianni, A. Gambi, R. Visinoni and S. Ghersetti, J. Mol. Spectrosc. 2;, 55 (1982). H. Burger, K. Burczyk, R. Grassow and A. Ruoff, J. Mol. Spectrosc. 2;, 55 (1982). H. Burger, K. Burczyk, D. Bielefeldt, H. Willner, A. Ruoff and K. Molt, Spectrochim. Acta 35A, 875 (1979). R.J.H. Clark and O.H. Ellestad, J. Mol. Spectrosc. 22, 386 (1975). 137 “'71))ng)uj))())))is