1055J. CHEM.SOC. FARADAY TRANS., 1994, 90(8), 1055-1060 High-resolution FTIR-Jet Spectroscopy of CCI,F, Don McNaughton,* Don McGilvery and Evan G. Robertson Centre for High Resolution Spectroscopy and Optoelectronic Technology, Department of Chemistry, Monash University, Wellington Road, Clayton, Victoria, Australia 3 168 An experimental system employing a cryopump has been assembled for high-resolution FTIR spectroscopy of species cooled in a supersonic jet expansion. Coupled to a Bruker IFS 120HR interferometer, it involves an external sample chamber and transfer optics allowing 11 optical passes of the infrared beam. Tests with CO and N20 indicate rotational temperatures in the range 6-30 K. By heating the nozzle, significant population of vibra- tional states can be achieved without sacrificing rotational cooling.The mid-infrared spectrum of Freon-12, CC12F2, has been recorded and the v8 C-type band at 1161 cm-' analysed. Rotational constants for C3%12F2 and C37C135CIF, have been calculated and the rotational temperature estimated at 40 K. Introduction Supersonic Jet Spectroscopy The development of high-resolution FTIR spectrometers has greatly increased the capability to explore the vibration- rotation structure of molecules. The size and complexity of molecules amenable to study by FTIR techniques is limited by spectral congestion and the problem of assignment of overlapping transitions. Free expansion of gases in a super- sonic jet achieves significant cooling within the degrees of freedom of the molecules (Trans< T,,< qib).This relaxation greatly increases the population of molecules in lower rota- tional energy levels at the expense of higher energy levels.The advantages this has for spectroscopy include simplifica- tion of the spectrum and enhancement of the signal intensity of low J, K transitions. A wide variety of spectroscopic tech- niques have been used in conjunction with supersonic jets.' FTIR spectroscopy, with the advantage of broad spectral coverage, was first used to probe a supersonic expansion by Snavely et al.2-5 who recorded spectra at 0.06 cm-' resolution. Quack and co-workers developed the technique, using a Bomem spectrometer to study a variety of molecules at up to 0.004 an-' resolution.'*6-'6 Barnes et al.investi-gated molecular clusters in a molecular jet' 7*18and more recently other groups have made use of FTIR-jet spectros-copy to study semi-stable species," the CN radical produced in a corona discharge2' and other species., 1-24 Dichlorodifluorornethane(Freon-12) Freon-12, or CCl,F, ,is heavily used in industry as a refriger- ant, aerosol, sterilant and in plastic foams. Of great concern is its effect on the ozone layer. A stable molecule, it has a long lifetime, and is not broken down in the troposphere. In the higher stratosphere it undergoes photolysis by solar UV radi- ation, releasing chlorine atoms capable of catalysing the breakdown of ozone.25 Freon-12 is estimated to account for over 40% of ozone depletion.26 Extensive use has been made of infrared spectroscopy in the detection of atmospheric CCI,F, .27 The molecular constants obtained from high-resolution measurements are useful in modelling the tem- perature profiles of atmospheric pollutants like Freon- 12.Several microwave studies have been carried out on CCl,F, , resulting in accurately known ground-state con-stants for the three most abundant isotopic forms, C35C1,F,, C37C135C1F2 and C37C1,F, These forms occur in the approximate ratios (9 :6 : 1). C37C135C1F, has C, symmetry while the other two isotopic species belong to the C,, point group. The principal axis orientation for CCl,F, is shown in Fig. 1. Apart from intensity measurements and atmospheric studies, several infrared investigations have focussed on vibration-rotation structure.A medium-resolution (0.1 cm -I) band contour was followed by an analysis of the v6 and v8 Q branches, identifying fundamentals and hot bands for the two most abundant isotopic species.34 In another band contour study, hindered by insufficient resolution, incorrect conclusions about band types were drawn.3 The 923 cm-' v6 band of CCI,F, (asymmetric CCI, stretch) has been studied in some detail by infared-microwave double and by diode laser spectroscopy of the mol- ecules in a jet.27 A diode laser was also used to examine under high resolution the rovibrational structure of the 1101 cm-' v1 band of CCl,F, (symmetric CF, stretch) at 200 K.39 Some measurements of the Q branch heads of v8 have been made,40 also with a diode laser, but no vibration-rotation analysis of this band has been published.We report the development of a multi-pass transfer optic system for FTIR spectroscopy of jet-cooled species and the subsequent high-resolution spectral analysis of the v8 C-type band of CCl,F, (asymmetric CF, stretch) at 1161 an-'. Design and Experimental Our apparatus has undergone some modification since we first experimented with jet-cooled spectroscopy. The first design involved mounting a Varian HV12 cryopump hori- zontally to the side of the sample chamber of the Bruker IFS 120HR interferometer and positioning the nozzle inside the instrument's sample chamber so that the FTIR beam focussed in the cold region of the jet.Like Ballard et ~1 a .~~ cryopump was chosen for its cost effectiveness and reduced size in comparison with the more conventional oil diffusion pump, roots blower pump and backing pump combination. Unfortunately, with this arrangement, the mechanical vibra- tions associated with the pump coupled to the spectrometer optics and resulted in spectral artifacts and increased noise. For this reason it was necessary to run the instrument in purge mode with the cryopump uncoupled. Although the signal to noise ratios of the spectra of simple molecules obtained in this manner were adequate for diagnostics, the centre of mass b Fig. 1 Principal axis orientation for CCI,F, spectra of more complex molecules suffered from the inherent low sensitivity of FTIR spectroscopy.An external sample chamber and transfer optics box were then constructed in order to decouple the cryopump physically from the interfer- ometer and allow a multiple pass of the infrared beam through the cold-jet region. After some experimentation with a double-pass system, a mirror assembly was built to allow 11 passes. The apparatus, shown in Fig. 2,includes an air cushion pressurized to ca. 500 kPa, effectively damping out most of the mechanical vibra- tions associated with our cryopump. The entire external system is also vibrationally isolated from the interferometer by a flexible bellows. Circular pinhole nozzles were manufac- tured from 6 mm Pyrex glass tubing and may be heated to 550 K by a heating jacket.The mirrors transferring the beam out of the interferometer sample chamber are positioned to give a slight beam waist in order to maintain minimal beam diameter throughout the transfer optic system. All the trans- fer mirrors except two in the external chamber may be tilted vertically and horizontally. The final detector focussing mirror has a focal length of 33 mm, serving to focus the nearly parallel beam of ca. 25 mm diameter to a small region on the Judson Infrared inc HgCdTe-InSb detector. The jet nozzle chamber is isolated from the spectrometer, optics box and detector by a KBr window. Most of the test spectra of N20 and CO were collected using the initially designed internally positioned jet, with backing pressures of ca.300 kPa and nozzle diameters in the range 120-240 pm. The room-temperature spectrum of CCl,F2 at 0.01 m-' unapodized resolution was recorded using a pressure of 13 Pa in an Infrared Analysis Inc. multiple-pass cell with path- length set at 4 m. Jet-cooled spectra of CCl,F, were recorded at 0.0034 cm-' unapodized resolution using a 700-1300 cm-' filter. Nozzle height was set so that the nozzle just began to impinge on the IR beam, a position which was found to maximize signal absorption by the sample gas. Commercially available CC12F2 was used at a stagnation pressure of 600 kPa with an unheated 240 pm diameter nozzle, resulting in a flow rate that actually overloaded the cryopump. Therefore, it was necessary to isolate the cryo- pump from the sample gas for ca.45 min between sets of two to four scans so that the pressure measured in the lower part of the chamber did not exceed 0.1Pa. 80 scans were co-added in this way. This arrangement was chosen in order to maxi- mize sample density and hence absorption for the weak P and R structure of the v8 band. A total of ca. 1.5 kg of Freon- 12 was used over the 3 h of scanning time required for this experiment. Such large quantities are often required for experiments of this nature. For this reason our system includes the facility for a liquid-nitrogen trap between the cryopump and the backing pump to enable recycling of gases. Results and Discussion Performance Testing Preliminary testing with CO in the original single-pass chamber achieved rotational temperatures down to 6 K by seeding with argon.Temperatures were obtained from Boltz- mann plots of ln{(J + J' + 1);) us. E. Varying the distance of the focussed FTIR beam to nozzle aperture yielded results similar to those of Quack and co-workers,' and the minimum temperature was observed at a distance of around 12 mm for a 110 pm nozzle. Cooling for N20 was not as efficient. Spectra of the vJ band of N20 recorded under different con- ditions are shown in Fig. 3. Fig. 3(a)is a spectrum of pure N20, while the spectrum in Fig. 3(b) is of N20 diluted to 30% in argon. Rotational temperatures in the ground vibra- J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 -sample inlet nozzle height adjustmentnozzle horizontal position adjustment -L rotary pump compressed air detector I interaction external adjustment af vertical and bo horizontal slant for 1st and last mirror Fig.2 FTIR-Jet experimental apparatus. (a) Side view. H, 100 R, 10 W resistive heating jacket. B, Flexible bellows. T, x-y translation stage. (b) Top view. F1, 121 mm focal length off-axis parabolic reflec- tor. F2, 33 mm focal length off-axis paraboloid mounted on x-y translation stage. Ml-M4, plane mirrors. F1 and M2 are mounted on an x translation state. tional state for these two spectra are calculated to be 27.0& 2.1 and 19.7f0.8K,respectively. As may be seen by comparison of Fig. 3(a) and (b), dilution in argon assists in the cooling of N20, but at the expense of the signal-to-noise ratio.This enhanced cooling with argon dilution can be J. CHEM SOC. FARADAY TRANS 1994, VOL. 90 Q) C c s1n m t r 1 me s1n 2240 2230 2220 2210 2200 wavenurnber/cm-' Fig. 3 v3 band of N20 recorded with the first experimental appar- atus. (a) Pure N20 at 300 kPa backing pressure, 180 pm unheated nozzle, 0.01 cm-' resolution, five scans. (b) Same conditions as (a) except N20 diluted to 30% in Ar. (c) Pure N20 at 300 kPa backing pressure, 120 pm nozzle heated to 180°C, 0.05 cm-' resolution, 10 scans. attributed to more efficient collisional relaxation. The spec- trum in Fig. 3(c) is of neat N20 where the nozzle has been heated to ca. 450 K. The (0110)-(0111)hot band centred at 2209.5 cm-is clearly evident.The rotational temperature is calculated to be 20.7 f0.7 K in the ground vibrational state and 23.6 f0.5 K for the v2 = 1 excited vibrational state. It is clear that the final rotational temperature is not greatly increased by moderate heating of the nozzle. By contrast the vibrational temperature, calculated from the intensity of the hot-band transitions relative to the corresponding fundamen- tal transitions, is 488 f32 K.For the unheated nozzle in Fig. 3(a) a vibrational temperature of 204 f15 K is obtained, a result comparable to the value 213 f7 K obtained by Wall- raff et al. for OCS.41 These calculations are only approx- imate, in that they assume a Boltzmann-type vibrational population distribution, whereas the vibrational states are far from being in equilibrium.Not only do the translational, rotational and vibrational temperatures in molecules cool to different extents in a free expansion, but there is evidence that the extent of vibrational cooling depends on the nature of the ~ibration.~'A torsional mode, for example, appears to couple strongly with rotation and cools almost as efficiently as rota- tional degrees of freedom.22 When the observation is taken into account that in our system the unheated nozzles them- selves tend to cool to below 263 K, it is evident that the spectra of Fig. 3(a)and (c) show little evidence of substantial vibrational cooling. This clearly demonstrates the inefficiency of vibrational cooling in comparison with rotational cooling, while showing that some heating of the nozzle should prove advantageous for the study of rotationally cool hot bands.Dichlorodifluoromethane The room-temperature spectrum and jet-cooled spectrum of the 1161 cm-' C-type band of CCl,F, are shown in Fig. qa) and (b).Those Q branches which appear only in the room- temperature spectrum are from vibrational hot bands.36 Q, C +.It.-l fsi 2 Y ,1150 1160 117d',, 1180 ,' wavenurnber/crn-' '\, 1155 1160 1165 1170 wavenurnber/cm-' Fig. 4 (a) Room-temperature spectrum of CC12F, at 0.01 cm-' unapodized resolution. (b) Observed profile of the vg band of jet-cooled CC12F2 at 0.0034 cm- ' unapodized resolution. (c) Simulated band profile based on addition of the C35C12F, and C37C135C1F2 predictions at 40K.Their absence in the jet-cooled spectrum is evidence for some degree of vibrational cooling for this molecule. This is con- firmed by the observation that in the v6 region of the cool spectrum around 923 cm-' the only hot-band Q branches observed are from v4, the lowest-frequency fundamental. A section of the cool spectrum is shown in Fig. 5. Assignment of transitions was achieved through Macloomis, an interactive program which displays peaks in Loomis-Wood (Fortran) format.42 This technique is suc-cessful for regularly spaced series of transitions. Developed for the spectral analysis of linear and symmetric-top mol- ecules, it has also been applied to asymmetric tops near the symmetric limit such as vinylamine (IC= -0.93).43 Note that even for a molecule as asymmetric as CC12F2 (K = -0.57) regular structure exists for higher values of K,, as may be J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 series immediately above corresponds to PP, transitions of C3’C13’C1F2. Series are displayed with different colours to distinguish them easily from each other, an essential feature when assigning densely overlapped spectra like those shown. The separation between transitions within a series corre- sponds to ca. B+ C.Because of the small difference in rota- tional constants of the two isotopomers, their series appear at different angles in a Macloomis plot. This regular structure XI I I allows a relatively straightfoward assignment of transitions using Macloomis. Transitions that appeared overlapped on the Macloomis display or asymmetry doublets calculated to be split by more than O.OOO4 cm-’ were assigned an experi- mental uncertainty of 0.002 cm-’,otherwise an uncertainty of O.OOO4 cm-’ (10% of linewidth) was used in the weighted least-squares fit.Following the assignment of transitions using Macloomis, 1156.8 1157.0 1157.2 1157.4 1157.6 the data were transferred to an asymmetric rotor fitting waven urnber/crn-’ program. Watson’s A-reduced Hamilt~nian~~ was employed Fig. 5 Upper trace; Part of the V’ band of CCl,F,. Lower trace; and the ground-state constants held to those of ref. 32,simulated spectrum based on the C3’Cl,F, and C37C135C1F2 iso- resulting in the rotational and centrifugal distortion con-topes only.stants given in Table 1. The magnitudes of the standard devi- clearly seen from the Macloomis plot in Plate 1. The red pP16 ations of the fits are less than 1.0, indicating a good fit to the series (A& = -1, AJ = -1, K: = 16) of C3’C1,F, is dis- data. This implies that the constants will accurately repro- played vertically in the centre, while the dark-blue diagonal duce transition frequencies, which is important if the con- Table 1 Fitted rotational and distortion constants (in an-’)for the vg band of CCl,F, using Watson’s A-reduced Hamiltonian vibrational state parameter/cm -C3’C1,F, C37C1 ’ClF, Vob 1 16 1.08498 (4) 1160.96467 (6) A -Bb 0.05646622 (59) 0.05730849 (104) Bb 0.08 123097 (43) 0.07948650 (71) B -Cb 0.01353579 (266) 0.01326170 (438) A 0.13769719 0.13679499 B 0.08799886 0.0861 1735 C 0.07446307 0.07285565 A,b 1.308 (66) x lov8 1.424 (111) x lo-’ AJKb -1.109 (178) x lo-’ -9.45 (294) x 10-9 AKb 6.021 (145) x lo-’ 4.995 (265) x lo-’ 8,’ 3.649558 x lo-’ 3.519435 x 6,‘ 4.461253 x lo-’ 4.789547 x lo-’ ground state A 0.1373908212 0.1364949781 B 0.0880166705 0.0861337079 C 0.0745079 1 17 0.072898585 1 AJ 1.495191 x lo-’ 1.441514 x lo-’ AJK -1.477762 x -1.458105 x lo-’ A, 5.284089 x lo-’ 5.281053 x lo-’ 6, 3.649558 x lo-’ 3.519435 x lo-’ 4.461253 x lo-’ 4.789547 x lo-’ 8,number of transitions 817 556 standard deviation of fit 0.8304 0.9901 correlation matrices: ~ ~~ C ’C1 ,F, VO A-B B B-C AJ AJK A, VO 100 3 -47 2 -33 2 -3 A-B 3 100 -82 52 -41 16 30 B -47 -82 100 -63 45 -6 -25 B-C 2 52 -63 100 21 -35 40 AJ -33 -41 45 21 100 -86 59 AJK 2 16 -6 -35 -86 100 -85 AK -3 30 -25 40 59 -85 100 ~~ C3 7C1 ’Cl F, VO A-B B B-C AJ AJK AK 100 -15 -28 -14 -38 19 -25VO A-B -15 100 -85 70 -26 -2 49 B -28 -85 100 -72 40 -2 -34 B-C -14 70 -72 100 23 -46 61 AJ -38 -26 40 23 100 -88 59 AJK 19 -2 -2 -46 -88 100 -85 AK -25 49 -34 61 59 -85 100 Ref.44.The numbers in parentheses are one standard deviation according to the least-squares fit in units of the least significant figure quoted. Ground-state constants are constrained to those of ref. 32. Note B = +(I3 + C). Correlation matrices for the fits are presented below the table.Fitted parameters. ‘Constrained to the ground-state value. Plate 1 Loomis-Wood diagram of the spectrum of CCl,F, in the region 1154.1-1159.9 cm-’. The PP,, series of C35C1,F, is displayed vertically in the centre, the diagonal series immediately above corresponds to pP,ztransitions of C3’C13’C1F2. D. McNaughton et al. (Facing p. 1058) J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 stants are to be used to model band contours for temperature analysis. Calculated rotational constants have been used to simulate the spectrum as a check on the assignment and in order to measure to rotational temperature in our spectrum. Although centrifugal distortion was not included, the spectrum is none the less well reproduced, as illustrated by the plots in Fig.4(c) and Fig. 5. The few spectral features that are unaccounted for are probably due to the less abundant C37C12F2 isotopomer. The only anomalous values amongst the fitted parameters are the A rotational constants for the v8 state which are 0.0003 cm-’ greater than the ground-state values for both isotopic species. This apparent increase in A rather than the expected decrease associated with a stretching mode may result from an a-axis Coriolis interaction with a nearby vibrational level such as (v2 + v3) or possibly a &-dependent Fermi resonance with a vibrational level such as (v2 + vg). These energy levels are shown schematically in ref. 39. The band centres for C3’C1,F2 and C3’C13’C1F2, calculated at 1161.08498 (4) and 1160.96467 (6) cm-’,respectively, are consistent with the pre- viously reported Q branch positions of 1161.07 and 1160.95 cm -from the medium-resolution Q branch analysis.36 The rotational temperature in the ground vibrational state was estimated at 40 K by comparison of the observed and calculated spectra using Q branch widths and P and R branch profiles.Lower temperatures could have been achieved through seeding in argon, but this would have the undesirable effect of decreasing absorption strength, and a rotational temperature of 40 K proved cool enough to resolve most of the rotational structure at 0.0034 cm-’. Although the simulated spectrum, calculated on the basis of a Boltzmann distribution at 40 K, shows little deviation from the observed spectrum, the apparent population distribution in the ground vibrational state should not be Boltzmann-like for a number of reasons.First, a small amount of warmer background gas is present in the sample chamber. Secondly, the design of the external sample chamber with its parallel beam means that the interaction zone is broad and molecules will exhibit a range of temperatures since they are sampled at a range of distances and angles from the nozzle aperture. Thirdly, relaxation is not complete in a molecular beam, so that even if the jet was probed at a well defined distance from the nozzle, the distribution would not be Boltzmann-like. Collision-induced rotational relaxation has been found to follow the dipole selection rules AJ = 0, f.1, AK = 0 and + ++-for symmetric tops.4s As Snavely et al.pointed out,2 the AK = 0 rule tends to cause molecules to relax within the same K-stack, and because the AK # 0 transitions are slower, the lowest J levels of each K-stack would be expected to be populated significantly. This is consistent with the observed pattern. Transitions are observed for relatively high K, values (up to K, = 25 in the ground state), but only corresponding to the lowest J levels in each K,-stack. In order to determine the rovibrational constants of the v8 band more accurately, higher J and K transitions might be obtained from room- temperature spectra and combined with those of the present study. Supplementary tables of assigned transitions of CC12F2 are available from the authors or from the British Library.? Conclusion An apparatus for high-resolution FTIR-jet spectroscopy has been constructed involving a cryopump and a multiple-pass configuration for greater signal absorption.Reasonable t Supplementary publication no. SUP 57001 (22 pp.), deposited with the British Library. Details are available from the Editorial Office. signal-to-noise spectra were recorded for the v8 band of CC12F2 despite the weak P and R structure of this band rela- tive to v1 and v6, the other fundamentals in this spectral region. The v8 band has been assigned using a Loomis-Wood interactive fitting program to select out the regular structure that occurs at higher values of K,. Molecular constants have been obtained through weighted least-squares fits of the observed transitions and used successfully to simulate the v8 band profile.This demonstrates the usefulness of FTIR-jet spectroscopy for resolving complicated overlapping tran-sitions for moderately sized molecules. Significant vibrational population can be achieved in a supersonic jet by moderate heating of the nozzle without adversely affecting rotational cooling. The financial assistance of the Australian Research Council and the Sir James McNeill Foundation is gratefully acknow- ledged. References 1 A. Amrein, M. Quack and U. Schmitt, J. Phys. Chem., 1988, 92, 5455. 2 D. L. Snavely, S. D. Colson and K. B. Wiberg, J. Chem. Phys., 1981,74,6975. 3 D. L.Snavely, K. B. Wiberg and S. D. Colson, Chem. Phys. Lett., 1983,%, 319. 4 D. L. Snavely, V. A. Walters, S. D. Colson and K. B. Wiberg, Chem. Phys. Lett., 1984, 103,423. 5 V. A. Walters, D. L. Snavely, S. D. Colson, K. B. Wiberg and K. N. Wong, J. Phys. Chem., 1986,90,592. 6 H. R. Dubal, M. Quack and U. Schmitt, Chimia, 1984,38,438. 7 A. Amrein, M. Quack and U. Schmitt, Mol. Phys., 1987,60,237. 8 A. Amrein, M. Quack and U. Schmitt, Z. Phys. Chem., 1987,154, 59. 9 A. Amrein, H. Hollenstein, P. Locher, M. Quack and U. Schmitt, Chem. Phys. Lett., 1987, 139,82. 10 A. Amrein, D. Luckhaus, F. Merkt and M. Quack, Chem. Phys. Lett., 1988,152,275. 11 A. Amrein, H. Hollenstein, M. Quack and U. Schmitt, Injured Phys., 1989,29, 561. 12 A. J. Ross, A.Amrein, D. Luckhaus and M. Quack, Mol. Phys., 1989,66,1273. 13 H. Burger, A. Rahner, A. Amrein, H. Hollenstein and M. Quack, Chem. Phys. Lett., 1989,156,557. 14 M. Quack, Annu. Rev. Phys. Chem., 1990,41,839. 15 M. Snels and M. Quack, J. Phys. Chem., 1991,%, 6355. 16 M. Quack, U. Schmitt and M. A. Suhm, Chem. Phys. Lett., 1993, 208,446. 17 J. A. Barnes and T. E. Gough, Chem. Phys. Lett., 1986,130,297. 18 J. A. Barnes and T. E. Gough, J. Chem. Phys., 1987,86,6012. 19 A. D. Walters, M. Winnewisser, K. Lattner and B. Winnewisser, J. Mol. Spectrosc., 1991, 149, 542. 20 B. D. Rehfuss, M. H. Suh, T. A. Miller and V. E. Bondybey, J. Mol. Spectrosc., 1992, 151,437. 21 J. K. Holland, M. Carleer, R. Petrisse and M. Herman, Chem. Phys. Lett., 1992, 194, 175.22 F. Melen, M. Carleer and M.Herman, Chem. Phys. Lett., 1992, 199, 124. 23 F. Melen, M. Herman, G. Y. Matti and D. McNaughton, J. Mol. Spectrosc., 1993,160, 601. 24 J. Ballard, D. Newnham and M. Page, Chem. Phys. Lett., 1993, 208,295. 25 F. S. Rowland, Annu. Rev. Phys. Chem., 1991,42,731. 26 D. G. Cogan, Stones in a Glass House, Investor Responsibility Research Centre Inc., Washington, DC, 1988, p. 37. 27 M. Snels and W. L. Meerts, Appl. Phys. B, 1988, 45, 27, and references therein. 28 F. Y. KO,F. Tsai and E. L. Beeson Jr., Bull. Am Phys. SOC., Ser. II, 1968, 13,253. 29 C. F. Su and E. L. Beeson Jr., J. Chem. Phys., 1977,1,330. 30 H. Takeo and C. Matsumura, Bull. Chem. SOC., Jpn., 1977, 50, 636. 31 R. W. Davis and M.C. L. Gerry, J. Mol. Spectrosc., 1983, 101, 167. 1060 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 32 R. A. Booker and F. C. De Lucia, J. Mol. Spectrosc., 1986, 118, 40 J. Orphal, B. Sumpf, K. Herrmann, V. V. Pustogov, F. Kuhne-548. mann, Proc. XII Int. Conf. High Res. Infiared and Mocrowave 33 S. Giorgianni, A. Gambi, L. Franco and S. Ghersetti, J. Mol. Spec., 1992, C19. Spectrosc., 1979, 75, 389. 41 P. Wallraff, K. M. T. Yamada and G. Winnewisser, J. Mol. 34 M. Morillon-Chapey, A. 0. Diallo and J-C. Deroche, J. Mol. Spectrosc., 1987, 126, 78. Spectrosc., 1981, 88,424. 42 D. McNaughton, D. McGilvery and F. Shanks, J. Mol. Spec-35 H. C. Jung, Bull. Korean Chem. Soc., 1988,9,275. trosc., 1991, 149, 458. 36 H. Jones and M. Morillon-Chapey, J. Mol. Spectrosc., 1982, 91, 43 D. McNaughton and E. G. Robertson, J. Mol. Spectrosc., 1994, 87. 163, 80. 37 H. Jones, G. Taubmann and M. Morillon-Chapey, J. Mol. Spec-44 J. K. G. Watson, in Vibrational Spectra and Structure, ed. J. R. trosc., 1985, 111, 179. Durig, Elsevier, Amsterdam, 1977, vol. 6, p. 1. 38 G. Taubmann and H. Jones, J. Mol. Spectrosc., 1986,117,283. 45 T. Oka, J. Chem. Phys., 1968,48,4919. 39 S. Giorgianni, A. Gambi, A. Baldacci, A. De Lorenzi and S. Ghersetti, J. Mol. Spectrosc., 1990, 144,230. Paper 3/07081F; Received 30th Nouernber, 1993