J. CHEM. SOC.FARADAY TRANS., 1994, 90(18), 2601-2610 2601 Microwave Spectrum of Sulfuryl Chloride Fluoride, S0,CIF : Structure, Hyperfine Constants and Harmonic Force Field Holger S. P. Muller and Michael C. L. Gerry Department of Chemistry, The University of British Columbia, 2036 Main Mall, Vancouver, B.C., Canada V6T 1Zl ~~ The pure rotational spectrum of sulfuryl chloride fluoride, SO,CIF, has been investigated in the frequency range 5.6-24.0 GHz using a pulsed molecular beam microwave Fourier-transform spectrometer. Between 48 and 170 lines of 8 to 26 rotational transitions have been observed for the six isotopomers SO,CIF, SO, 37CIF, 34S0,CIF, 34S0, 37CIF, SO'80CIF and SO "0 37CIF (unlabelled atoms indicate l60,"F, 32S and 35CI) in natural isotopic abundance.The rotational and quartic centrifugal distortion constants have been determined. ro TAP and r,-type I structural parameters have been evaluated. A harmonic force field has been calculated to derive ground-state average and estimated equilibrium geometries. The description of the normal modes in terms of internal coordi- nates is discussed. Chlorine and fluorine hyperfine structures have been resolved, allowing quadrupole coup- ling (including zbcfor the isotopomers containing "0) and spin-rotation constants to be determined. Variations in the chlorine quadrupole coupling constants with different isotopomers have allowed the CI quadrupole tensor to be diagonalized and indicate that its z-axis coincides with the SCI bond. Although sulfuryl chloride fluoride, SO,ClF, has been known for more than 50 years,' there is only one study, by Holt and Gerry, reporting its structure in the gas phase., Using a con- ventional Stark modulated microwave spectrometer, they determined rotational constants, chlorine quadrupole coup- ling constants and an effective distortion constant for each of the two most abundant isotopomers, S0,ClF and SO, 37ClF (unlabelled atoms indicate l60,19F, 32S and 35Cl). By assuming values for r(S0)and L(C1SF) from a comparison with the related molecules sulfuryl fluoride, SO,F,, and sul- fury1 chloride, S02C12, they were able to derive the remain- ing geometrical parameters.Recently Gombler investigated 32/34Sisotopic shifts of the "F NMR frequency of sulfur- and fluorine-containing com- pound~.~From these results he suggested that the SF bond in S02F2 should be longer than that in SO,ClF, in contrast to the conclusions from the microwave studies.More recently a single-crystal X-ray diffraction study of S0,XY (X, Y = C1, F) at low temperatures has yielded the precise geometrical parameters of these compounds in the solid state.4 Raman and IR spectra have also been re~orded.'.~ In addi- tion, Pfeiffer obtained a force field of S0,ClF by transferring force constants and structural parameters from S02C1, and S02F2.7 These force constants reproduced most of the observed vibrational wavenumbers moderately well. Better agreement was obtained when the assignments of two pairs of fundamentals in ref.5 were exchanged. In addition, the earlier assignment of the v6 band' was rejected and it was assumed that this fundamental is overlapped by v9. The assignments of Pfeiffer have been confirmed in a recent reinvestigation of the Raman and IR spectra.' At room temperature the microwave spectrum of S0,ClF is very dense, with many transitions between rotational levels at high J.However, it has recently been shown that, at the low rotational temperatures achievable in a supersonic jet (ca. 1 K), most of these rotational transitions have a very low inten~ity,~making the lines of the rarer isotopomers measur- able. The basic aim of the present work has thus been to measure transitions of several rare isotopic species (containing ''0 and 34S), using a cavity pulsed microwave Fourier-transform (MWFT) spectrometer," into which samples are injected as pulsed supersonic jets.From the resulting rotational constants the complete geometry has been derived. In addition, because of the high resolution available with the MWFT spectrometer, the hyperfine con- stants have been substantially improved. Finally, the harmo- nic force field has been refined using the centrifugal distortion constants. Experimental The spectra were measured in the frequency range 4-25 GHz using a cavity pulsed MWFT spectrometer. Samples were injected as pulsed jets of gas consisting of ca. 0.5-1% S0,ClF in neon at 1500-3000 mbar total pressure; rota- tional temperatures achieved were 51 K.Linewidths as low as 7 kHz full width at half maximum were obtained. It was possible to resolve (at least partially) lines ca. 5 kHz apart. The precision and accuracy of measurements for strong, well resolved lines are believed to be about f0.5 and f1 kHz, respectively. In order to minimize distortions due to overlap effects in power spectra, the frequencies of closely spaced lines were determined by fits to the time-domain ('decay') signals. The sample of S0,ClF was kindly provided by Prof. Dr. H. Willner, Universitat Hanover, Germany. It had been obtained by the combined synthesis of SO,F, and S0,ClF from S02C12 and NaF.13 Results Observed Spectra and Analysis S0,ClF is an asymmetric rotor close to the prolate sym- metric limit (K = -0.988).It has C, symmetry, with dipole components along the a and b inertial axes., The main fea- tures of its spectrum are extensive series of 6-type Q-branches Of the types JK,J-K-JK-l,J-K+l and JK,J-K+l -JK-1, -+ ,. These branches are partially overlapped at low J and K.Further overlap occurs because of the presence of the two abundant C1 isotopes (35/37Cl). In addition, there are several a-and b-type R-branches with sufficient intensity to be easily observed. All transitions show C1 nuclear quadru- pole hyperfine structure, which was partially resolved in the previous microwave study.' At the resolution available with the MWFT spectrometer the "F magnetic hyperfine struc- ture was expected to be observable. Initially, measurements were focused on the region 4-12 GHz.The lo,-Ooo and 1,'-Ooo transitions of SO'ClF and SO237ClF were measured first, in order to refine the chlorine nuclear quadrupole coupling constants. The anticipated "F hyperhe structure was also observed. It was concluded from the optimized microwave pulse length (n/2condition) of these transitions that the a-component of the dipole moment is about twice as large as the b-component. Further transitions involving J < 2 were predicted using the rotational constants of ref. 2,and were located within ca. 250 kHz of the predic- tions. Most of the lines were observable with one experimen- tal cycle. Since the hyperfine splittings due to the C1 nucleus are much larger than those due to the F nucleus, the assign- ments have been made using quantum numbers correspond- ing to the coupling scheme J + Zcl= F1,F1+ IF= F.In preliminary analyses the observed frequencies were fit to the hyperfine constants and the hypothetical unsplit line fre- quencies. The latter were then used to calculate rotational and some approximate centrifugal distortion constants (DJ, DJK,DK),from which further transitions were predicted and measured. Simultaneous fitting of these measured frequencies to the rotational, centrifugal distortion and hyperfine con- stants, using the program SPFIT,14 further reduced the uncertainties of these constants, but the two remaining quartic distortion constants, dl and d2, in Watson's S-reduction, were barely determined, and the precision of the Cl spin-r ot at ion coupling constants was only moderate.Uns pli t lines were given an uncertainty of 0.5 kHz (1.0 when only one Doppler component was observed). Unresolved groups of lines were given an uncertainty of ca. half the calculated split- ting, and each component was weighted according to its pre- dicted intensity; when the. latter was less than 0.65 of the strongest component, it was not taken into account. To improve the constants further, the frequency range of the experiments was extended to 24 GHz, with most mea- surements being of selected transitions between 18 and 24 GHz. Owing to the low rotational temperature of the jet, the line intensities rapidly decreased with J and K,, and only lines with J < 4 and K,< 3 were measured (a total of 155 and 170lines, of 25 and 26 rotational transitions of S02ClF and SOz3'C1F, respectively; a selection is given in Table 1 and a complete list is available from the authors).Inclusion of these transitions in the fits produced well determined values for the rotational, hyperfine, and most of the centrifugal dis- tortion constants. Reasonable values were obtained for dl and d2 despite their small magnitudes. F-Cl spin-spin coup ling constants were barely determined, and their values affected all other constants at most in their last quoted figures; consequently they were omitted in the final fits. The final spectroscopic constants of S02ClF and SO237ClF are presented in Table 2. The standard deviations of the fits, which are also in Table 2, are within experimental error.The constants are essentially uncorrelated, with the absolute value of only a few of the correlation coefficients being larger than 0.5.The largest correlation occurs between A and DK (CU. -0.82). Rotational transitions of the rarer isotopomers 34S02C1F, 34S0237ClF,SO 180ClF and SO l8037ClF, have been mea- sured for the first time in this work. Those of the first two were the dasiest to observe and assign, because 34S is more abundant than l80. Furthermore, because of the proximity of the S atom to the centre of mass, the shifts of the rotation- al constants on substituting 34S for 32S are relatively small. Using the structural parameters in ref. 2,lines of the lll-Ooo transition of both 34S02C1F and 34S0237ClF were found to be within 300 kHz of the predicted values.On the other hand, four sets of candidates were found in the regions pre- dicted for lo,-Ooo.Owing to the small changes in the rota- tional constants, it was expected that the set closest to the predicted frequencies was due to the 34S isotopomers. These J. CHEM. SOC. FARADAY TRANS.,1994, VOL. 90 Table 1 Observed frequency (MHz) of rotational transitions (selection) of S0,ClF and SO237ClF and residuals (kHz) ~~ S0,ClF SO, 37ClF transition" observed o-cu observed 04 101-000 1.5, 2-1.5, 2 5797.4754 -1.46 5645.2228 -1.55 2.5, 2-1.5, 1 5816.6061 -0.45 5660.2999 -0.29 2.5, 3-1.5, 2 58 16.6 135 0.09 5660.3070 -0.07 0.5, 1-1.5, 2 5831.9240 -0.81 5672.3702 -0.95 111-000 0.5, 1-1.5, 2 7971.4398 0.94 7895.9430 0.33 0.5, 0-1.5, 1 7971.4468 1.12 7895.9499 0.43 2.5, 2-1.5, 1 7978.5874 -0.02 7901 S739 0.15 2.5, 3-1.5, 2 7978.5994 -0.28 7901.5857 -0.29 1.5, 1-1.5, 1 7987.5148 -1.33 7908.6 1 12 0.26 1.5, 2-1.5, 2 7987.5233 1.73 7908.6 165 0.12 211-110 2.5, 2-1.5, 1 11315.3372 -1.60 2.5; 3-1.5; 2 11315.3458 0.47 0.5, 0-1.5, 1 11639.2436 0.71 11327.4094 1.40 0.5, 1-1.5, 2 11639.2494 -0.52 11327.4142 -0.99 1.5, 2-2.5, 3 11640.5091 0.23 1.5, 1-2.5, 2 -0.15 220-1 11 0.5, 1-0.5, 1 1 8068.17 19 0.53 3.5, 3-2.5, 2 181 50.5532 1.29 18073.2941 1.61 3.5, 4-3.5, 3 1 8 1 50.5 628 0.15 18073.3033 -0.18 2.5, 2-1.5, 1 18160.7709 0.38 18081.3449 0.7 1 2.5, 3-1.5, 2 18160.7838 -0.24 1808 1.3 576 -0.31 221-110 3.5, 3-2.5, 2 18137.8267 -0.66 18061.1949 -0.75 3.5; 4-2.5, 3 18 137.8400 -1.07 18061.2082 -1.18 2.5, 2-1.5, 1 18146.7658 -0.57 18068.2374 -0.62 2.5, 3-1.5, 2 18146.7838 -1.03 18068.25 19 -1.06 313-220 3.5, 3-2.5, 2 10853.5452 -0.44 101 60.8301 -0.95 3.5, 4-2.5, 3 10853.551 1 1.59 10160.8342 -0.34 2.5, 2-1.5, 1 10874.2296 0.25 10177.1290 0.29 2.5, 3-1.5, 2 10874.2343 -0.8 1 10177.1339 -0.24 4.5, 4-3.5, 3 10887.6855 0.981.39 10187.73684.5, 5-3.5, 4 -0.58 -0.56 1.5, 1-05, 1 10908.3267 0.01 1.5, 2-0.5, 1 10908.3501 -0.57 10204.0230 -0.29 404-313 5.5, 5-4.5, 4 21 116.8318 -0.38 20414.7762 0.18 5.5, 6-4.5, 5 -0.48 -0.19 3.5, 3-3.5, 3 21119.3816 0.14 0.63 3.5, 4-3.5, 4 -0.39 20416.7861 -0.13 413-312 3.5, 3-2.5, 2 23274.9520 -0.38 22652.0250 -0.34 3.5, 4-2.5, 3 23274.9571 -0.19 22652.0304 0.04 4.5, 4-3.5, 3 23275.8735 -0.34 22652.75 10 -0.55 4.5, 5-3.5, 4 23275.8784 -0.06 22652.7564 0.13 2.5, 2-2.5, 2 23290.5785 -0.11 22664.3384 -0.32 2.5, 3-2.5, 3 23290.5846 0.02 22664.345 1 0.27 -J&,,k,,/F'l,F' -F:, F; observed minuscalculated.assignments were confirmed with the resulting values of the C1 quadrupole coupling constants x,, , perpendicular to the C, plane, which, neglecting vibrational effects, should be the same as those of the corresponding 32S isotopomers; the values agreed to four significant figures. The remaining sets of lines were much further away and mostly showed values of xce significantly different from those of the 32S isotopomer.These lines were probably vibrational satellites of the "S iso-topomer; their assignment was beyond the aim of this study. Table 2 Spectroscopic constants' of isotopomers of sulfuryl chloride fluoride, S0,CIF rotational constants/MHz centrifugal distortion constants/kHz 5P A B C DJ DJK DK dl 4 4 pS0,ClF 5080.455 582 (90) 2912.846885 (40) 2899.926 403 (38) 0.47931 (103) 2.0282 (57) -1.5653 (147) 0.00336 (84) 0.00593 (76) so, 37~1~ 5080.472 921 (90) 2834.764 222 (37) 2822.517 563 (36) 0.46064(101) 1.9541 (52) -1.4510 (147) 0.00313 (83) 0.00573 (64) 8 34S02C1F 5078.573 758 (189) 2907.049 659 (63) 2893.493 326 (50) 0.47184 (135) 2.0414 (95) -1.499 (36) 0.00410 (174) 0.00545 (124) 34s0237~1~ 5078.590 843 (224) 2828.571 272 (80) 2815.727 184 (44) 0.4532 (33) 1.9575 (164) -1.410 (48) 0.00288* 0.00573' SO 180C1F 4893.158 903 (172) 2878.723 789 (80) 2823.393 752 (78) 0.4512 (32) 1.8893 (116) -1.430 (42) -0.0229 (27) -0.0063 (19) so 180 37~1~ 4893.084 694 (105) 2801.081 782 (62) 2748.638 373 (59) 0.4334' 1.806' -1.35' -0.0214' -0.00606 chlorine nuclear quadrupole coupling constants/MHz spin-ro ta tion coupling cons tan t s/kHz xaa x-lxal xbb L' M,(Cl) M,(Cl) Mm(C1) M,(F) Mbb(F) Mm(F) standard deviationkHz S0,ClF -76.551 60 (29) -5.10491 (61) 35.723 34 (45) 40.82826 (45) 0.330 (42) 1.860 (46) 1.740 (54) 12.247 (121) 3.252 (134) 8.176 (130) 0.72 SO, 37CIF -60.329 18 (28) -4.028 97 (65) 28.15011 (46) 32.17907 (46) 0.262 (43) 1.497 (44) 1.328 (51) 12.380 (112) 3.447 (117) 8.013 (123) 0.66 34S02C1F -76.574 70 (39) -5.087 38 (78) 35.743 66 (59) 40.831 04 (59) 0.451 (65) 1.920 (68) 1.609 (61) 12.29 (20) 2.81 (22) 8.06 (18) 0.84 34S0237ClF -60.34729 (52) -4.01407 (103) 28.16661 (78) 32.18068 (78) 0.118 (113) 1.501 (88) 1.495 (113) 12.54 (29) 3.48 (31) 7.13 (30) 0.65 SO "OClF -76.60849 (51) 2.17549 (110) 2.223 (32) 39.391 99 (80) 37.21650 (80) 0.051 (129) 1.812 (97) 1.698 (95) 12.24 (23) 6.53 (21) 5.01 (20) 0.88 SO "037ClF -60.37657 (54) 1.71742 (120) 1.768 (40) 31.04700 (87) 29.32957 (87) 0.183 (139) 1.557 (121) 1.360 (118) 12.43 (38) 5.80 (27) 5.19 (25) 0.75 Uncertainties reflect 16.'Fixed at values derived from the force field and other isotopomers, see text.Derived constants. Further transitions of the 34S isotopomers were measured and assigned by boot-strapping, following a similar pro- cedure to that used for the 32S species.Owing to the rela- tively low abundance of 34S, fewer lines were measured (for 34S02C1F and 34S0237ClF, a total of 111 and 71 lines, respectively, of 20 and 12 rotational transitions with J < 4 and K,< 2, of which a selection is given in Table 3). Their spectroscopic constants are in Table 2. For 34S0237ClF it was necessary to fix the constants dl and d, to values derived from the force field, scaled according to ratios of other isotopomers (Table 4). The rms deviations of the fits were again within experimental error. Measurements of the spectrum of SO I80C1F presented a particular challenge for several reasons.First, this isotopomer is less abundant by a factor of ca. 250 than SO,ClF, and a much larger number of cycles was necessary to obtain a satis- factory signal-to-noise ratio. Secondly, its transitions were much more difficult to predict than those of the 34S iso-topomers. This was mainly because the similarities of the masses of 0 and F atoms and the respective bond lengths to the S atom cause a large rotation (ca. 125", Fig. 1) of the b-and c-inertial axes around the a-axis on l80isotopic substi- tution, and small changes in the geometrical parameters had a large effect on this angle of rotation. Thirdly, this iso- topomer should thus exhibit a-, b-and c-type transitions, and it was not obvious a priori which of the latter two should be stronger.Furthermore, the uncertainty of the angle of rota- tion produced uncertainties in the predictions of B and C, and of x-. Fortunately, x-of S0,ClF is small; this is thus also the case for x-of SO '80C1, irrespective of the angle of rotation. Fourthly, because of the dense spectrum, accidental interference by lines of more abundant isotopomers may render the assignment of lines of SO l80C1F more difficult. Nevertheless, a successful assignment has been made. The a-type transition lol-O0, was found within 5 MHz of a pre- diction using the structure in ref. 2. It yielded an estimate of xu. Of the 2-1 a-type transitions 21,-1 was expected to be least affected by the presence of S0237ClF lines. The two strongest C1 hyperfine components, only slightly affected by x-, were found near 11 464 and 11 445 MHz, within 10 MHz of the prediction.With refined B and C values,.additional strong 2-1 a-type lines were easily found; these resulted in a reasonable estimate of x-, so that all strong components of these transitions could be observed. Finally, using distortion constants estimated from a simple force field, detection of 3-2 a-type as well as some b-and c-type transitions was straight- forward. In total, 95 hyperfine components of 19 rotational transitions with J < 3 and K,< 2 were observed; a selection is given in Table 5. From estimates of the optimized micro- \ c' \ \ Ib \\ F / / b ;/ / , Fig. 1 Projection of the S0,F moiety into the bc plane.The prin- cipal axes of S0,ClF (solid lines) and SO '*OCIF (dashed lines) as well as the substituted 0 atom are indicated. (The C1 atom is almost at the origin, below the plane of the paper.) J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 3 Observed frequencies (MHz) of rotational transitions (selection)of 34S02C1F and 34S0, "ClF and residuals (kHz) 3734S0,CIF 34~~, ~1~ transition" observed O-Ca observed 04 101-000 1.5, 2-1.5, 2 5785.2406 -1.39 5632.2360 -1.29 2.5, 2-1.5, 1 5804.3 774 -0.20 5647.3179 -0.34 2.5,3-1.5, 2 5804.3845 0.38 5647.3248 0.19 0.5,1-1.5, 2 5819.6993 -1.21 5659.3918 -0.48 1 11-000 0.5, 1-1.5, 2 7963.1 194 0.57 7887.2664 0.31 0.5, 0-1.5, 1 7963.1259 0.28 7887.2737 1.04 2.5, 2-1.5, 1 7970.2712 -0.26 7892.9006 -0.12 2.5, 3-1.5, 2 7970.2834 -0.27 7892.9 127 0.17 1.5, 1-1.5, 1 7979.2044 -0.85 7899.9416 -0.28 1.5, 2-1.5, 2 7979.21 15 0.83 7899.9474 0.28 202-1011.5, 1-0.5,0 11581.8784 2.47 11273.4474* 1.92 1.5, 2-0.5, 1 11581.8897 1.43 11273.4584 0.88 0.5, 0-0.5, 1 11600.9982 -2.27 11288.5210 0.84 0.5, 1-0.5, 1 11601.0110 -0.36 11288.5307 -0.09 2.5, 2-1.5, 1 1 1602.6431 0.42 11289.8 158 0.22 2.5, 3-1.5, 2 2.16 0.2311289.8230 -0.093.5, 3-2.5, 2 11602.6522* -0.63 3.5, 4-2.5, 3 1 1602.6592 0.45 11289.8291 0.24 1.5, 1-1.5,1 11616.3409 -0.07 1.5, 2-1.5, 2 11616.3475 0.71 0.5, 0-1.5, 1 11635.4627 0.81 0.5,1-1.5, 2 11635.4709 1.02 211-110 2.5, 2-1.5, 1 11600.0932 -0.84 11289.9669 0.25 2.5,3-1.5, 2 11600.1008 0.85 11289.9728 -0.14 1.5, 1-1.5, 1 11294.9973 -0.55 1.5, 2-1.5, 2 11295.0027 0.14 2.5, 2-2.5, 2 11610.2997 0.51 2.5, 3-2.5, 3 -0.35 0.5,0-1.5, 1 11615.4226* -0.71 1 1302.0438 0.83 0,5,1-1.5, 2 11616.4308* 0.97 1 1302.0496 -0.31 1.5, 2-2.5, 3 1 1616.6845 0.30 -0.34 1 1303.0420 0.24 1.5, 1-2.5, 2 -0.41 3.5, 3-2.5, 2 1 l305.O475* -1.22 1.96 3.5, 4-2.5, 3 0.5, 1-0.5, 1 11316.5362 -0.07 212-111 2.5, 2-1.5, 1 11573.0880 -0.42 11264.3625 -0.65 2.5, 3-1.5, 2 11573.0975 -0.37 11264.3713 -0.61 1.5, 1-1.5, 1 11580.3828 1.20 11270.1 109* 0.57 1.5, 2-1.5, 2 11580.3876 -1.22 11270.1 149* -2.09 2.5, 2-2.5, 2 2.99 2.5, 3-2.5, 3 11582.0252* 0.33 1.5, 1-2.5, 2 1 1589.3 165 1.12 0.69 11277,1518 0.31 1.5, 2-2.5, 3 0.22 0.5, 0-1.5, 1 11590.5966 -0.05 0.5,1-1.5, 2 1 1590.6063 -0.72 3.5, 3-2.5, 2 11592.2176 -0.05 11279.4420 -0.33 3.5, 4-2.5, 3 1 1592.2241 0.64 11279.4487 1.21 1.5, 1-0.5,0 11596.4601 -1.12 11282.7788 -0.75 1.5, 1-0.5, 1 11596.4675 -0.51 11282.7857 -0.42 1.5, 2-0.5, 1 11596.4800 -0.66 11282.7983 0.28 " See Table 1.wave pulse lengths of the 1-0 transitions (42 condition) it was estimated that pb zpc zpJ2. Using an rI.e structure (uide infu) determined from these five isotopomers, rotational constants of SO l8 037CIF were predicted. All other constants were estimated from the force field and other isotopic data. The high accuracy of these con- stants allowed some transitions to be observed, despite the very low isotopic abundance.The number of cycles required ranged from 400 (ca. 7 min averaging time, Fig. 2), to a J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 4 Comparison of measured vibrational wavenumbers" (cm- ') and centrifugal distortion constantsb (kHz) of S0,ClF with those calcu- lated from the force field SOzCIF so, 37~1~ 34S0,C1F 34~0~37~1~ SO '*OClF SO "0CIF obs. calc. obs. calc. obs. calc. obs. calc. obs. calc. obs. calc. 1230.1 1230.2 1230.1 1218.5 1218.5 1205.6 1205.5 826.5 826.3 826.3 818.3 818.3 822.9 822.8 632.1 632.1 630.7 622.0 620.5 626.5 625.0 503.0 503.0 502.5 500.1 499.6 495.9 495.4 422.0 422.8 415.0 416.0 422.8 415.9 419.5 412.7 295 294.4 292.8 294.0 292.4 290.4 288.8 1469.4 1469.3 1469.2 1448.4 1448.4 1449.8 1449.8 475.9 476.4 476.0 473.8 473.4 469.0 468.6 303 302.0 300.5 301.7 300.2 297.4 295.9 0.47931 0.48 107 0.46064 0.46038 0.47 184 0.47900 0.4532 0.4583 0.4512 0.4583 0.4334' 0.4385 2.0282 2.0199 1.9541 1.9591 2.0414 2.0203 1.9575 1.9591 1.8893 1.9052 1.806' 1.8327 -1.5653 -1.4865 -1.4510 -1.4054 -1.499 -1.4854 -1.410 -1.404 -1.430 -1.390 -1.35' -1.317 0.00363 0.00355 0.00313 0.00313 0.00410 0.00343 0.00288' 0.00288 -0.0229 -0.0123 -0.0214' -0.0114 0.00593 0.00675 0.00573 0.00604 0.00545 0.00681 0.00573' 0.00610 -0.0063 -0.00523 -0.00W -0.00467 * Ref.8. This work. Derived from the force field and scaled according to ratios of other isotopomers.maximum of 15000 (ca. 250 min). The deviation between tion, all other constants were affected to a certain extent experimental and predicted rotational constants was very (Q0.7 kHz), with values closer to those expected, especially in small: 27.9, 2.0 and 1.6 kHz for A, B and C, respectively. the case of the centrifugal distortion and the chlorine spin- Because only a few rotational transitions (eight, with J, K,< rotation coupling constants. Because of the limited input 2 and a total of 48 hyperfine components) were observed, data, the correlations were slightly higher than for the other centrifugal distortion constants were fixed at values derived isotopomers (the highest for SO "0ClF is -0.88 between A from the force field, scaled according to ratios of other iso- and D, ;for SO '*O37ClF -0.68 between B and C, because topomers (CJ: Table 4). A selection of the assigned transitions no distortion constants were fitted). The rms deviations is in Table 5 and the derived constants are in Table 2. became comparable with those of the other isotopomers.The initial fits to the l80isotopomers assumed that effects The contributions of xbc in the K, = 1 stack of SO "OCIF of the off-diagonal C1 quadrupole coupling constants could are for J = 1, ca. 5.51, 3.61 and 0.22 kHz for F, = 0.5, 1.5 and be neglected; however, the rms deviations were ca. 2 kHz, 2.5, respectively; they decrease rapidly for higher J. Although somewhat greater than the measurement uncertainties, and the levels J,, J-,and JK,J-K+ are much closer for K 2 2, also larger than the corresponding values of the other iso- essentially no perturbation was present in these levels topomers (< 1 kHz).In addition, very large deviations (of up (observed and calculated for K < 3 and J Q 4). Inspection of to 7 kHz) occurred for some of the 1 ,,-Oo0 and 1lo-Ooo lines. the matrix elements in the symmetric rotor limit showed that Since the levels l,, and l,, are relatively close, it was for every level connecting certain J,, -,with J,, -,+ , thought that the off-diagonal quadrupole coupling constant through &,& there was a corresponding level connected xbc might be contributing significantly to the frequencies, in through 4zt$d which cancelled the effect. spite of an anticipated small (absolute) value of 52.5 MHz.The only non-zero off-diagonal quadrupole coupling con- Inclusion of xbc in the fits yielded moderately well determined stant for isotopomers with two l60atoms is xab.Its effects values for both isotopomers (CJ: Table 2), of the right order of in the observed spectrum of S02ClF were <1.2 kHz, even magnitude and, as expected from the structure, approx- smaller than those of xbc for the isotopomers with "0. The imately in the ratio of the C1 quadrupole moments. In addi- resulting value, Ixabl= 5.77 (39) MHz, must be viewed cau- tiously, even though it is reasonably close to the expected 3.5,4-2.5, 3 value of 7.43 MHz (see Discussion on hyperfine constants). The standard deviation of the fit is reduced from 0.72 to 0.63 kHz, and all other constants are affected to within less thanA-2.5,2-105, ' twice their uncertainties.Similar results are obtained for the other isotopomers: Ixabl= 4.13 (57), 5.04 (48), 4.61 (72) and 4.73 (63) MHz for S0237ClF, 34S02C1F, 34S0237ClF and SO l80C1F, with the standard deviations reduced to 0.64, 0.79, 0.58 and 0.85 kHz, respectively. Although the absolute values of the correlation coefficients involving xab are small (<0.55), this constant has been omitted from the final fits because of its small contributions. Structural Parameters I I 1 Although precise isotopic data are now available, the deter- I 11 099.8 mination of the geometry of S02ClF was not straightfor- 11 099.7 frequency/MHz ward.The difficulties arose from several sources. Owing to the proximity of the S atom to the centre of mass, and also Fig*2 of the 202-101 transitionOf '*O37C1Fin natural because of the very small b-coordinate of the C1 atom, vibra- isotopic abundance, showing "F and "Cl hyperfine structure, dis- They are shown in com-played as a power spectrum. 400 averaging cycles were used. A stick tional effects are diagram indicates the calculated relative intensities and the line posi- parisons of the A rotational constant of SO2CIF and tions from the decay fit. The lines are Doppler-split by CQ. 57 kHz. S0237C1F, both with 32S and with 34S, where 2606 Table 5 Observed frequencies (MHz) of rotational transitions (selection)of SO "OClF and SO l8O37ClF and residuals (kHz) SO '80C1F so 180 37~1~ transition' observed o-c' observed O-C' 101-000 1.5, 2-1.5, 2 5686.8088 -1.45 5537.6527 -0.78 2.5, 2-1.5, 1 5705.9540 0.16 5552.7418 0.41 2.5, 3-1.5, 2 5705.96 1 1 0.33 5552.7485 0.5 1 0.5, 1-1.5, 2 5721.2835 -0.45 5564.8221 0.42 11 1-000 0.5, 1-1.5, 2 7706.6891 -0.42 7633.9492 0.13 2.5, 2-1.5, 1 7714.5768 0.60 7640.1643 1.28 2.5, 3-1.5, 2 77 14.5869 0.34 7640.1729 -0.70 1.5, 1-1.5, 1 7724.4 178 -1.39 7647.9213 -0.86 1.5, 2-1.5, 2 7 724.4249 1.12 7647.9271 0.25 116000 0.5, 1-1.5, 2 7762.5724 -0.21 7686.8272 -0.66 0.5, 0-1.5, 1 7686.8366* 2.65 2.5, 2-1.5, 1 7770.0144 -0.02 7692.6919 -0.20 2.5, 3-1.5, 2 7770.025 1 -0.58 7692.7022 -0.85 1.5, 1-1.5, 1 7779.3200 0.16 7700.0247 -1.18 1.5, 2-1.5, 2 7779.3257 0.86 7700.03 15 0.76 202-10 1 1.5, 1-0.5, 0 11383.9576 1.34 11083.3633 -0.05 1.5, 2-0.5, 1 11383.9700 0.66 11083.3765 0.69 25, 2-2.5, 2 2.33 2.5, 3-2.5, 3 11385.5992* -1.71 0.5, 0-0.5, 1 11403.0735 0.97 11098.4339 0.21 0.5, 1-0.5, 1 11403.0841 0.04 11098.4441 -0.58 3.5, 3-2.5, 2 11404.7383 0.68 11099.7465 1 .00 3.5, 4-2.5, 3 11404.7436 -0.27 11099.7503 -1.16 2.5, 3-1.5, 2 11404.7519 0.47 11099.7585 1.67 1.5, 2-1.5, 2 11418.4425 -0.54 211-110 2.5, 2-1.5, 1 11445.0859 0.80 11 140.4638 0.27 2.5, 3-1.5, 2 11445.0928 -0.94 11 140.471 1 -0.41 1.5, 1-1.5, 1 11452.1206 0.05 11 146.0083 0.58 1.5, 2-1.5, 2 11452.1264 -0.74 11 146.0136 -0.17 2.5, 2-2.5, 2 1.07 2.5, 3-2.5, 3 11454.3916* -1.29 3.5, 3-2.5, 2 11464.2253 0.09 11155.5506 -0.72 3.5, 4-2.5, 3 11464.2299 -0.57 11 155.5562 0.27 1.5, 1-0.5, 0 11 159.1988 -0.85 1.5, 2-0.5, 1 11468.8792 -0.17 11 159.2170 0.35 0.5,O-0.5, 1 1 1 166.9729 0.84 0.5, 1-0.5, 0 1 1166.9789 -0.69 0.5, 1-0.5, 1 11478.7404 0.58 11 166.9854 -0.28 212-1 11 2.5, 2-1.5, 1 11334.3863 0.82 11035.5450 0.16 2.5, 3-1.5, 2 11334.3921 -0.99 11035.5517 -0.70 1.5, 1-1.5, 1 1 1341.0344 -1.12 11040.7848 0.27 1.5, 2-1.5, 2 11341 .M26 1.31 11O40.7901 -0.15 2.5, 2-2.5, 2 1.03 2.5, 3-2.5, 3 11344.2295* -0.82 3.5, 3-2.5, 2 11353.5195 -0.77 11050.6287 -0.36 3.5, 4-2.5, 3 11353.5247 0.00 11050.6336 0.26 1.5, 1-0.5, 0 11358.7589 -0.53 11054.7514 -0.33 1.5, 2-0.5, 1 11358.7760 0.44 11054.7680 -0.03 0.5, 0-0.5, 1 1 1368.0748 -0.39 1 1062.09 15 -0.61 0.5, 1-0.5, 0 11062.1002 0.96 0.5, 1-0.5, 1 1 1368.0873 -0.84 1 1062.105 1 -0.03 a See Table 1.A(37Cl)> @$Cl), in contrast to the rigid rotor model. In addition, since F has only one stable isotope, it cannot be isotopically substituted. Several approaches have been taken to derive geometrical parameters. These have produced ro and r,-based param-eters,15*16 r,-based a ground-state average (rJ structure" and an estimate of the equilibrium (re)struc-ture.17 For the r,, r,-based and r, structures, least-squares J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 fits were carried out for the six parameters needed to describe the geometry completely. These were chosen to be r(SO), r(SCl), r(SF), 1/2L(OSO),and the angles between the bisector of ~(0S0)and the SCl and SF bonds, respectively.From these were derived the internal coordinates L(OSO), L(ClSO), L(FSO), and L(C1SF) along with their uncer-tainties. The program RU111J'6*'8 was used for the ro and ro-based structures. Initially, ro parameters were obtained by a least-squares fit of the moments of inertia derived from the effective rotational constants of Table 2, all given equal weight. The standard deviation for this fit was 0.0028 ut A', in agreement with cu. 0.003 u A2,expected under these cir- cumstances. In order to give the data more appropriate weight, the rotational constants were weighted according to the inverse squares of their experimental uncertainties (Table 2) for all subsequent structure calculations.The exception was SO ''0 37ClF, where the error bars were increased by a factor of 1.5 because spectral fits of the centrifugal distortion constants were not made. This procedure seems reasonable, because the relative uncertainties are of the same order of magnitude and reflect (roughly) the quality of the data. For the ro and r,-based structures the six parameters mentioned above were fitted to the principal planar moments PA, P, and P,, derived from the rotational constants. Because the C, symmetry of the molecule was a constraint, the values of P, for substitutions in the symmetry plane were omitted from the fit." The standard deviation of the fit is improved slightly (0.0019 u A2) over that obtained earlier.The geometrical parameters are given as the ro structure in Table 6,in com- parison with the approximate ro structure of Holt and Gerry.' There is agreement within the uncertainties of the latter. If vibrational effects could be reduced it should be possible to obtain more reasonable structural parameters and to reduce the standard deviation of the fit. Several r,-based pro- cedures are available to do this. They are based on the equa- tions Po= + E'U and where P and I refer to the principal planar and inertial moments, respectively, and g = a, b or c, to the principal axes of inertia. Although corresponding E'@ and (if9) are different for different isotopomers, they do not in general vary enor- mously. As a result, Rudolph has proposed a first-order approach by introducing isotopically invariant E'~ or E''~ as independent parameters in the fits.'5*'' The resulting struc- tures, called rp,eand r1,&,are identical when the same data are used in the fits: the different names indicate the fitted quantities.l5 An alternative approach is to fit the structural parameters to isotopic differences in principal inertial or planar moments.20 The resulting so-called 'pseudo-Kraitchman' structures, designated rAI or rApas appropriate, are the same as the rIv0and rp,estructures, although the methodology is different.' There are advantages in making a particular choice, however.Since the rI, and rp, fits produce values for E they are well suited for predictive purposes (e.g.to estimate rotational constants of hitherto unobserved isotopomers). Alternatively, the rA1 and rAp fits permit correction for 'Costain's (in which uncertainties in coordinates of substituted atoms are inversely proportional to their absolute t 1 u x 1.661 x lO-*'kg. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 6 Structural parameters" (pm, degrees) of S0,ClF r(w 140.8 (6)/ 140.711 (81) 140.877 (138) 140.876 (171) 140.877 (81) 140.721 (31) 140.543 r(SC1) 198.5 (15) 198.460 (130) 198.556 (125) 198.57 (29) 198.556 (81) 198.957 (52) 198.571 r(SF) 155.0 (20) 155.46 (23) 154.02 (6 1) 154.02 (46) 154.02 (37) 155.231 (95) 154.895 L(OS0) 123.7 (10) 123.941 (125) 123.30 (37) 123.31 (38) 123.30 (35) 124.039 (47) 124.039 L(ClS0) 107.5 (25) 109.177 (69) 108.974 (81) 108.99 (38) 108.974 (51) 108.979 (30) 108.979 L(FSO) 107.5 (20) 106.719 (121) 107.27 (21) 107.27 (21) 107.27 (21) 106.961 (52) 106.961 L(C1SF) 99.0 (30)' 97.906 (90) 98.09 (22) 98.04 (111) 98.090 (137) 97.673 (36) 97.673 See text for further details.Ref. 2. 'Coordinates and uncertainties of F fixed to values of rhp. With non-trivial first and second moment conditions. 'Estimates of equilibrium bond lengths, derived from r,, see text. f Assumed, see text. magnitudes). The method is described in greater detail in ref. 16; examples of these structure calculations were presented previously.' 'e2' Fits were made using the tr,e, rp,cand rap methods. The rr,e fit made to the first five isotopomers measured gave very good predictions of the rotational constants of SO l8037ClF, as indicated in the previous section.The fit to all six iso- topomers using the rp,e method reduced the standard devi- ation to 0.00006 u A' and yielded values for the &IArovibrational parameters of = 0.3577 (107), dB= 0.3017 (64) and dc = 0.3059 (64) u 8'. The error bars obtained for both the structural and the rovibrational parameters are probably artificially small, for the data set with the inclusion of SO'*03'ClF produced larger error bars than the set without it. In the ru fit, following the suggestion of Rudolph,16 Costain's error was accounted for using 6(Pg-P)= 0.003 u A'; where 6 denotes the difference between calculated and measured quantities, P denotes the planar moments of SO,ClF, P'g the corresponding moments for substituted iso- topomers. The resulting structure is presented in Table 6.It is preferred to the rr,e or rp,e structures, because the derived uncertainties are more realistic: in this case inclusion of SO ''0"ClF decreased the uncertainties slightly. The values of the r, and rp, parameters are essentially the same. The r,-based parameters were obtained using Rudolph's least-squares fitting program RU238J.I6 Initially, Typke used Kraitchman's equations to develop a least-squares pro- ~edure.~~ to take The method was extended by R~dolph,'~ symmetry considerations into account.'' In this program planar moments of inertia are fitted to the Cartesian coordi- nates of those atoms which have been substituted.To obtain a complete r,-type structure, positions of unsubstituted atoms, or of atoms close to inertial axes, can be fixed at values determined from alternative methods. In applying the method to S0,ClF several potential diffi- culties must be taken into account. The first two have been alluded to in the first paragraph of this section. In addition, the rotation of the b-and c-axes around the a-axis is ca. 125" on the substitution of one l6O atom by "0.In r,-type fits, vibrational effects on the structure are not compensated for when the rotation angle is so large.25 These effects have been accounted for by introducing the eg obtained from the rp,efit as constants in the r, fit (along with their uncertainties and correlation coefficients). Several calculations of r,-based parameters were carried out, of which the two most reasonable are described in detail.In both cases the coordinates of the F atom were fixed at the rAp values: a(F) = 73.81 (17) and b(F) = 136.57 (24) pm. This assumption is justified because r, structural parameters deter- mined by least-squares fitting of a large isotopic set converge to values from an rp,e or rap Costain's error was employed as described for the rap fit. The resulting geometri- cal parameters are given in Table 6 as r, values: their uncer- tainties are fairly large, mainly because the b(C1) coordinate is very small, thus incorporating a large uncertainty [-0.89 (365) pm]. The advantage of the least-squares method can be demonstrated best for this structure, when the error bars of a fit with five isotopomers (without S018037ClF) are com- pared with those from a fit with six.In the latter case the uncertainties of b(C1) are reduced by a factor of three, those of b(0)and 40)by 30%, and even b(S)is well determined. In the second r,-based fit the first and second moment con- ditions for the a-and b-axis and for the C, plane were also included. These had the effects of giving a well determined value of b(Cl), which agreed with the value from the rdp fit. The resulting geometrical parameters are presented in Table 6 as rl values. They agree with the values from both the rM and r, fits, though they have considerably smaller uncer- tainties. When the first and second moment conditions were taken into consideration to determine the F coordinates, these coordinates did not change with respect to the rAp values, again justifying the assumption made above.Note that accounting for the E# values was important in obtaining consistent r,-type structures: without their use deviations occurred between the r, and rl structures [e.g. 0.25 pm for <SF)]. Deviations were also obtained when the first and second moment conditions were used to locate the F atoms [e.g. 1 pm for r(SF)]. To obtain a ground-state average structure (rhof the main isotopic species, SO,ClF, a harmonic force field was evalu- ated as described in the following section. The harmonic con- tributions to the a-constants were subtracted off the effective rotational constants in Table 2, and the resulting BE were fitted to the geometrical parameters using Typke's program MWSTR.Isotopic variations in the bond lengths were accounted for using the equation17 6r, = 3/2a6(u2) -6K where (u') and K are, respectively, the zero-point mean- square amplitude of a given bond and its perpendicular amplitude correction, both obtained from the force field. The constants a are Morse anharmonicity parameters, which were approximated by values from the respective diatomics. Values of 2.072, 1.646 and 1.892 A-' for SO,26SCl (derived from ref. 27), and SF (derived from ref. 28), respectively, were used. Centrifugal distortion and electronic mass contribu-tions to the a-constant~~~ were neglected.The BB,values were weighted similarly to the BB, values. The resulting rz param-eters are in Table 6. Finally, equilibrium bond lengths were estimated from the r, structure according to ref. 17 rz = re + 3/2a(u2) -K (2) Changes in the bond angles between r, and re were neglected. The result is presented in Table 6 as re. When structural fits to the data of five (without SO l8O37ClF) and six isotopomers are compared, essentially no changes occurred for rAp and r:. ro and r, showed a greater dependence on the isotopic set (but within 1 a). J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 High correlation coefficients ( IcijI 2 0.9) were obtained for reproduced moderately well. For the refinement of the force all structure models.The most severe occurred for the ro and field the input data were weighted according to their experi- rz structures [-0.9997 for r(S0) and L(OS0)and -0.98 for mental uncertainties (0.1 or 1 cm-' for the vibrational wave- r(SF) and L(FSO)]. The value of most Icij I decreased for the numbers (cf: ref. 8) and three times the values of Table 2 for rp,c structure; however, E~ and E' are highly correlated the distortion constants). No attempt was made to estimate (0.9993),indicating that some of the high correlations are due anharmonic corrections of the input data.jO 13 force con- to vibrational effects. All the larger correlations are further stants were fitted, chosen because of their greatest sensitivity reduced for rhp, the largest occurring for r(SF) and L(OSO) to the input data with the largest deviations between mea- (0.96) and L(FSO)(-0.96)' respectively.For the r, structure, sured and calculated values (v8 and the distortion constants), inclusion of the data of S018037ClF reduced the corre- as indicated in Table 7. In the iterative procedure employed, lations substantially, but three lcijl were still larger than 0.9 a direct fit was first made, then the values offso, fSF andf,,,,, C0.996 for L(ClSO)/L(ClSF) 3. The inclusion of the first and were slightly adjusted to reproduce vl, v2 and vg better, and a second moment conditions reduced most of the correlations; new fit was carried out. All the resulting non-zero force con- the highest occurred between L(OS0)and L(FSO) (-0.96)/ stants are given in Table 7, along with the potential-energy L(C1SF)(-0.88) with all others smaller than 0.8.distribution. A comparison of the observed spectroscopic constants with those calculated from the final force field is Harmonic Force Field given in Table 4. This force field calculation confirms the assignments made A harmonic force field was calculated for S02ClF in order to in ref. 7 and is thus incompatible with that in ref. 5. However,determine the ground-state average structure r, (uide supra). the description of the vibrational modes in terms of internal Since the force field in the literature7 was obtained by taking coordinates is rather complex because of strong mixing, average force constants of S02C1, and S02F2, we have whereas that given in ref.7 is a simple one, based on the attempted to refine it by fitting the force constants to the potential-energy distribution in terms of symmetry coordi- vibrational wavenumbers of ref. 8 and to the distortion con- nates. vl, v2 and v7 are essentially unmixed and well described stants of this study. The refined force field should permit a as v,(SO,), V(SF) and v,(S02), respectively. Taking also into further confirmation of the assignment proposed by Pfeiffer. account the Cartesian displacement of the normal coordi- The refinement was carried out in terms of internal coordi- nates, v3 and vs can be rationalized as asymmetric and sym- nates, using Christen's programs NORCOR and NCA.29 metric combinations, respectively, of the S02F umbrella Of the ten internal coordinates seven are different [r(SO), deformation and the SCl stretching vibration.This can be r(SCI), r(SF), L(OSO), L(ClSO), L(FSO) and L(CISF)], symbolized as [d,,,(SO,F) + v(SCl)],, and [v(SCl)which result in 34 separate internal force constants (Table 7); + dUmb(SO2F)],, respectively. The remaining vibrations are this number is reduced to 27 when redundancy is taken into rather difficult to describe briefly in terms of internal or sym- account. Owing to the limited input data (see Table 4) not all metry coordinates. of the force constants could be determined independently. The r: structure was used to desribe the molecular Discussion geometry. Initial estimates of the force constants were taken Structure and Force Field from ref, 7; deformation force constants were normalized to 100 pm bond length, whereas in ref.7 the SO bond length Despite differences in approach and the inherent problems of was taken as the reference. With these force constants the calculating structural parameters of S02ClF the structures vibrational wavenumbers and distortion constants were presented in Table 6 are remarkably similar, especially when Table 7 Harmonic force constants" (100 N m-I) and potential-energy distributionb (PED)of S02ClF PED 10.85' 0.88 0.08 0.05 0.94 0.04 3.028' 0.40 0.09 0.54 5.03' 0.05 0.90 0.03 0.04 1.745 0.04 0.03 0.31 0.38 0.04 0.27 1.606' 0.23 0.03 0.12 0.5 1 0.16 1.34 1.788' 0.06 0.06 0.55 0.10 0.03 0.98 0.43 2.254' 0.16 0.04 1.03 -0.16' 0.25 0.028' 0.386 -0.15 0.13 -0.03 -0.056' 0.246' -0.05 0.379 -0.08 0.05 0.47 -0.21 0.09 -0.05 -0.30 0.379' 0.08 -0.28 0.04 -0.05 0.047' 0.489 0.07 0.04 0.15 -0.05 -0.41 0.404' -0.06 -0.06 -0.05 0.06 0.19 -0.36 0.129' -0.06 0.12 0.61 1' 0.18 0.06 -0.66 0.483' 0.15 -0.27 -0.12 0.415 -0.06 -0.04 -0.10 ~SO/OSO,~SO/CISO, ,~SO/FSO,fkp~Deformation force constants normalized to 100 pm bond length ;~SO/SCI,~SO/SF, ~~,CISO ,fSO/CSF &F/OSO &F/CISO fixed to zero, see text.Only contributions 20.03 are given. 'Fitted, see text. Adjusted, see text. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 the uncertainties are taken into account.Therefore, any of the structural methods may be regarded as a reasonable approx- imation to the equilibrium structure re. As may be expected,l63l8 the rAp and rk structural parameters are identi- cal within their quoted values. The additional assumptions for r: (fixing the F atom at the rAp coordinates and using the first and second moment conditions) are mutually compat- ible, showing that even for S0,ClF a reliable substitution- type structure is obtainable, despite the problems occurring for each atom. On the other hand, the ro and rz geometries (and therefore re as well) may be less favourable than the others because of the high correlations and a stronger depen- dence of the structural parameters on the isotopic set. and 34S0,37C1F, using the changes in x, and xbb.For iso- topomers containing 37Cl the constants were scaled by the ratio of the xcc values of the appropriate 35/37C1 isotopic pair. According to the r: structure, the a-axis is rotated by -0.023", 0.090' and 0.067", respectively, from that of S0,ClF. Using the method described in ref. 35, the angle between the z-quadrupolar axis and the a-inertial axis is then found to be 4.33 (53)", 3.55 (14)" and 3.53 (26)", respectively, in S0,ClF. The error bars reflect only the uncertainties in xaa and xbb, which are assumed to be much greater than those of the axis rotation in this structural model. These data strongly indicate that the z-axis coincides with the SCl bond, which is at an angle of 3.78 (11)" to the a-axis.This value agrees quite In spite of the limited data available in the previous micro- well with 8,,= 2.93 (20)" derived from xab of S0,ClF. The wave study,, the geometry was very reasonable. The present results are all within the quoted uncertainties in ref. 2. Note, however, that the angle of rotation of the b-and c-axes on the substitution of one l60atom by "0, which is very sensi- tive to the structure, is 123.0', while that from the structure in ref. 2 is 130.1". The bond lengths and stretching force constants of S0,ClF fit well in the series of related sulfuryl compounds (Table 8) and are as expected from the vibrational spe~tra.~,~,~~ Com-paring the 32/34Sisotopic shifts on the "F NMR frequency of a variety of sulfur fluorine compounds, Gombler suggested the SF bond in S0,ClF to be shorter than in S0,F,;3 this does not agree with both the previou~~*~,"*~~ and the present results.The bond angles follow mostly the expected trends in the series of S0,XY (X, Y = C1, F) compounds. Whilef,,, of S0,ClF was fixed between the values of SO,Cl, and SO,F,, the situation is less obvious for the other bending force con- stants (cf: Table 8). This is mainly due to the fact that for S0,ClF a greater number of interaction force constants has been fitted. Although some changes in the structural param- eters occur by going from the gas phase to the solid phase, data of S0,XY (X, Y = C1, F) from both phases agree well (see Table 8 and ref. 4). Hyperfine Constants The principal values of the C1 quadrupole coupling tensor have been evaluated.For isotopomers with C, symmetry xcc = xyy (representation 1'). Thus xcc should be the same for the isotopic pair S02ClF/34S0,ClF, as well as for the pair S0237C1F/34S0,37C1F. In both cases small deviations are within twice the combined uncertainties. Furthermore, for both pairs of isotopomers, S0,ClF/S0,37ClF and 34S02C1F/34S0,37ClF, the ratios of xcc , 1.268 783 (32) and 1.268 806 (49), respectively, agree very well with the expected value of 1.2688773 (15).34 To determine I,, and xxx the pro- cedure set out in ref. 35 was applied to the species retaining C, symmetry upon substitution, namely S0,37C1F, 34S0,C1F Table 8 Structural parameters (pm, degrees) and diagonal force con- stants" (100 N m-') of selected sulfuryl halides S02F2 SO,ClF, SO,Cl, r(S0)lfSo 139.7p1.22 140.9/ 10.85 141.8/10.53 r(SCl)If,C, 198.6/3.03 20 1.2/2.24" r(SF)IfSF 153.0/5.24 154.0/5.03 L(OSO)~~,~~122.6/1.90 123.3/1.75 1 23.5/ 1.60 L(ClS0)If,,,O L( FSo)/!fFSO L(XSY)If,s, 108.6/1.94 96.7/2.17 109.O/ 1.6 1 107.3/1.79 98.1/2.25 108.0/1.76' 100.3/1.90 Deformation force constants normalized to 100 pm bond length.r, structural parameters, ref. 33; force constants, ref. 7. This work, I: structural parameters. rgstructural parameters, ref. 32; force con- stants, ref. 7. e Exchange of wavenumber assignments of v, and vg leads to values off,,, = 2.45 andfc,,o = 1.55; see ref. 39. principal values of the quadrupole tensor are thus x,, = -77.045 (30), xxx = 36.217 (30) and xyy = 40.828 (2) MHz.The asymmetry parameter q = 0.0599 indicates a moderately cylindrically symmetrical SCl bond. The value for xzr agrees well with those derived from solid-state NQR measurements (78.7 MHz).~~Similar values were derived for other sulfur- chlorine compounds: x,, = -74.4 MHz and q = 0.085 for S0,Cl,,37 xzz= -71.5 MHz and q = -0.132 for S0,C1NC0,38 x,, = -63.8 MHz and q = -0.014 for SOC1, ,39 and x,, = -79.7 MHz, q = 0.200 for SCl, ?' The variations in x,, reflect the differences in the bond length^,^' those in q the different amounts of 7c bonding.42 The values of xii (i = a, b, c) for SO'80ClF and SO'8037ClF can be evaluated from the principal values.For the former, using the rk parameters, the results are xaa = -76.646, Xbb = 39.132 and xcc = 37.514 MHz. Assuming the SCl bond to be the z-axis, one gets xzz= -60.723 and xxx= 28.544 MHz for S037ClF, so that the results for the latter are xaa = -60.401, xbb = 30.829 and xcc = 29.572 MHz. For both isotopomers the agreement with the experimental values in Table 2 is reasonable. The values of x-and xbc of SO 180C1F and so 37ClF have been used to calculate the angle ex, in these iso- topomers. The values of 121.96 (17)" and 122.05 (27)" for SO 180C1F and SO l8037ClF, respectively, are very close to the angle of rotation around the a-axis on substitution of one l60atom by l80: 123.23' and 123.30", respectively, as might be expected from the quadrupole analysis given above.Although the magnitudes of the nuclear spin-rotation con-stants are much smaller than those of the C1 quadrupole coupling constants, they are in general well determined. The only exceptions are the values of M,,(Cl) for some of the rare isotopomers. To a first-order approximation, these constants are proportional to gN Bi, where gN is the nuclear g-factor of the respective nucleus and Bi the corresponding rotational con~tant.~'Since gN(F) is CQ. 5.257 and gN(3sc1) is ca. 0.548,42 it is reasonable that the values of M,AF) are considerably larger than their corresponding Mi,(Cl) values. However, while for M, of S0,ClF the ratio is ca. 4.7, M,,(Cl) and Mbb(F) are much smaller than expected from this simple cal- culation.This approach should hold strictly for the ratios of Mii (i = a, b, c) of S0,ClF and S0,37C1F for both nuclei (F and Cl). The experimental ratios of 1.26 (44),1.24 (7) and 1.31 (9) for Ma,, Mbb and M,, , respectively, of chlorine and 0.989 (19), 0.94 (7) and 1.020 (26) for those of fluorine are in moderate-to-good agreement with the expected values of 1.201, 1.234, 1.234, 1.000, 1.0275 and 1.0274, re~pectively.~' Similar agreements are obtained for the 34S isotopomers. Given the large rotation of the principal axis system around the a-axis upon l80substitution, the larger deviations of Mbb(F) and M,,(F) are not unexpected. Flygare has shown that the spin-rotation constants can be accounted for by two terms.43 One, a nuclear term, depends 2610 Table 9 Nuclear and electronic contributions (kHz) to the experi- mental spin-rotation constants (kHz) of S0,ClF nuclear electronic experimental contribution contribution ~~ M,(CO 0.330 -0.101 0.413 Mbb(C1) 1.860 -0.390 2.250 M,, (C1) 1.740 -0.391 2.131 M,(F) 12.247 -7.520 19.767 Mbb(F) 3.252 -1.636 4.888 M, (F) 8.176 -5.062 13.238 only on the molecular geometry; this has been evaluated for all components using the ri structure, and has been sub- tracted from the experimental values to give the electronic contributions (Table 9).Compared with the electronic contri- butions, the nuclear ones are quite large for the F nucleus and somewhat smaller for Cl. The ratios M, : M,, :Mcc for both contributions and for both nuclei are roughly those of the experimental spin-rotation constants.Similarly, the "F spin-rotation constants have been taken to calculate the average paramagnetic shedding cf) of the F atom.,, With the rk structure cf) = -589.8 ppm is obtained. The experimental chemical shift of S0,ClF is -99.1 ppm relative to CC1,F.43 With the absolute shielding of 188.7 ppm for F in CC13F,45 this gives CT:;) = 89.6 ppm. The diamagnetic shielding is calculated by ahF)= 0:;) -t~$ = 679.4 ppm. This value is comparable to that of SiF, (655 ppm) and the calcu- lated values for SF, (665 and 681 ppm for Fa, and Feq, respectively), but is smaller than that of SF, (747 ppm), and larger than that of NaF (524 ppm) or KF (553 ppm)?6 By analogy to "F, a value gP= -889 ppm has been derived for 35Cl in SO,ClF, similar to oP= -989 ppm in SOCl, .39 We are not aware of any 35Cl NMR measurement of S0,ClF which can be used for comparison with this value.We are indebted to Prof. Dr. H. D. Rudolph, Ulm, for pro- viding his structure programs, for many suggestions on their use, and for a preprint of ref. 16. We would also like to thank Dr. W. Jager for help during the experiments, Dr. D. Chris- ten, Tubingen, for help in using his force field programs, and Prof. Dr. H. Willner, Hanover, for providing the sample. Funding by the Natural Sciences and Engineering Research Council of Canada (NSERC) is gratefully acknowledged. References 1 H. S. Booth and C. V. Hermann, J. Am. Chem. SOC.,1936,58,63.2 C. W. Holt and M. C. L. Gerry, Chem. Phys. Lett., 1971,9,621. 3 W. Gombler, 2.Naturforsch. B Chem. Sci., 1985,40,782. 4 D. Mootz and A. Merschenz-Quack, Acta Crystallogr., Sect. C 1988,44,924. 5 R. J. Gillespie and E. A. Robinson, Spectrochim. Acta, 1962, 18, 1473. 6 T. Birchall and R. J. Gillespie, Spectrochim. Acta, 1966,22,681. 7 M. Pfeiffer, 2. Phys. Chem. (Leipzig), 1969, 240, 380, and refer- ences therein. 8 N. C. Craig and K. Futamura, Spectrochim. Acta, Part A, 1989, 45, 507. 9 N. Heineking, W. Jager and M. C. L. Gerry, J. Mol. Spectrosc., 1993,158,69. 10 T. J. Balle and W. H. Flygare, Rev. Sci. Instrum., 1981,52,33. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 11 Y. Xu, W. Jager and M. C. L. Gerry, J. Mol. Spectrosc., 1992, 155,403.12 J. Haekel and H. Mader, Z. Naturforsch., A. Phys. Sci., 43,203. 13 Ed. G. Brauer, Handbuch der Praparativen Anorganischen Chemie, Enke Verlag, Stuttgart, 1975, 3rd edn., Bd. 1. 14 H. M. Pickett, J. Mol. Spectrosc., 1991, 148, 371. 15 H. D. Rudolph, Struct. Chem., 1991,2,581. 16 H. D. Rudolph, in Advances in Molecular Structure Research, ed. I. Hargittai and M.Hargittai, JAI Press, Greenwich (USA), vol. 1, to be published. 17 (a)K. Kuchitsu, J. Chem. Phys., 1968, 49,4456; (b)K.Kuchitsu, T. Fukuyama and Y. Morino, J. Mol. Struct., 1968, 1, 463; (c) 1969,4,41. 18 K. Epple and H. D. Rudolph, J. Mol. Spectrosc., 1992,152, 355. 19 H. D. Rudolph, J. Mol. Spectrosc., 1981, 89, 460 and references therein. 20 R. H. Schwendeman, in Critical Evaluation of Chemical and Physical Structural Information, ed.D. R. Lide and M. A. Paul, Academic Press, New York, 1976, pp. 1-67. 21 (a)C. C. Costain, Trans. Am. Crystallogr. ASSOC., 1966, 2, 157; (b) B. P. van Eijck, J. Mol. Spectrosc., 1982,91,348. 22 (a) M. LeGuennec, G. Wlodarczak, W. D. Chen, R. Bocquet and J. Demaison, J. Mol. Spectrosc., 1992, 153, 117; (b) M. LeGuen- nec, G. Wlodarczak, J. Demaison, H. Burger, M. Litz and H. Willner, J. Mol. Spectrosc., 1993, 157,419. 23 V. Typke, J. Mol. Spectrosc., 1978,69, 173. 24 H. D. Rudolph, J. Mol. Spectrosc., 1981,89,430. 25 L. Nemes, in Vibrational Spectra and Structure, ed. J. R. Durig, Elsevier, Amsterdam, 1984, vol. 13. 26 K. Kuchitsu and Y. Morino, Bull. Chem. SOC.Jpn., 1965,38,805. 27 E. Tiemann, H. Kanamori and E. Hirota, J. Mol. Spectrosc., 1989,137,278. 28 Y. Endo, K. Nagai, Ch. Yamada and E. Hirota, J. Mol. Spec-trosc., 1983,97,213. 29 D. Christen, J. Mol. Struct., 1978,48, 101. 30 (a) H. J. Becher, Fortschr. Chem. Forsch., 1968, 10, 156; (b) H. S. P. Muller and H. Willner, J. Phys. Chem., 1993,97,10589. 31 E. A. Robinson, Can. J. Chem., 1963,41,3021. 32 M. Hargittai and I. Hargittai, J. Mol. Struct., 1981, 73,253. 33 K. Hagen, V. R. Cross and K. Hedberg, J. Mol. Struct., 1978,44, 187. 34 J. H. Holloway, Ph.D. Thesis, Harvard University, 1956. 35 N. Heineking and H. Dreizler, Z. Naturforsch., A Phys. Sci., 1992,47, 51 1. 36 (a)R. M. Hart and M. A. Whitehead, Trans. Faraday SOC., 1971, 67, 3451; (b) I. P. Biryukov and A. Ya. Deich, Zh. Fiz. Khim., 1972,46,2385; Russ. J. Phys. Chem., 1972,46,1362. 37 I. Merke and €3. Dreizler, Z. Natuforsch., A Phys. Sci., 1992,47, 1153. 38 0.L. Jo, J. D. Graybeal, F. J. Lovas and R. D. Suenram, J. Mol. Spectrosc., 1992,152,261. 39 H. S. P. Muller and M. C. L. Gerry, to be submitted. 40 I. Merke and H. Dreizler, Z. Naturforsch., A Phys. Sci., 1992,47, 1141. 41 V. E. Bel'skii, V. A. Naumov and I. A. Nuretdinov, Dokl. Akad. Nauk SSSR, 1974, 215, 355; Dokl. Phys. Chem. Proc. Acad. Sci. USSR, 1974,215,260. 42 W. Gordy and R. L. Cook, Microwave Molecular Spectra, Wiley, New York, 3rd edn., 1984. 43 W. Flygare, J. Chem. Phys., 1964,41,793. 44 P. A. W. Dean and R. J. Gillespie, J. Am. Chem. SOC., 1969, 91, 7260. 45 D. K. Hindermann and C. D. Cornwell, J. Chem. Phys., 1968,48, 4148. 46 (a)J. Mason, J. Chem. Soc., Dalton Trans., 1975, 1426, and refer- ences therein; (b) Adv. Inorg. Chem. Radiochem., 1976, 18, 197; (c) Adv. Inorg. Chem. Radiochem., 1979,22, 199. Paper 4/01862A; Received 29th March, 1994