Vapour Phase Raman Spectra of the Molecules MH, (M = C, Si, Ge or Sn) and MF, (M = C, Si or Ge) Ranian Band Intensities, Bond Polarisability Derivatives and Bond Anisotropies BY ROBERT AND ROBINJ. H. CLARK*S. ARMSTROKG Christopher Ingold Laboratories, University College London, 20 Gordon Street, London WC1H OAJ Received 5th March, 1975 The vapour phase Raman spectra of the molecules MH4 (M = C, Si, Ge or Sn) and MF, (M = C,Si or Ge) have been recorded with 488.0 and/or 514.5 nm excitation at pressures of 0.5-1.0 atrn and at temperatures of ca. 295 K. The intensities of all four Raman-active fundamentals of each niole- cule have been determined relative to that of the vl(nl)band of methane as external standard, and this has permitted the calculation of Raman scattering cross sections for each band.Molecular (%;.) and bond (Zhx)polarisability derivatives have been calculated. The cCLHvalues are similar, for a gjven M, atom, t,o the cCLF values and all follow trends established previously for other ligands (X), Orcx rn clsix < ZCeX < iiinx. Bond anisotropies (~Mx)have been deduced from the Raman intensities of the v2(e)fundamental of each molecule. This method only yields the modulus of the ym value, but, when taken in conjunction with Kerr effect results, it is concluded that ym is positive in each case. Quadratic force fields for each molecule are established on the assumption of the Wolkenstein intensity theory, and mean square amplitudes of vibration are also calculated. Many studies have been made on the Raman spectrum of gaseous methane,' studies which have been concerned primarily with analyses of the vibration-rotation structure of the Raman-active bands.Studies along similar lines have only very recently appeared on the other Group IV spherical top hydrides and deuterides, SiH, and SiD4,29 GeH, and GeD4,5 and SnH,.6 Several studies of the frequency-corrected molar intensities (i.e. Raman scattering cross-sections, RSC or do/d92 values) of the v,(a,) fundamental of methane have been carried out, and, as argued el~ewhere,~ the most probable value seeins to be 3.035 x cm2 mol-l sr-l for 514.5 nm excitation. This value is the average of that of Schrotter and Bernstein,8 as corrected by Bern~tein,~ as corrected by Holzer.' ' and that of Holzer and Moser,lo Intensity studies on the Raman-active fundamentals of the other spherical top hydrides have not previously been carried out. In this work, the RSC values for all four Raman-active fundamentals of silane, germane and staiinane as well as for the three non-totally symmetric modes of methane were determined by comparison with that of the vl(al) fundamental of methane. Such studies permit the calculation of both molecular (Ej) and bond (&x) polarisability derivatives, bond anisotropy derivatives (yhx), bond anisotropies (yMx), force constants based on the Wolkenstein force field, as well as mean square amplitudes of vibration. Owing to the fact that the intensity of a Raman band is determined by the square of a polarisability derivative, it follows that the sign of the latter is not determined by Raman studies.However, by linking the present Raman results with Kerr effect measurements l2 it is possible to determine the sign of yILIx. Siniilar studies have also been carried out on the spherical top tetrafluorides CF4, 11 RAMAN SPECTRA OF MH, AND MF4 SiF, and GeF,, for which the only previous Ranian intensity results relate to measure-ments on all four bands of CF4 '' and on the vl(al) bands of CF, 13* ',and SiF4.13 The present intensity results are, however, the first to be obtained for these molecules with laser excitation. EXPERIMENTAL PREPARATION OF COMPOUNDS The methane and carbon tetrafluoride were obtained from Matheson Co.Inc. Samples of silane, germane and stannane were prepared by the methods of Norman et aZ.15and of silicon tetrafluoride and germanium tetrafluoride as described by Hoffman and Gutowsky.l INSTRUMENTAL The Raman spectra were recorded by use of a Spex 1401 spectrometer with Coherent Radiation model 52 Arf and Kr+ lasers. The scattered radiation was collected at 90" and focused by afj'0.95 lens onto the entrance slit of the 0.75 m Czerny-Turner monochsomater after having been passed through a polarisation scrambler. The gratings (Bausch and Lonib 1200 lines mid) were blazed at 500 nm. The method of detection was photon counting in conjunction with a cooled RCA C31034 (grade I) phototube and linear display. The power available at 488.0 nm and 514.5 nm was approximately 1.3 and 1.6 W respectively.Band areas were determined either with a Kent Chromalog Two integrater or by the cut-and-weigh procedure. The area measurements are accurate to +lo % unless otherwise stated. All quoted relative molar band intensities have been corrected for the spectral response of the instrument. Quoted wavenumbers of band maxima are believed to be accurate to & 0.5 cm-1 (neon calibration lines). The cell arrangement (extra-cavity) was the same as that used previ~usly.~ The cell (volume ca. 60 cm3) was filled with methane to a pressure of ca. 1 atm. The Raman signal strength was then optimised and the vl(al) band centred at 2917 cm-l scanned 6-12 times. The cell was then evacuated and the second compound allowed in, also to a pressure of ca. 1 atm.After temperature equilibrium had been re-established, the same procedure was repeated for each band. The entire procedure was then repeated using the bracketing technique, an average of five different samples of each molecule being used. Possible decomposition of the sample in the laser beam was checked by measuring the area of its vl(al) band frequently throughout a run. Only for germanium tetrafluoride and stannane was decomposition detected (stannane exploded after a period of 2-5 h in the laser beam). However, by contrast with the observations of Oskam, who used in-cavity techniques, no gradual loss of intensity of the stannane peaks was noted during the course of a run. The laser beam traversed the cell once only.In the case of carbon tetrafluoride, the intensity of 2v2(A1) (which has p = 0.0) has been added to that of the vl(al) fundamental ;this was deemed to be the most satisfactory proce- dure for dealing with the Fermi resonance problem, as has been suggested e1~ewhere.l~ The same procedure has been adopted for silicon tetrafluoride where 2v4(A1) (which also has p = 0.0) is in Fermi resonance with vl(al). This adds 4.2 % in the first case and 6.8 % in the second case to the measured intensities of the vl(al) band. RESULTS The intensities of all four fundamentals of the tetrahydrides MH, (M = C, Si, Ge or Sn) and tetrafluorides MF4 (M = C, Si or Ge) have been measured relative to that of the v,(al) band of methane as external standard.As indicated in the intro- duction, the scattering activity of the standard has been taken to be 1226 x cm4 g-' mol-' (203.5 Nx cm4 g-') which is equivalent to a Raman scattering cross section of 3.035 x cm2 mol-' sr-l for 514.5 nm excitation, to an i?i value of 2.13 A2N3 g-* and thus to an i?& value of 1.07 A2. The results are given in table 1. R. S. ARMSTRONG AND R. J. H. CLARK TABLERELATIVE MOLAR INTENSITIES AND RAMANSCATTERING CROSS SECTIONS OF RAMAN-ACTIVE BANDS OF THE MOLECULES STUDIED molecuIe FIcrn-1 a ZzMi lZiM2 da/U CH4 2916.7 1.oo 3.035 x 1533.3 0.120 0.364 514.5 nm 3019.5 1.900.626 0.0051305.9 -0.015 SiH4 2185.7 2.59 10.0~x 10-30 972.1 0.82 3.19 488.0 nm 2189.1 1.04 4.04 913.3 -0,Ol -0.04 2155.7 2.65 8.04~10-30 972.1 0.85 2.58 514.5 nm 2189.1 1 .065 -0.03 3.23 913.3 -0.01 GeH, 21 10.6 3.28 12.7~x 10-30 930.6 1.19 4.63 488.0 nm 2111.5 5.33-0.02 1.37 821.O N 0.08 10-3021 10.6 3.30 10.0~ 930.6 1.23 3.73 514.5 nm 2111.5 4.16-0.02 1.37 821.O -0.06 SnH4 1907.8 5.01 15.2x 10-j' 753.3 2.57 7.80 514.5 nm 1905.1 2.22 0.02 -0.06 6.74 681.O CF4 908.4 0.310 0.904X 434.5 0.093 0.28 514.5 nm 1283.0 0.071 0.22 63 1.2 0.131 0.40 10-30SiF4 800.8 0.344 1.04~ 264.2 0.105 0.320 514.5 nm 1029.6 0.0473 0.144 388.7 0.108 0.328 GeF4 735.0 0.611 1.855x 202.9 0.420 1.275 514.5 nm 800.1 0.165 0.501 273.1 0.273 0.829 a The quoted wavenumbers come from ref.(1) (CH,), (2) (SiH4), (4) (GeH4), (6) (SnH,), and (19) (CF4, SiF4 and GeF4).b The scattering cross sections are quoted for 295 K. At this tempera- ture the 2917 cm-l band, vl(al), of the standard, methane, has da/dQ = 3.889 x for 488.0 nm excitation and 3.035 x cm2 mol-l sr-I for 514.5 nm excitation. RAMAN SPECTRA OF MH4 AND MF4 The relative molar intensities have been converted to polarisability derivatives by use of the relationship -=-[ ]12Ml fg2 45Ei2+7yi2 I1Mz 91 45Ei2-+7y12 where 1-exp(-hcv,/kT) 1-exp(-hcv2/kT) (2) 5; and ~f are, respectively, the mean molecular polarisability derivative and the aniso- tropy derivative with respect to the normal coordinate Qj, degeneracy gj,vo is the exciting frequency, and vj is the frequency shift of the normal modej. The Raman scattering cross-section for the jth fundamental (dal/dQ) is given by the expression7 (dOjldC2) = 0.969 44 x 10-37fj(SA)j (3) where (SA)j, the scattering activity, is defined to be (SA), = gj(45E5’++yi2).(4) Bond polarisability derivatives were calculated via the relationship where mxis the mass of the X atom in a molecule of general forniula MX4. The relative scattering activities of the different fundamentals, together with the mean molecular polarisability derivatives, molecular anisotropy derivatives, MX bond lengths, and MX bond anisotropies yMx(= all -EL) are given in table 2. This last quantity has been deduced from yi by way of the relationship 17*l8 3 Y; = e. (;) 2.-FREQUENCY-CORRECTED RELATIVE MOLAR INTENSITIES OF RAMAN-ACTIVETABLE FUNDA-MENTALS OF GROUP IV TETRAHYDRIDES AND TETRAFLUORIDES, AND MOLECULAR POLARISABILITY DERIVATIVES OBTAINED THEREFROM 1W2MI PlM2) polarisability derivatives Vl(U1) vz(4 v3(f2) v4(t2) c(y9 Y; Y; y; rMxc IYMX CH, 1.00 0.046 0.664 -0.002 1.07 0.82 2.54 0.12 1.091 0.32 SiH4 1.660 0.179 0.669 -0.002 1.38 1.62 2.55 0.14 1.480 0.85 GeH4 2.08 0.246 0.825 -0.003 1.50 1.89 2.83 0.18 1.527 1.02 SnH4 2.585 0.395 1.14 -0.003 1.72 2.40 3.33 0.16 1.701 1.455 CF4 0.062 0.0070 0.021 0.0161 1.14 0.32 0.455 0.395 1.323 0.645 SiF4 0.057 0.0038 0.011 0.0069 1.11 0.235 0.32 0.26 1.55 0.56 GeF4 0.091 0.010l) 0.0274 0.010~ 1.40 0.38 0.52 0.32 1.67 0.98 a Mean values for exciting lines Arf 488.0 nm and 514.5 nm.b The units of &hxare A2,off2, y 1 and yi areA28-3 (x N*),of rm are& and of (all -a&x are A3.C TablesofInteratomic Distances and Configuration in MolecuZesand Ions (Chem. SOC. Spec. Publ., No. 1 1 and 18, The Chemical Society, London, 1958 and 1965). Survey spectra of silane, germane and stannane are given in fig. 1-3. Scale expanded spectra of silane and germane are as good as, and of stannane are superior to, those obtained by Oskam and co-workers 2-5 who employed a more powerful laser than that used for this work, and also the in-cavity technique. The spectra of R. S. ARMSTRONG AND R. 3. H. CLARK the tetrafluorides (fig. 4-6) are similar to those obtained by Clark and Rippon,lg with one important exception. The unusual contour of the v3(t2)band of germanium tetrafluoride was reinvestigated in an attempt to understand the reason for the apparent very strong Q-branch.It became clear that germanium tetrafluoride 11xo.05 111 i/crn-' FIG.1.-Vapour phase Raman spectrum of silane at 295 K. The instrumental settings were as foi-lows :scanning speed (s.s.) 20 cm-l min-l, slit widths (s.w.) 200/300/200 pm, slit height (s.h.) 10 mm, time constant (t.c.) 0.25 s, gain 5 K (inset for vl, 100 K ; 1 K = 1000count s-ll n GeH, (464torr) iicrn-' FIG.2.-Vapour phase Raman spectrum of germane at 295 K. The instrumental settings were the same as for fig. 1, except for S.S. 40 cm-I min-'. attacks silicone grease during the course of the measurements to produce some silicon tetrafluoride and that the ~3(t2)band of the former is coincident with the v,(a,) band of the latter.This is evident because in an 1-Lscan of the v3(t,) band of ger- RAMAN SPECTRA OF MH, AND MF4 inanium tetrafluoride (fig. 6) the Q branch is barely evident whereas in an Itotalscan it is very pronounced. The contour of a non-totally symmetric band should, of course, be the same for an IL as an Itotalscan. Accordingly, suitable corrections (amounting to 10 %) were made to the Itotalspectrum of germanium tetrafluoride in the vl(t2) region to allow for the contribution made by the underlying vl(nl) band of the silicon tetrafluoride impurity. Additional weak (presumably impurity) bands at 676, 648 (medium), 616, 538, 478 and 390 cm-I also occur. SnH, (461 torr) I I I I, I ,I ,!I> 4 I I I , I I I 2100 2000 woo 1800 1700 900 eoo 700 600 ilcrn-' FIG.3.-Vapour phase Raman spectrum of stannane at 295 K.The instrumental settings were S.S. 40cm-' min-l, S.W. 150/200/150 pm, s.h. 10mm, t.c. 0.4s, gain 5 K (inset for vl, 50 K). CF, (760tom) G1crn-l FIG.4.-Vapour phase Raman spectrum of carbon tetrafluoride at 295 K. The instrumental settings were S.S. 40cm-l mine', S.W. 200/300/200 pm, s.h. 10 mn, t.c. 1 s, gain 500 count s-l (inset for vl, 5K). The situation restricts the accuracy of the intensity data on germanium tetra- fluoride to & 15 %. The present intensity results for carbon tetrafluoride differ substantially from those obtained by Holzer who used mercury arc excitation. Whilst the intensities of the vl(a,)and v,(e) bands are greater than Holzer's (by ca.25 %) those of the v,(t2)and R. S. ARMSTRONG AND R. J. H. CLARK v4(t2)bands are smaller (by ca. 65 % and 10 % respectively). Although no explana-tion for this difference can be offered, it is widely recognised that the presently used techniques are superior to those used previously. I I I I I 1 1 I If I 1 1 1 1 1 It00 1000 900 800 11 400 300 200 i/crn-l FIG.5.--B;apour phase Raman spectrum of silicon tetrafluoride at 295 K. The instrumental settings were the same as for fig. 4,except for the gain being 1 K (inset for vl, 10 K). Ge F4 '(760torr ) -1 ' ' I ' ' 800 700 300 200 100 v/cm-l FIG.6.-Vapour phase Raman spectrum of germanium tetrafluoride at 295 K.The instrumental settings were the same as for fig. 5. The diagram also includes an I1 scan, which demonstrates that the Q branch of the Y3(r2) fundamental is coincident with that of the v,(al)fundamental of the silicon tetrafluoride impurity. RAMAN SPECTRA OF MH4 AND MF4 DISCUSSION POLARISABILITY FUNCTIONS The bond polarisability derivatives (ELx)deduced as described above, and listed in table 2, lie in the order z& < ZSiH < ZLeH < ainH and && < 6$iF < a&& i.e. taken in conjunction with earlier results,' abx values are shown to increase in the order E;,(1.07) < E&(1.14) < E&(2.38) < EbBr(3.O3)* and ZiiH(1.38) 2 ZiiF(l.11) < EiiC1(2.68)< E4iBr(2.91)* < EkiI(4.21)" and E&,,(1.50) % EbeF(l.40)< E&.ec1(3.57) < ELeBr(3.61)*< &,,(5.02)* and Z&( 1.72) < EknC1(3.71) (all values are in A2;values known only for cyclohexane solutions are indicated by an asterisk).The interpretation of these trends is not simple, but the features of dependence of Ejitx values on both p, the fractional covalent character of the MX bond, and on some power function of the bond length seem clear.* The present values for both ELF and EkiF (1.14 and 1.11 A2 respectively, each & 5 %) are greater than the values previously foundby Long and Thomas13 (0.94-L 5 % and 0.90+_20% A2 respectively, after conversion to the same value for the intensity of the standard as used herein). Even after making due allowance for the estimated experimental error in each of the Ehx values, a small discrepancy remains, and it is concluded that the present measurements, being based on photon counting rather than photographic techniques, are the more accurate.Moreover, the previous measurements on silicon tetrafluoride are evidently subject to considerable uncertainty owing to attack by the compound on the mirrors. Two other values for &F for carbon tetrafluoride, again based on mercury arc measurements, appear in the literature.ll* l4 One of these l4 is 1.03A2& 5 %, after conversion to the present intensity scale, and thus it is in agreement (within the experimental uncertainty) with the present results. The other" is 0.96A2 (again after conversion to the present intensity scale), but the errors in this value were not specified.The parallel and perpendicular bond polarisability derivatives, ail and respec-tively, have been deduced from the relationships 20-23 atx = +(a;]+2a;) (7) and iyf = (2/J3)[L,,(ail -ai)+L,i(ail -aL)/?-], = 3 or 4 (8) * The semi-theoretical delta function expression for ZLx reducesto Zhx = (2/3>(Xtp/Zeffa,)(~n>r3, where x and Zeff are the geometric means of the electronegativities and effective nuclear charges, respectively, of M and X, Zeffin each case being taken to be the atomic number minus the number of inner-shell electrons, a. is the Bohr radius, p is the Pauling fractional covalent character, n is the MX bond order and r is the equilibrium MX internuclear distance.R. S. ARMSTRONG AND R. J. H. CLARK where L4iare entries in the t2block of the L matrix, which relates symmetry to normal coordinates (see later). The results are given in table 3, from which it is seen that xi] > a? and the ratio ctl/ajl is in the range 0.15-0.33, i.e. similar to that found for the MC1, MBr and MI bonds of molecular tetrahalides in the vapour phase.? Also in common with earlier results, cc\l is seen to be much larger and in general more sensitive to the nature of the chemical bond than is the case for a;; ail increases in the order MH > MF < MC1 < MBr < MI. TABLE3.-PARALLEL AND PERPENDICULAR BOND POLARISABILITY DERIVATIVES (A2)OBTAINED FROM THE WOLKENSTEINFORCE FIELD molecule (334SiH4 .1; 2.47 2.75 a; 0.37 0.70 a;/.;I 0.15 0.25 GeH4 3.01 0.74 0.25 SnH4 3.47 0.84 0.24 CF4 2.31 0.51 0.22 SiF4 2.02 0.66 0.33 GeF4 2.77 0.71 0.26 CALCULATION OF w FIELDS A description of the procedures used for the calculation of symmetry force con- stants from fhdarnental frequencies together with Raman band intensities has been given elsewhere.'* 20-23 Owing to sign ambiguities in the columns of the L matrices, corresponding to 180" changes of phase of the normal coordinates, the six- teen possible L matrices are reduced to four distinct ones.2o Of these, two are un- acceptable in that the off-diagonal elements Lijare comparable with (MH,) or larger than jMF4) the diagonal elements Lii ;moreover, the calculated values for the mean square amplitudes of vibration for these two solutions are inconsistent with the electron diffraction values.Of the remaining two matrices, for only one is the same set of and a1 values obtained from eqn (8) for both i = 3 as well as i = 4, and hence this solution is regarded as the only acceptable one.* The results are to be compared with those obtained by use of fundamental frequencies and Coriolis con- stants as constraints to the force field.16 The agreement between the two sets of force constants is not good, and this suggests (as indicated elsewhere) 24 that73 2og the Wolkenstein assumptions are not completely valid. Root mean square amplitudes of vibration have also been calculated and compared with some values obtained by electron diffraction measurements.The results seem satisfactory, when the insensi- tivity of these parameters to the force field is recalled. Both the force constants as well as the root mean square amplitudes of vibration are available on request. BOND ANISOTROPIES Bond anisotropies, as obtained from Raman intensity measurements on the v,(e) fundamental of each spherical top MX4 molecule, are determined in magnitude but not in sign (cf. table 2). Values for anisotropic bond polarisabilities, all and al, may be obtained by combining the two solutions for y with values of the sum, o! 11 +2x1 (bL+2bTin the context of Kerr effect studies). The latter are determined from 9 (a,]f2aJ = --EP (9)47cN * On this basis, the only acceptable soIution to the Wolkenstein force field calculations described in ref.(7) and (21) is that labelled W2 for all tetrahalides except TiCI4 and TiBr4, for which it is that labelled W1. RAMAN SPECTRA OF MH4 AND MF4 where is the bond electronic polarisation. may be obtained from R,, the bond refraction extrapolated to infinite wavelength, or approximately from 0.95 RD,where RDis the bond refraction for the NaD line. TABLE4.-ANISOTROPIC BOND POLARISABILITIES a OBTAINED FROM MEAN POLARISABILITIES AND BOND ANISOTROPIES FOR GROUP IV TETRAHYDRLDES AND TETRAFLUORIDES bond (all +2a_~) IYMXI all alia 11 a II aA da II C-H 1.955 0.32 0.865 0.545 0.63 0.44 0.76 1.73 Si-H 3.555 0.85 1.75 0.90 0.51 0.62 1.47 2.37 Ge-H 4.05 1.02 2.03 1.01 0.50 0.67 1.69 2.52 Sn-H 5.07f 1.455 2.66 1.205 0.45 0.72 2.175 3.02 C-F 2.12 0.645 1.14 0.49 0.43 0.28 0.92 3.29 Si-F 2.495 0.56 1.205 0.645 0.53 0.455 1.02 2.24 Ge-F 2.88 0.9s 1.61 0.6, 0.39 0.30 1.29 4.3 a In units of A3.b Obtained by application of eqn (9). CR. J. W. Le Fkvre, B. J. Orr and G. L. D. Ritchie, J. Chem. Soc., 1966, 273. d Derived from EP(Si-H) or EP(si-F), K. L. Wamaswarny and H. E. Watson, Proc. Roy. Soc. A, 1936, 156, 144. e From 0.95 RD (Ge-H), J. Satge, Ann. Clzim., 1961,6, 519. f From 0.95 RD (Sn-H), P. M. Christopher and J. M. Fitzgerald, Jr., Austral. J. Clzem., 1965,18, 1709. g Derived from EP(C-F), K. L. Ramaswamy,Proc. Indian Acad. Sci. A, 1935,2,630. h From 0.95 Rn (Ge-F), Comprehensive Inorganic Chemistry, ed. J. C. Bailar, H.J. Emeleus, R. S. Nyholm and A. F. Trotman-Dickenson (Pergamon, London, 1973), vol. 2, p. 23. The two sets of anisotropic bond polarisabilities appear in table 4 for each bond studied. Separate application of the two sets of data to appropriate molecules yields molecular anisotropy values which may be compared with those obtained experi- mentally from the Kerr effect. With this approach no firm distinction could be drawn regarding the sign of yCqH owing to the small absolute value thereof (0.32A3). However for Si-H, Ge-H, C-F and Si-F bonds, for which the necessary Kerr effect data already exist, there is indicated a clear preference for taking yMxto be positive. The detailed arguments are presented el~ewhere.~~ By analogy (and in the absence of supporting Kerr effect data), it is suggested that YMX values for Ge-F and Sn-H bonds also have positive signs.Thus it is concluded that all > a1 and that both all and a1 increase in the order CCCX< asix < aGeX < aSnX. One of us (R. S. A.) thanks the Royal Society for the award of a Commonwealth Bursary. The S.R.C. and the Administrators of the University of London Central Research Fund are thanked for financial support. A. Weber, in The Raman Efect, ed. A. Anderson (Dekker, New York, 1973), vol. 2, p. 543, and references therein. H. W. Kattenberg and A. Oskam, J. Mol. Spectr., 1974, 49, 52. D. V. Willetts, W. J. Jones and A. G. Robiette, J. Mol. Spectr., 1975, 55, 200. H. W. Kattenberg, W. Gabes and A. Oskam, J.Mol. Spectr., 1972,44,425. H. W. Kattenberg, R. Elst and A. Oskam, J. Mol. Spectr., 1973, 47, 55. H. W. Kattenberg and A. Oskam, J. Mol. Spectr., 1974, 51, 377.’R. J. H. Clark and P. D. Mitchell, J.C.S. Faraday 11, 1975, 71, 515. H. W. Schrotter and H. J. Bernstein, J. MuZ. Spectr., 1964, 12, 1. H. J. Bernstein, J. Mol. Spectr., 1967, 22, 122. W. Holzer and H. Moser, J. Mol. Spectr., 1964, 13, 430. 11 W. Holzer, J, Mol. Spectr., 1968, 25, 123. l2 C. G. Le Fkvre and R. J. W. Le Fkvre, in Physical Methods of Chemistry, ed. A. Weissberger and B. Rossiter (Wiley, London, 1972), vol. I, part IIIC, chap. VI,p. 399 ;R. S. Armstrong,R.J. W. Le F&vre and K. R. Skamp, J.C.S. Dalton, to be published. R. S. ARMSTRONG AND R. J. H. CLARK l3 D.A. Long and E. L. Thomas, Trans. Faraday Soc,, 1963,1026. l4 G. W. Chantry and L. A. Woodward, quoted by K. A. Taylor and L. A. Woodward, Proc. Roy. Soc. A, 1961, 264, 558. A. D. Norman, J. R. Webster and W. L. Jolly, Inorg. Synth., 1968, 11, 170. l6 C. J. Hoffman and H. S. Gutowsky, Inorg. Synth., 1953, 4, 145. l7 W. F. Murphy, W. Holzer and H. J. Bernstein, Appl. Spectr., 1969, 23,211. G. W. Chantry, in The Raman Eflect, ed. A. Anderson (Dekker, New York, 1971), vol. 1. l9 R. J. H. Clark and D. M. Rippon, J. Mol. Spectr., 1972, 44,479. 2o R. J. H. Clark and P. D. Mitchell, J. Mol. Spectr., 1974, 51, 458. 21 D. A. Long, Proc. Roy. SGC.A, 1953,219,203. 22 D. A. Long, D. C. Milner and A. G. Thomas, Proc. Roy. SOC.A, 1956,237,197. 23 G. W. Chantry and L. A. Woodward, Trurts. Faraday SOC.,1960, 5&,1110. 24 K. A. Taylor and L. A. Woodward, Proc. Roy. SOC.A, 1961,264, 55s. 25 R. S. Arrnstrong and R. J. H. Clark, J.C.S. Dalton, to be published. (PAPER 5 /442)