J. CHEM. SOC. DALTON TRANS. 1992 1Study of Dialkyltellurium Radical Cations by Electron SpinResonance Spectroscopy *Matthew J. Almond," Atiah Raqabah,b David A. Rice,a Martyn C. R . Symonsb and Carol A. Yatesaa Department of Chemistry, University of Reading, P.O. Box 224, Whiteknights, Reading RG6 2A0, UKDepartment of Chemistry, University of L eicester, University Road, L eicester LEI 7RH, UKThe ESR spectra are reported for the radical cations (CH,),Te+ and (C,H,),Te+ produced by exposure ofvery dilute solutions of (CH,),Te or (C,H,),Te in CFCI, to 6oCo y-rays at 77 K. The results show that thespecies are n radicals and that the unpaired electron is strongly confined to the tellurium 5p, orbital.As expected for such n radicals, g, > g, > g, z 2.00, e.g.for (CH,),Te+, g, = 2.26,, g , = 2.13, andg, = 1.99,. Linear correlations between the spin-orbit coupling constants and the g,, g, values andbetween the methyl-group proton hyperfine coupling constants and ionisation potentials for(CH,),X+ (X = 0, S, Se or Te) are presented. When more concentrated solutions of (CH,),Te inCFCI, were similarly irradiated some evidence in the ESR spectrum was found for the presence of theox three-electron bonded species [CH,),Te-Te( CH,),] +.Recently there has been a resurgence of interest in organo-tellurium compounds because of their wide-ranging appli-cations in the semiconductor industry,, in organic synthesis,'and as reagents in nuclear medicine., We have recently carriedout a thorough spectroscopic study of the compounds R,Te[R = CH,, C,H, or (CH,),C] utilising the techniques ofinfrared, Raman, ultraviolet-visible and multinuclear NMRspectroscopy together with mass ~pectrometry.~.~ One of theprincipal aims of our work was to search for weak adductsformed between R,Te and organocadmium acceptor molecules,such as Cd(CH,),, since such adducts may be importantintermediates in the formation of thin layers of the semi-conductor materials CdTe and cadmium mercury telluride bymetal organic vapour phase epitaxy (MOVPE).6 In this contextit is of interest to learn something of the structure and bondingof the RzTe molecules, and, as a part of this work, we undertooka study of the ESR spectra of the R,Te+ (R = CH, or C2H5)radical cations.These radical cations are of considerable interest in their ownright.To our knowledge there are no previous reports of theESR spectra of organotellurium cations. Indeed there has beenrelatively little work on any inorganic or organometallic cationscontaining heavy atoms. The series of cations (CH,),X+( X = 0. S or Se) have all been studied by ESR spectroscopywhen trapped in solid CFCl, at 77 K.7p9 Thus observation ofthe (CH3)zTe+ cation assumes an importance in completingthis series and in allowing us to study trends in the ESR spectraon descending the group. We also hoped that by recording thesespectra we might learn something of the nature of the highest-occupied non-bonding molecular orbitaI(s) of the R,Te+cations. As these cations are likely to have a similar structureto those of the neutral parent R,Te molecules, identification ofthe highest-occupied molecular orbital in R,Te allows sub-stantiation of the donor orbital involved in adduct formationwith organocadmium acceptor molecules.For .s t m i f f cations, e.g. H,Of, CH,', H2COf, etc., aconvenient way to observe their ESR spectra is to trap thespecies in noble-gas matrices." Such methods have beendeveloped by Knight,', where the cations are produced by avariety of methods immediately prior to or during deposition.For larger cations, such as the ones we describe here, themethod developed in part by one of US,'^ in which very dilutesolutions (eg. in CFC1, solvent) of substrates are frozen andthen exposed to ionising radiation, is more successful. Recently,such methods have been used to study dialkylmercury deriv-atives.', It is the view of many chemists that ionising radiationrepresents a 'sledgehammer' which will break molecules in-discriminately into a wide variety of products.However, it isbecoming increasingly apparent, that, for properly controlledsystems, the products formed are often simpler and more readilypredictable than those produced by high-energy ultravioletlight.7 Alongside a study of R,Te+ cations one of our aims wasto search for the o* radicals [R,Te-TeR,] + formed by reaction(1) which contain unusual two-centre three-electron bonds.R,Te+ + R,Te - [R,Te-TeR,]' ( 1 )The R,Te compounds discussed in this paper have previouslybeen studied by both ultraviolet--visible absorption and photo-electron spectro~copy.~.'~ By the latter method an ionisationpotential of 7.9 eV is obtained for (CH,),Te.Thus (CH,),Tefcations should readily be formed in CFCl, frozen solutionssince the ionisation potential for CFCl, is ca. 11.8 eV."Furthermore if any solvent interaction were to be seen in thissystem it is likely to be via I9F rather than via "Cl or 37C1.9ExperimentalMethod qfPreparation.-Samples of (CH,),Te and (C2H5)2-Te were kindly supplied by Professor D. J. Cole-Hamilton.These samples and CFCl, (Aldrich, ca. 99.5% pure) weresubjected to trap-to-trap distillation on an all-glass vacuum linebefore use. Very dilute solutions of (CH,),Te and (C,H,),Te inCFCl, (ca. 2 mmol dmP3) were prepared in silica ESR tubes(diameter 5 mm).The tubes were sealed under vacuum and,after cooling to 77 K, were exposed to 6oCo y-rays in a Vickradsource with doses in the 0.5-1.0 Mrad range.14.18Spectra-The ESR spectra were measured using a Varian E109 spectrometer at 77 K. Samples were annealed by allowingthem to warm in an empty Dewar with continuous monitoringof the spectra. They were recooled to 77 K when significan2 J. CHEM. SOC. DALTON TRANS. 1992Table 1 Experimental ESR parameters for R,Te+ radicalsHyperfine coupling/(;1 2 5 ~ ~ *Radical 'H X Y gx g , gz(CH,),Te+ 10 (6 H) &80 & 10 +90 _+ 10 +960 & 20 2.26, 2.13, 1.99,(C,H,),Te+ 10 (2 H) & 100 &- 20 3320 k 20 5950 5 20 2.25, 2.13, 1.99,20 (2 H)* The magnetic moment of lz5Te is negative.13305 GGain x 10(a 1, //' :------- - / ''""IsxQ Y1.95' 1 I I J0 5 10 15 20Spin-orbit coupling constant APlots of gx and g,, values against spin-orbit coupling constants Fig.2for the series of ions (CH,),X+ (X = 0, S, Se or Te)Fig. 1 Observed (a) and computer-simulated (b) X-band ESR spectraassigned to (CH,),Te+ radicals, produced by exposure of (CH,),Te inCFCI, at 77 K to ,OCo y-rays. The central features in (a) are ascribableto the presence of dimer cationsspectral changes were noted. In attempts to study the dimercations [R,Te-TeR,] +, more concentrated solutions were used(ca. 6 mmol dm-3).Results and DiscussionData.-The experimental results are summarised in Table 1.There are several problems in detailed spectral interpretation,the major one being the presence of central ( g = 2) featureswhich are in part due to dimer cations and in part to impurityspecies and radicals in the silica tube.Spectra.-By analogy with results for similar n: radi-c a l ~ , ' ~ ~ ' ~ ~ ' ~ we expect to find that g , > g,, > g , = 2.00, withseptet splitting for (CH,),Te+ from six equivalent protons.Since 125Te has I = 3 and is ca.7% abundant, there should besatellite septets flanking each major feature. For an electron in ap, orbital we expect A, z A, -g A, with A, positive. l 9 (N.B. themagnetic moment of '"Te is actually negative, but for ease ofpresentation of arguments leading to spin-density estimates thenegative sign has been omitted: thus for absolute data all signsneed to be reversed.) It should be noted that 123Te is in suchlow abundance (0.9%) that 123Te radicals make a negligiblecontribution to the spectra considered herein.components is such that all ++ features should be intense, since the x, y and z componentsare all close together.However, all -3 features should becorrespondingly weak, especially the -$(A,) feature which iswell removed from all others. This is shown in Fig. 1(6), which isa computer simulation based on our best fit for the low-fieldfeatures.The form of the g andThe g values.-The large shift in g, and g , from 2.00 is theresult of coupling between the singly occupied molecular orbital(SOMO) ('b]) and the a,(a,) and ay(b2) orbitals, and isrendered large by the high value of the spin-orbit couplingconstants for Te.Experimental g, and g, values are plotted as afunction of the spin-orbit constants for X (X = 0, S, Se and Te)for a series of Me,X+ cations in Fig. 2. There is a goodcorrelation for both g, and g y in this series, which suggests thatthe structures of all the radicals are very similar.There should also be a link between the g shifts and theionisation potentials obtained from photoelectron spectra.' 6320The shifts are expected to correlate approximately with theinverse of the differences in ionisation potentials between thecoupled orbitals. Thus the shift in g, of 0.26 correlates with1/2.46 eV = E(a, - b,) and the shift of 0.13 in g, with 1/3.43eV = E(b2 - bl). If the photoelectron spectroscopic resultsand the g , values are accepted, g, is expected to be greater( z 0.19) than is found ( z 0.13).This difference probably reflectJ . CHEM. SOC. DALTON TRANS. 1992 3Table 2 Proton hyperfine splittings for (CHJ2X+ radicals andionisation potentials (i.p.) for (CH,),X molecules (X = 0, S, Se or Te)X A('H)/G i.p./eV0 42.3 9.94s 21 8.7Se 15 8.2Te 10 7.95 17.8 8.6 9.4 10.2lonisation potential / eVFig. 3 'Trends in the proton hyperfine coupling constant A('H) for(CH,)2X+ ( X = 0, S, Se or Te)I 1 I I Ithe fact that photoelectron spectroscopy (PS) gives the verticalionisation potential whilst the g values relate to the fully relaxedcations.Proton hj3perfine coupling. Trends in A(' H) on going fromMezOf to MezTef are shown in Table 2 and Fig.3 as a func-tion of the first ionisation potentials of the parent molecules.There is a remarkably good linear correlation which lendsstrong support to the concept proposed many years ago2'that hyperconjugation (o-n: delocalisation) involving electrondonation from C-H o bonds into the half-filled p orbital of II-type radicals is in part responsible for 'H hyperfine coupling toP-protons (i.c). those on the C atom bound to the heteroatom).The steady fall in A('H) on going from 0 to Te seems to beprimarily controlled by the decrease in ionisation potential.I t is curious that the proton hyperfine splitting for the diethylderivative is also cu. 10 G, and the number of lines is ca. 7. Ifthere were free rotation of the CH,CH2 units about the C-Tebonds then a splitting of 10 G would be reasonable, but thereshould only be five lines.One possibility is that there is a smallcoupling t o '"F (from the CFCl, solvent) as is observed forMez%' cations." However, there is no sign of this for Me2Te+so it is unlikely to occur for Et2Tef. The explanation that wefavour is that there is a preferred conformation (I11 in Fig. 4)such that one C)-proton for each -CH2- group is fixed close tothe average 45 position, the other being close to the maximum-overlap site. In this case, couplings of ca. 10 and 20 G for thetwo pairs of protons give seven lines as required. This analysisgives a far better fit for the intensities of the septet features thandoes any model involving coupling to I9F.Ethyl substituents in radical cations often display restrictedrotation with no clear preference for specific orientations." InIEt20*XIEt2S'rnEt2Te*Fig. 4 Preferred orientations of the Et,X+ radical cations viewedalong the axis of one of the C-X bonds (X = 0, S or Te).For clarityonly one ethyl group is shown in each casethe present series, for Et20f the conformation must be close toI with two pairs of equivalent strongly coupled protons (onepair from each ethyl group) giving a value for A( 'H) of 68.7 G.7In contrast, only three lines (from two equivalent protons) havebeen resolved for Et2S+,8 the coupling being close to theaverage value. Thus the other two protons must be close to thenodal plane as in 11.These protons and the CH3 groups seem tohave changed places in Et2Te+, 111. We offer no explanation forthese curious switches in preferred orientations. The trendsseem to confirm that there is no strong hyperconjugationpreference for C-H over C-C o bonds in such systems.14The I2'Te hyperjbe data. Although a range of 12'Te featuresare clearly defined for both Me2Te+ and Et,Te+ by the protonsplittings [see Fig. l(a)], it is not easy to be certain of theanalysis because various predicted features proved to beundetectable. Nevertheless the results shown in Table 3 give riseto reasonable simulations of the observed features and showwhy the others were not detected. In order to obtain some ideaas to the significance of the hyperfine data it is necessary toallow for orbital magnetic contributions to the coupling.Thishas been done after converting the data into MHz and usingequations (2)-(4) l 9 which are thought to be sufficientlyaccurate in view of the uncertainties involved.A , = Ais0 + 2B(1 + ;Ag,) (2)A , = Ai,, + B(l - ;Ag,) (3)A , = Aiso + B(l - SAgJ (4)The results for the 2B term are reasonable if A, and A , aretaken to be either positive or negative (see Table 3). The valuesfor Ais0 are all reasonable for an R2Tef radical, and show thatthe n: structure must be correct. Hence it is not possible todetermine the signs of the x and y components.Using equations (2) + (3) or (2) + (4) should give the samevalue for Ai,, and 2B. In fact there are small differences, but theoverall results confirm that the spin density in the pz orbital isclose to unity as required.An average value has been used forthe results in Table 3.In calculating apparent orbital populations we have used twoalternative sets of atomic parameter^.'^,^^ Neither set is idealfor heavy atoms but both indicate the expected high degree oflocalisation of the SOMO.Dimer Cations [R,Te-TeR,] +.-Although R 2 0 + radicalcations show no tendency to form [R20-OR2]+ dimercations,' these form readily for R2S+ (ref. 8) and R2Se+ (ref. 9)either on annealing or when relatively concentrated solutionsare used.Our results show clearly that, on annealing, features grow inthe central region (expected for these o* three-electron bondedspecies) as those for the parent cations decay.Also, for moreconcentrated systems the relative yields of R,Te+ radicals arereduced and those assigned to the dimer cations are increased.Unfortunately we have been unable to unravel these centralfeatures. This failure arises, in part, because of overlappingimpurity lines and, in part, because of very poor resolution. W4 J. CHEM. SOC. DALTON TRANS. 1992Table 3 Isotropic ( A ) and anisotropic (2B) hyperfine couplingconstants for lr5Te after allowing for orbital magnetic contributions.There are two sets of results because the signs of the x and y componentsare unknown (the sign of A , must be negative in view of its largemagnitude). Also shown are estimated 5s and 5p orbital contributionsusing ( i ) the A' and 2B0 values given in ref.1 and ( i i ) the correspondingvalues given in ref. 9as2 (%I ap2A,,/MHz 2B/MHz ( i ) (ii) (i) ( i i )x,y, negative 1310 1361 4.8 2.4 81 103x,y, positive 1016 1655 3.7 1.8 65 79expect to have considerable overlap between x, y and zcomponents, and, for [Me2Te-TeMe,] +,' overlapping protonhyperfine components. The net result is unfortunately too poorfor good spectral interpretation, but it seems certain thatdimerisation does occur.AcknowledgementsWe thank Professor D. J. Cole-Hamilton for providing samplesof the R,Te compounds, also the University of Reading for astudentship (for C. A. Y.).References1 S. Patai and Z . Rappoport (Editors), The Chemistry of OrganicSelenium and Tellurium Compounds, Wiley, New York, 1986, vols.1and 2.2 See, for example, M. J. Almond, D. A. Rice and C. A. Yates, Chem.Br., 1988, 24, 1130; S. J. C. Irvine and J. B. Mullin, J. Cryst. Growth,1981,55, 107; J. B. Mullin and S. J. C. Irvine, J. Vac. Sci. Technol. A,1986,4, 700; B. J. Morris, Appl. Phys. Lett., 1986,48.3 F. F. Knapp, jun., K. R. Ambrose and A. P. Callaghan, J. Nucl. Med.,1980,21,251, 258.4 M. J. Almond, C. A. Yates, D. A. Rice, P. J. Hendra and P. T. Brain,J. Mol. Struct., 1990,239,69; M. J. Almond, C. E. Jenkins, D. A. Riceand C. A. Yates, J . Mol. Struct., 1990,222, 219.5 C. A. Yates, Ph.D. Thesis, University of Reading, 1991.6 J. Tunnicliffe, S. J. C. Irvine, 0. D. Dosser and J. B. Mullin, J. Crvst.7 M. C. R. Symons and B. W. Wren, J . Chem. Soc., Perkin Trans. 2.8 D. N. Rao, M. C. R. Symons and B. W. Wren, J . Chem. Soc., Perkin9 T. Clark, A. Hasegawa and M. C . R. Symons, Chem. Phys. Lett.,Growth, 1984,68, 245.1984,511.Trans. 2, 1984, 168 1.1985, 116, 79.10 L. B. Knight, jun. and J. Steadman, J . Chem. Phys., 1983,78, 5940.11 L. B. Knight, jun., J. Steadman, D. Feller and E. R. Davidson, J. Am.2 L. B. Knight, jun. and J. Steadman, J . Chem. Phys., 1984,80, 1018.3 L. B. Knight,jun., Acc. Chem. Res., 1986, 19, 313.4 M. C. R. Symons, Chem. Soc. Reu., 1984,13,393.5 A. Hasegawa, J. Rideout and M. C. R. Symons, J. Chem. Soc., Dalton6 S. Cradock and R. A. Whiteford, J. Chem. Sac., Faraduv Trans. 2,17 D. N. Rao, H. Chandra and M. C. R. Symons, J. Chem. Soc., Perkin18 M. Shiotani, Mugn. Reson. Rev., 1987, 12, 333.19 M. C. R. Symons, Chemical and Biochemical Aspects of Electron SpinResonance Spectroscopy, Van Nostrand Reinhold, Wokingham,1978.Chem. Soc., 1984, 106, 3700.Truns., 1990, 637.1972,28 I .Trans. 2, 1984, I20 1.20 K. T. Higa and D. C. Harris, Organometallics, 1989,8, 1674.21 M. C. R. Symons, J . Chem. Soc., 1959,277.22 J. R. Morton and K. F. Preston, J. Magn. Reson., 1978,30, 577.Received 26th Juljj 199 1 ; Paper 110383 1