J . Chem. Soc., Furuduy Trans. 1 , 1982, 78, 2205-2214 Radiolysis of Tetrachloromethane? BY MARTYN C. R. SYMONS* Department of Chemistry, The University, Leicester LEI 7RH AND EMANUELE ALBANO, TREVOR F. SLATER AND ALDO TOMASI Department of Biochemistry, Brunel University, Uxbridge, Middlesex UB8 3PH Received 25th September, 1981 Electron spin resonance studies of tetrachloromethane after exposure to 6oCo y-rays at 77 K reveal the formation of sCC1, and CCl:: radicals. On warming in the presence of spin-traps, or on irradiating fluid solutions, nitroxide radical adducts have been detected that are characteristic of CCl, and chlorine atom adducts. In the light of this evidence and that of other investigators a mechanism for the radiolysis of tetrachloromethane is postulated. In the presence of oxygen, *CCI, radicals are converted into C1,COO- radicals. The use of spin-traps to detect these radicals is described and evaluated. Mishra and Symons have analysed the solid-state e.s.r.spectra of 12CCl, and 13CC13 in various frozen media and the results make it clear that these radicals are non-planar, with extensive delocalization of the unpaired electron onto chlorine.’ Hesse et a1.2 have shown that CCl, radicals give rise to complicated e.s.r. spectra in pure tetrachloro- methane, and have suggested that there are two trapping sites for these radicals, which undergo ‘free’ rotation at 185 K. Other studies have concentrated on pulse radiolysis with optical dete~tion,~? and on the use of spin-traps to give stable nitroxide radicals where e.s.r. spectra give information concerning the original radical^.^? ti The former studies are not clear-cut because of the difficulty of assigning broad structureless ultraviolet absorption bands to specific intermediates.In the most recent work, it is suggested that a band at ca. 340 nm is due to CCl,+ radicals, another at ca. 475 nm is due to CCli+-Cl- ion-pairs, and a third in the 370 nm region is due to C1; ions., In contrast, only neutral radicals are detected in spin-trapping experiments. In this case, however, evidence for .CC13 radicals is clear cut because of the detection of well resolved 13C hyperfine ti Apparently the *CCl, radical has no absorption maximum above 230 nm,7 and hence is not readily detected in pulse-radiolysis studies on pure tetrachloromethane, which is optically black in this region.Our present aim is to use e.s.r. spectroscopy to distinguish between ionic and neutral radical mechanisms for the radiolysis of tetrachloromethane. Our interest in tetrachloromethane arises because it is a powerful hepatotoxic agent used in studies of liver CCl, radicals being implicated as significant intermediate^.^. 6 y lo EXPERIMENTAL Chemicals were of the highest grade available and were used as supplied by the manufacturers without further purification: tetrachloromethane, (B.D.H. AnalaR); N-t-butyl-a-phenyl- nitrone (PBN), (Aldrich); 2-methyl-2-nitrosopropane (MNP), (Aldrich) ; tetrachloromethane, 13C- CC1, (90.2 atom %), diluted as indicated, (British Oxygen Co Ltd). t Taken as Radiation Mechanisms, Part 22.22052206 RADIOLYSIS OF TETRACHLOROMETHANE The radiolysis of tetrachloromethane in the presence of the spin-trapping agent was performed in silica e.s.r. tubes exposed to a source of “Co y-rays either at 77 K or at room temperature. The mean dose was 10 krad. When MNP was used as the spin trap, experiments were performed in the dark, to avoid photolysis. Solutions were degassed by the freeze-thaw technique, or oxygenated by passing a stream of oxygen through the solutions prior to freezing. In some experiments liquid oxygen was added to the fine powder obtained by dropping tetrachloromethane and a solution of the spin-trapping agent into liquid nitrogen. Photolysis was performed using the full arc of a 400 W mercury source focused on the e.s.r. cavity at room temperature.Spectra were measured on a Varian E-3 spectrometer fitted with a variable-temperature cavity . RESULTS AND DISCUSSION SOLID-STATE STUDIES The most direct method of probing the mechanism of radiolysis is solid-state e.s.r. spectroscopy. Two well separated sets of features result, one set being that previously assigned to *CC13 radicals,’. and the other, shown in fig. 1 (a), being tentatively assigned to CC&+ radicals. The central features are due to *CC13 radicals and radicals in the e.s.r. tube, but the other features are due to a species containing strongly coupled 35Cl and 37Cl nuclei (35Cl and 37Cl have I = 3/2: the abundance of 35Cl is 75.4% and that of 37Cl is 24.6%). The possibility that this species is atomic chlorine can be rejected because the highest field line shows at least three features for 35Cl and 37Cl combinations.A two-chlorine radical such as C1; could be responsible for these features, but if it was we would expect only seven parallel features and gll would be ca. 1.95, whereas for Cl; or (RCFClR)+ radicals, gI1 should be close to the free-spin value (2.0023). If three equivalent chlorine atoms are involved then there should be ten sets of features. On this interpretation, gll = 2.005; this is a reasonable value for gll and lends support to the interpretation that the species CCli+. In this case, the high-field parallel (-9/2) line should have four components (3 35Cl; 2 35Cl + 1 37Cl; 1 35Cl + 2 37Cl; 3 37Cl) in the approximate ratios 9: 7: 5: 3. The first three components I 3200 G (9JOL GHz) FIG.1.-For legend see facing page.M. C. R . SYMONS, E. ALBANO, T. F, SLATER A N D A. TOMASI 2207 + 3200 G (9.109 GHrI t FIG. 1 .-First derivative X-band em-. spectra for tetrachloromethane after exposure to ‘j0Co y-rays at 77 K. (a) Showing features assigned to CCla cations; central features are due to .CCI, radicals and radicals in the e.s.r. tube. (b) Showing features for RO, radicals formed in aerated solutions after annealing. g1 are well defined, but the last is largely hidden under the more intense -7/2 feature. The nossibilitv of four eauivalent chlorine niiclei can he riiled niit hemiice then there W V U l U UL. JV111b lVw-llblu palallcl llllL.3 V U L 3 l U C LllC lllLC113G ytlycllulLulal I c a L U l c J , which were not detected.As is frequently the case for powder spectra of such radicals, analysis of the ‘perpendicular’ feature is more difficult. However, the analysis suggested in fig. 2 does accommodate the form of the major low-field lines quite satisfactorily, and, in particular, it accommodates the nearly isotropic nature of the + 5 / 2 set of lines, which is the most characteristic feature of the spectrum. We suggest that this species is CCl,+. The only other reasonable species having three equivalent chlorine atoms is *CCl,, the spectrum of which is quite different.l? Following the loss of an electron, the tetrahedral CCl; cation is bound to be distorted, according to the Jahn-Teller theorem, and our results suggest that this distortion is such that three chlorine atoms move so that the ‘hole’ is shared between them, leaving the fourth chlorine uncoupled.The resulting orbital is a linear combina- tion of 3p (n) orbitals on chlorine which are non-bonding with respect to carbon. A possible form for this orbital is shown, viewed along the axis of the unique chlorine, as structure I. structure I2208 RADIOLYSIS OF TETRACHLOROMETHANE I 3290 G ?) ( 0 "2) (- P ;2) !. FIG. 2.-For legend see p. 221 1 . Clearly the measured parallel and perpendicular components are not principal values for the individual 3p orbitals, but are intermediate between the true parallel and perpendicular values. Hence the estimated anisotropy should be smaller than the true anisotropy. If the data given in table 1 are analysed in the usual way" we obtain Aiso = 46.5 G and 2 B = 12.9 G, after correcting for the shift in gl. The correction for orbital magnetism is no greater than our experimental errors and is therefore neglected.(This analysis is based on like signs for A,, and A l . The choice of opposite signs leads to impossible values for calculated orbital populations.) These results can be converted into approximate orbital populations,ll giving ca. 2.3% s character and 13 % p character on each chlorine atom. The former value is high, but possible for spin polarisation. The latter value leads to a total spin density of 39%, which is far too low. As stressed above, this value should be low, owing to the fact that the experimental data are the values along the symmetry axes and not principal values. We conclude that the assignment of these features to CCI,+ cations is well supported and this is accepted in the following discussion.These features are much weaker thanM. C. R. SYMONS, E. ALBANO, T. F. SLATER A N D A. TOMASI 2209 1 3240 G ii r FIG. 2.-For legend see p. 221 1 . those for *CCl, radicals, which explains why they were not detected by others.2 Nevertheless, since the lines are far broader than those for *CCl,, the two species are present in comparable concentrations. Brede et al. have assigned a band at ca. 340 nm to the CCl,+ cation., That there should be a relatively low-lying optical band is supported by the presence of ultraviolet or visible transitions for the isostructural radicals Poi2- and SO;-. They stress that the species must be distorted in order to prevent rapid charge transfer CCl, + + CCl, z$ CCl, + CCl, +.(1) Our results establish the nature of this distortion. The nature of the 475 nm species is less clear. Their assignment is to the ion-pair CCl,+. * .Cl-, which might well give rise to an intense charge-transfer band. Alternatively, reaction with C1- could give the cr* radical, structure 11, which could exhibit an intense cr-+o* absorption.2210 RADIOLYSIS OF TETRACHLOROMETHANE I I 3LOO G FIG. 2.-For legend see facing page. TABLE 1 .-E.s.R. PARAMETERS hyperfine coupling constants/Ga radical spin trap 14N 'H other CClfb CC13C .13cc1, - 13cc1,d '2CC13 13cci3e c1 Clf Cl,COO Cl,COO - g { or C13C0 Cl3C0O ROO or ROii R0.j ROij - '2CC1,k PBN PBN MNP MNP PBN PBN MNP MNP MNP PBN PBN PBN PBN MNP 13.5 14.10 13.1 12.7 12.2 12.12 27.0 27.5 27.0 13.5 13.7 13.5 13.5-1 3.6 6.75 1.5 1.74 - 0.7 0.75 35Cl, 59.4, 40, is0 46.5 35Cl, 20, ca.0, is0 6.7 13c, 9.4 13C, 9.68 35c1, 2.25 35c1, 2.4 35c1, 6.1 35c1, 6.05 - 1.6 1.75-1.85 2.1 1.3 13c, 5.7 35c1, 0.6 a 1 G = Ref. (6). Ref. (13). f E. G. Janzen, B. R. Knauer, L. T. Williams and W. B. Harrison, J. Phys. Chem., 1970, 74, 3025. N. Ohto, E. Niki and Y. Kamiya, J. Chem. Soc., Perkin Trans. 2, 1977, 1770. i Ref. (17). IC Probably ClCONO(CMe,). T. Isotropic unless indicated. * g,, = 2.005, gL = 2.043. Ref. (1). Ref. (14). Ref. (15).M. C. R. SYMONS, E. ALBANO, T. F. SLATER A N D A. TOMASI 221 1 + 3290 G \ 1 + ' + '2C 0 0 1 (:I) C;:2JI2) (I,* .1J G2) FIG. 2.-First derivative X-band e.s.r. spectra for various spin-trap nitroxide radicals: (a) PBN + C1* [(PhCH(Cl)NO(CMe,)J, (b) PBN and '2CC1,, (c) PBN + * 13CC1,, ( d ) MNP+ - 12CC1, and(e) MNP + * 13CCl, in the presence of oxygen.Features a (a) and /? (a), (c) and ( d ) are due to PhCH(CCl,)NO(CMe,) and acyl nitroxides, respectively. However, we have not obtained any evidence for such species, nor have we been able to detect C1; anions, which were also postulated by Brede et al.3 C1 \ ci - CCl I c 1 / c1 structure I1 If oxygen is not removed from the system, a new species is detected after annealing,2 the e.s.r. spectrum of which is given in fig. 1 (b). This has gll = 2.037 and gl = 2.003, these values being characteristic of ROO- radicals. These must be C1,COO radicals since (Cl,C-CIOO*)+ or ClOO would have quite different spectra.12 USE OF S P I N TRAPS Two different traps were used, PBN and MNP (structures I11 and IV).PBN usually adds radicals ( R e ) to carbon to give a nitroxide, structure V, and MNP usually adds to nitrogen, giving structure VI. Ph CMe, \ / C=N '0 / H Me,C-NO PBN, structure I11 MNP, structure IV2212 RADIOLYSIS OF TETRACHLOROMETHANE 0 \ a / / \ R H-C-N Ph CMe, /* R-No \ CMe, structure V structure VI Radicals of structure V display hyperfine coupling to 'H, 14N and to atoms directly bonded to carbon that are P to the electron on nitrogen and interact via 0--7t overlap (hyperconjugation). Radicals of structure VI give a 14N triplet with hyperfine features due to coupling to a and nuclei in R. Thus, for example, l3Cc1, radicals of structure V should exhibit strong coupling to 13C but none to 35/37C1, whereas radicals of structure VI should give resolved coupling to both types of nuclei.Our results are summarised in table 1. Only PBN gave an adduct with chlorine atoms, and this was unstable above ca. 270 K. The spectrum is, however, quite distinctive [fig. 2(a)], showing features for 35Cl and 37Cl, and a large value of AiSO, as expected for D chlorine atoms with good a-n overlap. Thus the formation of chlorine atom intermediates is established. This accords with the optical detection of C1; radicals,, presumably formed from C1. and C1-. Also present in the e.s.r. spectrum are features for *CCl, adducts (a), which are stable and can be seen clearly at room temperature. This spectrum shows only coupling to 14N and lH [fig. 2(b)], but when 13C-enriched CCl, was used, another set of lines displaying coupling to 13C was also resolved [fig.2(c)]. The coupling of 9.4 G is again characteristic of a P-carbon atom with good a-n overlap. In contrast, with MNP the 12CCl, adduct gave clear coupling to all three chloride nuclei [fig. 2 ( d ) ] , a result which is diagnostic of * CCl, radicals. The coupling to 35Cl and 37Cl nuclei of 2.25 G agrees reasonably with the literature value1, (table 1). These results with spin traps confirm the formation of *CCl, radicals and establish the presence of C1* atoms. Taken alone, they would seem to support the homolysis (2) mechanism CCI, -+ CCl, + + e- (3) CCl; + + e- --+ (CCl,)* --+ *CCl, + C1 in which (CC14)* is an electronically excited molecule. However, the solid-state results show that *CCl, radicals are formed together with CCl;+ cations in the primary process, so this sequence is inadequate.To explain the presence of CClt cations, electrons must react elsewhere and we suggest that the reaction CCl, + e- --+ CC1, + C1- (4) well known in protic media, must also occur. However, this should prevent the occurrence of step (3). In that case, the simplest way of explaining yields of chlorine atoms comparable with those of *CCl, radicals is by the dissociation ( 5 ) CClt * cc1; + c1 This may be an overall reaction occurring by more complicated routes, but we stress that CCl; is isostructural with stable molecules such as SO,, so reaction ( 5 ) may not be too unfavourable. Also CCl; cations would be rapidly removed by reaction with C1- ions formed in step (4). In view of the evidence for CClgCl- ion-pairs,, and the expected rapid formation of ion-pairs in this low dielectric medium, the step (6) CCllCl- --+ CCl, + C1 .M.C. R. SYMONS, E. ALBANO, T. F. SLATER A N D A. TOMASI 2213 is a source of C1* atoms that avoids the formation of CCl; cations. Steps (2), (4) and (6) give all the required products, and so we favour these as the most probable process for the radiolysis of tetrachloromethane. REACTION WITH OXYGEN Our solid-state results confirm the rapid and irreversible reaction of CCl, radicals with oxygen. Reactions of ROO and ROO* radicals with spin traps show that MNP is by far the most useful, as a characteristic species with a greatly increased coupling to 14N is produced (ca.27 G). However, for ROO* radicals there is some confusion about identification, since the 14N coupling constants obtained in reactions with ROO. radicals14 are similar to those obtained with RO* radi~a1s.l~ The 14N isotropic coupling is greatly increased for nitroderivatives because these radicals are pyramidal at nitrogen whereas simple dialkyl nitroxides are planar, or nearly so.16 We expect that RO- and ROO- derivatives will have very similar properties in this respect, and hence that A(14N) will be similar. Evidence for RO- rather than ROO- adducts came from the observation of only one coupled 170 nucleus. However, it is not certain that both oxygen nuclei should give rise to detectable coupling. Whatever the correct identification, the detection of radicals having a large coupling to 14N (ca.27 G) remains diagnostic of ROO radicals in the present systems. In contrast, reaction of PBN with ROO* radicals gave rise to a species having coupling constants to 14N and lH similar to those for many other radicals, including *12CC1,, and we would hesitate to claim that the slight differences are diagnostic of ROB formation. However, using *13CC1, radicals plus oxygen, no 13C splitting was observed, showing that the trapped species was not CCl;. Merritt and Johnson17 have suggested that in this case also, RO- rather than ROO. radicals are trapped. Their results for these two types of radicals (table 1) are quite similar, and since our results lie between them it is impossible for us to decide which derivative we are studying.We have attempted to make the result more positive by using l'o-enriched oxygen. Unfortunately, the lines are broadened by spin exchange with dioxygen in these experiments, and we were unable to detect any coupling to 170 nuclei. FORMATION OF RCONO(R) RADICALS In many cases, especially in the presence of oxygen, we have detected radicals having A(14N) z 7G (/? in fig. 2). These radicals are almost certainly acyl nitroxide RCONO(R), radicals. These are frequently detected in reactions with spin traps and a variety of mechanisms have been proposed to explain their formation. We do not intend to discuss this further except to mention our results for *13CC1, radicals with MPN. In addition to a coupling of 6.75 G to 14N, we detected a small coupling of ca.0.6 G to one chlorine and a clear coupling of ca. 5.7 G to 13C [fig. 2(e)]. The radical is almost certainly C113CONO(CMe,), but we hesitate to propose a route for its formation. PHOTOLYSES In order to check the e.s.r. spectra of some of these spin-trap adducts, we also studied the photolysis of tetrachloromethane in the presence of PBN and MNP. The major species formed were CCl, adducts, but considerable yields of acyl nitroxides were also obtained. We thank the National Foundation for Cancer Research, the Bossolasco Foun- dation and the Wellcome Trust for financial assistance. We are also indebted to Dr K. A. K. Lott for providing e.s.r. facilities and for helpful advice.2214 RADIOLYSIS OF TETRACHLOROMETHANE S. P. Mishra and M. C. R. Symons, Radiat. Phys. Chem., 1975, 7, 617. C. Hesse, N. Leray and J. Roncin, Mol. Phys., 1971, 22, 137. 0. Brede, J. Bos and R. Mehnert, Ber. Bunsenges. Phys. Chem., 1980,84, 63. R. E. Buhler, Radiat. Res. Rev., 1972, 4, 233; Helv. Chim. Acta, 1968, 51, 1558. A. Tomasi, E. Albano, K. A. K. Lott and T. F. Slater, FEBS Lett., 1980, 122, 303. J. L. Poyer, P. B. McCay, E. K. Lai, E. G. Janzen and E. R. Davis, Biochem. Biophys. Res. Commun., 1980, 94, 1 154; J. L. Poyer, R. A. Floyd, P. B. McCay, E. G. Janzen and E. R. Davis, Biochim. Biophys. Acta, 1978, 539, 402. ’ B. Lesigne, L. Gilles and R. J. Woods, Can. J. Chem., 1974, 52, 1135. * R. 0. Recknagel, Pharmacol. Rev., 1967, 19, 145. @ M. U. Dianzani, in Biochemical Mechanisms of Liver Injury, ed. T. R. Slater (Academic Press, London, 1978), pp. 45-95. M. C. R. Symons, Chemical and Biochemical Aspects of Electron Spin Resonance (Van Nostrand Reinhold, London, 1978). R. S. Eachus, P. R. Edwards, S. Subramanian and M. C. R. Symons, J. Chem. SOC. A , 1968, 1704. lo T. F. Slater, Nature (London), 1966, 209, 36. l4 I. H. Leaver, G. C. Ramsey and E. Suzuki, Aust. J. Chem., 1969, 22, 1891. 1Q J. Pfab. Tetrahedron Lett., 1978, 843. l5 A. Mackor, Th. A. J. W. Wajer and Th. J. de Boer, Tetrahedron Lett., 1967, 385. l6 J. H. Sharp and M. C. R. Symons, Nature (London), 1969, 224, 1297. M. V. Merritt and R. A. Johnson, J. Am. Chem. Soc., 1977,99, 3713. (PAPER 1 / 1492)