J. CHEM. SOC. FARADAY TRANS., 1994, 90(5), 721-726 721 Spin Trapping of Radicals formed upon Irradiation of Organobromine Compounds with Low-energy X-Rays Valentin E. Zubarevt and Alexander Halpern lnstitut fur Nuklearchemie , KFA Julich , 52425 Julich, Germany Solutions of several organobromine compounds in methanol or benzene were irradiated with Mo-Ka, p X-rays (€ = 17.79 keV). In contrast with y-rays, such low-energy photons interact predominantly (ca. 89-95%) with bromine causing Auger cascades and eventually the formation of highly excited moieties, which are possible precursors of free radicals. The irradiation was followed by the detection of free radicals through spin-trapping and EPR spectroscopy. Radicals from both the solvent and the solute were identified and quantified.The yield of radicals normalized to unit energy deposited by secondary electrons was studied as a function of the solvent and the compound structure. In benzene, this yield is very low, indicating that the organobromine moieties largely escape fragmentation, since their excess energy is easily channelled into collective excitation of the n-electrons of benzene molecules. In methanol two factors contribute to a more abundant formation of radicals: (a) the lack of a 'deactivation ' mechanism of excited moieties which now tend to decay more freely, yielding aryl radicals and (b) the in situ radiolysis of methanol, yielding additionally methoxy and hydroxymethyl radicals. The in situ radiolysis follows the high-LET pattern.As the interaction of photons of energy up to 30 keV with atomic systems is nearly completely photoelectric it causes inner-shell ionization of the atoms affected. The creation of a vacancy in an inner-shell of an atom triggers an Auger cascade with the emission of, on average, several Auger and Coster-Kronig electrons of low energies. The process results in multiply charged ions and in a high density of electrons in the proximity of the photoelectric event. Thus, two physical processes, acting concomitantly, contribute to the molecular consequences of the Auger cascade :(a)effects associated with the high positive charge and its neutralization and (b) radi-ation effects due to the short-range electrons emitted (in situ radiolysis).The relative importance of these two processes depends on the system under study (aggregation state, molec- ular size and structure etc.),but regardless of the mechanism of action, extensive local molecular effects are anticipated.' Most experimental studies of the molecular consequences of the Auger effect in condensed phases have been based upon measurements of stable products in organic molecules or strand breaks in DNA. Alternatively, they may be based upon the analysis of the intermediate species, e.g. free rad- icals. However, at ambient temperature most of the radicals disappear within the duration of an experiment, so that the steady-state concentration is too low for EPR detection. This limitation can be overcome by using nitrone or nitroso com- pounds as scavengers.Both of these, referred to as spin traps, react rapidly with a variety of short-lived radicals to form relatively long-lived adducts which can be identified by EPR spectroscopy.2-4 In the present study, solutions of several organobromine compounds in methanol or benzene were irradiated with Mo-Ka, p X-rays (I? = 17.79 keV) followed by the detection of free radicals through spin-trapping EPR spectroscopy. Radicals from both solute and solvent were identified and quantified. An important feature of the experiments is that such low-energy X-rays are absorbed preferentially by bromine, whereas there is little absorption by other elements. This selectivity contrasts with y-rays which interact indis- criminately with the sample.For comparison, similar solu- tions were irradiated with 6oCo y-rays. Experimental Materials and Sample Preparation All reagents were purchased from Aldrich, except the spin trap N-benzylidene-tert-butylamine N-oxide (PBN) which was synthesized by one of us (V.E.Z.). They were of the highest grade and were used without further purification. Aliquots of a 1 mol 1-' organic bromide and 0.1-0.5 mol 1-' PBN in methanol, or 0.05 mol 1-' of a spin trap 2,4,6-tri- tert-butylnitrosobenzene (BNB) in benzene, were degassed by repeated freeze-pumpthaw cycles, sealed in Suprasil quartz tubes (id = 2 mm) and used for irradiation. The effective volume of the sample under irradiation was 40 pl. Irradiation Procedure and Dosimetry The primary beam from an Mo-anode tube (Siemens AgMo 61), operated with a Kristaloflex 4 generator (voltage 20 kV, current 60 mA) was used.Samples were placed in a quartz Dewar flask with cold water and irradiations were carried out at 0°C. The low-energy part of the continuous X-ray spectrum was fully absorbed by quartz (3.5 mm) and water (3 mm), so that practically only the intense Ka (17.44 keV) and KB (19.61 keV) characteristic lines, in the ratio 0.84 : 0.16, reached the samples. The cross-sections for the weighted mean energy of these two lines (17.79 keV) were used in the calculations. They were estimated by interpolation of the data in ref. 5 and 6. Radiation exposure was measured with a small (0.02 cm3) air ionization chamber for low-energy X-rays (PTW Freiburg) attached to a calibrated dosemeter PTW DL4.The photon fluence per rontgen in the quasi-monoenergetic X-rays was 4.5 x lo9 photons cm-* R- ',t and the exposure rate at the point at which the photons entered the liquid sample was 11.4 kR min-' (2.94 C kg- ' min-'). The yields of radicals were determined from the initial linear part of the dose curves (spin-adduct concentration us. irradiation time). Numerical data in Tables 2 and 3 (later) refer to 1 min irradiations with X-rays, which corresponds to the target sample being subjected to lOI3 photons. t Visiting scientist, on leave from the Chemistry Department of Moscow State University. t 1 R = 2.58 x C kg-'. 722 EPR Measurements The EPR spectra were recorded at -6.5 "C on a Varian E-9 X-band spectrometer operating at 100 kHz modulation fre- quency, supported with an HP 9285A microcomputer to save the spectra on a magnetic tape. A ruby monocrystal cemented inside the cavity served as a reference.Magnetic scans were calibrated using a solution of a freshly prepared sample of Fremy's salt. The spectra were first run at 0.4 G modulation amplitude, avoiding saturation. Absolute spin concentrations were estimated by integrating the spectra run at 2 G modulation amplitude twice using the Varian ESR-935 acquisition system. Solutions of known concentra- tion of the stable aminooxy radical TEMPO (tetramethyl- piperidine-N-oxyl) in methanol or benzene served as standards. The relative contribution of individual radicals to the overall spectrum was determined using a computer program described in ref. 7.The hyperfine splittings (hfs) were elucidated by a comparison of the simulated and experimen- tal spectra. Results The spectra observed while irradiating pure methanol or a solution of LiBr in methanol are a composite of the PBN adducts of methoxy and hydroxymethyl radicals and hydro- gen atoms. The spectra of the methanol solutions of phenyl bromide, 2-bromopyridine, 3-bromopyridine, bromotoluene or 5-bromopyrimidine exhibit, in addition, lines belonging to the respective aromatic radical. For example, Fig. 1 shows the spectrum obtained from irradiation of 1 mol 1-' 3-bromopyridine and 0.1 mol I-' PBN in methanol.The EPR parameters are summarized in Table 1. The total yields of the trapped radicals (the sum of radicals from methanol and an aromatic solute) and the relative contribution of individual species are given in Tables 2 and 3. Table 1 EPR parameters of the PBN spin adducts in X-or y-irradiated solutions of RBr in methanol RBr radical a,/G aH BIG AH/G ~ ~~ none H 15.75 8.43 0.5 CH,OH 15.15 3.52 0.9 CH30 14.38 3.0 0.5 LiBr" H 15.85 8.55 0.5 CH,OH 15.4 3.6 1.05 CH3O 14.5 3.2 0.55 Br 15.05 3.05 0.45 15.0 2.92 0.55 14.87 2.8 0.45 15.12 4.75 0.55 CH, CH, 14.75 2.85 0.5 -The increase in uN stems from the increased polarity of the solution. J. CHEM. SOC. FARADAY TRANS., 1994, VOL.90 Fig. 1 Simulated spectrum of 1 mol 1-' 3-bromopyridine and 0.1 mol I-' PBN in CH30H; best fit to the experimental spectrum. Table 2 Total yield of trapped radicals in 1 moll-' RBr-CH,OH \ [ ,N-O']/10-6 rnol I-' [PBN] = [PBN] = [PBN] = RBr 0.1 rnol 1-' 0.2 rnol 1-0.5 mol 1-' none 2.1 -2.3 LiBr 7.2 -7.4 11.4 11.3 12.8Q Br 11.0 11.5 13.4 QBr 6.2 7.8 9.8QrBr CH, 10.2bBr Br 7.8cH3tYcH'I CH, 8.3 Table 3 Relative contributions of individual species RBr H(%) CH,OH(%) CH,O(%) R(%) C6H5Br 4 25 44 27 C,H,NBr 4 22 49 25 C,H ,N,Br 4 25 49 23 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 4 Differential absorption of 17.79 keV X-rays in 1 moll-' RBr in methanol RBr sample thickness/g cm - (pe,/p)/cm2 g- none 0.161 0.517 LiBr 0.184 4.32 C,H,Br 0.178 4.02 C,H,NBr 0.178 4.02 C,H,N,Br 0.178 4.03 (p,,/p)/cm2 g-': H, 0.012; C, 0.24; N, 0.48; 0,0.85; Br, 42.0; Li, 0.012.The irradiation of 0.05 mol I-' BNB in pure benzene for up to 10 min did not produce any visible spectrum. The spectra observed when a 1 mol 1-' solution of phenyl bromide or 3-bromopyridine in C6H6 or C6D6 in the pres- ence of 0.05 mol 1-' BNB were irradiated exhibited only one adduct for each solute, independent of the solvent. These were aminooxy radicals (1) But 'But 1 where R = C6H5 (uN = 9.75 G, u,(o, p) = 2.75 G, uH(~)= 0.75 G) or CSH,N (multiplet, not all lines resolved). In the perdeuteriobenzene system the adducts with deuteriated phenyl radicals (triplet, uN = 9.5 G, AH = 1.5 G) were not present, indicating that the radicals detected stemmed from RBr, not from the solvent.The yield of the trapped phenyl radicals was 0.8 x mol I-' in C6H6 and 1.1 x mol I-' in C6D6, and that of the pyriminidyl radical 0.5 x mol 1-l in both sol-vents (1 min irradiation). Discussion Methanol Systems Interpretation of these results requires some knowledge of the initial photon interaction processes, of how the imparted energy leads to the formation of radicals, and how they are eventually converted into PBN adducts. As will be shown later, RBr contributes 89% of the total absorption. Scheme 1 gives a qualitative account of the elementary physical pro- cesses which follow the photoelectric event in RBr (R is an aryl radical). This scheme is adapted to the specific needs of the present study which was biased towards the observation RBr CH OH (thehalization, solvation) e-[R Br]'.-.a ring R autoionization es-, H, CH,O, CH,OH cleavage (slow electrons) Scheme 1 absorption (Yo) total absorption (%) C N 0 Br 8 17.4 -82.2 -55 1.9 -8.8 89.4 51 2.3 -8.8 88.9 51 2.2 0.17 8.8 88.9 51 2.2 0.3 8.8 88.7 of the radical species, not other intermediates (ions, excited states) or stable products.The left-hand side illustrates the processes associated with the charge following the Auger cascade; the right-hand side shows the in situ radiolysis (see Introduction).All primary radicals react with PBN in competition with methanol and RBr ; the respective reactions will be discussed later. Absorption Characteristics of 17.9 keV Photons: Initial Photon Interaction Processes In Table 4 the sample thickness, the mass energy absorption coefficients, and the percentage of photons absorbed in the sample are given. From the elemental composition of the material and the partial absorption coefficients, the differen- tial absorption in individual component elements was also calculated. The dominant mode of photon interaction is with bromine, although the ratio RBr : CH,OH in a 1 mol 1-' solution is 1 : 24.5. To understand the attractiveness of using low-energy X-rays note that for 1.25 MeV prays the mass energy absorption coefficient (pe,,/p) is 0.046 cm2 g-' for 1 mol I-' PhBr-CH,OH, and that these photons interact with the component atoms with relative probabilities that correl- ate well with the ratio of the number of electrons from each element (21% with C, 13% with H, 24% with 0,41% with Br).Thus, reducing the energy of the incident photons from the MeV to the keV region increases the interaction coeffi- cient significantly and drastically changes the topography of absorption. Low-energy electrons set in motion by photons traversing an absorber and interacting with the component atoms are the main agent through which the radiation effects arise. We estimate the number of electrons released in the bulk sample, and the energy they carry, by a two-step procedure.First, knowing the number of photons entering the target sample (1013),the percentage absorption and the contributions of the individual component elements to the total absorption (Table 4), we obtain the number of photoelectric interactions with each type of atom in the sample (Table 5). Then, we evaluate the number of photoelectrons and Auger electrons emitted per interaction from each element in question, and their energy. The product of these two numbers, summed over all elements, gives the total number of secondary electrons and Table 5 Number of photoelectric interactions with components of 1 rnol I-' RBr-CH,OH (bulk sample) no. of interactions x lo-', RBr C N 0 Br total x ~~ ~~~~~ ~ none 0.14 -0.66 -0.8 LiBr 0.10 -0.48 4.92 5.5 C,H,Br 0.12 -0.45 4.53 5.1 C,H,N,Br 0.11 0.0087 0.45 4.53 5.1 C,H,N,Br 0.11 0.015 0.45 4.52 5.1 724 their total kinetic energy.(Previously,' we applied the same computational procedure to estimate the electron energy deposited in brominated DNA upon irradiation with low- energy X-rays.) Interaction of an X-ray photon with C, N and 0 yields one photoelectron per interaction which carries away a large frac- tion of the initial photon energy. The energy of photoelec- trons (EpE= 17.79 -E, keV, where E, is the binding energy of the K shell) is 17.51 keV for C, 17.38 keV for N and 17.26 keV for 0.Moreover, since only 1s and 2s orbitals are avail- able in the electronic structures of these elements, and the radiative relaxation is very low (<0.3%), it is assumed that one Auger electron arises for each interaction. The energy of Auger electrons, estimated by the 2 + 1 approximation,' is 276.6, 393.4 and 515.9 eV for C, N and 0,respectively.The case of bromine is more complex. First, the K, L and M shells take part in the absorption in the ratio 0.864 : 0.1 16 :0.016.5 Thus, per interaction, there will be 0.864, 0.116 and 0.016 photoelectrons of 4.32, 16.0 or 17.5 keV kinetic energy, respectively. Secondly, and more impor- tantly, a large number of the Auger transitions are allowed, down to N orbitals. Thus, regarding cascades in many atoms one would find quite different electron spectra. Humm and Charlton' developed an algorithm to determine the Auger electron spectrum following photon absorption in bromine, which took into account electrons from every allowed tran- sition and rendered an accurate physical picture.Here we did not follow this computational route, assuming that its accu- racy is of minor importance for the interpretation of the experimental results which themselves represent averages over many different cascades. Instead, we adopted a simpli- fied procedure which accounts only for the predominant elec- tronic processes of formation of Auger electrons, but ignores subshell effects and Coster-Kronig transitions. In this way, we obtain, per photoelectric interaction in Br, 0.35 KLL, 1.30 L,MM and 1.33 M shell Auger electrons of energy 9.96, 1.38 and 0.23 keV, respectively (fluorescent yields : 0.60 for the K shell and 0.016 for the L shell; we ignored the re-absorption of the fluorescent X-rays within the sample).This procedure gave a smaller number of electrons per interaction, but it accounted for their total energy, which is our main concern. Thus, we consider a spectrum of electrons of energy ranging from 0.23 keV (Br M shell Auger electrons) to ca. 17.5 keV (photoelectrons from C, N and 0).Since these elec- trons traverse distances from 11 nm to some 7000 nm, they do not escape from the sample (size: 2 mm x 14 mm), but dissipate their energy within it. (This illustrates another feature of low-energy X-rays, namely that electron equi-librium exists even in the absence of photon equilibrium.) From the calculations based on the above data the follow- ing picture emerges.The irradiation of 40 p1 of methanol for 1 min with 17.79 keV X-rays releases 1.39 x 10" photoelec-trons of energy 17.51 keV, 6.58 x 10" photoelectrons of energy 17.29 keV, 1.39 x 10'' Auger electrons of energy 0.277 keV and 6.58 x 10'' Auger electrons of energy 0.516 keV. Hence, 1.4 x 1013 keV of electron energy is deposited. For 1 rnol 1-' RBr-CH30H solution (40 pl, 1 min irradiation), the number of photoelectrons would be : 0.12 x 10l2 of energy 17.506 keV, 0.45 x 10l2 of energy 17.286 keV, 3.90 x 10l2 of energy 4.32 keV, 0.47 x 10l2 of energy 16.0 keV and 0.09 x 10l2 of energy 17.5 keV; the number of Auger electrons would be: 0.10 x 10l2 of energy 0.277 keV, 0.48 x 10l2 of energy 0.516 keV, 1.70 x 10l2 of energy 9.96 keV, 6.41 x 10l2 of energy 1.38 keV and 6.51 x loi2 of energy 0.23 keV.Hence, the electron energy deposited in the sample is 6.1 x 1013 keV (3.6 x 1013 keV from photoelectrons and 2.5 x IOl3 keV from Auger electrons). J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 It is interesting to relate these energies with the number of radicals detected. From the data in Table 2 it follows that 5.3 x loi3 methoxy and hydroxymethyl radicals were detected in methanol. In 1 mol 1-' C,H,Br-CH,OH 2.9 x 1014 radicals were detected, of which 2.0 x 1014 were methoxy and hydroxymethyl radicals, the rest being phenyl radicals. Thus, in methanol we detected 0.38 radicals per 100 eV electron energy, and in 1 mol 1-' RBr-CH,OH 0.47 rad- icals (0.33 methoxy and hydroxymethyl plus 0.14 aryl radicals) per 100 eV electron energy.In reality, as follows from competition kinetics, 0.2 mol I-' PBN will trap only 60% of the phenyl radicals produced, the rest reacting with methanol (compare k, and k,, below), so that the corrected value for the overall phenyl radical yield is 0.23 per 100 eV. Note that these radical yields are lower than those observed when irradiating with ,OCo y-rays. This is consistent with the fact that the effectiveness of ionizing radiation of different qualities is determined largely by details of their track struc- ture.'O." In contrast with 6oCoy-rays when the spur separa- tion in the recoil electron track is so large (300 nm) that each spur can be considered as isolated, with low-energy electrons the track consists of a dense cylindrical core of reactive species.This favours their intra-track recombination before radicals have a chance to diffuse into the bulk of the solution and react with the scavenger. A similar effect of decreasing ion yield from liquid hydrocarbons with photon energy in the range 5-30 keV has been reported." Previously, the lower radical yields in water or methanol irradiated with tritium B-particles (e= 5.7 keV) as compared with ,OCo y-rays were ascribed to the different degree of separation or overlap of the spur^.'^,'^ Fate of the Multicharged Ion (Left-hand side of Scheme 1.) The positive charge acquired by bromine (on average 7 +) is redistributed quickly (lo-.15-s) between Br and the aryl moiety. The subsequent neutralization of the charge on bromine by catching electrons from the solvent molecules releases epithermal/thermal elec- trons by autoionization. Simultaneously, during the neutral- ization of the charge on the aryl moiety up to 100 eV of excitation energy is deposited in this moiety. This greatly exceeds the C-Br bond energy (ca. 293 kJ per mol of bonds), and even the atomization energy of the phenyl radical (5112 kJ mol-').15 One would expect such a highly excited moiety to fragment. Deutzmann and Stocklin16 reported that iodo- uracil labelled with a 12,1 Auger emitter efficiently breaks down following neutralization of the charge on the pyrim- idine ring.Destruction of the aromatic ring in iodobenzene and iodotyrosine was also observed."*'' There is no reason to expect that ring cleavage would be less significant in our systems. This may result in the formation of either non-radical products, which of course escape EPR detection, or non-aromatic radicals, the PBN adducts of which, if formed, are concealed in the spectra of CH30 and CH,OH adducts, merely increasing their detected yields. Although it is reasonable to expect that the Auger charge neutralization releases enough excitation energy to bring about the endothermic ring cleavage, the PBN adducts with aromatic radicals contribute appreciably (ca. 25%) to the EPR spectra (Table 3). This may indicate a stabilizing influ- ence of conjugated ring systems (degradation of excitation energy without scission) which prevents fragmentation.Another important event which should yield R' is the inter- action of solvated electrons with RBr (vide infra). Fate of Secondary Electrons; Spin-trapping Reactions Turning now to the right-hand side of Scheme 1 we introduce chemical phenomena into the discussion. Electrons released J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 in the Auger cascades, having been slowed down and sol- vated, cause radiolysis of methanol, eventually giving rise to H atoms, CH,O and CH,OH radicals and es- . These species are then available to react with the solvent and any solute present. The lifetime of a radical with respect to the scavenger reaction is z = (l/k)cs, where k is the rate constant and c, is the concentration of PBN.Since k for the reactions of the radicals in question with PBN are all between 2 x lo7 and 1 x 10' 1 mol-' s-', it is clear that the spin adducts were formed at 2lo-, s, i.e. long after the completion of the spur reactions (< s). The possible out-of-spur reactions in liquid methanol in the presence of PBN and RBr can be summarized as follows: es-+ CH,OH --* CH,O-+ H (1) k, = 8.5 x lo3 1 mol-' s-'; ref. 19 H + CH,OH + H, + CH,OH (2) k, = 2 x lo6 1 mol-' s-'; ref. 19 CH,O + CH,OH -+ CH,OH + CH,OH (3) k, = 2.6 x lo5 1 mol-'s-'; ref. 19 es-+ PBN + 0'-+ PhCH = NC(CH,), (4a) 0'-+ CH,OH -+ OH + CH,OH (4b) k,, = 1.4 x 10" 1 mol-' s-'; ref.20 k,, = 8 x lo8 1 mol-'s-'; ref. 21 H + PBN -+ P, (5) k, = 5 x 10, 1 mol-' s-'; ref. 3 CH,O + PBN -+ P, (6) k6 = 1.2 x lo8 I mol-' s-'; ref. 22 CH,OH + PBN -+ P, (7) k, = 4.3 x lo7 1 mol-' s-'; ref. 23 es-+ RBr --* R + Br-(8) R + CH,OH -+ CH,OH + RH (9) P1-P4 are spin adducts x 0'II PhCH-N-C( CH ,) , where X = H, CH,O, CH,OH or R (aryl radical). For PhBr k, = 8.7 x lo9 1 mol-' s-l (in pr~panol),~~ k, = 8 x lo5 1 mol-' s-',,' k,, = 6.9 x lo7 1 mol-'s-'. For other RBr the rate constants are not known, but they are likely to be close to the above values. Methanol solvates electrons with a solvation time of 11 ps at 293 K,26 so that an appreciable fraction of es- reacts rapidly in the spurs in the ps to ns range (inhomogenous kinetic stage).The yields of solvated electrons, G(es-), at 30 ps, 5 ns and 1 p after energy deposition by fast electrons are 3.4 & 0.3, 2.7 and 2.0, re~pectively.~~,~~The in-spur decay of es- is even faster under high-LET conditions, as in the present investigation. Thus, in reactions (l), (4) and (S), by es-is meant only the fraction of electrons that have escaped from the spurs and subsequently follow homogenous kinetics. In pure methanol solvated electrons produce H atoms and CH,OH radicals [reactions (1) and (2)]. The presence of PBN greatly suppresses these processes, since es- now react much faster with PBN with a rate constant 1.4 x 10'' 1 mol-' s-', ultimately giving CH,OH [reaction (4)]. Also other radical species react with PBN with rate constants 2-3 orders of magnitude greater than with methanol.Hereafter the emphasis is on PBN adducts with aryl rad- icals. As indicated above, these radicals may be initiated by two processes: by the neutralization of the primary multi- charged aryl moiety and by the interaction of es- with RBr. These two phenomena are interdependent, i.e. the number of multicharged ions and of es- alter concomitantly. Let us compare the total radical yields in 1 mol I-' LiBr-CH,OH and 1 mol 1-' PhBr-CH,OH (0.2 mol 1-' PBN) normalized to the same number of photoelectric events in Br, i.e. the same number of ejected secondary electrons, in which case the direct methanol radiolysis [the right-hand side of Scheme 1, reactions (1)-(3)] is independent of the Br target.These yields are 7.4 x 4.53/4.92 x lo6 mol 1-' = 6.8 x mol 1-' and 11.3 x mol 1-', respectively. The difference of 4.5 x lop6 mol 1-', an increase of 40%, represents the contribution of radicals other than those from direct meth- anol radiolysis and must be attributed to radicals from the aryl compound, i.e. resulting from the processes depicted on the left-hand side of Scheme 1 plus reaction (8). Since the contribution of the PBN-Ph adduct was found to be 27% (Table 3), the additional 13% is obviously due to CH,OH radicals from reaction (9). Similar calculations for the 2-bromopyridine system show an increase of 41.6%in the total radical yield, as compared with 25% yield of R, the addi- tional 16.6%being due to CH,OH from reaction (9).The presence of one or three methyl groups in the arene (bromotoluene or bromornesitylene) decreases the yield of the scavenged radicals (Table 2). Since the physical phenomena associated with neighbouring atomic centres in the molecule, such as the energy shifts in the Auger spectra, are far too small to influence the formation of intermediates or products, we assume that the effect is of a chemical nature. However, the alternative explanation that when orbital overlap is more pronounced, the intramolecular charge neutralization is greater should not be ruled out. The significantly higher radical yield from 2-bromopyridine than from 3-bromo-pyridine is indicative in this respect. Benzene Systems The lack of adducts with deuteriated phenyl radicals in RBr-C,D, systems provides important evidence that all detected radicals stem from the organobromine solute and not from the benzene solvent.This makes the data interpreta- tion more straightforward as compared with the methanol systems, as it is apparent that only those processes associated with the Auger charge (left-hand side of Scheme 1) are mani- fest, whereas the amount of in situ radiolysis is too small to be detected. This is obviously so because of the high radi- ation resistance to benzene. We now estimate the electron energy deposited in a sample of 1 moll -' C,H,Br-C,H, subjected to 17.79 keV X-rays for 1 min. The input data are as follows: sample volume 40 pl, sample thickness 0.192 g ern-,, photon flux 1 x loi3 photons, mass energy absorption coefficient 4.11 cm2 g- '.The calculated total absorption of X-rays is 54.6%,of which 95.2%is by bromine and 4.8%by carbon. Applying the same computational procedure as before, the estimated electron energy deposited in the sample is 6.14 x lo', keV. From the measured yield of phenyl radicals, 0.8 x mol 1-', it follows that in a 40 p1 sample 2 x lo1, Ph radicals were trapped. Thus, we detected 0.033 Ph radicals per 100 eV elec-tron energy. We assumed here that BNB trapped all the phenyl radicals while the reaction of phenyl radicals with benzene to form the phenylcyclohexadienyl radical was neg- lible. This assumption is justified, since the rate constant of the reaction C,H, + C,H, is 4.8x lo4 1 mol-I s-~,~'i.e.four orders of magnitude lower than the rate constant for the addition of phenyl radicals to a nitroso scavenger. It is interesting to note that the number of phenyl radicals per unit absorbed energy is a factor seven higher when the bromobenzene target is dissolved in methanol than when it is dissolved in benzene (0.23 :0.03),although an equal amount of electron energy is deposited in the target. This result can be understood in terms of intermolecular energy transfer. The inescapable explanation lies in the ability of benzene mol- ecules surrounding the highly excited bromophenyl moiety [Ph. .Br]* (vide supra) to take away the excess energy and to channel it into collective excitation of their own n-electrons.This deactivation mechanism may prevent the moiety from fragmentation. Note that, as commonly accepted, the radi- ation stability of aromatic compounds depends on the ability of the conjugated ring system to degrade excitation energy without concentration of energy in one bond which conse- quently breaks. In methanol systems the [Ph. * .Br]* moieties do not enjoy this 'protective' deactivation mechanism. A similar observation has been reported17 in a study of the fate of iodobenzene labelled with Auger-emitted '''1 in methanol, hexane or benzene. In the 3-bromopyridine-benzene system a slightly lower yield of the pyridine radical is observed than that of the phenyl radical in the PhBr-benzene system.This may be due to the fact that the replacement of one CH unit by an N atom in the ring increases the electron attraction in the position meta to Br. We do not elaborate upon this result which does not alter the general picture. Summary The following picture emerges from the present work. Low- energy photons used for irradiation of 1 mol 1-l solutions of organobromine compounds in methanol or benzene interact (by the photoelectric effect) predominantly with the com-ponent bromine, eventually forming highly excited species [Re. -Br]: surrounded by solvent molecules. It is important to note that the situation resembles that encountered in photochemistry (direct excitation of a solute), rather than in conventional radiation chemistry (the solvent is excited and energy is then transferred to the solute).The fate of these species depends on the environment. In a medium capable of rapid removal of a large amount of energy, such as n-electron solvents, e.g. benzene, a large frac- tion of highly excited states can relax instantaneously, pre- venting them from unimolecular decay and concomitant formation of aryl radicals. The situation is more complex in methanol systems. The deactivation of highly excited moieties via energy transfer is now inoperative, so that they tend to decay more freely. Moreover, secondary low-energy (high- LET) electrons interact efficiently with methanol (radiation- sensitive molecules, in contrast to radiation-resistant benzene molecules) and the in situ radiolysis becomes important.This follows the high-LET pattern, as testified by lower yields of radicals than those observed upon external irradia- tion with y-rays. The fate of a molecule undergoing the Auger effect, studied here in terms of the yield of radicals formed, was shown to depend on the molecular structure and, above all, on the nature of the medium. This is an aspect that has not been given adequate consideration previously. In particular, the J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 radiobiological damage caused by Auger emitters has received wide attention in recent years. 198*30 On the theoreti- cal level, the energy imparted by Auger and Coster-Kronig electrons to a selected volume of a cell is estimated and com- pared with the radiation effect (strand breaks, cell survival) thereon.However, the theoretical methods in use do not account for the physicochemical properties of a specific region of a cell, such as the degree of polarization or of con- jugation (i.e. the presence of polar or non-polar molecules and of delocalized n-bonds). Our results leave little doubt that these are important factors which may modify the final effect. References 1 A. Halpern, in Handbook of Hot Atom Chemistry, ed. J. P. Adloff, P. P. Gaspar, M. Imamura, A. G. Maddock, T. Matsu- mura, H. Sano and K. Yoshihara. Kodansha, Tokyo, VCH, Weinheim, 1992, pp. 550-570. 2 M. J. Perkins, Adv. Phys. Org. Chem., 1980,17, 1. 3 V. E. Zubarev, The Spin-trapping Method (in Russian), Moscow University Press, Moscow, 1984.4 E. G. Janzen and D. L. Haire, Adv. Free Radical Chem., 1990, 253. J. H. Hubbel, Int. J. Radiat. Isotope, 1982,33, 1269. E. Storm and H. J. Israel, Nuclear Data Tables, 1970, A7, 565. U. M. Oeler and E. G. Janzen, Can. J. Chem., 1986,51,1138. H. Menke, W. Kohnlein, S. Joksch and A. Halpern, Int. J. Radiat. Biol., 1991,59, 85. 9 J. L. Hum and D. E. Charlton, in DNA Damage by Auger Emitters, ed. K. F. Baverstock and D. E. Charlton, Taylor & Francis, London, 1988. 10 A. Chatterjee and J. L. Magee, in Radiation Chemistry, ed. Far- thaziz and M. A. J. Rogers, VCH, Weinheim, 1987, pp. 173-199. 11 D. T. Goodhead and H. Nikjoo, Znt. J. Radiat. Biol., 1989, 55, 513. 12 R. A.Holroyd and T. K. Sham, J. Phys. Chem., 1985,89,2909. 13 J. Kroh, B. C. Green and J. W. T. Spinks, Can. J. Chem., 1962, 40,413. 14 A. Halpern, Chem. Phys. Lett., 1984,103, 523. 15 R. T. Sanderson, Polar Covalence, Academic Press, New York, 1983. 16 R. Deutzmann and G. Stocklin, Radiat. Res., 1981,87,10. 17 M. Reiche, A.Halpern and G. Stocklin, Radiochim. Acta, 1988, 43, 191. 18 M. S. Berridge, V. W. Jiang and M. J. Welch, Radiat. Res., 1980, 82,467. 19 D. W. Johnson and G. A. Salmon, J. Chem. SOC., Faraday Trans. 1, 1977, 73,256. 20 V. E. Zubarev, R. Mehnert and 0. Brede, Radiat. Phys. Chem., 1992,39, 281. 21 F. Dainton, I. Janovsky and G. A.Salmon, Proc. R. SOC. London, Ser. A, 1972,327,305. 22 V. E. Zubarev, V. N. Belevskij and L. T. Bugaenko, Moskow Univ. Chem. Bull., 1975,30,28. 23 C. L. Greenstock and R. H. Wiebe, Can. J. Chem., 1982, 60, 1560. 24 W. V. Sherman, J. Phys. Chem., 1966,70,2872. 25 K-D. Asmus and M. Bonifacic, in Landolt-Bornstein, New Series, Springer-Verlag, Berlin, 1984, vol. 13, part b, p. 26. 26 W. J. Chase and J. W. Hunt, J. Phys. Chem., 1975,79,2835. 27 T. Sumiyoshi, K. Tsugaru, T. Yamada and M. Katayoma, Bull. Chem. SOC. Jpn., 1985,58,3079. 28 J. Hunt, in Adv. Radiat. Chem., ed. M. Burton and J. L. Magee, J. Wiley, Chichester, 1976, vol. 5. 29 A. MacLachlan and R. L. McCarthy, J. Am. Chem. SOC., 1962, 84,2519. 30 Biophysical Aspects of Auger Processes, ed. R. H. Howell, V. R. Narra, K. S. R. Sastry and D. V. Rao, American Institute of Physics, Woodbury, NY, 1992. Paper 3/06290B; Received 21st October, 1993