Yield and Decay of the Hydrated Electron in Proton Tracks A Pulse Radiolysis Study BY W. G. BURNS AND R. MAY Radiation and Surface Chemistry Group, Chemistry Division, A.E.R.E. Marwell, Oxfordshire, OX1 1 ORA AND G. V. BUXTON AND G. S. TOUGH University of Leeds, Cookridge Radiation Research Centre, Cookridge Hospital, Leeds LS 16 6QB Receitled 20th December, 1976 Direct observation of the time dependence of the yield of e, produced in water by nanosecond pulses of 3 MeV protons is reported for the first time. From the effect of added hydroxide ion, G(eJ is estimated to be at least 3.6 at lo-'' s, falling to 2.3 at lO-'s. The experimental measure- ments do not agree with predictions of the present diffusion model; modificatioiis to improve the model are discussed. Reaction (1) describes the radiolysis of water at about lo-' s after the passage of the ionising particle through the medium H20 --v-+ e,, H, OH, H02, M2, H202, H+.(1) Most of the G-values" of these species have been measuredl in steady state radiolysis experiments for a wide range of radiation LET (linear energy transfer). It is generally believed that the primary chemical species, e,, H, OH and H+, are generated inhomo- geneously in local volumes called spurs, blobs and tracks and that the products measured at s are those which exist when the primary species have either reacted together or diffused to a point where the probability of their reaction with a partner generated in the same local volume is negligible. On this basis the diffusion model has been developed with the aim of predicting both (a) the effects of scavengers2 and (b) the eEect of radiation type3 and LET on the steady state G-values of reaction (1).This is achieved by the inclusion of known or estimated values for some parameters (rate and diffusion constants) and selecting others (track or spur radii, initial G-values and energy per spur) to give agreement with experimental data. In the course of proceeding from the initial to the final state, the spatial distribution of each species modelled, and therefore the quantity of each present, is necessarily calculated from zero time onwards as a function of time. This predictive aspect of the model may now be tested since recent developments in pulse radiolysis methods * The C-value is the number of a species created or destroyed per 100 eV of absorbed energy.Steady state yields which apply to reaction (1) at low scavenger concentrations are written as G,; under other conditions yields are written as G(X).48 YIELD A N D DECAY OF THE HYDRATED ELECTRON I N PROTON TRACKS have made possible direct observation of e; from 10-I' s onwards for low LET radiation, typically G0.23 eV nm-l. Such experiment^^-^ clearly show that the early decay of e; is about an order of magnitude slower than that predicted by the diffusion model. No pulse radiolysis data have been reported for high LET radiation, so that the time profile of G(e;) is not known under such conditions. We are investigating, therefore, the pulse radiolysis of water and aqueous solutions at high LET by using 1 ns pulses of 3 MeV protons, which have a mean LET of 29 el? nm-l, to obtain information on the decay of e& and to test the diffusion model under these conditions.In this paper we report our preliminary findings which show features similar to those of the low LET data.4-7 EXPERIMENTAL Full details of the techniques and equipment used in this work will be described elsewhere,8 and only a brief description of the novel features is given here. 3 MeV protons were produced in nominally 1 ns pulses by the Harwell Van de Graaff IBIS facility9 at a repetition rate of 0.01, 0.1 or 1 MHz and a peak current of about 5 mA. The path length of 3 MeV protons in water is only 0.15 mm, so it was advantageous to restrict the sample to this thickness. This was achieved by irradiating the sample in the form of a streamlined jet flowing at about 8 m s-l across the proton beam as in fig. 1.The proton beam was focused at the jet into a 1 mm diameter spot, but slight instability in the beam transport system caused the position of this spot to wander within an area about 2 mm in diameter. Analysing light from a 2 mW He/Ne CW laser was focused with a 20 cm focal length lens so that it passed coaxially through the jet acting as a light pipe. The transmitted light intensity was measured with a photomultiplier and oscilloscope combination which had an overall risetime of about 5 ns. All solutions were prepared from distilled water and were irradiated under de-aerated conditions at ambient temperature. All solutes were AnalaR grade and were used as supplied.E from laser U lmrn I I FIG. 1.-Detail of proton pulse radiolysis system. A: Stainless Steel Jet, B: Spectrosil window, 1 mm thick, C: 633 nm light from He/Ne Laser, D: Water Jet, E: Proton Beam. RESULTS AND DISCUSSION Fig. 2 shows the formation and decay of the e& absorption signal produced in de-aerated water by a 1 ns pulse of 3 MeV protons. We attribute most of the time taken for the signal to reach its maximum value to the risetime of the detection system since the pulse width was 1 ns (fwhm), e i is expected to be formed in <4 ps,l0 and the major effect of the detection system risetime is to delay the signal by that amount, although some distortion of the early decay of the signal occurs. We have not deconvoluted the signal from risetime effects and, for the present, we shall assume thatW.G . BURNS, R. M A Y , G . V. BUXTON AND G . S . TOUGH 49 245 rnV 235 U 20 ns FIG. 2.-Oscilloscope trace from proton pulse irradiated deaerated water; ;1 = 633 nm; pulse width 1 ns. the signal maximum in fig. 2 occurs at 1 ns in order to place the ensuing decay on an absolute timescale. Evidently the decay of e; is very rapid for about 5 ns and then much slower for the next 100 ns or so. At this time under low LET conditions, the residual e& signal is generally identified4-7 with G,, in reaction (1). It seems reasonable, therefore, to adopt a similar procedure here for high LET radiation. No value of Ge, for 3 MeV protons appears to have been published, but a value of 1.07 can be interpolated from data on the dependence of Ge, on LET determined by Appleby and Schwarz" who used 7 x mol dm-3 nitrous oxide to scavenge e;.Equating this value with our measured absorption at 140 ns, we estimate that G(eJ = 2.35 6 0.20 at the signal maximum in fig. 2. Table 1 shows the effect of added solutes, which scavenge Hf and OH, on the yields of e i at the end of the pulse and at 70 ns after the pulse. 1 mol dm-3 OH- increased G(e4) significantly at both times, but addition of 1 mol dm-3 MeOH had TABLE ELECTRON YIELDS IN WATER AND AOUEOUS SOLUTIONS PULSE IRRADIATED WITH solution H2O 1 rnol dm-3 MeOH 1 rnol dm-3 OH- 3 MeV PROTONS G(e,)"lb AG(t)' AG(s)* end of pulse (1 ns) 2.35 & 0.20 1.18 k 0.19 1.17 f 0.28 - 2.67 f 0.17 1.20 k 0.25 1.47 f 0.30 0.32 f 0.26 3.58 f 0.31 1.86 i 0.27 1.72 f 0.41 1.23 & 0.37 after 70 ns 1 rnol dm-3 MeOH + 1 rnol dm-3 OH- 3.66 f 0.43 a Assuming G(e,) in H20 at 140 ns = 1.07 (see text).2.19 f 0.29 1.47 f 0.52 1.31 f 0.47 Quoted uncertainties are 1 standard deviation. AG(s) is the change in G(e&) at end of pulse produced by the scavengers. little effect on the yields in water or in 1 mol dm-3 OH- solution. At these solute concentrations the lifetime of H+ is s and that of OH is s. It is not surprising, therefore, that 1 mol dm-3 MeOH has no appreciable effect on G(ez) at the end of the pulse, but the large increase produced by 1 niol dm-3 OH- shows that reaction (2) is important between AG(t) is the change in G(eJ between end of pulse and 70 ns. s and low9 s. ea; + H+ -j H. (2) Table 1 also shows that the decrease in G(e;) after the pulse is rather similar for all the solutions, which suggests that reactions (2) and (3) are not very important in the decay of e; after s.Presumably the decay which we observed at these times is due largely to other reactions such as (4) and (5): ea; + OH -+ OH- (3)50 YIELD AND DECAY OF THE HYDRATED ELECTRON I N PROTON TRACKS Although we have only been able to measure G(ei) at a few nanoseconds we can get a good approximation of the yields at shorter times from the observed change in G(e,) and the lifetime of the scavenged reaction partner of the electron following the procedure used for low LET r a d i a t i ~ n . ~ Thus we equate the observed end of pulse G(ei) in 1 mol dm-3 OH- solution with G(eJ at lo-" s since this is approximately the lifetime of H+, We justify the use of this procedure since in the low LET case it provided good agreement with values subsequently measured by direct method^.^*^*'^ Fig.3 compares the time dependence of G(e,",) for water obtained directly for fast electrons'' and obtained as described above for 3 MeV protons. Two features are noteworthy. First, the initial values of G(ei) for the two qualities of radiation differ by about one unit, possibly because reaction (3) occurs before s; experiments to test this point are planned. Other possibilities are (a) that e; decays earlier than s in the higher LET case, (b) that reactions involving unsolvated (dry) electrons13 may reduce the initial G(e&) and (c) that the observed difference is due to the relatively slow response of the detection system.Deconvolution procedures are being in- vestigated. The second feature is that e; decays earlier and to a greater extent at a higher LET, in qualitative accord with diffusion model predictions. We now discuss 5 the implications of the results in fig. 3 on the diffusion model. 1 I I I 1 I 0 Cr 10 4 I I I I 1 10" lox) ld9 1 2 Ki7 106 t / s FIG. 3.-Decay of e, in proton and electron pulse irradiated water; 0-3 MeV proton pulses; this work, 0-fast electron pulses: Jonah et al., ref. (12). The model as developed by Schwarz2 predicts rather exactly the effects of scavenger concentration on low LET G-values. It is not satisfactory at higher LET, introduced into the model as the energy averaged LET, despite the attempt to reduce the dis- crepancies by including the influence of the impact parameter on spur separati0n.l' The Kupperman method l4 incorporates the distribution in the magnitude of energy loss events, which depends on the radiation quality, and if the modelling concept is valid this method should be more capable of simulating LET effects.The Schwarz method includes a similar consideration at low LET but does not take account of changes in energy degradation spectrum with radiation type. Parameters for both models, as used in our calculations, are given in table 2. Both methods fail to agree with experimental observation on the early decay of e; for low LET radiation, predicting too early a decay which can be corrected3 byW . G . BURNS, R. M A Y , G . V . BUXTON AND G .S . TOUGH 51 TABLE 2.-PARAMETERS OF DIFFUSION MODELS Schwarz model Kupperman model initial species G-value e , 4.78 H 0.62 H2 0.15 OH 5.70 H202 0 H + 4.78 energy per spur spur enlarge- ment factor diffusion constant/ cm2 s-l 4.5 x 10-5 7 x 10-5 5 x 10-5 2.8 x 10-5 2.2 x 10-5 9 x 10-5 62.5 eV 2.5 spur or track radius/nm 2.46 1.15 1.15 1.15 1.15 1.15 initial G-value 4.33 0.38 0 4.7 1 0 4.33 diffusion spur or constant/ track cm2 s-l radius/nm 4.5 x 10-5 1.875 8 x 0.625 2 x 10-5 0.625 10 x 10-5 0.625 - - - 1.4 x 10-5 48 eV 3.6 multiplying both the energy per spur and the spur radius by a spur enlargement factor (see table 2). The factor is smaller for the Schwarz model (2.5) than for the Kupperman model (3.6) because the spur size and energy are larger in the former case.Fig. 4 shows the decay predicted by the unmodified Schwarz model for water alone, and by the corrected model for water alone and water containing scavengers for OH and H+ 5 4 3 h v 1 0 e J Y 2 1 0 10" 10" lo" lo' t / s FIG. 4.-Comparison of observed and calculated decay of e, at low LET. (1) Unmodified Schwarz model, water alone; (2) expanded spur Schwarz model, water alone; (3) expanded spur Schwarz model, water + 1 mol dm-3 MeOH; (4) expanded spur Schwarz model, water + 1 mol dm-3 NaOH; ( 5 ) expanded spur Schwarz model, water + 1 mol dm-3 MeOH + 1 mol dmP3 NaOH; 0 Experi- mental data from ref. (12). separately and together. Agreement with experiment12 is good for water alone (see fig. 4) and is also reasonable for the case with alkali present.12 The predicted effect with OH scavenger present is discussed later.A suitable e i yield against time curve for 3 MeV protons could not be calculated by inserting the mean LET (29 eV nm-l) into the Schwarz model since, with the scavenger concentrations used by Appleby and Schwarz,ll a G,, of 0.55 was obtained compared with the interpolated experimental value of 1.07. The Kupperman model52 YIELD AND DECAY OF THE HYDRATED ELECTRON I N PROTON TRACKS was, therefore, applied using the energy degradation spectrum obtained by the BurchIS method, the Kupperman rate constant^'^ and the parameters in table 2. The Ganguly- Magee-Schwarz equations were used, thereby avoiding a possibly arbitrary join between spherical and cylindrical geometry, but incurring the approximation of modified prescribed diffusion. The results (see table 3) were 0.64 for the normal spur, and 1.11 for the expanded spur, in better agreement with experiment.However, to avoid excessive computation in a preliminary comparison with experiment, the method used was to select a mean LET for the Schwarz model (12.3 eV nrn-l) which gives the experimental value of G e ~ = 1.07 when the model contains the scavenger concen- trations used by Appleby and Schwarz." With these parameters the model gives predictions within 12% of values of the other measured yields (those of H, H2 and H2Q2) interpolated for 3 MeV protons. Fig. 5 shows the calculated e i decay so obtained for the conditions indicated, together with the experimental points from fig. 3. The plotted points agree better with the unmodified Schwarz model, curve 1, than with the expanded spur curve 2, although the points for times below 5 ns may have to be revised upwards, as discussed earlier, so that they approach closer to curve 2.5 1 I I I 4 0 1 I I I I 1 16'' l P lo9 Xie 1 6 ~ li? t / s FIG. 5.-Comparison of observed and calculated decay of e, at high LET. (1) Unmodified Schwarz model, water alone; (2) expanded spur Schwarz model, water alone; (3) expanded spur Schwarz model, water + 1 mol dm-3 MeOH; (4) expanded spur Schwarz model, water + 1 rnol dm-3 NaOH; ( 5 ) expanded spur Schwarz model, water + 1 mol dm-3 MeOH + 1 mol dm-3 NaOH; 0, experimental data-this work. There are a number of reasons, however, why the enlarged spur is not a satisfactory solution to the problem of these discrepancies.When applied to the effects of scavengers on low LET steady state yields, it causes changes in G-values of l0-15%, vitiating the original very good agreement of the Schwarz model.2 In addition it causes the changes resulting from increased LET to occur more slowly (table 3). This latter effect may be advantageous, especially for the Schwarz model. However, the large spur increases an already significant error in unacceptably overpredicting the effect of OH scavenging. In the low l2 and high LET cases the large effect predicted fur 1 rnol d ~ n - ~ methanol (Comparable with that for 1 mol dm-3 hydroxide) is not observed experimentally. This discrepancy is due to the predominance assigned byW. G. BURNS, R . MAY, G . V . BUXTON AND G .S . TOUGH 53 the geometry of the model to reaction (3) compared with (2). This predominance, which is marginally less in the Kupperman model, and which our results suggest should be inverted, arises because of the higher rate constant of reaction (2) and because the higher diffusion constant of H+ compared with OH quickly reduces the concentration of H+. Another modification which should now be considered, in view of this discrepancy, is to change the shape of the initial electron distribution so as to give initially a low concentration at the spur or track centre and higher concentration zones displaced from the centre. In the early stages of track expansion OH + e; will then be sup- pressed relatively more than e; + H+ since the faster diffusing H+ will have a greater probability of occupying the zones of higher concentrations of e;.Preliminary calculations with a two species model suggest that such trends can be observed, in addition to a delay in the decay of the species with low initial concentration at the spur or track centre, which is also a desirable feature of any improved model. Further work on modelling will be reported elsewhere. TABLE 3.-DEGRADATION SPECTRUM FOR 3 MeV PROTONS AND G(e&) CALCULATION FOR SCAVENGER CONCENTRATIONS OF APPLEBY AND SCHWARZ." (fis the fraction of energy deposited by ionizing particles in the energy range given) PROTONS mean total local f G(e,) f x G(eJ G(e,) f x G(e&) energy energy/ LET/eV LET/eV normal spur expanded spur range/MeV MeV nm-' nm-l 2.6-3.0 2.2-2.6 1.8-2.2 1.4-1.8 1 .o-1.4 0.6-1 .O 0.2-0.6 0 .2 2.8 12.5 2.4 14.1 2.0 16.1 1.6 19.05 1.2 23.5 0.8 31.5 0.4 49.6 0.1 93.2 6.90 7.79 8.94 10.65 13.22 18.0 29.13 72.70 0.0736 0.0737 0.0740 0.0745 0.0750 0.0762 0.0789 0.0520 I= 1.25 0.0920 1.17 0.0869 1.10 0.0814 1.00 0.0745 0.85 0.0638 0.70 0.0533 0.47 0.0371 0.23 0.0119 f x G(eJ 0.501 1.90 1.80 1.72 1.60 1.44 1.20 0.92 0.54 0.1398 0.1327 0.1273 0.1 192 0.1080 0.09 14 0.0726 0.0281 0.8191 ELECTRONS (produced by protons) range/keV energy/keV LET/eV normal spur expanded spur energy mean local .f G(eJ f x G(eG) G(eJ f x G(eJ nm-I - 4.8-6.4 3.2-4.8 1.6-2.4 2.4-3.2 1.2-1.6 0.8-1.2 0.6-0.8 0.4-0.6 0.3-0.4 0.2-0.3 0.15-0.2 0.1-0.15 0-0.1 5.6 4.0 2.8 2.0 1.4 1 .o 0.7 0.5 0.35 0.25 0.175 0.125 0.05 2.46 3.30 4.34 7.00 11.00 16.6 25.3 36.6 50.4 66.6 81.0 90.2 93.0 0.0005 0.0036 0.005 1 0.01 23 0.01 18 0.0223 0.0192 0.0337 0.0279 0.0468 0.0357 0.0577 0.1459 1.88 0.0009 2.38 1.75 0.0063 2.27 1.55 0.0079 2.15 1.25 0.01 54 1.90 0.98 0.01 16 1.57 0.73 0.01 63 1.36 0.53 0.0104 1 .00 0.40 0.0135 0.80 0.32 0.0089 0.67 0.24 0.01 12 0.56 0.20 0.007 1 0.49 0.18 0.0104 0.45 0.17 0.0248 0.43 Z f x G(e,) 0.1445 electrons Zf x G(eJ 0.6455 electrons and protons 0.0012 0.0082 0.0109 0.0234 0.01 85 0.0303 0.0192 0.0269 0.01 87 0.0262 0.01 75 0.0259 0.0627 0.2896 1.108754 YIELD AND DECAY OF THE HYDRATED ELECTRON I N PROTON TRACKS CONCLUSIONS The decay of the hydrated electron formed by 3 MeV protons in water alone and containing scavengers for OH and H+ has been observed on the nanosecond time- scale, and some implications for the subnanosecond timescale have been derived.The decay is observed to be delayed compared with predictions of the current diffusion model, but the modification of the model (use of enlarged spur) which is successful in bringing the observed and calculated decays together for low LET appears not to be so successful for 3 MeV protons. We thank the S.R.C. for the award of a studentship to G. S. T., and Dr. F. Wilkin- son for helpful discussions on the signal detection equipment. J. W. T. Spinks and R. J, Woods, An Introduction to Radiation Chemistry (Wiley-Interscience, London, 2nd edn, 1976), p. 258. H. A. Schwarz, J. Phys. Chem., 1969,73, 1928, and references therein. A. Kupperman, in Physical Mechanism in Radiation Biology, ed. R. D. Cooper and R. W. Wood, (Technical Information Centre, Office of Information Services, USAEC, 1974). J. K. Thomas and R. V. Bensasson, J. Chem. Phys., 1967,46,4147. G. V. Buxton, Proc. Roy. SOC. A, 1972, 328, 9. R. K. Wolff, M. J. Bronskill, J. E. Aldrich and J. W. Hunt, J. Phys. Chenz., 1973,77, 1350. C. D. Jonah, E. J. Hart and M. S. Matheson, J. Phys. Chem., 1973, 77, 1838. A. T. G. Ferguson, Contemporary Phys., 1964, 5, 269. * W. G. Burns, R. May, G. V. Buxton and G. S. Tough, to be published. lo P. M. Rentzepis, R. P. Jones and J. Jortner, J. Chem. Phys., 1973, 59, 766. l1 A. Appleby and H. A. Schwarz, J. Phys. Chem., 1969,73, 1937. l2 C. D. Jonah, M. S. Matheson, J. R. Miller and E. J. Hart, J. Phys. Chem., 1976,80, 1267. l3 W. H. Hamill, J. Phys. Chem., 1969, 73, 1341. l4 A. Kupperman, in Radiation Research, ed. G. Silini (North-Holland, Amsterdam, 1967), p. 212. P. R. J. Burch, Rcdiation Res., 1957, 6, 289; Brit. J. Radiol., 1957, 30, 524.