摘要:

Pulse Radiolysis Studies of the Reactivity of the Solvated Electron in Ethanol and Methanol* BY IRWIN A. TAUB MYRAN C. SAUER JR. AND LEON M. DORFWAN Argonne National Laboratory Argonne Illinois U.S.A. Received 10th June 1963 By means of the pulse radiolysis technique a short-lived transient species has been observed in irradiated deaerated ethanol and methanol exhibiting an optical absorption throughout the visible and near infra-red. This transient is suggested to be the solvated electron on the basis of the nature of the spectrum the reactivity with hydrogen ion and with various organic electron acceptors, and the formation of mononegative ions of some of these acceptors. The absolute rate constants have been determined for the reactions of the solvated electron with hydrogen ion oxygen and benzyl chloride in ethanol and methanol.The diphenylide ion was found to be short-lived in ethanol. The absolute rate constant for the first-order decay of the diphenylide ion has been determined. The theoretical considerations of electron solvation in polar liquids such as ammonia 1 and water 2 apply to other polar liquids. Thus in ethanol and methanol, for which the values of the static dielectric constant at room temperature are 25 and 33 respectively 3 (large compared to unity) one may expect a priori that the physical process of electron solvation will take place. Measurements of the die-lectric dispersion in these alcohols,4 which indicate that both liquids have a relatively high atomic polarization as does water lead to this conjecture.The natural lifetime of the solvated electron if solvation does occur will then be determined largely by the specific rate of removal of the electron in chemical reactions with the solvent. The rates of these reactions with the solvent may vary widely from one liquid to another as is evident from a comparison of the natural lifetime of the solvated electron in ammonia 5 and in water.6 If this lifetime is sufficiently long the species may be amenable to direct observation. Some evidence for electron solvation in organic glasses,7 consisting of compounds having lower dielectric constants than the alcohols has been obtained in flash photochemical experiments with solutions of the alkali metals. These experiments have been extended to the liquid state? Observation of an optical absorption and formation of mononegative ions of solute molecules in y-irradiated organic glasses,g including ethanol has been presented as evidence for electron solvation.The possible role of electron solvation in the radiation chemistry of liquid ethanol has been discussed,lo and a mechanism involving negative polarons has been sug-gested. Recently experimental evidence 11 indicating that the solvated electron is a precursor of hydrogen in the radiolysis of methanol has been obtained by the use of scavengers which are effective electron acceptors. The present paper is concerned with the application of the pulse radiolysis technique 12- 13 to this question in the case of liquid ethanol and methanol. The primary objective has been the determination by this fast reaction method of the specific reactivity of an observed transient which from its spectral and chemical characteristics is suggested to be the solvated electron.* based on work performed under the auspices of the U.S. Atomic Energy Commission. 20 I . A . TAUB M. C. SAUER JR. AND L . M. DORFMAN 207 EXPERIMENTAL The technical details of the pulse radiolysis method have already been presented.13 Some additional details are given in the later reports 143 6 on ethanol and aqueous ethanol solutions. Only the briefest description will be presented here along with those experi-mental conditions which are unique to the present investigation. PULSE IRRADIATION A 15 MeV electron beam from the linac was used throughout. In the spectrophoto-graphic experiments a 5 psec pulse was used.In the spectrophotometric experiments in which some of the rate curves had a total duration of 2 p e c or less a 0.4psec pulse was used. The pulse currents were in the range 40-150 mA. A 0-4 psec pulse at 100 mA delivers a dose of approximately 1 . 4 ~ 1017 eV/g. The electron beam had an incident diameter of 16 m and an emergent diameter of about 18 mm for a 4 cm long cell. The cells used were cylindrical quartz cells as in previous work. SPECTROPHOTOGRAPHY The four-fold reflection system with collinear analyzing light beam and electron beam 13 was used throughout. The Jarrell-Ash 2.25 m grating spectrograph fl24 was used. Eastman-Kodak spectroscopic plates type 103-F were used over the region 450-670 mp. A single flash of the xenon flash lamp was sufficient to produce satisfactory spectra on these plates.Over the region 550-850 mp Eastman Kodak spectroscopic plates type I-N, were used. These plates were sensitized in aqueous ammonia followed by an alcohol wash and air-dried with a blower. Considerable difficulty was encountered in obtaining sensitized plates of uniform density without blotching. Five flashes of the spectroflash lamp were needed to obtain spectra of satisfactory density with these plates for use on the densit ometer . SPECTROPHOTOMETRY A 1 P28 photomultiplier tube was used to monitor the light from the steady lamp, which was an Osram mercury lamp type HBO 10711. The visible absorption was usually monitored at 5461 A although a few runs were also done at 5770-5790& Since the ab-sorption is extremely broad a 5 mm aperture in front of the photomultiplier was used, giving a band width of 75 A.MATERIALS The ethanol was U.S.I. Absolute Pure Ethyl Alcohol U.S.P.-N.F. Reagent Quality, obtained from US. Industrial Chemicals Co. The methanol was Anhydrous Methyl Alcohol Analytical Reagent obtained from Mallinckrodt Chemical. The methanol was fractionally distilled in the following manner. About 11. methanol to which had been added 1 ml of concentrated sulphuric acid and 4 to 5 g of 2,4-dinitrophenylhydrazine, was placed in the still pot of a Hasteloy B (#1979) Podbielniak still 8 nlmx 12 in. After refluxing for about 1 h the still was run at maximum take-off rate and the first 110 ml were discarded. The next 500-6OOml were collected and used.This purification was found to be necessary since the distillation substantially increased the half-life of the solvated electron in degassed neutral solution. In the purified methanol the half-life of the transient was fully an order of magnitude longer than for the rate curves for reaction with added solutes. The ethanol in most runs was used without further purification on the basis of the observed relatively long half-life of the transient in neutral solution. Some ethanol was purified with a drying agent prepared from 5 g of magnesium powder 60 ml of ethanol and 0.5 g of iodine. To this was added 900 ml of ethanol which was then refluxed for 1 h and then distilled. This procedure reportedly reduces the water content to less than 0.02 %. In the one case in which runs were carried out with this purified ethanol there was no difference in the measured rate constant.Other compounds used in this work were benzyl chloride CP Baker’s Analyze 208 SOLVATED ELECTRON IN ETHANOL AND METHANOL Reagent ; diphenyl from Matheson Co. ; naphthalene Baker’s Analyzed Reagent ; anthra-cene Scintillation Grade from Reilly Tar and Chemical Corp. The degassing technique was the same as described previously,6 which involved sealing off the cells under vacuum except for the experiments with oxygen in solution. For the solutions containing oxygen for which it was necessary to determine the concentration of dissolved oxygen it was desirable to use a cell having no gaseous volume. Conse-quently. cells having capillary leads and ground glass caps adapted from a description of an irradiation syringe,ls and using the degassing method described there were used ex-clusively in the rate studies with oxygen.The solutions containing oxygen were made up by admitting a known pressure of oxygen to the degassed alcohol maintained at about -778°C. The solution was then brought back to room temperature and the cells which had previously been flushed with helium were filled by forcing the liquid from the degassing bulb 15 into the cells under helium pressure. These cells had no gaseous volume being filled entirely with the alcohol solution. Acidified alcohol solutions which were used in determining the rate of reaction of the solvated electron with the hydrogen ion were made up from HC1 and H2SO4 in the follow-ing manner.Anhydrous HCl was bubbled through ethanol for about 15 min and through methanol for about 5 min to make up stock solutions which were roughly 1-2 h4. Solutions where then made up by micro-pipetting from this stock solution to give solutions in the concentration range 10-4-10-5 M. Sulphuric acid stock solutions were made up from a weighed quantity of concentrated acid. There was thus an uncertainty of a few perccnt in the sulphuric acid content. The solutions for kinetic studies were similarly made up by micro-pipetting from the stock solution. In all cases appfopriate corrections were applied to the degassed solutions for volume loss during degassing. The benzyl chloride stock solutions were made up both gravimetrically and volu-metrically. In calculating the concentration of benzyl chloride in the dilute solutions the volume correction for solvent loss in degassing was taken into account.ANALYTICAL The oxygen content of the oxygenated solutions was determined by extraction of the gas from the alcohol in a vacuum system followed by analysis on a gas chromatograph. The oxygen and dissolved helium were removed by means of a Toepler pump and forced into a U-tube between two stopcocks. This gas was then analyzed chromatographically on a column packed with molecular sieve 13X which separated oxygen and nitrogen. The amount of nitrogen was an indication of the extent of air leakage into the cell. This gener-ally amounted to less than 5 x 10-6 M 02. In several runs duplicate cells were filled and the oxygen concentration determined both before and after a run.The acid content of the stock solutions was determined by diluting the stock solution 100-fold in water (the resulting aqueous solution being roughly 2 x 10-2 M) and determining the pH with a Radiometer model 4 pH meter. RESULTS AND DISCUSSION The experimental observations which have been carried out at the time of pre-paration of this paper include the spectrophotographic recording of the transient spectrum in deaerated ethanol and methanol the determination of the absolute rate constants for the reactions of the transient with hydrogen ion oxygen and benzyl chloride in both alcohols and an investigation of the reactions in solutions containing electron acceptors such as naphthalene and diphenyl. In the last system, these studies involve not only the primary transient but also observation of transient species formed from the solute molecules.SPECTRUM Spectrophotographic observations made during and within 2 or 3 psec after a 5 psec electron pulse (current -80 mA) show a transient optical absorption i I . A . TAUB M. C . SAUER JR. AND L . M. DORFMAN 209 deaerated ethanol extending from about 300 mp to our long-wavelength limit of observation at about 860 mp. The spectrum shows a broad peak at about 700 mp with a shoulder at about 520 mp as determined from densitometer tracings. This is shown in fig. 1. The location of these maxima must be regarded as approximate since " H and D " corrections have not been applied to the plates. Moreover the peak is not at all sharp and the most obvious feature of the spectrum is the absence of any pronounced structure.Deaerated methanol also shows a strong absorption in the visible but the location of any maximum has not yet been established. wavelength mp FIG. 1 .-Absorption spectrum of the solvated electron in irradiated ethanol. The half-life of the transient in deaerated neutral ethanol following a 0.4 psec pulse at 110 mA (which corresponds to 1-5 x 1017 eV/g) is approximately 3 psec. The decay curve under these conditions is shown in fig. 2. This curve undoubtedly contains a contribution from the reaction with the counter-ion. The lifetime in basic solution is considerably longer as indicated in a single experiment but a com-plete investigation of the natural lifetime in strongly basic solution such as we have reported for the hydrated electron,6 has yet to be carried out.From the rate curves we estimate the product of the molar extinction coefficient at 5461 A and the G-value &GeSo,. These estimates give : EG,- = 1 to 1.5 x lo4 M-' cm-' mo1./100 eV EG,- = 0.8 to 1.5 x lo4 M-' cm-' mo1./100 eV sol ethanol methanol sol In the absence of iuformation on G,;- we are unable to estimate the oscillator strength. The region of the absorption spectrum and the maximum at about 700mp are very similar to the absorption spectrum of the hydrated electron in water.16 17 The transient absorption is eliminated in acidic alcohol solutions. Moreover the specific reactivities of the transient with hydrogen ion and with oxygen are nearly identical to the reactivities of the hydrated electron 6 18 with these stable species.The transient reacts rapidly with electron acceptors such as benzyl chloride 210 SOLVATED ELECTRON IN ETHANOL AND METHANOL naphthalene and diphenyl. For diphenyl which has been studied in detail the product has been identified as the mononegative ion. These observations indicate that the transient is the solvated electron e&. time FIG. 2.-Rate curve for the disappearance of the solvated electron in deaerated pure ethanol at 5461 A following a 0.4 psec electron pulse. RATE CONSTANTS The absolute rate constants which have been determined are shown in table 1. In all cases the rate curves have been observed at a concentration of the added reactant such that the decay rate of the electron is fully an order of magnitude faster than in the absence of the reactant and is furthermore pseudo-first-order.TABLE ABSOLUTE RATE CONSTANTS FOR REACTIONS OF THE SOLVATED ELECTRON IN ETHANOL AND METHANOL AT 23°C reaction solvent rate constant M-1 sec-1 x 10-10 es; + HZ01 C2HsOH 2-0 f0.4 G I + 0 2 C2H50H 1.9 f0.3 e&+ C6H5CH2Cl C2H5OH 0.51 f0-12 %;l+ 0 2 CH30H 1.9 f0-4 e&+ C~H~CHZCI CH30H 0*50f0*12 es;1+ HA CH3OH 3.9 f0.9 The reaction of the solvated electron in ethanol with oxygen, &+ 0 = o,, has been determined over a concentration range of oxygen from 4 x 10-5 to 11 x 10-5 M. The initial electron concentration was varied 2-fold. Typical decay curves are shown in fig. 3 and fig. 4 for the reaction in ethanol. A typical first-order rate law test of such rate curves is shown in fig.5 in which is presented a plot of the logarithm of the optical density as a function of time. This is a representation of the integrated form of the pseudo first-order differential rate expression where Dt = loglo (Io/It,-) and k' is the pseudo first-order rate constant. The results give rate constants of (1.9 & 0.3) x 1010 M-1 sec-1 and (1-9 0.4) x 1010 M-1 sec-1 at log, D' = k'ti2-303 FIG. 4.-Rate curve for the reaction of the solvated electron with oxygen in ethanol. The sweep-rate is 1 psec/ large division. The curve was obtained at 5461 A. The pulse current is approximately 80 mA. [To facepage 210 I . A . TAUB M. C. SAUER JR. AND L . M. DORFMAN 21 1 23°C for the reactions in ethanol and methanol respectively.The slightly larger error limit with methanol stems largely from the fact that only three separate deter-minations were made compared with ten with the ethanol. These rate constants 1.0 4" 0.87 2 G 0.7 7 0.66 'I p sec. time FIG. 3.-Rate curve for the reaction of the solvated electron with oxygen in ethanol following a 04 psec pulse at 60 mA. The initial oxygen concentration was 46 x 10-5 M. The rate curve was obtained at 5461 A. 8 L 0-07 001 I I I 0 I 2 time psec FIG. 5.-Test of first-order rate law for tke reaction of the solvated electron with oxygen in ethanol. This is a logarithmic plot of the optical density as a function of time. are the same as that of the hydrated electron with oxygen for which a value of 1.9 x 1010 M-1 sec-1 has been reported.18 The rate of reaction with hydrogen ion was determined over a concentration range of the hydrogen ion of 0.24 x 10-4 to 1-2x 10-4 M in ethanol and 0.4 x 10-212 SOLVATED ELECTRON I N ETHANOL AND METHANOL to 1 x 18-4 M in methanol using HCl.A somewhat smaller range was covered using HzS04. The hydrogen ion concentration in the HCl solutions was calculated on the assumption of complete dissociation 19 of this acid in the alcohols at the low concentrations used. In ethanol the absolute rate constant for reaction of the electron with hydrogen ion is found to be (2.0+0.4) x 1010 M-1 sec-1. The sulph-uric acid solutions on the assumption of only the first dissociation to H+ and HSOz, give (2.1 f0-3) x 1010 M-1 see-1 in agreement with this value. The rate constant for the reaction in methanol was found to be (3-9f0.9) x 1010 M-1 sec-1.We cannot explain the apparent higher rate constant for the reaction with the hydrogen ion in methanol. A small correction generally 5 to 10 % for the decay in the pure solvent thus including the contribution of the reaction with counter-ion was applied to all the rate curves. The electron reacts rapidly with benzyl chloride. The rate constants in ethanol and methanol were found to be (5.1 & 1.2) x 109 M-1 sec-1 and (50+ 1-2) x 109 M-* sec-1 respectively at 23°C. In these solutions we find the benzyl radical as an intermediate as established by its transient absorption spectrum in agreement with previous work.20 The reactions of this transient are under investigation in our laboratory.In solutions of triphenyl chloromethane we similarly find the tri-phenyl methyl radical. AROMATIC ANIONS The transient absorption spectra and kinetics in ethanol solutions of aromatic molecules such as naphthalene diphenyl and anthracene are being investigated. The initial objective was to obtain supporting evidence for the identity of the solvated electron through the observation of the mononegative ions of the hydrocarbons which are presumably formed (as for the diphenylide ion) in the reaction : The ultra-violet and visible absorption spectra of these mononegative ions are known.21922 The solvated electron reacts rapidly with these electron acceptors, and with the diphenyl we have identified the diphenylide ion from the peak at 400 mp and the wide band in the visible showing two peaks at 610 and 635 mp.This cor-responds with the spectrum published,al in which peaks are shown at 400 617 and 637 mp. As expected since the electron is the precursor these transients are eliminated in acid solution. In the experiments with diphenyl and naphthalene the simultaneous decay of the absorption at 5461 A and the formation of a second transient at 3130 A has been recorded using a double photomultiplier arrangement. Such simultaneous disappearance and appearance curves are shown in fig. 6. The decay curve at 5461 A contains a contribution from both the solvated electron and the diphenylide ion the former presumably predominating at low concentration of diphenyl ca. 10-5 M with an increasing contribution by the diphenylide ion as the concentration is increased.The formation curve at 3130 A consists primarily of an absorption of a transient formed from the diphenylide ion probably by proton capture from the solvent to form a hydrogen adduct. Such a process has been suggested to occur in methanol solution.11 The near identity in the rates of the formation and disap-pearance curves in fig. 5 clearly establishes the precursor relationship of the transients. The observation made both spectrographically and spectrophotometrically at 4047A that the diphenylide ion in ethanol is short-lived is in contrast with the long lifetime of the aromatic anions in solvents such as tetrahydrofuran. This difference in Lifetime is attributed to the protonic character of the ethanol. Th 1. A .TAUB M. C. SAUER JR. AND L . M. DORFMAN 213 decay of the diphenylide ion in ethanol is first-order. The rate constant for the reaction ClzHlo + C2H50H determined at diphenyl concentrations greater than 10-3 M was found to be (4-1 k0.7) x l o 5 sec-1 at 23°C expressed as a first-order constant. Work on the protonation reactions of other aromatic anions is continuing. time FIG. 6.Cimultaneous disappearance and formation curves following a 0.4 psec pulse at 5461 A (upper curve) and 3130 8 (lower curve) in ethanol solution of diphenyl. The upper curve contains a contribution from both the electron and the diphenylide ion. The lower curve represents a transient formed from the diphenylide ion. The base line for the upper curve is two divisions from the bottom of the grid.We are indebted to a number of our colleagues for technical assistance. In particular we thank Mr. Douglas Harter in this regard. The linac was operated by Mr. B. E. Clifft. We are grateful to Dr. R. Platzman and Dr. W. H. Hamill: with whom we have had a number of illuminating and stimulating discussions. 1 Davidov Zhurn. Expt. Theor. Fizik 1948 18 no. 10 913. 2 Platzman Basic Mechanisms in Radiobiology (Nat. Acad. Sci. Publ. no. 305) 1953 22. 3 Buckley and Maryott Nat. Bur. Stand. circ. 589 1958. 4 Lane and Saxton Proc. Roy. SOC. A 1952,213,400. 5 Symons Quart. Reu. 1959 13 99. 6 Dorfman and Taub J. Amer. Chem. Soc. 1963,85,2370. 7 Linschitz Berry and Schweitzer J. Amer. Chem. SOC. 1954,76 5833. 8 Eloranta and Linschitz J. Chem. Physics 1963 38 2214. 9 Ronayne Guarino and Hamill J. Amer. Chem. SOC. 1962 84,4230. 10 Hayon and Weiss J. Chem. SOC. 1961 3962. 11 Baxendale and Mellows J. Amer. Chem. SOC. 1961,83,4720. 12 Matheson and Dorfman J. Chem. Physics 1960,32 1870. 13 Dorfman Taub and Buhler J. Chem. Physics 1962,36 3051. 14 Taub and Dorfman J. Amer. Chem. Soc. 1962,84,4053. 15 Senvar and Hart Proc. 2nd U.N. Int. Con$ Peaceful Uses of Atomic Energy 1958 29 19. 16 Boag and Hart Nature 1963,197,45. 17 Hart and Boag J. Amer. Chem. SOC. 1962 84 4090. 18 Gordon Hart Matheson Rabani and Thomas J. Amer. Chem. SOC. 1963,435 1375. 19 MacInnes The Principles of Electrochemistry (Reinhold Publishing Corp. New York 1939), 20 McCarthy and MacLachlan Trans. Faraday SOC. 1960 56 1187. 21 Balk Hoijtink and Schreurs Rec. truu. chim. 1957 76 813. 22 Paul Lipkin and Weissman J. Amer. Chem. SOC. 1956 78 116. p. 366

ISSN:0366-9033    

DOI:10.1039/DF9633600206

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

Nature and Reactivity of the Primary Reducing Species in the Radiolysis of Aqueous Solutions BY G. SCHOLES, M. SIMIC AND J. J. WEISS Laboratory of Radiation Chemistry, School of Chemistry, The University, Newcastle-upon-Tyne Received 24th June, 1963 A study has been made of aqueous solutions of various organic solutes irradiated with C060 prays in the presence of N20 and C02, which can act as scavengers for radiation-produced reducing species. The yields of the radiation products have been determined as a function of solute con- centration and of pH. The results clearly indicate the high specific reactivity of N20 and of C02 towards negative polarons (H20)-, and the conversion of these polarons into oxidizing species by reaction with N2O. The results further support the independent formation of two reducing species, i.e., of the negative polarons and of a species reacting as hydrogen atoms.The reactivities of both these species with different organic and inorganic solutes have been investigated. Chemical evidence 1-6 has shown that radiation-produced electrons constitute the major fraction of the reducing species in aqueous media, a conclusion conkmed by physical measurements.7 These species should be regarded as negative polarons,8 represented here by (H20)-. Man and Scholes 3 showed that, in addition to the negative polaron, another species is formed independently, in a yield G-0-6, pre- sumed to be hydrogen atoms and formed from excited water molecules according to H20*-,H+O€€. The existence of this latter species was confirmed by the work of Rabani and Stein? These hydrogen atoms are designated here by Ha. EXPERIMENTAL Irradiations were carried out with C060 y-rays, the dose-rate ( 2 .4 ~ 1017 eV ml-1 min-1) being determined by the Fricke dosimeter, taking G(FelI1) = 15.5. All solutions were prepared with triply distilled water. Medical-grade N20 was purified by several distillations on a vacuum line, prior to introduction into the radiation vessel containing the degassed solution. C-14-labelled sodium bicarbonate solutions were prepared by the addition of C-14-labelled sodium carbonate, of known activity, to solutions containing an excess of sodium bicarbonate. Under these experimental conditions, the activity was essentially present as C14-bicarbonate. Gas-analyses were carried out mass spectrometrically after trapping out all condensable gases by liquid nitrogen.The nitrogen and hydrogen yields are estimated to be accurate to f 3 %. ~ The yield of labelled oxalate was determined as follows. A known excess of oxalic acid was- added and total oxalate was then precipitated as the calcium salt in the presence of acetic acid+acetate buffer. The precipitate was filtered, dried and weighed, and the activity measured by standard counting techniques. RESULTS AND DISCUSSION EXPERIMENTS WITH N20 Dainton and Peterson 4 have investigated the y-radiolysis of aqueous N20 solutions and shown that this solute has a relatively high reactivity towards the 214G . SCHOLES, M. SIMIC AND J . J . WEISS 21 5 negative polaron, leading to the formation of nitrogen, viz., N2O+(H20)-+N;!+OH-+OH. (1) Hydrogen atoms, on the other hand, react only very slowly with N20.This system therefore is a useful one for confirmation of the presence of these two reducing species. Solutions of N20 (1-6 x 10-2 M) containing various concentrations of aliphatic alcohols were irradiated with C060 prays, and the yields of hydrogen and nitrogen were determined from the linear yield against dose plots. Fig. 1 shows the results obtained with isopropanol. The yields of hydrogen vary from G = 0-94 to G = 1.22 as the isopropanol concentration is varied from 10-3 to 1 M and are clearly in excess of the molecular hydrogen yield (G;,). If it is as- sumed that all of the negative polarons react with N20 according to reaction (1) and only Ha reacts with isopropanol to give molecular hydrogen, the measured hydrogen yield corresponds to the sum: G(H2) = G;";,+G(Ha).The value of Gg2 in 1.6 x 10-2 M N20 solutions has been determined by irradiation in the presence of cupric ion (copper sulphate). Under these conditions, a yield of hydrogen of G(H2) = 0.35 was observed, which was independent of cupric ion concentration over the range 2-5 x 10-4-4.75 x 10-3 M. Thus, on the above assumption, the yield of Ha varies from G = 0.59 to 0-87 under the conditions of the experiments given in fig. 1. Part of this increase at high isopropanol concentration may only be apparent, since the alcohol may also scavenge hydrogen atoms which may otherwise contribute to the molecular yield. Under the above experimental conditions, the presence of N20 leads to a reduc- tion in GE, from 0.45 to 0.35.Since its reaction with hydrogen atoms is generally considered to be slow, this observation is also consistent with the view that the negative polarons contribute to the molecular yield. It has, in fact, been previously suggested 1 that molecular yield hydrogen is formed according to (H20)-+ (H20)-+H2 +20H- (2) and this has been recently confirmed by Dorfman and Taub.10 The yields of nitrogen recorded in fig. 1 are a composite function of the scavenging by N20 of negative polarons present in the bulk of the solution as well as some of those leading to molecular hydrogen, and possibly also some normally undergoing primary recombination, according to : (H20)- + (H20)++2H20 (3) (H20)- + OH+ OH- + H20.(4) The contribution to the observed yield of nitrogen by scavenging of some of the molecular yield precursors is equivalent to a G(N2) = 0.20. With regard to the recombination reactions (3) and (4), G(N2) in N20 + isopropanol solution appears to increase as the alcohol concentration is increased. The marked dependence of the yields of nitrogen on N20 concentration in pure N20 solutions,4 also probably indicates scavenging of some negative polarons normally recombining with the oxidizing species. The N20 + isopropanol system, therefore, clearly demonstrates the existence and reactions of the two reducing species. Addition of another acceptor for negative polarons to this system should lead to a simple competition between the acceptor and N20 for (H20)-, and hence to a decrease in the yield of nitrogen. Likewise, addition of another acceptor for Ha should lead to a competition reaction with isopropanol, and, if the gdded acceptor reacts with HG so as not to give molecular216 RADIOLYSIS OF AQUEOUS SOLUTIONS hydrogen, there should be a decrease in the hydrogen yield.It has been found that all these conditions can be realized using cupric ions as the additional solute. 3- 2- E z! Q) x *.+ E 2 2 # h .d I 0 - 10 -3 lG2 10 -I I [isopropanol] (mole/l.) FIG. 1.-Effect of isopropanol concentration on the yields of nitrogen and hydrogen in the y- radiolysis of aqueous solutions of N20 (1.6 x 10-2 M) at pH-6. 17, nitrogen ; 0, hydrogen. I I I 1 - I 2 3 4 5 OL C [Cu2+] (rnolej.) x lo3 FIG. 2.Effect of cupric ions on the yields of nitrogen and hydrogen in the 7-radiolysis of aqueous solutions of isopropanol(10-1 M) and N20 ( 1 .6 ~ 10-2 M); pH-6; 0, nitrogen; 0, hydrogen. Fig. 2 shows the yields of nitrogen and of hydrogen on irradiation of N20 (1.6 x 10-2 M) +isopropanol (10-1 M) solutions containing varying amounts of cupric ions.G . SCHOLES, M. SIMIC AND J . J . WEISS 21 7 The simple competition kinetics between reaction (1) and the reaction, C U ~ + + (HzO)--+Cu+ + H20. ( 5 ) can be expressed by the relationship : The experimental results satisfy these conditions as is evident from fig. 3, and lead to a value of k~/,k1 = 4-7. 0 0 L-.; 2 I 3 4 I J 5 [Cu2+] (mole/l.) x 103 FIG. 3.-Competition plot for the nitrogen yields in the cupric ionfN20 + isopropanol system. Representing the reactions of Ha as those of hydrogen atoms, viz., H + (CH&CHOH-,H2 + (CH3)zCOH (6) H+CU~+-,CU++H+, (7) one obtains the equation : 1 1 k,[Cu2+] =- +- 1 G(H2)- G& G(H") G(H") k,[isopr.]' where G(H2) is the observed hydrogen yield.Taking GZ, = 0.35, the experimental hydrogen values obey this relationship (fig. 4), the plot giving an intercept corres- ponding to G(Ha) = 0.66 and a slope corresponding to a value of k,/k6 = 12.4. The effects of pH on the yield of the gaseous products in the N20 +isopropanol system have also been investigated.11 Increase of hydrogen ion concentration leads to a decrease in G(N2) and a corresponding increase in G(H2). This is the result of the reaction (H20)-+H30f+2H20+H (8) competing with reaction (1). From these data, a value of k&l = 1.7 has been obtained, in agreement with other workers.4, 12 In alkaline solution, the N20 + isopropanol system exhibits a radiation-induced chain reaction leading to high yields of acetone and of nitrogen. For example,21 8 RADIOLYSIS OF AQUEOUS SOLUTIONS at pH 13, a value of G(N2)-50 was obtained.It appears that the isopropanol radical-ion can interact with nitrous oxide giving an oxidizing radical, viz., 0- 0- 1 I CH3C CH3 + OH-CH34iCH3 + H20 (9) H 0- I CH3C CH3 + N20 + H20+CH3COCH3 + N2 + OH+ OH- with a chain-breaking process, e.g., the disproportionation of two alcohol radicals giving acetone and isopropanol. Solutions of N20 containing one of several other aliphatic alcohols have also been investigated. As with isopropanol + N20 solutions, where the negative polarons react essentially with N20, yields of nitrogen in the range G(N2) = 3.0-3.3 (10) I 0 I 2 3 4 5 [Cu2+](moIe/l.) x 103 FIG.4.-Competition plot for the hydrogen yields in the cupric ion+NZO+isopropanol system. were observed. The yields of hydrogen, however, vary, depending upon the reactivity of the organic substance towards He. This is evident from the data of table 1 where hydrogen yields from aqueous solutions of methanol, isopropanol and t-butanol at various concentrations are given. TABLE 1.-YIELDS OF HYDROGEN IN THE 7-RADIOLYSIS OF AQUEOUS SOLUTIONS OF N20 (1.6 x 10-2 M) AND ALIPHATIC ALCOHOLS : pH- 6 methanol t-butanol isopropanol yield aIcohol yield alcohol yield (moles/l. ) G(Hd (moles/I.) G(Hd (moles/I.) G(Hd 10-3 0 . 6 1 10-3 0.42 10-3 0.94 alcohol 10-2 0.94 10-2 0.54 10-2 1 -02 10-1 1 -05 10-1 0.88 10-1 1.10 1 1 -22 The suggestion 4 that the product of the reaction of negative polarons with N20 is an oxidizing species has been confirmed by an examination of the product yieldsG .SCHOLES, M. SIMIC AND J . J . WEISS 219 from methanol + N20 solutions. The following are the yields observed on irradi- ation of 10-1 M methanol-1.6 x 10-2 M N20 solutions : G(N2) = 3.10 ; G(H2) = 1.05 ; G(H202) = 0.45 ; G (formaldehyde) = 0.13 ; G (glycol) = 3.3+_0.2 (deter- mined with periodate 11). For the mechanism : OH + CH30H+*CH2OH Ha+ CH30H+*CH2OH + H2 2*CH2OH-+HCHQ + CH30H 2CH2OH- (CHZOH)~ rhere the oxidizing radicals reacting with methanol come both from the and from reaction (l), it follows that 2G(glycol) + 2G(formaldehyde) = G(0H) + G(Ha) = 7.2 & 0.4.Assuming that for every (H20)- reacting with N20 a corresponding number of oxidizing species from the water reacts with methanol and also that an additional number of OH radicals equivalent to Ha reacts with the alcohol, it can be shown that G(OH) = 6.7. Thus, G(OH)+G(Ha) = 7.4 agrees well with that from the organic products yield found above. Solutions of N20 at neutral pH are useful for the measurement of the reactivities of various compounds towards the negative polarons. Substances which compete efficiently with N20 for (H2O)- lower G(N2) and the relative rates compared to N20 can thus be obtained, as for example, with acetone and chloroacetate. In this way, it has been found that the anions of carboxylic' acids, e.g., formic acid and acetic acid, cannot compete with 1.6 x 10-2 M N20 and therefore must have relatively low reativities towards the negative polarons.EXPERIMENTS WITH c02 The influence of COz in the radiolysis of aqueous solutions has been ascribed to the reaction : 3, 139 14 (H,O)-+CO,+COT +H,O. Production of the COT radical-ion in the irradiation of aquo-organic systems can lead to the formation of carboxylic acids.14 In general, the carboxylation process depends upon the formation of a free radical from the organic solute (RH), e.g., by the process, RH+OH+R*+H20, (16) R*+CO,+RCQ, (17) and association of the free radical with COT according to A variety of organic solutes have been carboxylated in this way. Many of the ex- periments were carried out with carbon dioxide or bicarbonate labelled with carbon- 14, this technique being used to facilitate identification and analysis of the carboxylated products. Irradiation of aqueous solutions of methanol (10-3 M) with C060 y-rays gave glycollic acid.Under similar conditions, irradiation of aqueous ethanol gave some lactic acid, acetic acid solutions some malonic acid, solutions of methane some acetic acid and from formic acid solutions containing C14-labelled CO2, labelled oxalic acid was obtained. In general, the extent of carboxylation is markedly220 RADIOLYSIS OF AQUEOUS SOLUTIONS pH-dependent, falling off rapidly in acid solution. This is not unexpected since reaction (8) competes with reaction (15). In preliminary studies 15 of the extent of carboxylation in methanol + C02 solutions the maximum yield was G(glycol1ic acid) = 2.3, obtained on irradiation of 10-1 M methanol solutions saturated with C02 and adjusted to pH4.5-5.5.This seemed indicative of a somewhat lower negative polaron yield than in alcohol solutions irradiated in the presence of N20. However, a re-investigation of this system has shown that the lower glycollic acid yield results from a loss of some of the COT radical-ions due to the process, COY +H202+C02+OH+OH-. (18) There is, in fact, a low stationary concentration of hydrogen peroxide in solutions irradiated in the neutral pH range. Correction for the extent of reaction (18) leads to a minimum value of G(C0;) of 2.7-2-8. In addition, as in the &O+methanol system, Ha is also observed in the presence of C02, reacting with methanol to form hydrogen.1 - 0 % to4 lo3 lo2 lo-’ t sodium bicarbonate (mole/l.) FIG. 5.-Irradiation with C060 y-rays of deaerated aqueous solutions of sodium formate in the presence of sodium bicarbonate. Dependence of the yield of C-14-labelled carboxyl group in oxalic acid on the concentration of C-14-labelled bicarbonate; 0, 10-3 M formate; 0, 10-1 M formate. With formate solutions, relatively high carboxylation yields have been obtained. Fig. 5 shows the results using HCOONafNaHC1403 solutions. The data are given in terms of C1402, incorporated into the radiation-produced oxalic acid. The ex- periments show that it is the COa (in equilibrium with the bicarbonate ion) rather than the bicarbonate ion itself which reacts with the negative polarons. Such a view is also supported by similar results obtained with methanol+HCO; solu- tions.16 In the formate system, oxalic acid is presumably produced by the process 17 2c0, -+(coo-)2.(19)G . SCHOLES, M. SIMIC AND J . J.. WEISS 221 Isotopically-labelled COT radical-ions result from reaction (1 5 ) whereas unlabelled ones are formed mainly according to OH + HCOO--+H20 + COO- (20) Hence both -0OC14 COO- and -OOC14 C14OO- wifl be formed. In 10-1 M formate solutions, values up to G(C14O;) 21 4.5 have been observed (fig. 5). Since formate has a relatively high reactivity towards the oxidizing species formed in the radiolysis of water,l8 these high yields may be due to an increased availability of negative polarons as a result of some competition by formate with the recombin- ation reactions (reactions (3) and (4)).In all the systems examined, no carboxyl- ation occurs in strongly alkaline solutions. This is understandable if the negative polarons can only react efficiently with C02. When irradiations are carried out in the presence of organic acceptors which can be dehydrogenated, the competition of reactions (8) and (15) manifests itself in a pH-dependent hydrogen yield.19 Allan et aZ.,19 from studies of the yields of hydrogen from irradiated ethanol+ COz solutions, reported that k8/k15 1: 3. There- fore, kl/kls N 1.8. Dr. A. Appleby of this department has recently measured the yields of nitrogen on irradiation of solutions containing various amounts of N20 and COa, obtaining directly a value of kl/k15=2-2, which is in good agreement with the above ratio.CONCLUSION The systems described above can be readily interpreted on the basis of the presence of two reducing species. The fact that Ha reacts with Cu2+ ions, as well as with ferricyanide,g means that the appearance of this species is not specifically associated with the presence of organic compounds as such. The identification of the dehydrogenating species as a hydrogen atom appears to be supported by kinetic evidence 9 as well as by its reactions with OH- ions.20. 21 These hydrogen atoms could be formed either, as originally suggested,3 from excited water molecules, H20* +H + OH, or from the reaction, (H20)- + H30++H+ 2H20, occurring in the spurs.22 The existence of Ha in systems containing oxygen has been questioned; Czapski and Allen23 concluded that in the y-radiolysis of solutions of oxygen and hydrogen peroxide, the reducing radicals behave as if they were all (H20)-.On the other hand, Hummel and Allen24 in an investigation of the yields of hydrogen on y- irradiation of aqueous ethanol solutions containing oxygen, concluded that there was a competition between ethanol and oxygen for a reducing species, but were unable to decide whether this was the negative polaron or a hydrogen atom. This dehydrogenation of an organic solute in the presence of oxygen has recently been investigated using formate solutions,25 the latter solute being chosen since its reaction with the negative polaron does not lead to the formation of hydrogen? The results indicated that it was Ha which was competing between formate and oxygen.Some questions still remain concerning the yields of the reducing species in solutions containing oxygen and investigations along these lines should lead to information which may help to elucidate the problem of the exact nature and distribution of these hydrogen atoms. We acknowledge with thanks the financial support of this investigation by the U.S. Department of the Army, through Contract No. DA-9 1-591-EUC-2750.222 RADIOLYSIS OF AQUEOUS SOLUTIONS 1 Hayon and Weiss, Proc. U.N. Int. Conf. Peaceful Uses Atomic Energy (Geneva, 1958), 29, 80. 2 Baxendale and Hughes, Z. Physik. Chem., 1958, 14, 306, 323. 3 Allan and Scholes, Nature, 1960, 187, 218. 4 Dainton and Peterson, Nature, 1960,186,878 ; Proc. Roy. SOC. A, 1962,267,443. 5 Hayon and Allen, J. Physic. Chem., 1961, 65, 2181. Czapski and Schwarz, J. Physic. Chenz., 1962, 66, 471. 7 Hart and Boag, J. Amer. Chem. SOC., 1962, 84, 4090. Keene, Nature, 1963, 197, 47. 8 Weiss, Nature, 1960, 186, 751. 9 Rabani, J. Amer. Chem. SOC., 1962, 84, 868. Rabani and Stein, J. Chem. Physics, 1962, 37, 1865. 10 Dorfman and Taub, J. Amer. Chem. SOC., 1963, 85, 2370. 11 Munday, unpublished resdts. 12 Jortner, Ottolenghi and Stein, J. Physic. Chem., 1962, 66, 2037. 13 Getoff, Scholes and Weiss, Tetrahedron Letters, 1960, 18, 17. 14 Simic, Scholes and Weiss, Nature, 1960, 188, 1019. 15 Appleby, Holian, Scholes, Simic and Weiss, Proc. 2nd Int. Congr. Radiation Res. (1962). 16 Appleby, Ph.D. Xhesis (University of Durham, 1963). 17 Czapski, Rabani and Stein, Trans. Paraday SOC., 1962,58,2160. 18 Rabani and Stein, Trans. Faraday Soc., 1962,58,2150. 19 Allan, 6etoff, Lehmann, Nixon, Scholes and Simic, J. Inorg. Nucl. Chem., 1961, 19, 204. 20 Allan, Robinson and Scholes, Proc. Chem. SOC., 1962, 381. 21 Rabani, personal communication. 22Lifshit2, Can. J. Chem., 1962, 40, 1903. 23 Czapski and Allen, J. Physic. Chem., 1962,66,471. 24 Hummel and Allen, Radiation Res., 1962, 17, 302. 25 Scholes and Simic, Nature, 1963, 199, 276. 26 Smithies and Hart, J. Amer. Chem. SOC., 1960, 82, 4775.

ISSN:0366-9033    

DOI:10.1039/DF9633600214

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

Scavenger Studies in the Radiolysis of Aqueous Ferricyanide Solutions at High pH BY G. HUGHES AND C. WILLIS Dept. of Inorganic, Physical and Industrial Chemistry, The University, Liverpool Received 23rd May, 1963 A study has been made of the radiation-induced reduction of alkaline, aqueous solutions of potassium ferricyanide. From the effect of methanol on the reduction yield, G(0H) and the sum, G(H)+2G(Hz02), have been obtained. Two separate processes for the production of mole- cular hydrogen peroxide are indicated from results obtained with added potassium ferrocyanide. The reactivity of the ‘‘ OH ” radical is consistent with its being present as 0-. This may be pro- duced either by ionization of OH or from some other precursor. Although much of the work done on the radiation chemistry of aqueous systems has hitherto been confined to studies of acid or neutral solutions, the study of alkaline solutions is currently receiving increasing attention.The radiation chemistry of aqueous acid solutions of ferrous and ferric ions has been investigated exhaustively 1 and the reactions occurring in these systems have been elucidated. It seemed desirable therefore, in the light of this knowledge, to investigate an iron system which could be studied over the whole range of pH. The ferro-ferri-cyanide system appeared to be the most convenient. In an earlier study 2 it had been shown that whereas in acid solution ferrocyanide is oxidized, in alkaline solution ferricyanide is reduced. A suggested reaction scheme to account for the results in acid solution has since been proposed.3 Some pre- liminary results obtained from a study of alkaline aqueous ferricyanide solutions 4 have been shown to be substantially in agreement with results obtained from the ferricyanide+NZO system 5 but to differ markedly from those obtained from the PtD+PP and TeIV+Tem systems.6 We here report a more detailed investigation of the alkaline ferricyanide system.EXPERIMENTAL All irradiations were carried out using y-radiation from a 100 curie 137Cs source. The experimental procedure is similar to that outlined previously.3 A.R. methanol was further purified by distillation from H2SO4 and 2,4-dinitro-phenylhydrazine. Ferricyanide was determined from its optical density using a Unicam S.P. 500 spectrophotometer at 420 mp where E = 1000. Formaldehyde was determined by the chromotropic acid method.7 Control experiments established that the reduction of ferricyanide by hydrogen peroxide under all conditions was rapid and stoichiometric according to the equation The thermal oxidation of methanol and formaldehyde by ferricyanide was shown to be negligible at the concentrations and irradiation times used.223224 RADIOLYSIS OF FERRICYANIDE SOLUTIONS RESULTS AND DISCUSSION From the known chemistry of ferro-ferri-cyanide solutions and the products believed to be produced on the radiolysis of water, the possible reactions occurring in the radiolysis of oxygenated alkaline ferricyanide solution are : H,O-+H, OH, H,, H202 (1) H+ Oz+H02 H + Fe(CN); - 3 H + + Fe(CN)z- Reactions of HO2 and H202 with ferrocyanide in 0.1 M NaOH may be excluded.H, OH and HO2 are used only as formal representations of the species occurring. It is believed that in neutral solutions and in solutions of higher pH, most, if not all, of the so-called hydrogen atoms are present as solvated electrons.8 Eqn. (2)-(4) might then be more properly written as eaq + 02-+ 0, e,, +Fe(CN):--+Fe(CN);- 0; + Fe(CN);- + 0, + Fe(CN)z-. The pK for the dissociation of H02 H 0 2 s H + + 0 ; is 9 - ~ 3 so that, in 0.1 M NaOH, any HO2 produced as such would react as 0;. In this discussion the term OH radical will be used, unless indicated otherwise, to cover both the radical OH or any equivalent species. The overall reaction is, however, independent of the precise nature of H or HO2 and irrespective of whether I3 atoms react via (2) or (3), and HO2 via (4) or (9, the net reduction yield is given by G(-Fe(CN);-) = G(H)+2G(H20,)- G(0H).(9) The disappearance of ferricyanide is proportional to dose and independent of ferricyanide concentration as shown in fig. 1. G(-Fe(CN)i-) is 1.89&0.05 both in oxygenated and deoxygenated solution. This is to be expected if the only reactions of the so-called hydrogen atoms are (2) and (3) (or (2a) and (3a)), since both equally lead to reduction of ferricyanide. A similar value for the reduction yield of ferri- cyanide solutions appeared during the course of this work.10 Addition of chloride ion has no effect on the reduction yield. In 1 M NaOH, G(-Fe(CN)z-) is increased to 2*12&0.10. This would indicate that there is a slight effect of alkali concentra- tion leading to (a) an increase in G(H) and G(H202) and/or (b) a decrease in G(0H).On addition of methanol, the following additional reactions are possible : H+CH30H+Hz + CH,OH (10) OH+ CH30H-,H20+CH20H (1 1) CHZOH + 0, -+ CHZO + H02 (12) (13) CH, OH + Fe(CN);- --* CH,O + H+ + Fe( CN): - .G. HUGHES AND C. WILLIS 225 If hydrogen atoms are present mainly as solvated electrons, reaction (10) is un- likely to occur appreciably. Its occurrence, however, would not affect the reduction yield since each CHzOH radical leads ultimately, either via reaction (12) or (13) to the reduction of a ferricyanide ion. If all OH radicals produced react via (1 1) then G(-Fe(CN):-) = G(H)+ G(OH)+2G(H2O2). dose x 10-19, eV/d FIG. 1 .-Irradiation of ferricyanide solutions.Solutions are 10-3 M K3Fe(CN)6 in 0.1 M NaOH except where otherwise indicated : 0, oxygen- ated solution ; x , deoxygenated solution ; El, 10-2 M C1-, oxygenated solution ; +, 10-1 M CI-, oxygenated solution ; 8, 5 x 10-4 M K3Fe(CN)6, oxygenated solution. dose x 10-18, eV/ml FIG. 2.-Irradiation of ferricyanide and methanol solutions. Oxygen saturated solutions containing 10-3 M K#e(CN)6 ; 0, 10-1 M CH30H, 10-1 M NaOH ; El, 1 M CH3OH, 10-1 M NaOH; +, 10-1 M CH3OH, 1 M NaBH; X, 1 M CH3QH, 1 M NaOH. As may be seen from fig. 2, the disappearance of ferricyanide is again proportional to dose. G(-Fe(CN)z-) in 0.1 M NaOH is 7-82+_0.10 and is independent of €I226 RADIOLYSIS OF FERRICYANIDE SOLUTIONS methanol in the range 0.1 -1 -0 M methanol.Within experimental error, G(--Fe(CN)i-) is the same in 1.0 M NaOH. However, it is doubtful whether the small change in radical yields indicated by the experiments in alcohol-free solutions would be detected under these conditions. From experiments on the effect of alcohols on the radiation induced oxidation of ferrous ion in acid solution, chloride ions exhibit a protective effect.11 The enhanced oxidation yield due to alcohols has been attributed to the reaction (1 5 ) OH+ RCH,OH+H,O + RCHQH, and subsequent formation of peroxides from RCHOH. Chloride ions protect via the competing reaction The chlorine atom thus produced then reacts with ferrous ion. At equimolar con- centrations, chloride ion is able to protect significantly against ethanol indicating OH+Cl---+OH- +a.(16) dosex 10-18, eV/n:l FIG. 3.-Efiect of chloride ion on reduction yield. Oxygen-saturated solutions containing 10-3 M K3Fe(CN)6 and 0-1 M NaOH : El, 10-1 M CH30H ; 0, 10-1 M CH3OH, 10-1 M C1- ; x , 1 M CH30H, 10-1 M CI-. that at these concentrations, chloride ion competes with ethanol for OH. It would be expected that chloride ion would compete more effectively with methanol. If reaction (16) were to occur in the ferricyanide system and be followed by the reaction, then the reduction yield should decrease towards the value observed in the absence of alcohol. The results of fig. 3 indicate that chloride ions exhibit no protective effect, suggesting that the species of OH present in alkaline solution reacts much less readily with chloride ion than its counterpart in acid solution.Using the results of fig. 1 and 2, from eqn. (9) and (14), it may be deduced that in 0.1 M NaOH Cl+ Fe(CN):-+Cl- + Fe(CN);-, (1 7) G(0H) = 2.97; G(H)+2G(H202) = 4.85. These values are substantially the same as those observed in acid solution12 and are in agreement with the more detailed yields deduced from a study of the alkaline ferricyanide + NzO system. G(OW agrees reasonably well with the value observed from the Ptn + PtIV system 6 but G(H) +2G(H202) as deduced from thisG . HUGHES AND C . WILLIS 227 latter system is only 2-94. In solutions containing M methanol, G(CH10) is 3.2. Since this is not greatly different from G(0H) it is unlikely that reaction (10) is occurring to any significant extent. The effect of added ferrocyanide has been investigated and the results for the irradiation of oxygenated solutions are shou7n in fig.4. In so far as the reduction yield is decreased, ferrocyanide must be reacting with species which formerly led ls 0 -4B "6 D I I I 2.0 4.0 4 - 0 8.0 10.0 [&Fe(CN)b] x 102, M FIG. 4.--Effect of ferrocyanide on reduction yield. 18-3 M K3Fe(CN)6 in oxygen saturated 0.1 M NaOH ; 0, 1 M CH30H ; 0, no CH30H. to reduction of ferricyanide. atoms in acid solution is well known, it is unlikely that its counterpart Although the oxidation of ferrous ions by hydrogen H + H+ + Fe(CN);- -+ H, + Fc(CN)i - (18) HO,+Fe(CN):--+HO, +Fe(CN)z-, (19) could take place, particularly in alkaline solutions. possibility, though this reaction seems less likely if the intermediate is 0,.Control experi- nieiits on the reaction of hydrogen peroxide with mixtures of ferro- and ferri-cyanide in 0.2 M NaOH, in which system HO2 (or 0;) is believed to be an intermediate, showed that all HOz (or 02) reacted directly with ferricyanide and that the con- tribution then of reaction (I 9) could be ignored. It would seem that the only mechan- ism by which the reduction yield may be decreased is in scavenging of the mole- cular hydrogen peroxide by the ferrocyanide. If the formation of molecular hydrogen peroxide in the system may be represented formally by the reaction, and QH radicals are scavenged by the ferrocyanide, as in reaction (7), then radicals scavenged in this way lead to oxidation rather than reduction via the formation of molecular hydrogen peroxide.If the yield of molecular hydrogen peroxide in the absence of ferrocyanide is denoted as usual by G(HzQz), and that scavenged by ferrocyanide by Gs(Hz02) then, in the absence of methanol, Reaction with H02 is a OH+ OH-+'--I,Q2, (20) G(--Fe(CN);-) = G(H)+2G(H,Q2)- G(OH)-4G,(H2O2), so that the reduction yield is especially sensitive to scavenging of the molecular hydrogen peroxide. The dependence of Gs(H202) on ferrocyanide concentration, as calculated from the results of fig. 4 is given in table 1.228 RADIOLYSIS OF FERRICYANIDE SOLUTIONS If it is accepted 5 that G(H202) is ~ 0 . 7 , then two processes must be contributing to the formation of molecular hydrogen peroxide, the one whose precursor is readily scavenged by ferrocyanide and TABLE DEPENDENCE OF Gs(H202) ON the other, whose precursor is only sca- CONCENTRATION OF FERROCYANIDE venged with difficulty, if at all, at higher [F~(cN)~,-I x 103, M G d S ~ o i ) concentrations of ferrocyanide.Two similar processes have been suggested for the formation of molecular hydrogen in 1-5 -08 5.0 -15 acid solutions 13 Studies of scavenger 10.0 -20 20.0 -25 effects on molecular hydrogen peroxide 50.0 -31 in hydrogen peroxide solutions 14 indicate 100.0 -33 that in acid and neutral solutions, only a fraction, G = 0-4-0.5, of the molecular hydrogen peroxide is directly scavengeable. Our results indicate that the yield of molecular hydrogen peroxide readily scavengeable by ferrocyanide is ~ 0 . 3 . It is unlikely that any effect due to scavenging of the molecular hydrogen peroxide by methanol would be noticed in this system since any radical precursors scavenged by methanol would still lead ultimately to reduction of ferricyanide. However, in 10-3 M &Fe(CN)6, containing 10-2 M K4Fe(CN)6 and M CH30H in 0.1 M NaOH, G(-Fe(CN)i-) is 7.1 and is independent of methanol concentration in the range 0-5-1-0 M.This yield is significantly lower than the value of 7.8 observed in the absence of added ferrocyanide. The discrepancy is approximately the same as that produced by 10-2M K4Fe(CN)6 in the absence of methanol. The same effect is observed at lower ferrocyanide concentrations as shown in fig. 4. As will be shown later, k7/kll = 0.98, so that if scavenging of the molecular hydrogen per- oxide involved reaction with OH in the spur, then, under the above conditions, scavenging would be predominantly by methanol and it would therefore be expected that G(--Fe(CN);-) = 7.8.It would appear that scavenging of the precursor of the molecular hydrogen peroxide by the competing methanol and ferrocyanide involves different reacting species to those occurring in the bulk of the solution and that ferrocyanide is a much more efficient scavenger than methanol for those species. If the scavengeable yield of hydrogen peroxide were formed via (22) H2 0' + H, 0 * + Hz02 f 2H+, then it is possible that such a precursor could be readily scavenged by ferrocyanide, H2 0' + Fe(CN):- -+ H, 0 + Fe(CN):-, but not by methanol. Some different precursor, possibly an excited molecule, would be responsible for the less readily scavengeable yields of hydrogen peroxide, 2H20*+H2O2 +H2.(24) At high methanol concentrations, the lowering of G(-Fe(CN)i-) by ferro- cyanide can be accounted for entirely by its effect on the molecular hydrogen peroxide yield. In fig. 4, the difference between curves 1 and 2 is equal to 2G(OH). This is independent of ferrocyanide concentration within the range measured. Since G(0H) is constant while Gs(Hz02) increases, then whatever the precursor of the molecular hydrogen peroxide scavenged in the spur by the ferrocyanide, it cannot at the same time be also a precursor of the OH radical in the bulk of the solution.G . HUGHES AND C. WILLIS 229 At intermediate methanol concentrations, competition of methanol and ferro- cyanide for the OH radical may occur: OH+ CH30H+H20 + CH,OH (11) (7) OH + Fe(CN):- -+ OH- + Fe(CN);- The results obtained are given in table 2.ferrocyanide, then If reaction (.7) were the only effect of However, molecular hydrogen peroxide may also be scavenged though this effect will be determined by ferrocyanide concentration only. Under these conditions, it follows that G(-Fe(CN);-) = G(H)+2G(H202)- G(0H)-4G,(H20,)+ (26) 2k1 [CH, OH]G(OH) kl 1[CH3 OH] + k,[Fe(CN):-] TABLE 2.-REDUCTION YIELDS AT INTERMEDIATE METHANOL CONCENTRATIONS [CH3OH] X 102, M 100.0 50.0 10.0 2.0 2.0 2.0 1.0 [Fe(CN):-] x 102, M G(-Fe(CN)%-) 1.0 7.13 1.0 7.06 1.0 6.36 0.5 6-06 1.0 5.35 2 . 0 3.87 1.0 3-74 The reduction yield in the absence of alcohol is given by G(-Fe(CN);-) = G(H)+2G(H202)- G(OH)-4G,(H2O2), (21) 2k1 [CH, OH]G(OH) :.G,(--Fe(CN):-) - Go(-Fe(CN)2-) = kll[CH,OH] +k,[Fe(CN)z-]' (27) where G1(-Fe(CN)z-) and Ga(-Fe(CN):-) are the reduction yields in the presence and absence of alcohol respectively at the same ferrocyanide concentration. Eqn. (27) may be rearranged to give k , [Fe(CN)g-] = l + 2G(OH) GI(-Fe( CN); -) - Go(-Fe(CN): - ) kllCCH3QHI - A plot of 2G(OH)/(G1(-Fe(CN)z4) - Go(-Fe(CN);-)) against [Fe(CN);-]/[CH@H J is shown in fig. 5 and is linear. From fig. 5 it may be deduced that k7/kll = 0.98. Neglecting the effect of ferrocyanide on the molecular hydrogen peroxide does not seriously affect the value of k7/kll, though a less satisfactory kinetic plot is obtained. Relative rate constants for the reaction of OH radicals with substrates have been obtained previoudy though at a lower pH than in this work.15 Thus,it has been shown for the reaction, OH+CH3CH2OH+H2O+CH3CHOH, (29) k7]kZ9 is 9 over the pH range 6-10-5.From the known reactivities of methanol and ethanol, k,/kll at these lower pH would be greater. From work on Fenton's reagent,l6 it can be estimated that k29/k111: 2 in acid solutions. If the rates of alcohol230 RADIOLYSIS OF FERRICYANIDE SOLUTIONS reactions are independent of pH, then a value of k7/kll- 18 would be expected. Preliminary experiments on acid solutions of ferrocyanide and methanol indicate an even higher value.17 Thus there is a marked difference in reactivity between the species of OH present at pH 13 and that in acid solutions, such that at pH 13, reaction with methanol is considerably enhanced relative to that with ferrocyanide or chloride ion.The species involved in acid solutions is the actual OH radical.18 Earlier work had suggested that for the ionization of the OH radical 19 OH+ 0- +H+ (30) pK 21 10, so that at pH 13,O- should be the predominant species. Using appropriate thermochemical cycles, estimates of heats of reaction indicate that whereas there is little difference between reactions of OH or 0- with methanol, the reaction of either ferrocyanide or chloride ion with 0- is much less favoured than that with OH. 6 0' - q , , I , $ 0.2 0.4 0 . 6 0.8 1.0 Q. % [Fe(CN) :-I/ [CH30Hl FIG. 5.-Kinetic plot for reaction of OH with CH3OK and Fe(CN):-. pK for the ionization of OH is not yet known with sufficient accuracy. Un- certainties in thermodynamic data lead to theoretical estimates 179 19 of pK from 8 to > 15.If the higher value of pK were correct, 0- must be produced from some radiation-produced precursor. In this case, the effect of pH could be represented as HzO+ + HzO+H30+ + OH at low pH, (31) and H20+ + OH--+H30+ + 0- at high pH. (32) However, the precursor H20+ must be different from that involved in the formation of molecular hydrogen peroxide. This latter may be an excited ion. If 0- is pro- duced via ionization of OH, a possible precursor of OH would be an excited water molecule This might also be one of the precursors of molecular hydrogen peroxide since it is only scavenged with difficulty, if at all, by ferrocyanide. A study of the effect of pH on k7/kll might yield information on the ionization of OH. This should dis- tinguish between the possible modes of formation of 0-. H,O*+H+OH. (33) One of us (C. W.) thanks the University of Liverpool for a maintenance grant during the tenure of which this work was carried out.G. HUGHES AND C . WILLIS 23 1 Allen, Hogan and Rothschild, Radiafiotz 1 Allen and Rothschild, Radiation Res., 1957, 7, 591. 2 Tarrago, Masri and Lefort, Cornpt. rend., 1957,344,244. 3 Hughes and Willis, J. Chem. SOC., 1962, 4848. 4 Hughes and Willis, Proc. 2nd Int. Con$ Radiation Research, Harrogate, 1962, to be published. 5 Dainton and Watt, Nature, 1962, 195, 1294. 6 Haissinsky, J . Chim. physique, 1963, 60? 402. 7 Bricker and Johnson, Ind. Eng. Chem. (Anal.), 1945, 17,400. 8 Allan and Scholes, Nature, 1960,187, 218. 9 Uri, Chem. Rev., 1952, 50, 375. 10 Masri and Haissinsky, J. Chim. physique, 1963, 60, 397. 11 Dewhurst, Trans. Faraday SOC., 1952, 48, 905. 12 Hochanadel and Lind, Ann. Reo. Physic. Chem., 1956, 7, 91. 13 Hayon, Nature, 1962, 194, 731. 14 Anbar, Guttman and Stein, J. Chem. Physics, 1961, 34, 703. 15 Rabani and Stein, Trans. Faraday SOC., 1962, 58, 2150. 16 Waters, Disc. Furaday SOC., 1947, 2, 179. 17 Hughes and Willis, unpublished work. 18 Hummel and Allen, Radiation Res., 1962, 17, 302. 19Hart, J, Amer. Chem. Soc., 1953,75, 6169. Res., 1957, 7, 603. Rothschild and Allen, Radiation Res., 1958, 8, 101.

ISSN:0366-9033    

DOI:10.1039/DF9633600223

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

GENERAL DISCUSSION Prof. J. L. Magee (University of Notre Dame) said I would like to make a few remarks which bear on the applicability of the concept of temperature in the tracks of high energy particIes. What I shall say is largely theoretical and more or less summarizes my own ideas of the deposit of energy and the early history of a track.1 Because high-energy charged partides lose about equal amounts of energy in low-energy (resonant) losses and higher energy processes (delta rays) we like to picture the track as a string of beads with branches which are themselves similar strings of beads. For low L.E.T. particles such as electrons formed in absorption of C060 y radiation the beads (spurs) are well-isolated from each other regardless of any detailed mechanisms which may be operative.Unless otherwise stated we shall be concerned with a single spur of a low L.E.T. track in a medium such as a liquid. The state of the system formed by the energy deposit is very far from equilib-rium in a thermodynamic sense and all processes which occur will naturally tend to make the system approach equilibrium. At some later time a spur will appear as a region which has a somewhat elevated temperature. For this purpose sometimes a Pt scale (similar to the pH scale of hydrogen ion concentra-tion) is used Pt means the negative logarithm of the time in sec. The earliest time which is significant for a spur which accepts about 50 eV from a fast electron we can estimate from the uncertainty principle AtAE-ti. A lapse of time of about A t - 10-15/50-2 x 10-17 sec is required before it can be known that the particle lost the energy in some vicinity near the particle path.This puts us be-tween 17 and 16 on the Pt scale of time. There is a corresponding uncertainty of the location of the energy and this is what is called a " delocalized excitation ". The subsequent fate of this energy depends upon the system and there are many significantly different possibilities. The development of a spur in molecular liquids probably follows a similar pattern in most cases and we are thinking of this system. At the earliest times (i.e. Pt> 15) the energy is entirely in electronic motion. It migrates in space and makes transitions from one state to another. In a general way we may say that a one exciton state was initially formed and that transitions to states containing larger numbers of excitons occur.Frequently the expression " cascading " of electronic states is used to describe this period. The tendency is to spread the excitation to more electrons. The earliest description of this period (used TO years ago) was given in terms of a secondary electron which has kinetic energy equal to 50 eV minus the ionization potential. Such a low-energy electron was thought to produce a short track losing its energy in further ionizations and excitations. A more sophisticated description is now being attempted. In this early stage all of the energy can be said to be electronic since the nuclei are frozen in their initial positions. On the other hand it is not profitable to attempt to assign an electronic " temperature " to a spur although this may be possible in very high L.E.T.tracks. The transformation of the electronic energy during this period certainly depends upon the nature of the medium. We have thought that in molecular materials the description was quite similar regardless of state (i.e. liquid glass or crystalline) but it has always been evident that in principle such differences were possible. While the energy is in several excitons migration distance may be much larger in 1 see also Magee Ann. Rev. Physic. Chern. 1961 389. 232 It is interesting to follow the events as they develop in time in a spur GENERAL DISCUSSION 233 crystalline materials than in liquids and the processes which alter such excitons can also be different.No substantial amount of energy can be said to be in vibrational energy before the order of time v-1 of the vibrations but in such time there is every reason to believe that a large fraction of the energy will be in such motion. The mechanisms of vibrational excitation involve the Franck-Condon effect accompanying electronic transitions oscillation of the electronic charge (i.e. what might naively be thought of as impacts of slow electrons) and transfers of excitation from more highly ex-cited molecules. The vibrational energy of molecules is transferred in a perfectly natural way into rotations and oscillation of molecules on the time scale of such motions. TABLE PROCESSES IN SPUR processes electronic Pt scale of time 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 initial state excited cascading of elect.states r I parent spur recapture of electrons I I I migration of electronic 1 energy I t I luminescence thermal motion of molecules reactions other actions excitation of vib- ion-molecule re-rational energy dissociation and radical formation excitation of rota-tional energy molecular oscilla- temperature estab-tion Iished diffusion starts bubble formed in t r 1 I cylindrical -1 track reac- cylindrical track bubble chamber spur reaction at low L.E.T. I spur cooled off tion cooled off j tracks overlap at moderate radi-ation rate After the lapse of time of the order of several times the period for jiggling motion of a molecule (10-12 sec) it is possible to assign a temperature distribution to a spur with a certain amount of validity.Before such time it can be done if at all in a very restricted sense. Of course a certain amount of energy will remain in mole-cular dissociation electronic excitation etc. not in equilibrium at this temperature. The magnitude of the initial temperature depends upon the size of the region sharing the energy. There is no real theory for this. The size certainly depends upon the maximum distance the electronic excitation phase extends. It wa 234 GENERAL DISCUSSION suggested in ref. (1) that the region for a 50 eV spur has a radius of about 30 A. This gives a maximum temperature excess of about 50°K. The time constant for the heat flow estimated in terms of constants obtained from observations on macroscopic matter is given by z M r,2/4D M 9 x 10-14/4 x 10-3 M 10-11 sec so that the temperature does not remain elevated very long in the spur.After 10-9 sec the excess temperature is about a degree. The table gives a summary of various events which occur in the development of a track. It has been customary in diffusion models to neglect excess temperature. Al-though this is probably a good procedure for low L.E.T. radiation it is not as clear for high L.E.T. radiations and is almost certainly incorrect for fission fragments. This discussion is in agreement with the theory of bubble chamber operation.1 No bubbles are seen for low L.E.T. radiation for spurs with less than about 103 eV. The initial temperatures and sizes of the heated region estimated by Seitz are in agreement with the present discussion.It is interesting to note that the first theories of the bubble chamber involved electrostatic forces which were supposed to be due to charge separation. At the present time however it is universally believed that bubbles result from heat alone and that no significant charge separation is involved. Dr. A. J. Swallow (Christie Hospital and HoIt Radium Inst.) said I would like to comment on reactions (14) and (17) in Fueki and Magee’s paper. According to McGrath and Norrish 2 the sequence, 0(11))+ 0,+20; 0; + o3+2o2 + o(loj, occurs in the photochemistry of ozone where 0; is vibrationally or possibly elec-tronically excited. Moisture increases the chain length from 8 to 130 and eliminates the 0; through the sequence : 0(1D)+H20+20H OH + 03-+H02 + 0 2 H02 + 03-+OH + 202.There seems to be no reason why these reactions should not all occur in radiolysis, in which case need we look further for an explanation of the radiation-induced chain decomposition ? In our experience on irradiated air ozone measurements are extremely irre-producible even under apparently very carefully controlled conditions.3 Con-sequently it is doubtful whether the oxygen system is likely to be very valuable for a comparison of theoretical prediction with experimental results. Prof. J. L. Magee (University of Notre Dame) said Dr. Fueki and I were familiar with the work of McGrath and Norrish. At very low concentrations of ozone, i.e. in the range of a few parts per million of interest in our paper the electronically excited species are quenched before they can collide with ozone and it is impossible for Dr.Swallow’s scheme to account for any significant fraction of the ozone de-composition. A chain involving electronically excited species is certainly impossible. We have made a quantitative investigation of the negative ion chain suggested in the paper and it seems that this chain is almost certainly involved. We were aware of the contradictory nature of various sets of experimental results and considered it a challenge. A careful investigation of the reaction mechanism was clearly indicated and we hope that our paper will be considered as a good start. 1 Seitz Physics Fluids 1958 1 1. 3 Less and Swallow to be published. 2 McGrath and Norrish Nature 1958 182 235 GENERAL DISCUSSION 23 5 Dr.W. G. Bums (A.E. R.E. Harwell) (cornnzunicated) The calculations of the initial G values of the primary species (table 4 of Fueki and Magee's paper) have been carried out for pressures up to 100 atm. These G values depend on the numbers of primary species formed per 100 eV and the probability of initial ion recapture of electrons. At such high pressures random recombination of the initial ions and electrons formed in the track might be expected to become im-portant with increasing L.E.T. and since first-order reactions of these species are considered (table 1) the probability of recapture should be dependent on L.E.T. Is this an important effect? Prof. J. L. Magee (University of Notre Dame) said Generally speaking as L.E.T.increases the effective coulomb force on ionized electrons will become larger due to the contribution of the neighbouring spurs and the escape probability of electrons will be decreased. Where the average interspur distance in a track is smaller than R (the distance at which the potential energy is equal to -kT in an isolated spur) this effect is very significant. A quantitative study has not been made. In our case at high pressures the situation mentioned above is favourable to the decrease of escape probability of electrons. Nevertheless the G-value for ozone formation would not be much changed by this factor. Using our mechanism, variation of the escape probability (Ep) from zero to 0.4 changes the value of G(O3) by less than 20 %. Dr.G. R. A. Johnson (University of Newcastle-upon-Tyne) said An investiga-tion of the pray induced formation of ozone from oxygen gas carried out in col-laboration with J. M. Warman has shown that the initial G(O3) = 12.8+0.4 is independent of pressure between 100 and 760mmHg and of the temperature be-tween 90 and 290°K. This value of G(O3) in the gas phase may be compared with the value G(03) = 12.9+0.4 in liquid oxygen at 90°K obtained by Riley.1 Appar-ently G(O3) does not decrease in going from the gas at 100 mm to the liquid con-trary to the predictions of the Fueki and Magee calculation. G(O3) from oxygen gas irradiated at 290"K decreased with increasing dose a fall-off from the initial value being observed at conversions as low as 10-4 mole %. This dependence of G(O3) on dose has been observed by earlier workers2 and it has been assumed by Fueki and Magee and others3 that a chain-decomposition of ozone is involved.However although there is evidence4 that a chain-decom-position occurs when pure ozone is irradiated it is unlikely that a similar mechanism will be responsible for ozone decomposition during the initial stages of oxygen radiolysis when ozone is present at a concentration of less than 10-3 mole %. The dependence of G(O3) on dose may be interpreted in terms of a non-chain, back-reaction involving reaction of ozone with some precursor of ozone. The nature of this back-reaction remains in doubt. However the reaction O + O3+2O2, where the O-atom is either in the 3 P or the 1D states cannot be involved since there is evidence 5 that such a reaction would be unfavourable compared with the reaction, 0 4- 0 2 4- 02-+03 + 0 2 , at the ozone and oxygen concentrations at which the back-reaction occurs.1 Riley J. F. private communication. 2 cf. Lind Radiation Chemistry of Gases (Reinhold Publishing Corporation New York 1961), 3 Kircher McNulty McFarling and Levy Rad. Res. 1960 13 452. 4 Lewis J. Physic. Chem. 1933 37 533. 5 cf. Kaufman Progress in Reaction Kinetics (Pergamon Press London 1961) p. 3. p. 83 236 GENERAL DISCUSSION Our results have also shown that the back-reaction is temperature dependent. At 90°K the yield of ozone was directly proportional to dose up to the highest dose used corresponding to a conversion of 2 x 10-2 mole %. Prof. J. L. Magee (University of Notre Dame) said The high-pressure limit in our calculation is 100 atm while the pressure range in Johnson and Warman’s experimental work is 100-760 mrn Hg over which range the initial G(03) does not change significantly in our calculation.There are two experimental investiga-tions 1 9 2 giving G(O3)x6 in liquid oxygen. We do not think that G(03) in liquid oxygen is well established at the present time. If the high initial value of G(03) in liquid 0 2 is verified it must indicate a significant alteration of the primary processes as compared with the gas phase. Our chain mechanism is an ionic one. Since the stationary concentration of the ions will be much higher than that of the oxygen atom it is possible that the ionic chain mechanism will be responsible for ozone decomposition even when ozone is present at a low concentration.In liquid oxygen the stationary concen-tration of the ions is expected to be much lower and also some of the ion-molecule reactions which contribute to the ionic chain may be temperature dependent or be inhibited due to the change in energetics in liquid oxygen. These effects will possibly result in an increase of the stationary concentration of ozone in liquid oxygen. Dr. A. Henglein (Mellon Inst. Pittsburg) said How much may ionic processes contribute to the formation of ozone? If Dr. Magee assumes that all ions are formed in their ground state the ion-molecule reaction 0; * + 02-+03+ + 0 is excluded from the proposed reaction scheme. It has been shown by Cermak and Herman 3 that this excited ion molecule reaction takes place in a mass spectro-meter and some evidence for a lifetime longer than 10-6 sec of the Oi* ion has been obtained in our laboratory.4 In a real system in which excited oxygen ions are formed part of the ozone may well be produced by that ion-molecule reaction.Prof. J. L. Magee (University of Notre Dame) said The cross-section of the ion-molecule reaction mentioned by Dr. Henglein is relatively very small N 10-16 cm2 as reported by Cermak and Herman.5 In the low-pressure region a certain amount of 0; is formed by this reaction and produces a slight increase in the ozone yield, but even here quenching of the excited ion probably dominates. In the high-pressure region ion clustering is more important and determines the ozone formation yield.Prof. A. Charlesby (Roy. Military CoZZ. of Sci. Shrivenham) said In a number of radiation-induced reactions critical concentrations often arise which correspond to distances of about l0OA. Examples are the maximum concentration of radicals revealed by e.s.r. in many irradiated solids and the concentration of radical scav-engers beyond which little further change is obtained. Some of these concentra-tions may perhaps be explained in terms of the diffusion distance of an H atom, or in other ways. Does Prof. Magee think that this may also be the radius of the delocalization volume over which energy is initially spread and that this energy then focuses itself down on the most suitable electron in this volume? Furthermore how does the delocalization theory following Fano tie up with the simple uncertainty principle? If a particle travelling with a speed close to 1 Kircher McNulty McFarling and Levy Rad.Res. 1960 13,452. 2Brown and Wall J. Physic. Chem. 1961 65 915. 3 Cermak and Herman J. Chim. Physique 1960,57,717. 4 Henglein 2. Naturforsch. 1962 17a 37. 5 Cermak and Herman J. Chim. Physique 1960,57,717 GENERAL DISCUSSION 237 that of light (e.g. a fast electron) loses about 30 eV during a collision the corres-ponding uncertainty in time is about 10-16 to 2 x 10-17 see and the distance travelled 60 A or more. This in itself prevents us from saying that 30 eV have been trans-ferred to a given electron in a solid or liquid. Prof. J. L. Magee (University of Notre Dame) said It has been recognized that chemical effects of small components (i.e.scavengers) could be brought about through their role in the localization of energy of a delocalized electronic state. A number of examples which could be explained in this manner have been proposed but as far as I know it is still just speculation. The establishment that this effect is responsible would be difficult although some day it may be possible. The uncertainty principle can be said to cause the initial delocalization of the electronic state produced by the loss process itself. This principle is applicable in the same way to loss processes in any state of aggregation. The subsequent transformations of the electronic energy depend very much upon the magnitude and type of the electronic interaction energy in the system.In a one-component system uncertainty with respect to the site of localization of energy utilized in chemical reaction is unimportant. On the other hand in a two or more component system there are many non-equivalent sites and the manner of localization can make differ-ences in the subsequent chemical reactions. I think the suggestion most commonly made regarding the effect of a small component is that it " triggers " the cascade of electronic states so that it is inevitably involved in the later chemical stages. Prof. F. S . Dahton (University of Leeds) said In considering recently the results of some experiments on the y-radiolysis of liquid GO2 at - 51 6°C and 7 atm pressure made by Dr. R. L. S. Willix in Leeds I had come to essentially similar conclusions to those of Dr.Dominey. The principal facts and conclusions are the following. Liquid C82 is decomposed into CO 0 2 and 0 3 only. Initial values of G(CO)=5 and G(02)~0.25 fall and rise respectively to constant values of 3.5 0.3 and 0.65 i 0.2 at doses 3 1020 eV/ml where G(03) = 0.7 k0.2 so that within the experimental error G(C0) = 2G(02)+3G(03). No suboxide was detected nor was there any effect of dose-rate. These results are explicable on the basis that the primary act can be summarized by C02(lX)-+C0(1C) + O(3P or 10). Reaction of O(3P) + C02 -+GO + 0 2 can be ignored because E> 9 kcal and O(1D) + C02-CO + 0 2 is spin for-bidden and has been shown by Taube to be slow at room temperature. Further since k~o+o+M-rOz+M)>k~CO+O+M-tC02) the 0 atoms do not react with CO (this has been confirmed by demonstrating the absence of oxygen isotope exchange when mixtures of C1602 and CIS02 are irradiated) but form oxygen.When [02] reaches a value of about 10-7 M oxygen atom recombination gives way to formation and decomposition of ozone regulated by the reactions 0 4 - 0 2 4 - bb03-t-M O+03+202. However if the primary act is the same in the gas phase at pressures of C02 greater than a few mm Hg these arguments are equally valid and we expect G(-CO& >3-5 whereas gaseous C02 is noted for its radiation stability. Hence we conclude that there is a fundamental difference between the mechanisms of radiolysis of liquid and gaseous carbon dioxide. In the former the L.E.T. is higher and the lifetime of the positive and negative ions is short whereas in the latter the charge separation and ion lifetimes are larger.If CO can be rapidly oxidized by any of the ions formed before charge neutralization then a rapid back reaction will exist in the gas phase which will be absent in the liquid. The ions concerned cannot be CO; or COY and must therefore be O$ or 0 2 which could be formed by charg 238 GENERAL DISCUSSION transfer from CO$ (Dominey's reaction (7)) and by direet electron capture by 0 2 . Provided either reaction (I) or reaction (2) or both can occur rapidly and if AH(eV) Q:+CO-+O++CO (or CO; see Johnson) (1) +0.9 GO; or 0; or 0; can also react with NO2 and SO2 then the major features of the radiation chemistry of liquid and gaseous CO2 can be accounted for. Of the two reactions I prefer (2) as being energetically the more feasible.Ds. D. A. Dominey and T. F. Palmer (A.E.R.E. Harwell) said The general mechanism put forward by Prof. Dainton is iil agreement with our own views on this subject but we feel that there is as yet insufficient evidence to identify the actual process by which CO is oxidized to CO2. Rcactions which are endothermic be-tween ground-state species have been shown to occur efficiently due to the parti-cipation of excited ions e.g., Although the reaction between CO and 0; is endothermic it cannot be ruled out. Dr. W. Wild (A.E.R.E. Harwell) said The parent-ion CO; is easily the pre-dominant feature of the mass-spectrum of carbon dioxide. Brown and Miller2 have reported a photo-multiplier study of a strongly quenched light-emission arising fron a-particle irradiation of COa.This was ascribed by them to the Angstrom bands of C0* (1 2") formed by dissociative excitation of carbon dioxide. Some years later Ds. €3. Brocklehurst 3 photographed the eniission by exposing the gas to an intense beam of 50 kVp X-rays. A strong band at 2890A and a single pro-gression of bands between 3100 and 4400 A was observed. These could be identified as transitions from the CO,' levels ZZ; and 2nu to the ground state ZITs. These band systems are observed in the negative glow of discharges in COz and it is not surprising that they were observed here. Using data of the first-named authors, he concluded that - 15 % of the CO,' ions are found in an excited state. The simple Stern-Volmer quenching found by Brown and Miller probably involves conversion of electronic into some form of thermal energy the energy iiivolved (up to 4.3 eV) is smaller than that required for bond breaking (5.7 eV to form CO-++ 0 ; 5-3 eV for CO+O+).Dr. I$. S. Corney (G.E.C. Lrd. Wernbley) said In contrast to the observation of Prof. Harteck concerning the radiation-induced decomposition of carbon dioxide at high pressure I should like to say that we could detect no such decomposition at pressures up to 7 atm at a temperature of 100°C using pile radiation to a dose of 2.9 eV/C02 molecule.4 More recent work has given the same result at doses up to 22 eV/C02 molecule. Dr. Dominey has observed in his present paper a real difference in the values obtained for G(exchange) in a gamma source (G = 4.2) and in the reactor (G = 3.0).This difference is in the same sense as the earlier observation for the CQz+NOz system where in the gamma source G(C0) = 3.5 and in the reactor G(C0) = 2.8. We have observed a similar difference in the radiation-induced reaction of C02 with graphite under conditions where the gas is mainly confined within the pores of the graphite. In the gamma source G(C0) = 2.3 whereas in the reactor a range of G(C0) values between 0.8 and 2-3 have been obtained with a mean of 1.5. Another O~+CO-+Q-+CQ (2) -1.0 Oi+H2+HOi+H.1 1 Hamill and Kubose J. Anzer. Chem. Soc. 1963 85 125. 2 Brown and Miller Tram. Faraday Soc. 1957 53 748. 3 Brocklehurst U.K.A.E.A. Research Group Report C/R 2669 to appear shortly. 4 Copestake and Corney Nature 1961 191 1192 GENERAL DISCUSSION 239 example of this difference qpears on comparing the value of G(C02) = 2.3 ob-tained by Clay Jolinson and Warman for the radiolysis of CO in a gainma source with our own value of G(C02) = 1.2 for the radiolysis in the reactor.These values indicate that there is a real and systematic difference between the effccts of gamma and reactor irradiation in these gas phase systems an effect which has been noted in condensed phase systems. Prof. F. S . Dainton (University of Leeds) said Can Dr. Beck tell us whether his investigations extend over a sufficient number of alkanes and aromatic hydro-carbons to justifsr any generalization that the value of y the ratio of neutral molecule decompositions per ion fragments is likely to be generally larger for the former than for the latter group of compounds? 1 ask this question because of the marked differences between the known radiation chemical characteristics of these two classes.Secondly do the benzene data prechde the possibility that electronically excited benzene molecules have lifetimes before decomposition which are longer than the time limit set by his experimental arrangement and would it be possible to vary this time ? Dr. D. Beck (Freiburg) (comnzuizicated) Regarding Prof. Dainton’s first question, propane n-butane n-hexane and benzene are the only substances for which y has been measured. y is the ratio of the number of neutral molecule decoinpositions per ion produced i.e. including parent ions. Particularly for unsaturated as well as aromatic hydrocarbons it is thus not a good measure of the ratio of neutral to ion decompositions because in these substances the parent ion is quite often dominant.For instance for benzene the ratio of neutral to ion decompositions (i.e. excluding parent ions) may well be close to unity within the large limits of error for this substance. Secondly the data presented in the paper I think do not give evidence for or against sufficiently long-lived excited states of benzene. Measurements to give such evidence are possible but were not made at the time when we investigated benzene. Varying the time limit of the apparatus is feasible between very narrow limits only depsiiding on the molecule studied. Because of this restriction we have not yet seriously considered the chances of an experiment of this kind.Dr. P. Ausloos (National Bzlreau of Standards) said There is some inconsistency concerning the niodes of decomposition of neutral n-butane and propane. For instance according to Dr. Beck free radicals phy a very minor role in ths decom-position of neutral n-butane molecules while recent detailed studies by Dr. Okabe and Becker have shown that in the vacwni ultra-violet photolysis using the xenon and krypton resonance lines the ratio of free radicals to ‘’ molecular ” products is 0.73 and 1.33 respectively. The increase of the relative importance of free radical formation with an increase in the degree of excitation is a general phenomenon, which has for instance also been observed for propane. My second comment concerns the role of excited molecules in the gas-phase radiolysis of hydrocarbons.The following comparison of the yields of the “ mole-cular ’’ products from n-butane with the yields of the major ion (C3I-1;) clearly shows that neutral excited molecule decompositions are considerably less important in II-BUTANE C 3 q - CH4 C2H6 C2H4 C3H6 mass spectrophotometer 1 100 122 46 262 46 radiolysis 2 100 16 22 44 10 1 Beck this Discussion. Beck and Niehaus J. Chem. Physics 1962 37 2705. 2 Borkowski and Ausloos J . Chem. Physics 1963 39 818 240 GENERAL DXSCUSSION the radiolysis than under the experimental conditions used by Dr. Beck. It should, however be noted that the mass-spectrometric study was carried out at a pressure of 10-5 mm as compared to a pressure of 3 cm in the radiolysis study.Dr. T. J. Hardwick (Gulf Res. and Development Co.) said In connection with Dr. Beck’s paper I should like to draw attention to two other pertinent results. McNesby 1 has studied the photolysis of propane in the far u.-v. and has concluded that most evolved hydrogen gas originates by molecular elimination. Yang 2 has decomposed propane gas by y-radiolysis and using hydrogen atom scavengers, has determined the ratio “ molecular ” hydrogen gas/hydrogen atoms to be 0.30. The following table summarizes the data obtained for the decomposition of pro-pane and indicates the source of the products measured. experimenter products collected ratio ionic from H2/H intermediates decomposition removed of excited molecules Beck Yes Yes 17 no Yes 0.30 McNesby none Yes > 20 A reasonable conclusion to be drawn from these data is that in radiolysis most of the molecular hydrogen is generated from molecules in highly excited but un-charged states whereas the precursors of hydrogen atoms are predominantly molecules which have undergone ionization and subsequent charge neutralization before decomposing.A similar case can be made with the more limited results using n-butane. Dr. F. S. Dainton (University of Leeds) said Dr. Futrell has stressed the im-portance of 1-3 elimination of methane from alkanes which is referred to by Dr. Ausloos. What is the reason for this? Is it possible that the molecular structures of the C3Hi ion is such that both C1 and C3 are electron-deficient and are joined by a hydrogen bond of a type somewhat analogous to that found in the boranes? I think I have seen this suggestion made but I do not recall the author.H H Dr. J. H. Futrell and T. 0. Tiernan (Aerospace Res. Lab. Wright-Patterson Air Force Base Ohio) said We have studied dissociative charge exchange reactions of propane and of the deuterated propanes CH3CD2CH3 and CD3CH2CD3 with the rare gas ions.3 We record here some results pertinent to molecular elimination processes in the gas-phase radiolysis of propane. Our experiments with propane yielded charge exchange data in excellent agreement with Pettersonn and Lindholm 4 and served to validate the technique which was used while data on 2,2-dideutero-propane served mainly to complement the results for 1 1,1,3,3,3-hexadeutero propane.1 Okabe and McNesby J. Chem. Physics 1962,37 1340. ZYang J. Amer. Chein. SOC. 1963 84,719. 3 Futrell and Tiernan J. Chem. Physics (to be published). 4 Pettersonn and Lindholm Arkiv F’sik 1963 24,49 GENERAL DISCUSSION 24 I The data were obtained with a single-stage 180" Dempster-type magnetic mass-spectrometer operating the ion source in the manner first developed by Cermak and Herman 1 and subsequently employed by Henglein and Muccini.2 This tech-nique involves formation of ions by electron impact outside the usual ion source region and accelerating these ions into the source in a direction perpendicular to the usual ion trajectory. Thus complete discrimination against primary ions may be achieved, and only secondary ions as from charge exchange are extracted and focused by the mass spectrometer.Hence the technique is particularly well suited to the study of dissociative charge transfer processes A++BCD = CD++B+A (1) where in the present case A represents a rare gas atom and BCD the partially deuter-ated propane molecule. Data obtained in a series of measurements with CD3CH2CD3 are summarized in table 1. In columns (a) are the experimental data which have been corrected TABLE 1 .--CHARGE EXCHANGE MASS SPECTRA OF 1,1,1,3,3,3-HEXADEUTERO PROPANE WITH RARE GAS IONS Xe Kr Ar mle species (4 (b) (4 (6) (4 (h) 18 27 28 29 30 31 32 44 45 46 47 48 49 1.0 1.0 1.8 1.0 20-0 20.4C 7.0 4.8 44.2 46.9 1.8 1.7 1-5 0.6 8.7 9.7 1.9 1.9 0.7 0.4 2.7 2.4 4.2 4.0 14.0 13.1 8.2 8.2 31.0 31.2 8-3 8.4 16.0 16.8 7-2 3.9 4-1 3.7 58.4 61.9 13.7 14.6 1.3 1.0 3.3 2-9 3.2 3.3 8-2 8.3 2.3 2.3 4-9 5.4 2.0 2.1 1.1 1.0 2.4 2.6 a as measured ; C from results on unlabelled propane contribution of vinyl ions CzDi to this m/e is approxim-b corrected for 98 % deuteration ; ately 1 %.for C13 content and for the mass-dependent relative multiplier response observed with our apparatus and in columns (b) these results are corrected for the observed 98 atm per cent deuteration and renormalized.* Only isopropyl ion and ethylene ion yields are materially affected by this correction factor. Of interest is the forma-tion of ethylene ion from the parent molecule ion of propane which is accompanied by the elimination of molecular methane, * This correction may be illustrated for C3D6Hf propyl ion at m/e 49 for which the probabilities of actual d-6 and d-5 ions being formed are (0.98) 6 and 6(0-98) 5 (0-02) 0-886 and 0.108 respectively.Thus the measured ion abundance at ni/e 49 is multiplied by (0.1 1)/(0.89) and the result is both added to mje 49 and subtracted from the m/e 48 to correct for lack of total deuteration in the specified positions. A similar correction is applied in turn to each ion in the spectrum. 1 Cermak and Herman Nucleonics 1961 19 106. 2 Henglein and Muccini 2. Nuturforsch. 1962 17a 452 242 GENERAL DISCUSSION The excitation function (breakdown curve) for this process was previously established by Pettersonn and Lindholrn for ordinary propane 1 and is reproduced as the solid line in fig.1. The vertical lines are the recombination energies of the rare gas ions used in the present study and we expect (except for isotope effects) that the charge exchange spectra will yield points falling on the breakdown curve for ordinary propane. From table 1 it is clear that the mechanism of ethylene ion formation involves at least two reactions 1,3 elimination CD3CH2CDi = C2D2Hf +CD4 (3) and 1,2 elimination of methane, Appearance potentials for both processes were determined and are reported in table 2. Since the appearance potential of ethylene ion from ordinary propane TABLE 2.-APPEARANCE POTENTIALS FOR SELECTED IONS ion mle appearance potentials eV from c3Hg : 27 1 5 - 1 28 11.7 29 1 2 . 3 from CD3CH2CD3 : 30 1 1 . 8 3 1 1 2 . 2 32 1 2 .3 corresponds to mass 30 from the deuterated compound the lowest energy process involves 1,3 elimination of methane as first proposed by Schissler Thompson and Turkevich.2 The corresponding 1,2 elimination process has an appearance potential of 12.2 eV which is above the lowest recombination of energy of xenon at 12- 13 eV. Thus in the charge exchange experiment with xenon all the mass 31 ethylene ion is forined by the upper 2P+ state of the ion which is formed with a relative proba-bility of one-third.6 Normalizing the spectrum at this energy therefore requires multiplying the measured abundance by 3 which gives a point falling on the solid curve corresponding to ethylene ion from ordinary propane. Therefore the 13.44 eV recombination energy of xenon is not responsible for any significant amount of 1,3 methane elimination and the measured abundance of mass 30 ethylene ion should be multiplied by the corresponding normalization factor of 3/2 at the recombin-ation energy of the ZP state.This also yields a point falling on the solid curve as shown in fig. 1. The other points for these two processes are determined by the appearance potentials for the two elimination mechanisms. Thus we can indicate with reasonable precision the individual breakdown curves for the two types of mechanisms involved in ionic decomposition of propane to form ethylene ion and methane. The 1,2 elimination process begins at its appearance potential and reaches the solid curve at some point below the upper state energy level for xenon while 1,3 elimination falls off in corresponding manner.The sharp peak of the break-down curve of Pettersonn and Lindholm,l therefore represents the low energy 1,3 elimination process which falls off quite sharply as the 1,2 elimination process 1 Pettersonn and Lindholm Arkiv F'sik 1963 34 49. 2 Schissler Thompson and Turkevich Disc. Faraday Soc. 1951 10,46. 3 Futrell and Tiernan J. Chem. Physics (to be published) GENERAL DISCUSSION 243 becomes important. This behaviour is sketched in dotted lines in fig. 1 . Deduc-tions from the krypton and argon data are somewhat more ambiguous than xenon because the possibility of processes producing ions of different structure but same nominal mass complicates the interpretation. Howevcr the upward break in the ethylene curve implies a third process and the present data and those obtained for 2,2-dideutero propane suggest that thc third process produces priniarily C2H2Di ethylene ions.1 This postulated third process is also sketched in the figure.I I I i I 5 x e + 2P 2 800 '2 '/2 -I I 1 recombination energy eV FIG. 1 .-Breakdown curve for ethylene ion from propane. Vertical lines represent recombination energies for doublet ground states of rare gas ions. It is also of interest to consider molecular hydrogen elimination steps in the detailed fragmentation mechanisins for tlie deuterated propanes. The major source of hydrogen involves ethyl ion the major ion in the mass spectrum of propane de-caying to vinyl ion with the elimination of hydrogen. C,H,+ = C,M,+ +M ( 5 ) Comparing the data in table 1 with data on ordinary propane and 2,2-dideutero propane and making the first approximation estimate that the amounts of various isotopic ions of the same structure formed by charge exchange are the same in all three cases it is possible to estimate the relative probability of E2 HD and D2 elimination.From such considerations we first estimate the contribution from other isotopic species to the measured argon spectrum of table 1. We estimate that C2D3f is produced to the extent of approximateiy 3.5 % and CZD~H; 4 %. The residual concentrations then lead to an estimate for the relative probability of H2, HD and D? elimination of 0.24 0-58 8.18 from the CzD3E-I; ethyl ion dissociating to vinyl ion in reaction (5). For the recombination energies of argon approxim-ately SO % of the ethyl ions initially formed have dissociated in this manner.2 The second important soiirce of molecular hydrogen involves the decomposition of propyi ion into ailyl ion plus hydrogen.The data for ordinary propane demon-strate that at the recombination energies of argon all of the propyl ions have de-cayed to allyl. The presence of 2.9 o/ C3D3Hl for d-6 propane in table 1 therefore Ar + HOR 1 Prof. Lindhom has reached similar conclusions regarding the detailed mechanisms of ethylene-ion formation in a preliminary report at the ASTM Committee E-14 meeting in San Francisco, May 1963. Futrell and Tiernan J. Chern. Physics (to be published,) 244 GENERAL DISCUSSION indicates the existence of a propyl precursor which lost a hydrogen atom from the primary carbon.The remaining isotopic ally1 ions may be formed from either n- or s-propyl precursors and lack of information on isotope effects the possibility of HD exchange in secondary ions and an incomplete description of the detailed dissociation mechanism make it impossible to conclude to what extent s-propyl ions are formed as a transient intermediate in this high energy region. The abun-dance of C3DS from d-6 propane appears too large to reconcile with all propyl io:i precursor involving I -position elimination. The relative amount of the two primary processes cannot be estiinated from the present data however so thc rela-tive probability of isotopic hydrogen elimination froin the two propyl ion structures cannot be deduced. Nevertheless it is clear that HD elimination is an important process in the decomposition of propjjl ions.The only other numerically significant source of hydrogen in the ionic frag-mentation of propane is the decomposition of ethylene ion to acetylene ion plus hydrogen. C,H,+ = C,H,+ +H (6) This reaction is somewhat less important as a hydrogen source than the two already discussed. A rough estimate from the present data is that H2 and HD elimination from the ethylene ions which consist of a mixture of C2H2Df and C!2D3Hf are of about equal importance. Thus for the overall ionic decomposition of syrn-metrical d-6 propane ND elimination is an important process substantially exceeding the amount expected from statistical considerations for the parent molecule. This may be contrasted with the recent results of Okabe and McNesby con-cerning the vacuum ultra-violet photoiysis of the deuterated 'propanes.1 They concluded that at all wavelengths studied elimination from the 1,l and 2,2 positions account for most of the inolecular hydrogen produced while 1,2 eliinination is a relatively minor process.In addition they demonstrated that molecular methane elimination froin neutral propane occurs mostly via a 1,2 process in contrast to the present results that 1,3 elimination is rather more important in the unimolecular dissociation of propane ions. These observations are consistent with the present findings of Ausloos et al. and as pointed out in their paper may be attributed to the rather larger amounts of energy deposited in the neutral molecule when it undergoes decoinposition by these molecular elimination steps than is required for the corresponding decompositions of the ion.It is therefore suggested that a critical comparison of details of fragmentation processes for the ionic decomposition path with corresponding decoinposition processes for excited neutral molecules may lead to a better understanding of the relative importance of ionic and neutral decomposition processes in the radiation chemistry of simple gases. The authors gratefully acknowledge helpful discussion of these problems with Dr. Vladiinir Cermak and Prof. Einar Lindholm and Prof. Arnim Henglein and are indebted to Dr. Cermak amid Dr. Lindholm for access to unpublisheddata. We further wish to acknowledge many helpful discussions of this and related problems with Dr.Peter Ausloos. Ds. J. Futrell (Air Force Base Ohio) said Recent theoretical work in mass spectrometry of paraffin molecules by A. E. Wahrhaftig and M. Vestal (personal communication) indicates that the processes occurring may be generally classified as proceeding via " loose " complexes or " tight " complexes-the latter character-ized by a larger quantum of energy in the critical oscillator and a higher mean fre-quency. Qualitatively a low activation energy seems to be associated with the 1 Okabe and McNesby J. Chem. Physics 1962,37 1340 GENERAL DISCUSSION 245 “ tight ” activated complex configuration and a high activation energy with “ loose ” configurations. The low energy ethylene ion process involving 1,3-elimination of methane involves such a “ tight ” activated complex configuration in theoretical calculations of the breakdown curve.One means of visualizing such a constrained configuration is a cyclic complex perhaps involving a hydrogen-bridge structure as suggested by Prof. Dainton. One further bit of evidence suggesting a cyclic structure or some other structure in which all carbons are equivalent was reported by Stevenson 1 in which he found that Q of the methyl ions from propane 2-C13 were n7/4 16 C13Hl. It is obvious to me how the structure suggested (which at present is short of one hydrogen a typographical error) would predict 1,3-elimination of methane. It may be useful to think of the transition state in these terms but in view of the limited information available much speculation on fine details of the mechanism may be only that.Dr. W. P. Hauser (Mellon Inst. Pittsburg) said I would like briefly to com-ment on some results which I have obtained from the radiolysis of methane which are relevant to Dr. Hummel’s paper. C14 labelled methane at 100 mm pressure and in the presence of small quantities of unlabelled ethylene (between 0.2 and 0.02 %) was irradiated with 2-8 MeV electrons. At low conversions the unlabelled ethylene was expected to protect the active ethylene produced during the radiolysis. As-suming a yield of ethane of 2 molecules per 100 eV absorbed the yield of active ethylene was found to be 1.3 molecules per 100 eV absorbed which is in agreement with the value reported by Dr. Hummel. When unlabelled methane with trace quantities of C14 labelled ethylene (-0.006 %) was irradiated a yield of 0.6 molecule per 100eV absorbed was obtained for the production of ethane from ethylene in the methane radiolysis.Dr. P. Ausloos (National Bureau of Standards) said In the communications of Dr. Huminel Dr. Hauser and Dr. Rowland the role of ionic reactions is somewhat underestimated. Recent studies in our laboratory have shown that the addition of trace amounts of foreign compounds to methane strongly affects the product distribution mainly because the CHS and C2Hf ions can readily undergo proton and hydride transfer reactions with saturated and unsaturated hydrocarbons. For instance,2 upon addition of 0.01 72 C3Hg to CD4 the following reactions occur: CD + CD,-+C,D; + D, CD + CD,-+CD; + CD, C,D; +C,H,+C,D,H+C,H~ CD; + C,H,+CD,+ CD,H+ C,H: C,H; + C3H8-+C2H6 + C3H;.It may be concluded that addition of olefins even in trace amounts will have a profound effect on the product distribution. In this connection recent studies carried out in our laboratory have shown that in the absence of additives and at conversions of 0-005 % or less C2Hi and CHf will react with accumulated products instead of undergoing neutralization. NO or 0 2 will effectively remove the C2Hf and CHZ ions from the system by reaction. Dr. R. W. Hummel (U.K.A.E.A. Wantage) said It is certainly true that the addition of acetylene ethylene or propylene to methane introduces factors whose 1 Stevenson J. Chern. Physics 1951 19 17. 2 Ausloos Lias and Gorden J. Chem.Physics to be published 246 GENERAL DISCUSSION overall effect on the radiolysis cannot be accurately assessed. However since the unsaturates are themselves products of methane radiolysis the perturbations will presumably be ones of degree only. Examination of the products resulting from such additions is not only possible but also yields what appear to be significant data for mechanistic consideration. On the other hand the addition of other foreign gases such as nitric oxide and iodine introduces complications both of degree and of kind the effects of which are certainly not fully understood even in systems where ions are absent and the chemical products of which are at least in NO systems, almost entirely unknown. Prof. F. S. Dainton (University of Leeds) said Dr.Hummel points out in his paper that methylene is an important intermediate in the radiolysis of methane. Can he tell us the relative proportion of these which are formed in the singlet and triplet states and is the triplet state formed directly or by collisional deactivation of the singlet? I imagine that studies of the degree of stereospecificity in the addition of methylene to say cisbutene-2 and of the effects of pressure and oxygen on this would provide the necessary information. Dr. R. W. Hummel (U.K.A.E.A. Wantage) (communicated) In reply to Prof. Dainton as was reported at the meeting we have carried out some exploratory experiments in which cis or trans butene-2 (1-10 cin Hg) was added to methane (80 cm Hg). After irradiation both cis and trans 1,2-&methylcyclopropane were found among the products the isomer corresponding to the butene isomer being 3- or 4-fold in excess.Addition of several mm pressure of 0 2 to the mixture con-taining trans butene-2 reduced the cis cyclopropane yield but affected the trans cyclopropane yield only slightly. Further work is in progress but at present the facts indicate that both singlet and triplet CH2 are present in irradiated 02-free CH4. At 80 crn pressure some collisional deactivation of singlet CH2 to the triplet state might be expected but little or no isomerization of the initially excited cyclo-propane. The work is not yet quantitative and Prof. Dainton’s question cannot be answered except l o say that most of the CH2 must be produced in the singlet state.Positive identification of triplet CH2 as an initial radiolytic fragment probably is not possible by this technique since at low pressures the excited cyclopropane is able to isomerize and at high pressures singlet CH2 is collisionally converted to the triplet. Prof. A. Charlesby (Roy. Military Coll. of Sci. Shrivenham) said In y-irradiated paraffins there are three major changes when oxygen is absent cross-linking, degradation and trans-unsaturation. To discover whether these can be appor-tioned as between ionic and excitation reactions we have carried out parallel work with u.-v. irradiation of polyethylene incorporating a sensitizer (benzophenone or hexachlorbenzene). After absorption of a u.-v. quantum the sensitizer reacts with the polymer by a radical mechanism being itself destroyed in the process.At the wavelength used (2537 9.) no ionic reaction can occur but nevertheless with benzophenone all three changes are observed in about the same ratio as with gamma irradiation. (With hexachlorbenzene where several radicals are produced together the pro-duction of trans-unsaturation is enhanced about fourfold.) There seems no reason therefore to ascribe a specific ionic character to any one of these reactions. The relatively high efficiency of formation of transvinylene unsaturation by reaction with a neighbouring sensitizer radical is surprising in view of the need to eliminate two hydrogen atoms simultaneously. The alternative explanation whereby energy is transferred from sensitizer to polymer causing the Toss of a hydrogen molecule with formation of unsaturation seems unlikely in view of the low amount of energy available GENERAL DISCUSSION 247 Dr.L. M. Dorfman (Argonne National Laboratory) said I have a question and a comment pertaining to the paper by Dr. Collinson et al. This is specifically concerned with the comparison of the electron multiplication results for butane with our results 1 for the photolysis at 1470 A and the conclusion implied in this comparison. The product distribution in the photolysis (row 1 of table 1) is sub-stantially different from that in mixture 111 run 6 ; do you believe that they corres-pond closely as you state? More important however I suggest that the implication inherent in your con-clusion does not have a sound theoretical basis. Different excited electronic states of the butane molecule need not be expected to give the same decomposition pattern.In the direct photolysis the excitation is almost surely to an optically allowed state. In your electron multiplication experiments on the other hand since excitation by low energy electrons plays a major role there may be a high probability of excitation by electron exchange. Thus you may have substantial formation of optically for-bidden states (either spin-forbidden or symmetry-forbidden depending upon the existence of the particular low-lying states with butane). This important difference in the nature of the electronic excitation would very likely contribute to a different mode of decomposition than occurs in the photolysis. Dr. E. Collinson (University of Leeds) said Dr.Dorfman has perhaps misinter-preted the phrase “correspond most closely with ”. By this we do not mean “ correspond very closely with ” but merely that the degree of correspondence between the yields from the xenon photolysis and those from run 6 though not high in the absolute sense is higher than that for any of the other runs if ratios of all products taken in pairs are compared. Ethylene is a marked exception and it is perhaps unfortunate for the purpose of this comparison that ethylene was chosen for normal-ization since the other products then appear to compare unfavourably. We agree that our technique is likely to give a variety of excited states of both higher and lower energy than that arising from the photolysis of 1470A. For this reason and also because our technique will always give rise to ionization as well as excitation, an exact correspondence with the u.-v.work can never be expected and only trends were sought. At this level the suggestion that an increasing extent of production of excited states in our system leads to an increasing proportion of ethylenic products is reflected in the (admittedly limited) published results with which any comparison could be made. With regard to our assumed ion distribution Dr. Futrell has a valid point; the more appropriate results to which he refers were not available to us at the time of writing. Prof. F. S. Dainton (University of Leeds) said Have Dr. Allen and Dr. Hummel tested the Debye expression for the ion recombination coefficient in relation to the temperature dependence which it predicts ? Dr.A. 0. Allen (Brookhaven Nat. Lab. N.Y.) said No we have not studied effects of changing temperature. Dr. G. R. Freeman (University of Alberta) said In answer to Dr. Dainton’s question about the effect of temperature on G(free ions) G(free ions) appears to be nearly independent of temperature over an 80°C range. The temperature depend-ence of the induced conductance is mostly due to the temperature variation of the ion mobilities. Prof. J. L. Magee (University of Notre Dame) said The G value ~ 0 . 0 9 for charge separation reported by Allen and Hummel is clearly consistent only with 1 Sauer and Dorfman J. Chem. Physics 1961,35,1497 248 GENERAL DISCUSSION escape of electrons from delta rays and agrees with an estimate made by this author 1 obtained by counting delta rays having energy greater than 300eV.If charge separation is found to have a greater value than this e.g. Gz50.2 as suggested by ionic polymerization initiation,z or by other conductivity measurements 3 this means that even lower energies must be included. Although ranges of such low energy electrons are not known with great certainty in any medium estimates of the range of a 300 eV electron in the system of interest 4 5 give values of less than 100 A. In any case it appears that charge separation can result from delta rays which have so little energy that they are essentially spurs themselves rather than branches in the track. The mechanism by which this separation is effected must be con-sidered as not understood at present.It has generally been assumed that the energy loss mechanism for low-energy electrons i.e. in the neighbourhood of an electron volt would occur through ex-citation of single vibrational quanta at a time. Recently it has been discovered that low energy electrons can excite several quanta at a time (up to seven in gaseous Nz) through the transient formation of negative ion states.6.7 In many molecular liquids it is likely that such low-energy negative ion states are available and that this mechanism is operative. On the other hand the low-energy levels of negative-ion states are expected to vary greatly from one molecule to another. It is likely that rather great variations in the G value of charge separations can be found for various molecular liquids.Dr. G. R. Freeman (University of Alberta) (partly communicated) By numerical and graphical analysis of the delta-ray spectra in Lea,8 I estimated that to obtain a value of G(free ions) = 0-2 each delta-ray with an energy greater than 30 or 40 eV would have to result in one pair of free ions in a liquid hydrocarbon.9 However, in the analysis I had considered only the delta-rays generated by the primary elec-trons. By including the second generation of delta-rays (i.e. those produced by the primary delta-rays) only electrons with energy greater than 60-70 eV would be required to escape their parent spurs. It is much more difficult to include the third and fourth generations of delta-rays in the calculations but if this were done the minimum energy of escape would probably be about 100eV.This still seems to be too low according to conventional ideas about the ranges of electrons in matter.* I have made crude calculations from another point of view. If it is assumed G(tota1 ionization) = 3 in liquids one can use a Boltzmann factor 10s 11 to calculate a sort of average initial separation distance y for the thermalized electron-ion pairs. The fraction of ions that escape initial recombination is given by G (free ions)/G (total ionization) = 0.2/3 = exp( - r/y), where r is the distance at which the electron-ion coulombic interaction energy equals kT. In cyclohexane at 22"C r = 280& so y = l0OA. Alternatively Hamill et aZ.12 report that an electron ejected during irradiation of an organic glass at 77°K encounters about 103 molecules before recombining with its positive hole.If the 1 Magee Ann. Rev. Physic. Chem. 1961 12 389. 2 Charlesby this Discussion. 4 Lea Actions of Radiations on Living Cells (The MacMillan Company New York 1947). 5 Seitz Physics of Fluids 1958 1 1. 6 Schultz Physic. Reu. 1959 116 1141. 7 Chen and Magee J. Chem. Physics 1962,36 1407. 8 Lea Actions of Radiations on Living Cells 2nd ed. (Cambridge University Press 1955). 9 Freeman J. Chem. Physics 1963 39 988. 10 Onsager Physic. Reu. 1938 54 554. 11 Freeman J. Chem. Physics 1963 39 1580. 12 Hamill Guarino Ronayne and Ward this Discussion. 3 Freeman J. Chem. Physics 1963 39 988 1580. 6 Schultz Physic. Rev. 1959 116 1141 GENERAL DISCUSSION 249 electron undergoes a more or less random walk during these encounters the maximum separation distance attained by the electron-ion pair during the process i.e.the distance at which the electron approaches thermal energy would be about 102A ( z 3 x (lOy x molecular diameter). These calculations are all very crude and might be without significance but two completely different and independent calculations each lead to an estimate of about 100 A for the average initial separation distance of the thermalized electron-ion pairs. The value is very large and if it is valid I do not understand how such separations are attained by so many electron-ion pairs. Prof. F. S . Rowland (University of Kansas) said Dr. Schmidt-Bleek and Dr. Koyama and I have made some studies of the radiolysis of methane which tend to be complementary to those reported by Dr.Hummel. We have studied the reactions affecting unsaturates present during radiolysis. These experiments have involved CH4 radiolysis in the presence of trace quantities of CHz=C14H2 or HC=C14H. Through radio gas chromatography we can then follow alterations in the chemical identity of the C14 activity. Several of the products can be simply interpreted in the following way hydrogen atoms from the radiolysis of methane do not react readily with the parent compound and hence react very preferentially with olefinic traces and convert the ethylene-Cl4 to ethyl radicals. These labelled ethyl radicals then undergo combination reactions with other radicals present in the system from radiolysis. The chief labelled products formed are propane-Cl4 and n-butane-Cl4, indicating the presence of CH3 and C2Hs radicals in the steady-state radiolysis system.The former arises through radiolysis of the CH4 while the ethyl radicals are formed predominantly from H atom addition to unlabelled ethylene. Ethane-CI4 is also found and in larger quantities than can be expected from known dispropor-tionation/combination ratios. The excess indicates another source of labelled ethane. The subsequent comment by Dr. Ausloos on ionic reactions perhaps identifies the source of this " excess " C14-ethane. Small yields of (214-cyclopropane are also observed increasing linearly with dose as is the case for the labelled alkanes. The source of the cyclopropane is presumably the reaction of radiolytic methylene with ethylene-Cl4 : CH2 CH2-C14H2.CH2 + CHz=C14H2-+ A We looked also for the propylene-C14 expected if the cyclopropane-C14 had under-gone isomerization. However propylene is even more sensitive to H atom attack than ethylene and its yield would not increase linearly with time and is too low to measure accurately. After continued irradiation the labelled ethylene completely disappears and the amount of activity found as n-butane-Cl4 begins to decrease. This result indicates essentially complete removal of olefinic compounds such that the radiolytic H atoms then can react effectively with secondary C-H positions in the accumulated products. Similar experiments with HC = C14H showed some formation of propyl-ene-Cl4 the methyl-vinyl-Cl4 combination product in analogy with the ethylene-Cl4 experiments.The inclusion of 0 2 during radiolysis shows two regions of effects. At low 0 2 concentrations the H atoms continue to attack the ethylene-Cl4 but the resulting ethyl radical reacts with 0 2 and the radioactivity is diverted into oxygenated products. At higher 0 2 concentrations the H atoms themselves react with 0 2 and the olefinic compounds are not removed 250 GENERAL DISCUSSION The inclusion of 0 2 and HCr C14H during CH4 radiolysis shows two interesting effects. First the most prominent new labelled molecule is C.140 indicating a com-plex oxidation of HC = C14H. Secondly the trace quantities of acetylene apparently act as a catalyst for the reaction of CH4 with 0 2 for the 0 2 disappears much more rapidly than in the absence of acetylene.The effect is catalytic since the removal of 0 2 occurs much more rapidly than the oxidation of acetylene on a total number of molecules basis. This combination by acetylene of catalytic effect on hydro-carbon oxidation plus slow oxidation itself has many similarities with the chemical behaviour observed in hydrocarbon combustion. Dr. G. R. Freeman (University of Alberta) said It appears from various state-ments in the paper by Dr. Allen and Dr. Hummel that my note in the J. Chem. Physics 1 was misunderstood. The graph in the note was of conductance against voltage and they seem to have interpreted it as current against voltage. I would like to present further work that will perhaps help Dr . Allen and Dr. Hummel feel less dubious about the physical significance of the value of G(free ions) that appeared there.0 10 20 0 10 20 30 kV FIG. 1.-Current against voltage as a function of dose rate. P dose rate (eV/ml sec) t("C) A 6 1-20 x 1015 21 B 7 1-03 x 1014 26 C 8 1 . 6 6 ~ 1013 26 D 9 8.9 x 1011 24 All the results in fig. 1-4 were obtained with cyclohexane in a cell with platinum parallel plate electrodes. The electrode areas were 7-1 crn2 and the distance between them was 0.43 cm. In fig. 1 is shown a series of current against voltage (across the cell) curves obtained at dose rates that vary over a range of three orders of magnitude. The dose rates used by Allen and Hummel were rnosily lower than my 1 Freeman J. Chern. Physics 1963 38 1022 GENERAL DISCUSSION 25 1 FIG. 2.-Conductance against voltage as a function of dose rate.dose rate P (eV/ml sec) A 11 1 . 2 0 ~ 1015 B 12 1.03 x 1014 C 12 1.66 x 1013 D 13 8.9 X 1011 20 18 16 14 12 34 32 12 10 8 6 20 16 12 8 4 0 5 10 15 20 25 kV FIG. 3.-Current against voltage at low voltages. dose rate = 1-20 x 1015 eV/ml sec 2 4 6 8 10 252 GENERAL DISCUSSION lowest by about an order of magnitude so their current against voltage plots would show even more curvature than that in fig. lD if measured over the same voltage range. Fig. 2 shows the conductance against voltage curves that correspond to the results in fig. 1. The changing shape of the curves with changing dose rate is partly due to the competition between the second-order neutralization of the ions in the body of the liquid and the first-order removal of the ions at the electrodes.The positive slope 1 10 100 1000 l0,OOO volts FIG. 4.-Conductance against log (voltage). dose rate = 1.20 x 1015 eV/ml sec of the conductance curve at high dose rates (fig. 2A) is due to the fact that the number of electrons that escape their parent spurs increases with increasing field strength when the field is greater than N 1 kV/cm. Onsager 1 derived an equation to explain the effect of electric field strength on the amount of initial recombination of ion pairs generated in high pressure gases by P-rays or X-rays. I have extended the equation to explain the effects in liquid hydrocarbons.2 A plot of current against voltage at fields of only a few V/cm is shown in fig.3 for the dose rate 1-20 x 1015 eV/ml sec. At this high dose rate the curve is linear up to several 1CO V. The intercept of 0.5 V was also obtained in the absence OF radiation when the irradiation cell was replaced by a 2 x 1010 ohm resistor. It appeared to be generated by the Beckman micromicroammeter on certain sensitivity settings when it was shorted across high resistances. The metal-to-metal connec-tions may also have contributed to the effect. A similar intercept was observed by Allen and Hummel and it might have had similar causes. If the conductance of cyclohexane at a dose rate of 1-20 x 101s eV/ml sec is calculated from the expression C = amps/(vclts-0-5) to correct for the voltage intercept in fig. 3 the conductance against voltage curve in fig.4 is obtained. I have determined the value of G(free ions) in a number of hydrocarbons whose viscosities varied over a factor of 540.3 All give G values of about 0.1 or 0-2 at zero electric field strength. 1 Onsager Physic. Rev. 1938,54,554. 2 Freeman J. Chern. Physics 1963 39 O00. 3 Freeman J. Chem. Physics 1963,39 988. Cyclohexane and n-hexane each gave values of 0.2 GENERAL DISCUSSION 253 Dr. G. R. Freeman (University of Alberta) said In answer to Prof. Burton’s question about whether the currents were due to electrode effects the currents are not due to electrode effects because I have obtained the same calculated values of G(free ions) using parallel-plate electrodes of either platinum or mild steel with different electrode dimensions and separation distances and also using a nichrome cylinder with a tungsten wire running down its axis.Dr. A. 0. Allen (Brookhaven Nut. Lab.) said We have made some improve-ments in the measurements of ion mobilities and rate of disappearance of ions since submitting our manuscript. The experimental values of k / p about 1.4 x 106 V cm, have been found to be too high. An appreciable fraction of the ions were being removed by the electrodes. To minimize this effect (which is first-order in ion con-centration) Vc/Z was kept small and the initial ion concentration large. The new value of k/p for the recombination of the ions in hexane is l.O( k0.1) x 10- 6 V cm, which is in good agreement with the theoretical value (0.96 x 10-6). 5 10 15 sec FIG. 1. Mobility measurements have been made by the method of Gzowski in which the ions are produced in the solution by a narrow sheet of X-rays.The X-rays are delivered in a short pulse. After the pulse a field is applied to the cell and the current due to the movement of the ions in the field is measured. Superimposed on the ion current is another time-dependent current originating in the glass walls of the cell. The second current may be observed by applying a field to the cell in the absence of radiation as shown in the lower curve in the figure. With the radiation pulse we observe the middle curve. The difference between these two curves should be the ion current and is given as the upper curve in the figure. Thus the actual mobility curve is much closer to the expectations described in our paper.With prope 254 GENERAL DISCUSSION electrical shielding of the glass the extraneous current can be avoided ; however up to now mobilities have been measured while the effect was present. The new mobility experiments give ,D+ = 6 9 ( +0-6) x 10-4 and p- = 1.0(+0.2) x 10-3 cni2 V-1 sec-1. There is stiil some irreproducibiiity in the conductivity data. A careful study of this poirit is being carried out. From our data up to now we find 7i2/it= 3-O( k 1 -0) x 10-21Q-2 cm-2 A-1 which gives G = 0.09 + 50 %. Dr. 6. R. Freeman (University of Alberta) said It appears that the polymer-ization of cyclopentadiene is caused by the “ free ions ’’ that are generated in the liquid by the radiation. The average lifetime of the free ions calculated from my work in cyclohexane at the dose rates used by Williams and co-workers is of the order of 10-2 sec.This is an order of magnitude greater than the lifetime of the ions in cyclopentadiene estimated by Williams et al. One cannot conclude there-fore that ions have intrinsic lifetimes of much shorter duration in cyclohexane than in cyclopentadiene. Their calculated ion lifetimes in cyclohexane are only valid if ammonia is capable of scavenging 5 6 G units of ions in cyclohexane. I would like to ask the authors if they have found yields of I-ID in their more recent work that are higher than those reported in the table of their paper. Dr. Ff. Williams (University of Tennessee) said At the time our paper 1 was written we did not have access to the results obtained by Allen and Huinmel,2 and by Freeman.3 This Discussion now provides an opportunity to attempt some comparison between our findings from chemical studies and the new information which has been gained from physical measurements 29 3 on liquid hydrocarbons.Our reported yield of ionic chain initiators for the polymerization of cyclopenta-diene G = 0.20 is reasonably close to the G yield of separated ion pairs in other hydrocarbons viz. 0.09 +044 (Allen and Hummel) and cu. 0-2 (Freeman). More measurements have confirmed G = 0-20kO-07 and we estimate the uncertainty in the absolute value to be less than a factor of two. Now if free ions are responsible for the polymerization we must conclude that termination of the chains occurs mainly by impurities since the rate of poly-merization of the “ pure” nionorler depends on (dose rate)o-9 instead of (dose rate)o.5 as for bimolecular termination between oppositely charged ions.From the absence of any acceleration in the yield-dose plots to the gel point of the “ pure ” monomer we previously inferred 1 that diffusion-controlled termination whether by impurities or by combination of charged species was improbable. However, we know that an extremely mobile gel is formed at very low conversion of the ‘‘ pure ” monomer so in the absence of any data on the change in viscosity of the liquid within the gel we now think that the precise significance of the linear plots is debatable. Assuming free ions let us calculate the iinliting Go (-monomer) wnich would obtzin in the absence of any impurity. According to the usual stationary-state treatment, W,CC5I-I61 C (-monomer) = ( k2 Ri)+ ’ where l i 2 is the rate constant for charge combination and Ri is the rate of ionic initiation.Using the following values G = 0-2, (calculated from the paper of Allen and Humme12 using their revised value for the kz = 1.0 x 1012 1. mole-1 sec-1 1 Busler Martin and Williams this Discussion. 2 Allen and Hummel this Discussion. 3 Freeman J. Chem. Physics 1963 39 988 GENERAL DISCUSSION 255 sum of mobilities u = 17+3 x 10-4 cm2 V-1 sec-1 and k& = 0.96 x 10-6 V cm), Ri = 1.0 x 10-7 moles 1.-1 sec-1 (calculated from the highest dose rate we have used viz. 1.8 x 1018 eV ml-1 min-11, k = 3.1 x 106jGx = 1-5 x lo7 1. mole-1 sec-' (ref. (I)), we obtain Go (-monomer) = 1.2 x 105 which is higher than our observed value by a factor of six.Hence this quantitative application of the free-ion theory is also consistent with the predominance of impurity termination in our experimsnts. It is also of interest to calculate the concentration of an impurity [Y] which is necessary to scavenge half of the propagating ions; this is given by the relation where k is the rate constant for the reaction between Y and the propagating ion. Assuming k = lQ9 1. mole-1 sec-1 and the previous values for Xi and k2 [Y] is evaluated to be 3 x 10-7 M or a mole fraction of 2-3 x 10-8. Extrapolation of the results in fig. 3 of our paper 1 to the observed Go (-monomer) of 20,000 would in-dicate an impurity level equivalent to an ammonia mole fraction of 1.6 x 10-7 so we conclude that under these conditions about 90 % of the ions undergo termin-ation by impurities; then it follows that the mean lifetime z of the ions during propagatioii is about a factor of ten less than the lifetime with respect to charge recombination.In reply to Dr. Freeman the validity of our calculated ion lifetimes in cyclo-hexane is not critically dependent upon the maximum yield of scavengeable ions. For example if this maximum G yield of HD is only 0.56 the ion lifetimes would only be increased by a factor of ten to the range 10-7-10-8 sec. Rather it is the dependence of the yield on scavenger concentration which is the main factor in determining the ion lifetime. We have now embarked on a detailed study of this system. The results under discussion show a marked dependence of the HD yield upon ND3 concentration at about 10-2 M.Earlier we calculated that 3 x 10-6 M concentration of ammonia would scavenge 90 % of the free ions so it appears that some of the ions which are intercepted at the high concentration of scavenger in cycloliexane are not free ions. This conclusion does not seem surprising. In summatioil our results show that the ions responsible for the polymerization of cyclopentadiene are produced in low yield but they are long-lived ( N i0-3 sec) and probably free; whereas at least some of the scavengeable ions in cpclohexane are not free and have much shorter lives ( N 10-9 sec). These observations may provide a partial answer to the question raised by Dr. Samuel for it is important to distinguish between (i) the ionic polymerization characteristic of cyclopenta-diene,l isobutene,29 3 cr-methylstyrene,4 P-pinene,s isobutyl vinyl ether 6 with a kinetic chain length (k.c.1.) of 103-105 molecules and where the ionic lifetime is indisputably long ( N 10-3 sec) and (ii) the short chain polymerization of l-hexene,7 l-hexadecene (liquid),S which typically affords dimers trimers and tetramers.I am unaware of any experiments involving the effect of added ionic scavengers in low concentra-tion on this second class. To my mind there is no barrier to accepting the view 79 8 1 Busler Martin and Williams this Discussion. 2Davison Pinner and Worrall Proc. Roy. Soc. A 1959 252 187. 3 Collinson Dainton and Gillis J. Plzysic. Chem. 1959 63 909. 4 Best Bates and Williams Trans. Faraday SOC.1962 58 192. 5 Bates Best and Williams J. Chem. SOC. 1962 1531. 6 Bonin Calvert Miller and Williams submitted to J. Polymer Sci. B. 7 Chang Yang and Wagner J. Amer. Clzem. Soc. 1959 81 2060. 8 Collinson Dainton and Walker Trans. Faraday Soc. 1961 57 1732 256 GENERAL DISCUSSION that short chains can be propagated with high (G- 3) initiation efficiencies €or there is no reason why rapid condensation reactions between an ion and neighbouring molecules cannot precede neutralization ; the situation is analogous to that of high scavenger concentration as discussed above. On the other hand it seems that for case (i) the free-ion model provides the most satisfactory explanation to date. The reason for the absence of long chains in irradiated 1-olefins is probably related to a low kp value for these monomers since they are not easily polymerizable to high molecular weight products by conventional cationic initiation.Finally it seems that we need to know mwli more about the actual nature of the ions (carboniuni or ion-radical) which are responsible for these reactions. Prof. M. Magat (Faculte' des Sciences de Paris) said I would like to make two remarks concerning different points mentioned in the discussion. The first remark concerns the life-time of a " free " electron before it is recaptured by a positive ion. This life-time is widely different when the electron recombines with the ion from which it was abstracted (" parent ion ") and where it recombines with another positive ion present in the solution. It was established by Magee that in order to recombine with the parent ion the electron has to become thermalized i.e.its kinetic energy has to become of the order of kT when at a distance smaller than a critical distance re. NOW how far an electron travels before it is thermalized depends essentially on the rate of its energy loss and more specifically on the rate of energy loss of subexcitation electrons. The mechanism is more or less understood for energy loss due to the excitation of molecular vibrations. It is a relatively rapid process. The situation changes when the energy of the electron drops below the energy of the lowest vibrational frequency Le. below about 0-1 eV. Then for polar molecules the electron dissipates its energy inducing vibrations and rotation. It was shown by Frohlicli and Platzman that the rate of energy loss is related to the dielectric relaxation time and becomes much slower than in the higher energy ranges.It is this process which determines the time spent in thermalization and hence how far the electron travels before being therrnalized. But this theory is valid only for polar molecules. For non-polar media there exists for the moment no theory of energy loss by epithermal electrons and it is impossible to predict at what distance from the parent ion the electron will become thermalized. One could envisage the possibility that the electron loses its energy by orienting dipoles it induces in non-polar molecules particularly along the axis of the largest polarizability the mechanism being akin to the Kerr effect in non-polar substances but there is no theory of the process and no evaluation of even order of magnitudes is possible at present.The second remark concerns " collective excitation ". The notion of collective excitation was introduced if I remember correctly by Pines for the case of metals. This collective excitation was due essentially to the interaction of an individual electron with all free electrons of the metal. It was shown experimentally that losses did occur in some metals at energies predicted by the theory. Later on, Fano generalized the theory to include molecular crystals and by extension to liquids. However some limitations appeared. In order to have collective ex-citations in this case it is necessary to have low-lying excited levels of high oscillator strength.There are some experimental indications that collective excitation may occur in such compounds as benzene polystyrene and the like. But there are no indications whatsoever concerning saturated hydrocarbons. Until the existence of energy loss in saturated compounds due to collective excitation is shown by experi-ments on thin foils it is premature to try to explain everything by this mechanism as seems to be the general tendency GENERAL DISCUSSION 257 Dr. A. H. Samuel (Stanford Res. Inst. California) said Galimba Wilson and I 1 have irradiated solutions of stannic chloride in toluene at dry-ice temperatures in the presence and absence of ethylene and nitric oxide. We found yields of ortho- and meta- and/or para-chlorotoluenes which rose with SnC14 concentration to reach constant values at about 10 % SnC14.The total chlorotoluene yield at these stannic chloride concentrations approaches G = 0.2. Gas and polymer were observed but not measured. The yields were not affected by dose temperature (up to 25"C) and the presence or absence of ethylene and NO. We believe that we are observing an ionic reaction with SnC14 as a rather poor scavenger for the ions produced with G z 0.2 in irradiated toluene. When ethylene is present alkyl-ation yields are very small (G<0.01). This seems to indicate that the ions are not conventional carbonium ions. We offer this reaction as an additional piece of evidence that ionic yields in many organic liquids are small. It seems that the main question now is why some liquids (toluene n-hexene isobutene cyclopentadiene) have small ionic yields while others (1 -hexene 1-octene 1 -hexadecene) have large ones (G = 3).Dr. W. Wild (A.E.R.E. Harwell) said The arguments advanced by Dr. Williams for the choice of ammonia in his study apply equally to the water molecule with the possibility that its lower proton affinity will result in less efficient retardation by the processes described. Following on the work of Bates,z Mr. J. R. Brodie (1962) has re-examined the effect of water on the radiation-induced polymerization of cyclopentadiene. Water is less efficient than ammonia and its effect is given ac-curately by the equation, where Ro and R H ~ O are the uninhibited and retarded rates at the same radiation intensity. A similar behaviour was found earlier with a-methyl-styrene73 when the suggestion was made that the retarder prevented propagation by reaction with the initiating ion.Could Dr. Williams say whether he has revised his opinions about the effective reaction step? Have enough molecular weight data been obtained in the present study to support the adoption of the Mayo-Overberger relationship, which must need modification for low ammonia concentrations. All his experi-ments were done under conditions of marked inhibition i.e. G(-CsH6) < 500, whereas in our experiments this figure was always exceeded. This may explain why our polymers even those corresponding to cO.01 % conversion were insoluble in a variety of boiling solvents. This insolubility pre-vented measurement of the degree of polymerization and thus the possible con-tribution of chain transfer.The possibility does however remain that differing chemistry may affect the form of the rate equations and the properties of the re-sultant polymers when using different inhibitors. The equivocal answers given by infra-red and n.m.r. studies of the products contribute little information on this question. Dr. Ff. Williams (University of Tennessee) said It is of interest to compare the results presented by Dr. Wild 4 on the retardation of the radiation-induced polymerization of cyclopentadiene by added water with the effect of ainmonia re-ported in our paper.5 Two salient facts emerge from this comparison (i) in the concentration range up to 0.036 M water is a much less efficient retarder per 1 to be published ; work supported by Aeronautical Systems Division.2 Bates private communication. 3 Best Bates and Williams Trans. F'araday Soc. 1962 58 192. 4 Wild this Discussion. 5 Busler Martin and Williams this Discussion 258 GENERAL DISCUSSION molecule than ammonia; (ii) different quantitative relations appear to hold for the dependence of polymerization rate on retarder concentration in the two cases. Dr. Wild's results suggest a close parallel with previous work on the effect of water on the polymerization of a-methylstyrene.1 These differences may be explained by two considerations. First termination by exothermic proton transfer from a carbonium ion could well require the particip-ation of more than a single water molecule. We discussed this point in detail when we first reported 2 the effect of water on radiation-induced ionic polymerization.On the other hand monomeric ammonia would be expected to function as a more efficient base than monomeric water since the proton affinity of ammonia exceeds that of water by some 20 kcal/mole. Secondly whilst the coincidence in the ultra-violet absorption spectrum of ammonia vapour and of 0.02 M solutions of ainmonia in n-hexane suggests the absence of appreciable association for dissolved ammonia in organic solvents,3 there is evidence4 that dissolved water forms associated or " polymeric '' aggregates in hydrocarbons such as benzene and toluene and that the fraction present as aggregated water-relative to the total water content-increases markedly as the overall water concentration is increased.Under these circumstances it is easy to envisage that the effective concentration of retarder is the sun1 of a power series in water concentration so that the general form of the results is consistent with the proposition that oligomers of water terminate the propagating ion. In the light of this possible explanation the failure of the a-methystyrene + water system to obey a conventional kinetic description can be understood and this calls into question the alternative model which has been advanced 1 to explain the experimental results. However for the a-methylstyrene system it should be emphasized that a clear-cut decision between rival explanations is rendered particularly difficult by the occurrence of spontaneous chain transfer from the propagating ion to the monomer; in these circumstances the molecular weight of the polymer is insensitive to the concentration of water.On the other hand spontaneous chain transfer does not seem to be a complicat-ing factor in the polymerization of cyclopentadiene and our recent work indicates that the molecular weight of the polymer approximates to a linear dependence on the reciprocal of the ammonia concentration which accords with the kinetic description given in our paper.5 From fig. 3 we see that it represents only a small contribution (-0.8 %) to the G yield for the " pure " monomer. At high concentration of ammonia (> 10-3 M) the results are consistent with the idea that ammonia not only suppresses the growth of long chains but also curbs the growth of short chains which would otherwise (at low [NH3]) undergo independent termination by some undefined process.Again, it seems necessary to invoke at least two types of ionic processes which are kinetically distinguishable.6 Prof. F. S. Dainton (University of Leeds) said I agree with Dr. Williams that the retarding action of ammonia on the polymerization of cyclopentadiene is strong evidence of the cationic nature of the polymerization. There is a corollary to this conclusion namely that ammonia should be a good solvent for radiation-induced anionic polymerization. Dr. T. Skwarski has recently shown this to be the case. The significance of the empirical Gc(-C5H6) is not immediately obvious. 1 Best Bates and Williams Trans. Faraday SOC. 1962 58 192.2 Bates Best and Williams Nature 1960 188 469. 3 Stevenson Coppinger and Forbes J. Amer. Chem Soc. 1961 83 4350. 4 Gordon Hope Loan and Roe Proc. Roy. SOC. A 1960 258,215. 5 Busler Martin and Williams this Discussion. 6 cf. Williams this Discussion GENERAL DISCUSSION 259 At - 60°C free radical polymerization of vinyl compounds in ammoniacal solution does not occur but dissolved acrylamide which cannot polymerize cationally is rapidly polymerized by y-irradiation. Moreover the kinetic character istics of this reaction exclude the possibility of a free-radical mechanism. From other evidence we know the primary processes to be summarized by (1) and we conclude that the initiating species is the solvated electron and that neither NH3 not NH; is sufficiently acidic to sur-render a proton to the growing polymer anion.Dr. Ff. Williams (University of Tennessee) (communicated) The remarks by Prof. Dainton on the radiation-induced anionic polymerization of acrylamide in ammonia are of interest in connection with some of our recent work on isobutyl vinyl ether. In this latter case we have reason to believe that the situation is the converse of that described by Prof. Dainton in that the propagating entity is a protonated molecule while the inactive negative species is a solvated electron. Our evidence for this interpretation is based upon (i) the retarding effect of ammonia, and (ii) the formation of a blue colour 1 when the monomer is irradiated in the glassy state at -196°C. Further we find that the rate of polymerization Rp is propor-tional to (dose rate)o.Gs and from this we infer that there is a substantial degree of bimolecular termination between unlike charges.Other studies of radiation-induced ionic polymerization in liquids have generally found Rp to depend approx-imately on the first power of dose rate but this could be due to termination by impurities2 Since our results for isobutyl vinyl ether suggest that ion-pair separ-ation may be a prerequisite to polymerization I should like to ask Prof. Dainton whether he has come to any definite conclusion about the nature of the termination step in the acrylamide-ammonia system. Prof. F. S. Dainton (Lee& University) (communicated) In reply to Dr. Ffrancon-Williams the kinetics of the polymerization of acrylainide in liquid ammonia at - 60°C indicate that the most probable termination reaction is either ml+ NHt; Pj+NH3 or mT+NHz+Pj+x- where X- is unknown and may be NH-.TO satisfy the observed kinetics NH,+ and/or NH2 both of which are derived from the parent ion NH; must be separated from m?. An interesting corollary to this conclusion is that addition of NH; ion by diminishing [NH,+] and converting NH2 into the less efficient terminating species NH- causes a marked increase in the rate of the radiation-induced reaction. Prof. M. Burton (University of Notre Dame) said The comment to which Dr. Williams refers was evoked by his paper but the criticism was not intended to be limited to that specific case. In general whenever an additive is added to a con-densed system there is the possibility that the additive may change the site of the favoured localization of energy.Of course one is repeatedly confronted with the fact that this method of determination of what is going on may be the best avail-able at the moment. However such a method of testing a mechanism is not ideal. A much superior method for studying mechanism is that involving e.g. e.s.r. or light-absorption techniques. In such case it would be hoped that the device em-ployed for the observation of intermediates would not affect the course of reaction. More importantly it appears that such devices would not affect the site of localiz-ation of energy. My comment is addressed particularly to that point. I certainly have not attempted to dictate restrictions which should be applied to experimental work ; I have attempted to indicate ideal situations toward which experimentalists should strive.N H ~ ~ ~ ~ ~ - s NH,(+NH,+ i % N&b (1) 1 cf. Hamill Guarino Ronayne and Ward this Discussion. 2 Williams this Discussion 260 GENERAL DISCUSSION Dr. F f . Williams (University of Tennessee) (partly communicated) In relation to our paper,l I find it difficult to accept the validity of Prof. Burton's criticisms2 about the use of additives. It is well recognized that many " free-radical scavengers " can react by a variety of paths in radiation-induced reactions,3 but for reasons already given in our papers,ls4 the chemistry of ammonia in cyclopentadiene and cyclo-hexane suggests little ambiguity. Since the cationic polymerization of pure cyclo-pentadiene can be regarded as an example of a general acid-base reaction the effect of ammonia and other amines serves only to demonstrate the acidic nature of the intermediate species in the reaction.Because the recognition of an acid requires the participation of a base the usefulness of such chemical reasoning is lost if we insist upon the rigid exclusion of additives. In general it seems unduly restrictive to consider the chemistry of intermediate species without recourse to the informa-tion which can be gained by examining the course of birnolecular processes with well-ch osen additives. Dr. W. Van Dusen (Sandia Lab. Albuquerque New Mexico) said Recent measure-ments in this laboratory on the HD yields in the y-radiolysis of the liquid cyclo-hexane + heavy water system suggests the possible occurrence of a proton transfer reaction (1) analogous to the one observed by Williams 5 for the liquid cyclohexane + ammonia system.The heavy-water solubility in cyclohexane is less than 1 mole % but to insure maxi-mum mixing of the liquid layers the heavy-water concentration has been varied from 1-35 mole %. G(HD) varied from 0.1 1 to 0.20 the HD yield being equal to the increase of total hydrogen over the value for pure cyclohexane. For reaction (1) to occur it is necessary for the proton afhity of DzO to be equal or greater than 167 kcallmole assuming that the proton affinity of DzO is similar to that of H20.6 Tal'roze and Frankevich7 have observed that the H30+ ion of water produced in the mass spectrometer is enhanced by the addition of am-monia to the system.From this result it appears reasonable that the proton affinity of water will be similar and that reaction (1) is exothermic. Neutralization of HD2O+ would lead to formation of H and D atoms probably in a 1 2 ratio, followed by hydrogen atom abstraction from cyclohexane. The above arguments are based on gas-phase results but should hold for the liquid phase. One might inquire why the yield of HD does not increase with increasing amounts of DzO. This may be due to the combination of two reasons (a) although the solub-ility of D20 in cyclohexane is limited the maximum G value of 0-2 obtained for the HD yield may represent a maximum attainable even if the D20 solubility were greater and (b) the nature of the ionic species involved.Magee 8 has predicted that the yield of free electrons produced by delta rays should approximate GxO.2. This should produce a corresponding yield of free cyclohexane ions which should have a longer life-time than a cyclohexane ion undergoing geminate recombination with its electron a condition which should favour ion-molecule reactions such as (1). C-C,H;~ +D2O-+c-C,HI1 +HD,O+. (1) * This work was performed under the auspices of the United States Atomic Energy Commission. 1 Busler Martin and William this Discussion. 2 Burton this Discussion. 3 Williams Quart. Reu. 1963,17 101, 4 Bonin Busler and Williams J. Amer. Chem. Soc. 1962 84,4355. 5 Busler Martin and Williams this Discussion. 6 Lampe and Field Tetrahedron 1959 7 189. 7 Tal'roze and Frankevich Dokl.akad. Nuuk S.S.S.R. 1956 111 376. 8 Magee Ann. Rev. Physic. Chem. 1961,12 389 GENERAL DISCUSSION 26 1 Allen 1 and Freeman 2 have obtained G values of -0.2 for ions in y-irradiated n-hexane. It is interesting to note the similarity of yields obtained by the above-mentioned workers.3.19 2 Prof. F. S. Dainton (University of Leeds) said Cundall and Griffiths draw the conclusion from the fact that ferric chloride has no effect that the isomerization in benzene is not a free-radical process. This argument seems to me to be invalid since the concentrations of ferric chloride used (see fig. 3 of their paper) were never greater than about 7 % of the butene-2 and there is ample evidence to show that under these conditions any radicals formed would react with the butene rather than the ferric chloride.Dr. R. B. CundalI (University of Nottingham) said In reply to Prof. Dainton in our system radicals produced by the radiolysis of the solvent can react by R+C4&-+C4&R (1) or R + C4H8 + RH + C4H7 (2) R-I-C6H6+RH+C6H5 (3) FeC13 + R+FeC12 etc. (4) R+R-+R2 (5) or R + Rjolefine + paraffin. (6) and in the presence of ferric chloride Isomerization might result from reaction (6) or as a result of hydrogen abstraction by the methallylic radical or by the reverse of reaction (1). The last two possibilities can be excluded since they will be too slow to compete with ( 5 ) and (6). From the data of Buckley and Szwarc4 and those of James and Steacie 5 we can calculate that the rate constants for reactions (1) and (2) must be about 40 1.mole-1 sec-1 for cis-2-butene (abstraction and addition occurs with almost equal probability). The recombination-disproportionation rate constants k5 + kg are probably about 1010 1. mole-1 sec-1.6 The rate constant for sec-butyl radical attack on ferric chloride is probably about 6.3 x 103 as can be deduced from the results of Bamford Jenkins and Johnston 7 for polymerization initiator radical removal. The steady-state concentration of alkyl radicals can be calculated (assum-ing G (radical) in benzene is 2 and that H atoms are rapidly scavenged by the olefine and benzene) to be 1-9 x 10-9 mole 1.-1. Under these conditions the relative rates of ((I) + (2)) ((5) + (6)) and (4) would be 16 19 and 34.7 and ferric chloride would be expected to have a marked effect.None was found though there was a considerable reduction in the amounts of side products such as n-butane and butene-1. A recent paper by Kuntz and Mains 8 re-examines the mercury-sensitized photo-lysis of liquid alkanes which we refer to in our paper in connection with the possi-bility of the disproportionation reaction (6). They have shown unlike the earlier work of Phibbs and Darwent9 that disproportionation occurs with about 40 % 1 Hummel and Allen this Discussion. 2 Freeman J. Chem. Physics 1963 38 1022. 3 Busler Martin and Williams this Discussion. 4 Buckley and Szwarc Proc. Roy. SOC. A 1957,240 396. 5 James and Steacie Proc. Roy. SOC. A 1958,244,289 297. 6 Fessenden and Schuler J. Chem. Physics 1960 33 935. 7 Bamford Jenkins and Johnston Trans.Faraday Sac. 1962 58, 8 Kuntz and Mains J. Amer. Chem. SOC. 1963 85 2219. 9 Phibbs and Darwent J. Chern. Physics 1950,18,679. 1212 262 GENERAL DISCUSSION probability. This does not mean however that the isomerization in our system results to any appreciable extent in this way. Mr. Griffiths has carried out experiments in which benzoyl peroxide and azo-bis-isobutyronitrile were thermally decomposed in benzene + butene solutions at concentrations of 10-3 M and 75°C. Some isomerization occurred but only to an extent comparable in amount with the other side products expected (and found) from radical attack. Isornerization under these conditions was no greater than 10 %, of that found under comparable radiolysis conditions. This agrees with the fact that Gkom.in paraffn solution is less than in benzene in spite of the higher free radical yields in radiolysis. Dr. R. G. Kaufman (University of Notre Dame) said The results of Cundall and Griffiths in fig. 5 show that all curves appear to have about the same slope, i.e. k' is the same in each case. I would suggest rearranging their equation to 1 1 1 k M + k 1 G,+G,( k )m9 where Go is the maximum value of G(isom) and identical to k'. Go is obtained by graphing l/G(isom) against 1/[C4&] and extrapolating from data at low con-centration which is preferable to evaluating Go (or k') from the highest measured concentration. Using the data of fig. 5 yields a common value for all additives, Go = 2.5. Go(c+t) = Go(t+c) suggest an equal probability in forming either isofiier from the activated state and therefore a G for the active species in benzene of 5.0.Because most of the additives quench singlets as well as triplets let us consider the analogous photochemical reaction where R is [C4H8] and M is quencher : h v B+1B 1B+B 1B+3B 3B + R-+B + P 3B+M+B+M *B+M+B+M 3B+B which gives 1 1 1 k'M+k' 1 - -+-@-a a( 4 k ; ')i? where If the state of benzene were significant in radiolysis by analogy Go should also change upon addition of a singlet quencher. An approximate upper limit for concentration may be calculated ; k; + k; N 5 x 107 sec-1,1 (k&; + k;) = 0.322 k; for anthracene (est.) 1010 1. mole-1 sec-1 est. maximum error in Go about 20 %. Solution of the equations x + y = 1 and ( l / d ) x + ( l / a ) y = 1*2(l/a') where a' = (kL/ki+ki) gives 1.8 % as the maximum contribution of 1BzP benzene or based on a G value of 5-0 for the active species the G for 1Bzu in the radiolysis of benzene is less than 0-1.It also follows that the lBzu state is not a precursor of the active species as in photolysis. a = ki/(k; + k; + k;M). 1 BurtonIand Dreeskamp Elektrochem. 1960 64 165 GENERAL DISCUSSION 263 A second consideration for the active species is the 3BIU state. The measured value of the ratio (kk/ki) for cis-2-butene is less than 10-3 moles/l. under conditions where there is no light intensity effect.1 The analogous ratio in radiolysis (k4/k2) from the data of Cundall and Griffiths is 41-5 x 10-3 mole/l. and over a 25-fold change independent of dose rate. Identical results were also obtained at 2-1 x 1021 eV/l.mole (8 times their highest dose rate). There is therefore a marked difference between the 3Blu state and the active species in radiolysis. An insignificant dose rate effect over a factor of 200 does not preclude possible spur reactions such as triplet-triplet quenching which would shorten the apparent lifetime of a triplet species. Nevertheless the difference is large enough to suggest the active species is not the 3Blu state. These experiments have not eliminated charged species which could contribute to isomerization. By way of speculation I would suggest the possibility that the active species could be a non-optical excited state. There is no fundamental reason why electron excitation of liquid benzene which could be the mechanism in radio-lysis should give only optical states.Dr. R. B. Cundall (University of Nuttingham) said Mr. Grifiths and I have also examined the isoinerization of 2-butene in benzene and obtained results similar to those quoted by Dr. Kaufman. They do however support our contention that the isomerization of butene in both the radiation and photochemical systems is brought about by the same mechanism viz. triplet-triplet interchange. It is the close analogy between the two systems which lead us to our interpretation of the energy transfer process. We have carried out experiments on the photosensitized rearrangement over the same butene concentration range as used in the radiation system. Plots of G and rate values for the two systems have the same form and plots of G-1 and (rate)-1 against (conc.of butene)-l have the same slope viz. k4k2 (on our designation), within the limits of experimental error. The slopes determined are 5.9 x 10-2 and 6.6 x 10-2 in the radiation and photochemical systems. If as seems likely process (2) occurs at every collision then kq should be in the range k4x 1.2 x 108. This suggests a low value for the lifetime of the benzene triplet (10-8 sec) as compared with rigid glass and crystalline state phosphorescence measurement. Recent ex-periments of Dubois and Wilkinson 2 have given 10-8 sec as an upper limit for the lifetime of benzene triplet in hydrocarbon solution. The data of Dr. Kaufnian suggest a slightly longer lifetime but not comparable with that found in thc solid state. This value agrees more closely with the lower limit of 10-6 set by Lipsky 3 for the benzene triplet lifetime.Impurity quenching may be significant and explain the discrepancy between our results and Kaufman’s. It would be interesting to have data in the concentration dependency range of the radiation system with Dr. Kaufman’s solutions. It is evident that the lifetime of the 3Blu state benzene is profoundly affected by conditions peculiar to the liquid state. We have carried out fluorimetric observations on benzene + 2-butene mixtures in hexane solution and find that concentrations of butene-2 much in excess of the concentrations used in the radiolytic and photochemical work failed to quench the singlet state emission of the benzene. This shows that it is not able to react with or deactivate the lBzu state of benzene to any appreciable extent.This ob-servation provides a practical answer to the remark of Mr. Katsuura. The effect he describes might well be significant in the solid state but in solution where various 1 Kaufman and Hamill to be published. 2 Dubois and Wilkinson J. Chim. Physique 1963 38 2541. 3 Lipsky J. Chem. Physics 1963 38 000 264 GENERAL DISCUSSION other processes can occur it must be relegated to a very secondary role. Merrifield 1 also predicts that the ultra-violet absorption spectrum of the impurity should be modified by the impurity-exciton band interaction. This does not seem to occur under our conditions. Dr. J. G. Burr (North American Aviation Science Centre) said The importance of cis-butene-2 isomerization by transfer of positive charge from ionic intermediates to the butene is probably supported by the observations which we have made of the isomerization of cis-butene-2 during irradiation of solutions of this olefin in various carbinols.The data are shown in table 1 taken from a forthcoming pub-lication by F. C. Goodspeed and myself. The amount of isomerization is about the same in all of the alcohols and about the same as that observed in acetone and in benzene. Triplet states of the carbinols are not likely to be long enough lived to permit much sensitization by them of butene isomerization but it has been shown by Baxendale and co-workers that positive ionic species are very important in the radiolysis of alcohols. The probable importance of such ionic species for isomeriz-ation of butene in isopropanol is emphasized by our observation that the presence of a few per cent water in the isopropyl alcohol raises G(trans-butene) to high values, 8-1 5.TABLE 1 .-ISOMERIZATION OF CIS-BUTEM-2 IN IRRADIATED ALCOHOLS alcohol % butene G (cis-butene-ttrans-butene) methanol 0-5 0.94 ethanol 0-5 0-96 isopropanol 0.5 0.87 t-butanol 0.5 2.1 * acetone 0.5 0.64 benzene 0-36 0.8 5 * This alcohol may be a little wet. We also think that these observations in addition to the appropriate remarks made by Cundall about the possibility of multiple mechanisms for butene isomeriz-ation practically destroy the usefulness of butene isomerization as a detector of triplet states and populations in irradiated systems unless it is already firmly known that triplet states are important in the radiolysis.Dr. R. B. CundalI (University of Nottingham) said I agree with Dr. Burr’s interpretation of his results on the isomerization of 2-butene in the presence of carbinols. A negative ion mechanism is not to be disregarded under these con-ditions. I think that the butene isornerization can still provide a useful detector of triplet states in irradiated systems if appropriate tests are also carried out for ionic mechanisms along the lines which we attempted in our paper. Dr. M. A. Golub (Stanford Research Institute) said In connection with the paper by Cundall and Griffiths,Z I report some results obtained by Dr. Jeanne Danon 3 and myself on the radiation-induced isomerization of cis- and trans-polyisoprenes and squalene the latter being a hexaisoprene in which the internal double bonds are all in the trans configuration.The present work on the isoprenic molecules is a sequel to earlier studies on polybutadiene 4 which was found to isomerize with high yields either in benzene solution or in the solid state. We now can compare the isornerizability of the very small olefin butene-2 studied by the British workers, 1 Merrifield J. Chem. Physics 1963 38 920. 3 NATO Fellow at Stanford Research Institute 1962-3 from Laboratoire de Chimie Physique, 4 Golub J. Amer. Chem. SOC. 1960 82 5093; J. Physic. Chem. 1962 66 1202. 2 this Discussion. FacultC des Sciences Orsay (S. et O.) France GENERAL DISCUSSION 265 with that of the higher molecular weight squalene on the one hand and with that of the macromolecules polybutadiene and polyisoprene on the other.The pertinent G values for cis+trans and transjcis isomerization of these various compounds are listed below. approx. energy conc. in ~ 6 % G (isom) Gp (isom) G transfer factor = by weight solution pure compound G ‘/Gp- 1 cis-polybutadiene 1 % 0.66 8.0 66 trans-polybutadiene 1 % 0.33 4.0 33 cis-polyisoprene 1 % -1.0 10 100 trans-polyisoprene 1 % -1.0 10 100 cis-butene-2 - 1 % 2.2 14 (ref. (1)) 220 squalene 10 % 2.6 - 0-83 26 4-0 (ref. (2)) trans-butene-2 -1 % 1.8 4-1 (ref. (2)) 180 7 7 9 9 30 15 54 44 Since the G,(isom) values were based on the energy absorbed by the whole solu-tion we examine the corresponding Gi values calculated in terms of the energy absorbed only by the solute as though no intermolecular energy transfer occurred in the solutions.Since these G values are all significantly higher than the Gp values for the pure compounds there must be considerable net energy transfer from benzene molecules to the various olefinic solute molecules in order for the latter to isomerize with such a high effective yield in solution. As suggested previously,3 and now advanced also by Cundall and Griffiths,4 it is tempting to assume that a large part of the energy transfer from solvent to solute proceeds from the benzene triplet (3.6 eV) to the low-lying triplet state of the vinylene unit (3-2 eV). On an energy basis charge transfer as well as transfer between higher excited states might also be involved since benzene and the olefinic groups have similar ionization and excitation potentials.As a result of very rapid internal con-version processes affecting highly excited states whether formed directly or through neutralization of ions the precursor in the isomerization may again be a low-lying excited state of the olefinic group. Whatever the detailed mechanism for energy transfer in the benzene solutions of olefins it is clear that the latter compounds undergo an important benzene-sensitized isomerization over and above that charac-teristic of the pure compound. Moreover the various compounds mentioned with the exception of squalene isomerize in the pure state with roughly comparable yields while all of them display similarly high G values for isomerization in benzene.The energy transfer factors shown are not unreasonably large in view of the sensitized fluorescence studies of Kallman and Fur~t.5~6 At the moment we have no ex-planation for the observation that pure squalene isomerizes much less efficiently than any of the other compounds in the pure state. Nor can we account for pre-liminary results indicating that cis-4-methylpentene-2 isomerizes either not at all or at most very slightly.7 The G values given above for polybutadiene have been revised downwards from earlier published values 3 due to newer more accurate infra-red analysis of this material. On this point we note that in contrast to the situation with polybutadiene, the isomerization of polyisoprene and squalene is difficult to follow by means of 1 Cundall private communication.2 Kaufman J. Physic. Chem. 1963 67 1671. 3 Golub J. Amer. Chem. SOC. 1960 82 5093; J. Physic. Chem. 1962 66 1202. 4 this Discussion. 5 Kallmann and Furst Physic. Rev. 1950,79 857; 1951 81 853. 6 Furst and Kallmann Physic. Reu. 1954 94 503. 7 Golub unpublished results 266 GENERAL DISCUSSION infra-red spectroscopy; in our work with these compounds we made use of high resolution n.m.r. spectroscopy according to the technique described recently,1 which involves the measurement of the relative areas of the cis- and trans-methyl proton peaks. With regard to the specific excited states of benzene involved in the energy transfer process further study of the radiation-induced isomerization of polybutadiene in benzene solution and in the presence of certain additives whose lowest excited states are somewhat below that of the olefinic unit should be very informative.This sort of thing has already been considered by Cundall and Griffiths for the benzene+ butene system where anthracene and naphthalene e-g. were found to reduce the G(isom) value.2 Complementing that approach is a study now in progress on the vacuum photochemical isomerization of polybutadiene involving light quanta with energies below and above the ionization potential of the olefinic group it is hoped that such work will throw further light on the fate of both ionized and higher excited states in polybutadiene and hence also on the number of primary excitations per ionization in this polymer. Just as Cundall and Griffiths observed iodine to have no effect on the butene-2 isomerization in paraffin solutions and only a small inhibiting effect at high con-centrations in benzene I was unable to detect any iodine sensitization of the photo-chemical or radiation chemical isomerization of pqlybutadiene in benzene but instead at high iodine concentrations I even observed a similar reduction in iso-merization yield compared to the iodine-free solution.3 I agree with their conclusion that free radicals cannot be involved in the isomerization but rather excited and ionized states.Significantly with the partially deuterated polybutadiene, -CH2-CD=CD-CH2-, there was a negligible amount of double bond shift through a reaction such as -CH2-CD=CD-CH2-+-CH2-CD=CD-CH-~ -CH2-CD-CD = CH-Clearly if free-radical reactions were important in the radiation chemistry of pofy-butadiene a substantial amount of -CD=CH- units would be formed but this was not the case.4 Dr.C. H. Krauch (Max-Planck Inst. Mulheim) said It is surprising to me that Cundall and Griffiths did not observe dimeric products in reaction mixtures by careful analysis for we observed in the same systems stereoselective formation of cyclobutadimers from olefins. We irradiated solutions of cyclic olefins as indene or 1,2-dihydronaphthalene or dimethylmaleic anhydride in aromatic solvents and observed the formation of trans-head-to-head-cyclobutadimers with G-values from 0.5-3 depending on the system.5 Addition of aromatic carbonyl compounds such as benzophenone raised the G-values. We studied cyclic olefins but Holroyd also found some cyclobutane formation by irradiating ethylene.Because of the stereo-selectivity of the reaction we think that the mechanism goes via short-lived adduct biradicals with chemical selectivity of termination. Such a mechanism should also be considered for your cis-trans-isomerization. Dr. R. B. CundaU (University of Nottingham) said We did not identify any dimeric products in our reaction mixtures. Any dimer which may have been pro-duced could not have amounted to more than a small percentage (probably 2 or 3 % 1 Chen Anal. Chem. 1962,34 1793. 2 this Discussion. 4Golub J. Arner. Chern. Soc. 1960 82 5093; J. Physic. Chem. 1962 66 1202. 5 Krauch Metzner and Schenck Natnrwiss. 1963,50 210. 3 Golub unpublished results GENERAL DISCUSSION 267 at most) of the products.A similar finding was made in the photochemical system. I do not consider the dimerization reported by Dr. Krauch is in opposition to our findings; in fact the reverse. Similar results are found in photochemical systems and the possibility of dimerization depends upon the phase as well as the particular olefine examined. We find no evidence for the photosensitized dimerization of butadiene in the gas phase 1 whereas in solution vinyl cyclohexane and divinyl cyclobutane are produced.;! Cyclopentadiene dimerizes both in the gas phase and in solution. The ap-unsaturated ketones differ widely in behaviour. The most consistent explanation of these facts is to imagine excitation of the olefine by an energy transfer process which produces the triplet state as we have suggested.The work of Hammond and his collaborators 3 and e.s.r. studies provide strong support for this view.4 Dimerization in the irradiation of pure olefines is probably brought about by an ionic mechanism. We have evidence for octene formation in the irradiation of pure cis and trans-2-butene. Mr. K. Katsuura (Northwestern University) said Dr. Cundall said that the energy of the first singlet excited state of benzene cannot be transferred to butene because the energy of the first singlet excited state of butene is higher than that of benzene. Generally when the lowest excitation energy state of the impurity molecule is greater than that of the matrix molecule the presence of the impurity molecule will be responsible for the creation of an energy level in the crystal above that of the exciton band.However if the interaction between the impurity molecule and the molecules of the matrix is strong enough then another impurity level (singlet) may appear slightly below the level of the exciton band.5 These energy states of the crystal will be concentrated about the impurity molecules. When the electrons of the matrix molecules are raised to the energy levels of the exciton bands they will be in an energy state above that of the state created by the impurity molecule and they can then make a transition to the energy state of the crystal created by the presence of the impurity molecule. Dr. T. J. Hardwick (Gulf Res. and Development Co.) said In connection with effects of L.E.T. in hydrocarbons and other organic liquids it is necessary to con-sider the consequences of the large specific energy deposition at high L.E.T.The passage of an ionizing deuteron will result in an energy deposition of say 10 eV/a of track length and according to the Magee model in a track radius of about 20A. For a liquid such as n-hexane it can be calculated that this is equivalent to about 17 eV for each molecule in the track. There will be a distribution of energies among these track molecules. The energy deposition will occur in a time less than 10-14 sec, and attention is drawn to the subsequent steps of energy dissipation. Within about 10-12 sec charge neutralization will have occurred and through intermolecular col-lisions and internal conversions most of the energy will have been degraded to heat.The loss of energy from the track can occur in three ways (a) by photon emission, which may account for about 2 of energy loss (b) by exciton transfer which also accounts for a very small fraction of the energy loss and (c) by the transfer of kinetic energy from one molecule to another by collision. In this last case the maximum rate of energy dissipation will be equal to the velocity of sound which for n-hexane However we cannot reach this conclusion so easily. 1 Fletcher to be published. 2 Hammond Turro and Fischer J. Amer. Chem. SOC. 1961 83,4674. 3 Hammond and Hardham Pruc. Chem. Suc. 1963 63 and earlier papers. 4Brandon Gerkin and Hutchinson J. Chern. Physics 1962 37 447. Siege1 and Eisenthal, 5 Katsuura Pruc. Physic.SOC. Jcpan 1962 17 833. Katsuura and Inokuti to be published J. Chem. Physics 1963 38 2785. in J. Physic. SOC. Japan 1963. Merrifield J. Chem. Physics 1963 38 920 GENERAL DISCUSSION 269 10.4 electron % cyclohexane in cyclopentane (Gextrap. = G,b,,/electron fraction C-CgH12) we found 4.8 mole/100eV of cyclohexane. From 4.9 electron % cyclo-hexane in cyclopentane containing 10-2 M iodine G(C-C~H~O),,~~~P. was 4.2. On the basis of the agreement reported between the yields of C-C~D~O and D2 and despite quantitative differences in yields of corresponding products from c-CgH12 and C-CgD12 there is poor agreement between the extrapolated yields Gl(D) = 1-6 and G(c-C~HI~) = 4.2. On the other hand ~G(C-C~H~~),,,~~. Eg(HD) in agreement with eqn. (1) and (2). Dr. R.H. Schuler (Mellon Inst. Pittsburgh) said In order to make the record complete on the question of the existence of He atoms as an important intermediate in hydrocarbon radiolysis I would point out that in irradiated liquid methane hydrogen atoms are directly observed in spin resonance experiments.1 The signals due to these hydrogen atoms disappear upon addition of an olefin to the system and are replaced by those of the hydrogen atom adduct of the olefin. Although hydrogen atoms are not observed in other systems (presumably due to the inevitable presence of olefins in the e.s.r. experiments as they are carried out) the addition of C2D4 to the sample results in the appearance of the lines of CzD4H. in the spectrum. Dr. G. R. Freeman (University of Alberta) said Dyne Denhartog and Smith interpret their results on the basis of one primary active species and one “ protection ” reaction.There is a considerable amount of published evidence that the inhibition of the radiolysis of liquid hydrocarbons cannot be explained in this way using homogeneous solution kinetics.2-4 The shape of the curve in fig. 6 of the present paper by Dyne and co-workers is also consistent with this conclusion (GI@) is reduced much more rapidly than is G(tota1 hydrogen)). There are at least two possible explanations (i) there are two or more kinetically distinguishable activated species which lead to the formation of hydrogen or (ii) the inhibition reactions do not obey simple homogeneous kinetics. Does Dr. Dyne favour one or the other of these possibilities or has he another explanation ? Although there is considerable evidence for the existence in irradiated liquid hydrocarbons of some scavengeable species such as hydrogen atoms or electrons, the hydrogen atom idea has not yet been able to explain the results of Dyne and co-workers.Dr. P. Ausloos (National Bureau of Standards) said We agree with Dr. Dyne et al. that mixture-law deviations in the gas-phase radiolysis of hydrocarbons occur to a lesser extent than in the liquid phase. This has been demonstrated in earlier studies.5 For instance in the radiolysis of CgH12 + C3Dg and C(CH3)4 + C3Dg mixtures G(C2H4)/G(C2D4) and G(CH4)/G(CD4) were shown to be independent of the relative concentrations of the two compounds. Values of the ratio G(Hz)/G(D2) however do not follow in many cases the mixture law.This is not surprising if one considers that hydrogen is often a product in ion-molecule reactions. This could account for the results shown in fig. 9 in the paper by Dyne et al. Al-t validate Dr. Dyne’s C6D12 results obtained with low percentages of C6D12 as a rule it is more advisable to test the mixture law on products such as methane and ethylene which can in many cases be more un-ambiguously ascribed to unimolecular elimination processes. I would also like to point out that the differences between gas and liquid phase are not unexpected if one considers that as shown by the mass spectral patterns 1 Fessenden and Schuler J. Chem. Physics 1963 38 2147. 2 Freeman J. Chem. Physics 1960,33,71; 1962 36 1542. 3 MacLachlan J.Amer. Chem. SOC. 1960 82 1005. 4 Hardwick J. Physic. Chem. 1961 65 101 ; 1962 66 291. 5 Lias and Ausloos J. Chem. Physics 1962 37 877 270 GENERAL DISCUSSION obtained at 10-6 sec and the 10-10 sec fragmentation calculations of Vestal et al. referred to in our paper as well as by published experimental results from our laboratory,l in the gas-phase radiolysis of hydrocarbons parent ions constitute only a small fraction of the total number of ions present in the system while in the liquid phase fragmentation of parent ions does not occur to any great extent. It may thus be said that it is incorrect to question Hardwick‘s charge-transfer meclian-ism in liquid alkane mixtures on the basis of the fact that such processes are not observed in the gas phase. NEOPENTANE 2.05 0.46 1-35 0-57 1.0 0-45 1.40 0.81 0.32 0.42 1-35 0-76 1 -PENTENE 1.0 1-68 0.4 1 0.46 0-33 1.64 0.42 0.97 Dr.J. G. Burr (N. American Aviation Science Centre) said The detector system employed by Dyne et al. seems to show that the processes producing HD, those producing D2 and those producing hydrogen from the solvent are all affected equally by each of the four additives which were studied. The clear implication can be drawn from this work that either hydrogen does not arise from CsD12 or the solvents by hydrogen atom processes or that the additives employed -among them benzene and cyclohexene-do not react with hydrogen atoms from the C6H12 or the solvent. Therefore a clear-cut demonstration that hydrogen atoms are produced from cyclohexane and that they react with benzene or cyclohexene under conditions similar to those used by the authors should suggest a re-inter-pretation of the results from their system.Mere observation of products which could be formed via hydrogen atom process-such as cyclo-hexane from radiolysis of solutions of cyclohexene-14C in cyclohexane (under study by M. Cher in this laboratory) is not conclusive since such products could conceivably arise by reaction of an excited (triplet?) olefin molecule with another hydrocarbon molecule. We believe that we have achieved the desired demonstration by observing the hydrogenltritium isotope effect in the formation of “ addition ” products during the radiolysis of certain solutes dissolved in tritiated hydrocarbons. Reaction of an excited solutc molecule with a tritiated hydrocarbon would give products showing a “ normal ” isotope effect i.e.products with a lower specific tritium activity than the tritiated solvent but addition of hydrogen and tritium atoms to the solute could in principal show a “ reverse ” isotope effect. The existence of a “reverse” isotope effect for hydrogen atom additions to A clear-cut demonstration of this sort is difficult to achieve. 1 Lias and Ausloos J. Chern. Physics 1963 39 818. 2 Lias and Ausloos J. Chem. Physics 1962 37 877 GENERAL DISCUSSION 27 1 unsaturated hydrocarbon molecules is indicated in our study of the polymer pro-duced in the radiolysis of benzene- 14C-cyclohexane (tritiated) solutions. The tritium content in the polymerized benzene corresponds to an apparent 24 hydrogen atoms from cyclohexane incorporated in each benzene molecule.An isotope effect in favour of tritium atom addition must be involved. The high concentration of tritium in the benzenoid polymer mentioned above now appears to be the result of successive additions of two or three hydrogen or tritium atoms from the tritiated cyclohexane to the benzene or its polymer. The magnitude of the “reverse” effect has been established by measuring the tritium activity of the cyclohexane and cyclohexene produced in the radiolysis of cyclohexene + n-octane (tritiated) solutions. Formation of these products apparently begins with the addition of H or T atoms from the tritiated n-octane to the cyclo-hexene to give cyclohexyl radicals. If the normal hydrogen/tritium isotope effect in the hydrogen abstraction reactions of cyclohexyl radicals and n-octyl radicals is relatively large (kh/kt - lo) the specific tritiun 1 Lctivity in the cyclohexane will be about the same as that of the cyclohexyl radicals formed in the addition reaction.The specific activity in cyclohexane 1.25 mc/g atom hydrogen relative to that in n-octane 3.20 mc/g atom hydrogen then corresponds to a “ reverse ” isotope effect kt/kh-4. Measurement of the specific activity of the bicyclohexyl yet to be accomplished will give a more reliable estimate of the size of this isotope effect. The high concentration of tritium in the benzenoid polymer mentioned above now appears to be the result of successive additions of two or three hydrogen or tritium atoms from the tritiated cyclohexane to the benzene or its polymer.We think that this demonstration that hydrogen atoms presumably formed in the radiolysis of cyclohexane can add to dissolved benzene points up the intuitive feeling that many of us have about the system discussed by the authors-namely, that it needs to be connected in some way with the rest of radiation chemistry. The results obtained from it are too absolute and too conclusive; it does not seem reasonable that the additives studied should only transfer energy and never react with hydrogen atoms nor does it seem reasonable that no hydrogen atoms are formed in the radiolysis of cyclohexane. Prof. S. Lipsky (University of Minnesota) said Dr. Merklin and myself have recently studied the dependence of G(H2) both on dose and on benzene concentra-tions for liquid cyclohexane methylcyclohexane and ethylcyclohexane.For cyclo-hexane G(H2) exhibits only a slight dose dependence in agreement with earlier work. Yields at zero dose (extrapolated) and at 1.2 x 1022 eV/l. are respectively 5-83 and 5.82. For both methylcyclohexane and ethylcyclohexane however much more severe dependence on dose is found. With methylcyclohexane G(H2) is re-duced from an extrapolation value at zero dose of 6-2 to 5-85 at 1.2 x 1022 eV/l. For ethylcyclohexane these values are 6-1 (extrapolated) and 5.05 (at 1.2 x 1022 eV/l.). At higher doses G(H2) from methyl- and ethylcyclohexane becomes less sensitive to dose and in fact approaches the dose dependence exhibited by G(H2) fro M c yclohexane.We find similar differences between these cycloalkanes in their response to low concentrations of benzene. G(H2) from methyl- and ethylcyclohexane decreases sharply with the addition of traces of benzene whereas cyclohexane is much less affected. For example the addition of 0.05 M benzene decreases G(H2) by about 6.4 18 and 27 % for cyclohexane methylcyclohexane and ethylcyclohexane respec-tively. Above 0-05 M benzene G(H2) from methyl- and ethylcyclohexane decreases much more gradually and approaches the dependence on benzene concentration of G(H2) from cyclohexane 272 GENERAL DISCUSSION The behaviour of methylcyclohexane and ethylcyclohexane seems to be most plausibly explained by invoking thermal H atoms as precursors of a portion of the observed H2 yield and assuming that these are efficiently removed by the scavenging action of benzene and where dose dependence radiolytically produced olelins.The absence in cyclohexane of the marked dose dependence and sensitivity to benzene suggests accordingly that in this system neither olefins nor benzene protect by a scavenging mechanism. If thermal H atoms are produced& cyclohexane and are precursors of H2 it would be exceedingly unlikely that benzene would compete for these H atoms less efficiently with cyclohexane than with methyl- or ethyl-cyclohexane. We have therefore concluded that either thermal H atoms are not produced in the radiolysis of cyclohexane or if produced play an insignificant role in H2 production. Dr. A. Henglein (MeZZun Inst. Pittsburg) said The radiation chemistry of mixtures of cyclohexane and carbon tetrachloride has thoroughly been studied by Mr.Heckel in our laboratory.1 Some of his results appear of interest with respect to the question of chemical or physical action of carbon tetrachloride on the radio-lysis of cyclohexane. The amount of hydrogen formed in dilute solutions of cyclo-hexane in carbon tetrachloride is proportional to the electron fraction of cyclo-hexane. This shows that hydrogen is produced here by a unimolecular decomposition of C6H12 molecules. G(H2) is equal to 0.3 molecules per 100eV directly absorbed by the solute under these conditions where each C6H12 molecule is surrounded by CC14 molecules. If there is a physical protection of cyclohexane by carbon tetra-chloride it is certainly not efficient by 100 %.G(dicyclohexy1) and G(cyc1ohexane) both decrease with increasing CC4 concentration and level off at concentrations of a few mole per cent of CCl4. Since we could show that CC4 rapidly reacts with C6H11 radicals this effect may be explained by the scavenging of these radicals before they disproportionate or combine. However a ratio of the rate constants of disproportionation and combination of 2.0-2.5 is obtained from these experi-ments which seems to be much too high. It must therefore be concluded that the formation of dicyclohexyl and cyclohexane is partly affected by carbon tetrachloride in a mechanism where no cyclohexyl radicals are involved. G(HC1) at low concentrations of CCl4 was found to be somewhat higher than G(-H2).This indicates that HCl is not only produced by H-atom scavenging by carbon tetrachloride. A possible explanation of this phenomenon is the reaction of excited cyclohexane molecules with carbon tetrachloride which are deactivated in pure cyclohexane. All these results indicate that carbon tetrachloride strongly affects radical and H-atom reactions in cyclohexane although a certain interaction with the precursors of these radicals and atoms must be postulated too. Dr. T. J. Hardwick (Gulf Res. and Development Co.) said The results of Dyne’s paper are self-consistent and the conclusions drawn from them appeared justified. However unless the mechanism of cyclohexane radiolysis is unique among saturated hydrocarbons it is necessary that these conclusions be applied to the large body of experimental results on the radiolysis of liquid alkanes.On so doing it is difficult to account for many of the observed chemical syntheses without assuming the chemical reaction of a hydrogen atom with a scavenger. Two examples will suffice. (i) In the radiolysis of n-hexane containing small amounts of cyclohexene, Dewhurstz found cyclohexane among the products and we have confirmed this. Dewhurst suggested that hydrogen atoms add to cyclohexene to form cyclohexyl radical and that subsequently this radical forms cyclohexane by disproportionation with another alkyl radical. If as Dyne suggests the decrease in the radiolytic 2 Dewhurst J. Physic. Chem. 1958 58 15. 1 paper submitted to Ber GENERAL DISCUSSION 273 hydrogen gas yield on the addition of cyclohexene is the result of a physical process, it is difficult to write a mechanism for cyclohexane formation.(ii) Holroyd 1 has performed many experiments in which radioactive ethylene was added to an alkane solvent prior to radiolysis. The resulting product distribu-tion can be explained only if an intermediate ethyl radical is present resulting pre-sumably from the scavenging of a hydrogen atom by ethylene. It would follow from Dyne’s results that ethylene like cyclohexene decreases the hydrogen gas yield by a physical process and ethyl radicals should not be formed from dissolved ethylene. Dr. C. D. Wagner (Shell Development Co. Emeryville) said While the paper by Dyne Denhartog and Smith demonstrates that energy or charge transfer is im-portant in the radiolysis of alkane mixtures their conclusion that additives including cyclohexene reduce the total hydrogen yield exclusively by a physical mechanism, ignores published data on the role of hydrogen-atom scavenging.In addition, data obtained in our laboratory to be published shortly shows that 5 % I-pentene in n-butane reacts with a very high G-value (44 at a dose of 60 Mrads) to form a distribution of products that is nearly that to be expected by radical combination. The G-values for radicals combining appears to be Me = 0.16 Et = 0.53 n-Pr = 0.04 n-Bu = 0-66 S-Bu = 0436 I-pentyl = 0.10 2-pentyl = 0.55 3-pentyl = 0.0, pentenyl (1 -vinylpropyl) = 0-36. The principal C5 radicals formed (2-pentyl and pentenyl) are exactly those expected from addition of hydrogen and abstraction by hydrogen atoms.One can dismiss this result only by saying either that (i) the above Cs structures can also be formed from an energy-rich pentene molecule by abstraction or frag-mentation processes or (ii) if the 2-pentyl is formed by hydrogen atom addition, these hydrogen atoms somehow would not have formed hydrugen in the absence of the pentene. In our opinion both of these alternatives are unlikely. Dr. W. G. Burns (A. E. R.E. Harwell) (communicated) 1 In explaining the observed kinetics of the formation of D2 and HD from mixtures of n-hexane (H) cyclohexane (CH) and cyclohexane-dlz (CD). Dr. Dyne has suggested the existence of three simultaneous equilibria between H* C** and CD I should like to ask if all three equilibria are essential or weether an equilibrium between H* and CH* brought about through the intermediacy of CD.is all that is required: Such a postulate is incompatible with the observed results. products cH* +prodUcts /” % R H” CD* .1 D2 + HD. Dr. A. J. Swallow (Christie Hospital and Holt Radium Inst.) said It has been assumed by many speakers that energy absorbed from secondary electrons is par-titioned in a mixture according to the electron fraction of the components. How-ever it is important to distinguish between the electrons in inner shells and the valence electrons. Where an inner electron is excited the event is followed within about 10-15 sec by the emission of an Auger electron or for heavier atoms by a cascade of Auger electrons. Such electrons carry away most of the energy initially transferred to the atom or molecule.The Auger electrons can excite or ionize several valence electrons but they do not have sufficient energy to excite or ionize 1 Holroyd and Klein Int. J. Appl. Rad. Isotopes 1962 13 493 274 GENERAL DISCUSSION inner electrons like the ones to which they owed their origin. For oxygen e.g., the excitation of one inner electron would be followed by the emission of an Auger electron with an energy of about 500 eV and this would excite or ionize the valence electrons of about 20 more molecules. Much of the ionization and excitation in mixtures is caused by electrons whether Auger electrons or others with an energy of less than 100 eV 1 and these do not have sufficient energy to excite inner electrons.Bearing in mind these considerations we can see the limitations of the " electron fraction law". The energy absorbed from y-rays by Conipton scattering is par-titioned precisely according to electron fraction if by " energy absorbed " we mean energy transferred to the medium in the form of kinetic energy of Compton electrons. So far as the Compton electrons are concerned these partition their energy in the first instance according to electron fraction but if we take into account the Auger effect and low-energy secondary electrons then energy is partitioned in a homo-geneous mixture according to the valence electron fraction of the components. The general rule then is that energy is partitioned in homogeneous mixtures accord-ing to the fraction of valence electrons possessed by the various components.However, this rule is not necessarily obeyed by mixtures containing aromatic compounds.2 Prof. M. Magat (Faculte' des Sciences des Paris) (communicated) I entirely agree with Dr. Swallow that for mixtures the division of energy absorbed by each con-stituent according to their electron density has no theoretical basis whatsoever. But his proposal is also debatable. A theoretically satisfactory calculation could be performed only if all the values of frequencies and oscillator strengths of the two compounds of the mixture were known-which is never the case. Dr. Fano suggested in Paris that the energy be divided up according to the van der Waals cross-section of the constituents using the 2/3 root of the molar volume but to my knowledge lie has not published a detailed justification for it.One can be satisfied that for molecules containing only elements of the first period the three empirical methods mentioned here lead to practically identical results which is not the case if heavier elements (e.g. halogens) are present. Dr. P. J. Dyne and Dr. D. R. Smith (Chalk River) said We do not view the comments by Hardwick Wagner Schuler Lipsky Burr and Chariesby as being criticisms of our paper but as restatements of our conclusion viz. that there is serious conflict between our experiments and the " conventional wisdom " of H atom reactions in irradiated alkanes. While we claim that the additives used in our experiments do not reduce the hydrogen yield from irradiated n-hexane and n-pentane by hydrogen-atom scavenging it is still possible that hydrogen atoms are present in these and other irradiated alkanes and that they may be detected in reactions with suitable solutes.If so then the additives used in our study interact more strongly with the precursors of €4 and D atoms (CW CD*> than with the H atoms themselves. The simplicity of our results may indeed be deceptive but chemical evidence for H atoms and their reactions can also be deceptive. Only recently the existence of OH-radicals in irradiated water was presented as evidence for the presence of H atoms (if one fragment of a molecule is found the other fragment must also be there) as was the formation of h-xz by a bimolecular combination (what two things can com-bine to give €32 if they aren't H atoms?).Our term " physical interaction " is certainly not perfect and Prof. Schenck's points are well taken. Interpretive terms such as " energy transfer " " charge However the coiiflict is not absolute. 1 cf. Platzman Int. J. Appl. Rod. Isotopes 1961 10 116. 2 cf. Lamborn and Swallow J. Physic. Chem. 1961 65 920. Swallow Proc. Int. Con$ Rad. Res. (Natick 1963) GENERAL DISCUSSION 275 transfer ” and “ hydrogen atom scavenging ” may have hypnotic qualities which paralyze the critical faculties. The term “ reactivity transfer ” has the merit of being defined operationally. There are of course other possible mechanisms for the production of H2 and HD. Prof. Schenck has proposed one and Dr. Toma, another. These can only lead to a re-interpretation of our data if they also include a mechanism for the formation of D2 which is first order with respect to C-CgDiz.Our arguments would carry little weight if they were confined to variations in the HD yield. For the most part these additives were studied at low concentrations where G(H2)>2; this is the region of the low concentration interaction mentioned by Dr. Freeman. At higher concentrations where G<2 an abrupt change in slope is observed in fig. 6. This could be interpreted as a change in mechanism but in our opinion these and other isotopic experiments do not substantiate the separation of the yield into two parts.1-3 Dr. Ausloos and Burns have drawn attention to the paradoxical evidence on the mixture law in the vapour phase. We have little to add to our comments in the paper.Our isotopic experiments (fig. 7 and 8) show that the interaction studied by this technique is absent in the vapour phase mixtures studies. There are however, other interactions with hydrocarbon mixtures observable at higher solute concen-trations which are not detected by the isotopic technique (fig. 9 and lo). The yield of hydrogen found by Henglein in dilute solutions of cyclohexane in carbon tetrachloride is paralleled by our observations on solution of CsD12 in ben-zene 4 and by Burton Chang Lipsky and Reddy 5 who found a yield of hydrogen evolved from cyclohexane of 0-5 molecule/iOO eV absorbed in the cyclohexane. This is discussed further in ref. (4). The results of Dr. Henglein’s study of the cyclohexane +carbon tetrachloride system appear to be similar to those obtained by Dr.Stone and myself on the same systern.6 At the lowest solute concentrations used (- 0.004 M) scavenging of cyclohexyl radicals occurred without a corresponding decrease in the hydrogen yield. In C6H12 + CC14 mixtures all the radicals are scavenged at the lowest concentration, a chain reaction ensuing with CHC13 C6H11Cl and C2C16 as major products : ‘C6Hll+ CCl4+C6H11Cl+.CCl3 CC13 + C6H12 -+ CHC13 + *C6€111 2*CC13+C2CLj. No chain reaction occurs with either CHC13 or CDCl3 as solutes and their scavenging efficiency is less than that of CC4. At higher solute concentrations G(H2) is reduced. Up to 0.04 M all solutes reduce G(H2) to the same extent but above this concentration CC14 is more effective, CHC13 and CDC13 having the same efficiency.This differing efficiency between cc14 and both chloroforms cannot be due to the scavenging of thermal hydrogen atoms, H + CHC13-+H2 + CC13 H + CDC13-+HD + CC13, since a negligible amount of HD is formed in C6H12+CDC13 mixtures. 1 Dyne and Jenkinson Can. J. Chem. 1961 39,2163. 2 Stone and Dyne Rad. Res. 1962 3 353. 3 Dyne Smith and Stone Ann. Rev. Physic. Chem. 1963. 4 Dyne and Jenkinson Can. J. Chem. 1962,40 1746. 5 Burton Chang Lipsky and Reddy Rad. Res. 1958 8 203. 6 Stone and Dyne Can. J. Chem. submitted for publication 276 GENERAL DISCUSSION Again in agreement with Dr. Henglein we find that G(HC1) is greater than AG(H2) (the reduction in hydrogen yield). At 0.004 M CCl4 G(HC1) is 0.9 and AG(H2) is 0-54. At a higher concentration of CC4 0-2 M G(HC1) = 4.3 and AG(H2) = 3-2.These observations are inconsistent with the formation of HCl by H + CC4 -+ HCl + CCl3. In reply to Dr. Swallow the electron fraction happens as a special case to be the correct parameter for calculating energy partition in mixtures of alkanes since cyclohexane and cyclopentane (for instance) differ only in the number of electrons associated with a CH2 group. The collinearity of the points in fig. 7 and 8 is only obtained if yields are calculated on an electron-fraction basis and not on a mole-fraction basis. We disagree with Dr. Burns that the equilibrium H*+C* in our triangular dia-gram is unobservable (and hence unnecessary) because our experiments always involve a third entity CD. The diagram represents a radiation steady-state and not a thermodynamic equilibrium (to which Dr.Burns’ arguments would apply).I The equilibrium H*+C* plays an important part in determining the activation of CD to form C p . We also have evidence of this interaction in systems which do not contain a deuterated hydrocarbon? In any case if this interaction exists between two hydrocarbons one light and the other deuterated it is difficult to see why it should not occur between two “ light ” hydrocarbons. Our comment about the approximate equality of the yields of D2 and C-C~D~O has led to some misunderstanding. As Dr. Toma said we reported elsewhere,z that the yield of c-C6D10 is in fact greater than that of Dz. Our phrase “ as great as ” was meant to imply that the product yields were comparable (not differing by an order of magnitude) and consistent with our mechanism.A serious inconsistency with our mechanism would have appeared if the initial D2 yield had been greater than the c-C~DIO yield. In this reference we also discussed the possibility of C6D10 and HD being formed through the step C6D12-+C6Dio+2D which would con-tribute to the observed inequality of the C6D10 and D2 yields. Dr. Toma’s sug-gestion that the D atoms in this dissociation are “ hot ” (unscavengeable) is of interest. In our opinion simple comparisons between solutions of C-CgD12 and c-C6H12 in cyclopentane are invalid. There is a large difference in reactivity in the two cyclohexanes as is shown in extrapolated yields measured in cyclopentane solutions (Gextrap in Dr. Toma’s notation g in ours); g(C6D10) = 0-9 g(CsH,o) = 8.9 re-spectively.Simple comparison of the yields of isotopically substituted liquid products with HD are again complicated by change of identity reactions *C6Dii +CsHi2-)C6DiiH+C6Hii. In cyclohexane mixtures this reaction makes the HD yield some four times greater than the sum of deuterium deficient products (C6Dlo and C6Dll . c6&l). Dr. M. A. Golub (Stanford Research Institute) said With reference to the inter-esting paper by Fessenden and Schuler,3 I would like to call attention to the striking similarity between the extent of energy transfer observed in their ethane +ethylene system with that found in the work of Cundall and Griffiths 4 on benzene+butene solutions as well as that observed in our work on benzene solutions of polybutadiene,5 1 I m indebted to Dr.G. J. Mains for bringing this point to my attention. 2Dyne and Denhartog Can. J. Chem. 1963,41 1794. 3 Fessenden and Schuler this Discussion. 4 Cundall and Griffiths this Discussion. 5 Golub J. Amer. Chem. Soc. 1960,82 5093 ; 3. Physic. Chem. 1962 66 1202 GENERAL DISCUSSION 277 polyisoprene,l and squalene.2 Thus the yield of vinyl radicals from ethylene in a 10mole % solution in ethane was found 3 to be 8 times that observed in pure ethylene indicating considerable intermolecular energy transfer from ethane to ethylene. Analogously polybutadiene and polyisoprene were found to isomerize in 1 % benzene solutions with yields based on energy absorbed only by the solute, which were 8 and 10 times respectively those of the pure polymers.Squalene and butene-2 in benzene on the other hand gave even higher energy transfer factors, 30 and 15-55 respectively. Thus there is highly efficient transfer of excitational and/or ionization energy from either an aliphatic or an aromatic structure to an olefinic group. Such energy transfer can be considered probable inasmuch as both the ionization and excitation potentials of olefinic groups are lower than the corres-ponding ones for paraffins and are either lower than or comparable to those in benzene.4 5 Indeed the very high yield for isomerization of cis-polybutadiene in the solid state was explained on the grounds that the double bonds apart from being excited directly were also excited indirectly through efficient intramolecular energy transfer from the methylenic groups in the polymer.Thus assuming about two excitations and one ionization per 30 eV deposited anywhere in the polymer the potential iso-merization would be around 10. The actual initial G(cis+trans) value was found to be 8 in polybutadiene and 10 in polyisoprene. Presumably other processes must be operative here in order to have such an efficient utilization of the energy intro-duced into the solid polymer such as the one advanced by Lorquet ElKomoss and Magee 6 in which electronic excitation initially localized in one of the double bonds migrates along the chain inducing isomerization of several of them. At any rate there appears to be little doubt that most of the excitational energy de-posited in the methylenic groups is ultimately transferred to the double bonds which can then isomerize.This view thus finds strong support in the work of Fessenden and Schuler on the ethane + ethylene system. Prof. A. Charlesby (Roy. Military Coll. of Sci. Shrivenham) said Some results we obtained with irradiated polyethylene appear relevant to the behaviour of mixture of paraffins and olefins. On gamma or electron irradiation of polyethylene in the absence of oxygen trans-vinylene unsaturation is produced with an initial G value of about 1.7. At much higher doses the G value decreases progressively so that a maximum trans-vinylene concentration in the neighbourhood of 2 x 10-4 mole/g is reached. This maximum may be ascribed to some back reaction induced by radiation which destroys unsaturation.With alpha radiation the initial G value is substantially unaltered but the maximum concentration reached is much greater (about 10 x 10-4 mole/g) corresponding to a smaller back-reaction. Vinyl unsaturation which is reduced on irradiation is also less affected by alphas. We ascribe this back-reaction to the attack of H atoms on the unsaturated groups. In the alpha track the H atoms formed in close proximity can often react with each other leaving fewer to attack unsaturated groups outside the primary track. Numerical estimates can be obtained by using the reaction constants deduced by Hardwick for addition of H to unsaturation and for abstraction by H from a saturated chain. Thus about 40 % of H atoms abstract from neighbouring chains while still “ hot ” whereas the remaining 60 % lose their excess energy rapidly and only 1 Golub and Danon to be published.2 Danon and Golub to be published. 3 Fessenden and Schler this Discussion. 4 Cundell and Griffiths this Discussion. 5 Golub J. Amer. Chem. SOC. 1960 82 5093; J. Physic. Chem. 1962 66 1202. 6 Lorquet ElKomoss and Magee J . Physic. Chem. 1962,37 1991 278 GENERAL DISCUSSION abstract at a considerable distance from their origin (after roughly 104 collisions). Only radicals formed in the second process could be readily scavenged by low con-centrations of additive. That the back-reaction is greater with gamma or electron radiation than with alphas is shown by the following observation. If a specimen is exposed to alpha irradiation until maximuin transvinylene unsaturation is reached and then re-exposed to gammas the unsaturation level decreases.This would be difficult to explain in terms of an energy transfer process. Dr. G. R. Freeman (University of Alberta) said The C2H5 and C2D4H yields in the C2H6+C2D4 mixtures (fig. 1 of Fessenden and Schuler’s paper) and the C2H3 yields in the C2H6+C2H4 mixtures (fig. 2) are consistent with the idea that both a scavenging reaction of some sort and an activation transfer reaction occur to significant extents in these systems. With regard to the value of q = 10 in eqn. (l) do Fessenden and Schuler mean to imply that the energy absorption cross-section of ethylene is ten times greater than that of ethane? Is it possible that the energy might be initially distributed roughly according to the electron fractions of the components and then transferred from ethane to ethylene ? Prof.C. E. Klots (Florida State University) said The results of Dr. Schuler have served to raise the problem of initial energy partition in the radiation chemistry of binary mixtures. A treatment has been presented 1 which offers some hope of clarifying this problem at least in the gas phase where it may have meaning. Thus in ethane + ethylene mixtures a weighting factor of 1.08 & 0.1 is indicated for the initial energy partition closely matching the relative electron numbers (1 8/16). This result suggests that Fessenden and Schuler are justified in seeking elsewhere for an understanding of their observations. That this apparent validity of the electron fraction parameter is only fortuitous may be demonstrated by reference to argon+ propane mixtures; here an energy partition factor of 0-16 is indicated in marked contrast to the relative electron numbers (1 8/26).Dr. A. J. Swallow (Christie Hospital and Holt Radium Inst.) (communicated): For argon and propane the valence electron numbers are in the ratio 8/20 giving an energy partition factor of 0-4. Thus in this case valence electron fraction gives a better estimate of energy partition than electron fraction. For ethane + ethylene mixtures as for other mixtures containing only elements of the first period the difference is very slight as Prof. Magat has pointed out. Dr. J. H. Futrell and Dr. T. 8. Tiernan (Aerospace Res. Lab. Ohio) said In the paper by Collinson Todd and Wilkinson an attempt is made to assess the im-portance of ionic processes initiated by argon metastables using an ion distribution obtained with 11.6 V electrons.This is of course incorrect as an electron can transfer any amount of energy up to its maximum energy whilst a metastable atom can transfer only a quantum of energy corresponding to an electronic transition. Recently a charge exchange determination of the break-down curve for n-butane has become available,2 and a more accurate estimate of the ionic fragmentation for 11.6 eV excitation is possible. From these experiments the ratios of C3HT, C4Hf, C3H and C2Hi become 80 5 12 3. This will change the calculated yields for the ionic mechanism in Collinson’s table 1 mainly in decreasing the relative yield of ethane while increasing all other products relative to ethylene.One generalization 1 would draw from this research is the striking change in product yields with fairly modest changes in experimental parameters ; i.e. with 1 Klots J. Chem. Physics 1963 39 1571. 2 Chupka and Lindholm Arkiu F’sik 1963 to be published GENERAL DISCUSSION 279 small changes in the average energy deposited in the molecule. This would seem to be characteristic of experiments performed near the threshold for competing processes. This work shows a corollary behaviour to low energy electron impact studies in mass spectrometry where violent changes in the cracking pattern occur over the range of electron energy a few volts above the ionization potential for complex molecules. TABLE VARIATION OF MASS SPECTRUM OF PROPANE WITH VOLTAGE per cent of total ionization m/e 70 V 180 V 260 V 510 v 590 V 15 3.2 2-6 2.4 2-0 1.7 26 3.4 2.9 2.7 2.3 2.2 27 13-2 12.7 12.5 12.3 12.3 28 16.6 17.1 17-4 18.0 18.0 29 28.3 29.8 30.5 31.4 31.5 39 6.2 5-3 5.0 4.7 4.6 40 0.9 0-8 0.8 0.8 0.8 41 4.1 4.2 4.2 4.4 4.4 42 1.8 1.8 1.9 2.0 2.0 43 6.8 7.1 7.3 7.6 7.7 44 8.1 8.6 8.7 9.2 9.4 This energy region below 50 eV has been of greatest interest to mass spectro-scopists and the higher energy region where the cracking pattern is relatively struc-tux eless is essentially unexplored.This region is however of interest in conven-tional radiation chemistry as a significant portion of the energy deposited in the system must be from electron impact with energetic secondary electrons.1 Ac-cordingly we have made a brief study of the mass spectra of several paraffin hydro-carbons recently over the energy range 50-6OOV.The results for propane are typical and are given in table 1. In contrast with mass spectra at low energy and with the chemical study of Collinson et al. the changes in pattern illustrated are slight annd are monotonic as a function of energy. The small changes observed may be described with reference to a simplified mechanism for propane ion fragmentation : (1) C3H8 + e-+C,Hz * + 2e The general behaviour of the cracking pattern with increasing energy below 50V is a shift from left to right in the above scheme as the more endoergic reactions become important. The amount of excess energy deposited in the molecule ion-hence available for more extensive fragmentation-increases monotonically with energy.At higher energies the cracking pattern shifts in the reverse direction as shown in table 1. The shift apparently is from CsHi+C3HZ-+C3Hf from C~H~+CZH: 1 Plqtzman Int. J. AppL Radiation Isotopes 1961 10 116 280 GENERAL DlSCUSSION and from C2H; + CzH,+. The methyl ion reaction (9) above has not been characterized in detail but is known to be a high energy process. It also decreases with energy in table 1 while the molecule-ion increases. Thus a paradox results in that in the 50 to 700 V range the changes in mass spectra can be interpreted as resulting from a gradual decrease in the amount of energy deposited in the molecule ion as the energy of the ionizing electron increases.This effect seems to be quite general for hydrocarbons. We have observed it for methane ethane ethylene n- and isobutane in addition to propane over this energy range and results at kV energies 1-4 for several molecules are consistent with our data. Simple considerations of the electron impact process suggest that the probability of depositing large amounts of energy should decrease with increasing electron velocity as is observed because of the decreasing time interval in which an interaction is effective. However it is somewhat surprising that this effect should manifest itself in such a low energy region as that studied. The changes are subtle and one may therefore reasonably expect that the ionization and dissoci-ation processes produced by fi rays and high energy S rays are quite similar to those produced by 50-100 V electrons.This observation contrasts sharply with studies of mass spectra near the appearance potentials of various dissociation processes and with the present results of Collinson Todd and Wilkinson. Dr. G. R. Freeman (University of Alberta) said The calculation of an exciton range involves some sort of physical model of the exciton. There seem to be at least two models for excitons. One is that a certain amount of energy is deposited in a spot and that it spreads as a spherical wave. In this case the amount of energy per element of volume might be expected to decrease as the inverse square of the distance from the point of origin. Another model is that an exciton moves rapidly as an energy packet from molecule to molecule.The time spent on each molecule is so short that the atomic motions within the molecule are altered only slightly or not at all. According to the first model the effective range of the exciton might be calculated using the cube root of the volume of solvent that contains an average of one solute molecule when the concentration is sufficient that a certain fraction of the exciton energy is trapped by solute. In the second model the exciton might be considered to migrate by a random walk in which case the range of the exciton would be roughly calculated using the square root of the mole ratio of solvent to solute at a concentration where a certain fraction of the exciton energy is being trapped, multiplied by the solvent molecular diameter.It appears from the ranges reported by Dainton and co-workers that they used a cube-root calculation. A square-root calculation would increase their range values by a factor of five or so. The important point is that the range of the exciton in certain crystals described in Dainton’s paper is greater than the range of the electrons in the glasses described by Hamill and co-workers.5 Dr. J. Cunningham (1.1.57. Res. Znst. Chicago) said We have investigated the effect of impurity centres viz. univalent silver ions upon the radiolysis of ionic nitrates in which like some of the aromatic systems where Prof. Dainton observed strikingly high efficiency of radiation reduction of FeC13 molecular species are ar-ranged in crystalline array. The first-order rate constants for radiolysis of silver-doped KNO3 at 300°K were 7 14 17 and 23 % lower than that of the pure material 1 Kebarle and Godbole J.Chem. Physics 1962 36 302. 2 Melton J. Chem. Physics 1962 37 562. 3 Monahan and Stanton J. Chem. Physics 1962,37,2654. 4 Vestal M. personal communication. 5 this Discussion GENERAL DISCUSSION 28 1 for approximate Ag+ concentrations of 4 x 10-4 10-3 4 x 10-3 and 10-2 M. For KNO3 containing 10-2M of molecular oxygen product a (30+10) % decrease in rate constant has been observed by several workers. The initial rate constant for radiolysis of AgN03 is also (30$-10) % lower than expected from the relationship between “ free-space ” and GNO; values for other univalent nitrates. Previous electron-spin resonance and optical-absorption studies have provided good evidence that electron attachment to NO; does occur at 77°K and we have lately observed that the Ag+ concentrations stated above suppressed electron-excess bands assigned to NO$- and NO;- in favour of absorption associated with Ago.Although all these results are consistent with the hypotheses that (30+ 10) % of radiolytic dissociations at 300°K proceed via dissociative electron attachment (DEA) and can be suppressed by scavenging of each low-energy electron by one Agt- or 0 2 impurity centre per 100-500 sites visited an alternative explanation involving energy localization around impurity centres cannot be rejected. Dr. Katsuura said that he and Dr. Inokuti have considered the possibility for localization of quanta of excitational energy upon impurity centres in a one-dimensional chain of coupled species.They concluded that such localization can indeed occur and may be important in explaining the results of Siege1 and Fluornoy on H atom production from 99 % deuterated ice. Exciton movement in ionic crystals has been known for many years and conditions for exciton localization may exist at Agi- impurities in KNC$/Ag+. On this explanation of the effect of impurity centres our experimental results would require that the probability for dissociation of NO; ions situated close to impurity centres would be lower than for the normal lattice i.e. that localization in KNO3 results in de-excitation. However the exact nature of the interaction between host molecular species on which energy (including superexcited states) is localized and the impurity centres causing the localization would depend on the properties of host and impurity species and would determine whether sensitization (as in Prof.Dainton’s results) or protection (as in our results) is observed. Dr. G. 0. Phillips (University College Cardifl) said Our studies on the decom-position of a-D-glucose by C060 y-radiation in the solid state provide further evidence for the existence of a mechanism of energy transport which is facilitated by a highly ordered crystalline system. The behaviour of a-D-glucose on y-irradiation in the polycrystalline syrup and freeze-dried states is shown below. The appreciably state conditions G(acid) C(hydrogen) polycrystalline air 13-2 3.8 in vacuo 13.3 in C02 13-0 SyrUP in vamo 6.4 4.7 freeze-dried in vacuo 4.0 4.0 higher acid yield for the polycrystalline material compared with the syrup is ana-logous to the enhanced yield observed by Dainton et al.for crystalline solutions of type B. The behaviour of the freeze-dried state however is not explicable on similar considerations. X-ray powder photograplp of polycrystalline and freeze-dried samples have shown that the same chemical and crystallographic forms are present giving identical lattice parameters and spacings to those previously reported for a-D-glucose.192 Previously it was suggested that point defects or dislocations introduced into the lattice by rapid crystallization during freeze-drying was the 1 Wyckoff Crystal Structures 5 61 62A (Interscience 1953). 2 Index to X-ray Powder Data File card 1-0374 (Amer.Chem. Soc. 1961) 282 GENERAL DISCUSSION most likely explanation for the difference in behaviour of the two polycrystalline samples.1 Recently using electron microscopy,2 it was possible to demonstrate the existence of a large number of defects in the freeze-dried crystals. The photo-graphs (fig. 1) show electron micrographs of positive carbon replicas prepared from normal polycrystalline (fig. 1 a) and freeze-dried crystals (fig. 16). The large number of defects in the latter indicate a highly strained and imperfect crystal lattice. Such lattice imperfections appear to act as barriers which hinder the transport of energy, resulting in appreciably less radiation decomposition after freeze-drying. It may be demonstrated that energy transfer is facile in such carbohydrate crystals using the crystalline complexes which the cyclic Schardinger dextrins form with aromatic molecules.The Schardinger dextrins are cyclic molecules composed of six or more a-D-glucopyranose units linked 1-4 as in amylose (fig. 2).3 The space within the j-cyclic Schardinger dextrin (7a-~-glucose units) e.g. is sufficient to accommodate a molecule of p-nitrophenol. If such a complex is irradiated with C060 y-radiation (dose 1 -93 x 1020 eV/g) the characteristic spin resonance spectrum of the irradiated uncomplexed Schardinger dextrin (fig. 3a) disappears and only the ill-defined spec-trum which is observed after the irradiation of p-nitrophenol alone to an equivalent dose is recorded (fig. 3b and c) The same stoichiometric uncornplexed mixtures SO gauss I (a) (4 (4 FIG.3.-Electron spin resonance spectra of (a) 7-irradiated /l-cyclic Schardinger dextrin; (b) p-nitrophenol ; (c) complex of p-nitrophenol and &cyclic Schardinger dextrin. of the P-dextrin and nitrophenol after irradiation do not show this behaviour and give after irradiation to 1.93 x 1020 eV/g an electron spin resonance spectrum iden-tical with irradiated P-dextrin. .It would appear therefore that complex formation is essential for energy transfer from the a-D-glucose moiety to the p-nitrophenol. Such an energy transfer results in significant chemical protection of the carbohydrate. 1 Phillips and Baugh Nature 1963,198 282. 2 Photographs and carbon replicas prepared by Miss Anna T. Moore of the Southern Regional 3 French Adv.Carbohydrate Chem. 1957 12 189. Research Laboratory U.S. Department of Agriculture New Orleans La FIG. la.-Positive carbon replica of anhydrous polycrystalline a-D-glucose crystal ( x 23,000). FIG. 16.-Positive carbon replica of anhydrous polycrystalline a-D-glucose after freeze drying [To face page 282. ( x 5,000) (markers correspond to 1 micron) FIG. 2.-Model of /I-cyclic Schardinger dextrin (background squares correspond to 5.5 A) GENERAL DISCUSSION 283 Dr. W. G. Bums (A.E.R.E. Harwell) (communicated) I wish to raise three points relating to the elegant demonstration of efficient transfer of absorbed radiation energy to additives in crystalline aromatic systems (type B behaviour). (i) Three conditions perhaps not all separable experimentally seem to be fulfilled for type B behaviour.The systems (a) are aromatic (b) probably have emission bands overlapping the absorption spectra of the additives (c) have a first singlet excited state which does not readily decompose in the condensed state. The authors have stressed the importance of condition (b) but condition (c) may also be im-portant since under this condition the propagation of energy would not be in com-petition with dissociation. Voevodskii and Molin 1 have pointed out that this con-dition is probably responsible for the normally low decomposition G values for aromatic substances. The stability of excited aromatic molecules may be further manifested by the generally strong dependence of aromatic decomposition G values on the L.E.T.of the radiation the high G values at high L.E.T. being due to co-operative effects of neighbouring excitations in competition with first-order de-activations. (ii) The results are possibly analogous with the situation in crystalline systems with small concentrations of impurities e.g. KH in KBr or KC1,2 where small concentrations (10-3-10-4 mole ratio) of the additive can cause large changes in the efficiency of F-centre formation on irradiation. (iii) The authors imply that G(lBzu) in liquid benzene is C4 whereas there is evidence from scintillation studies 3 that this G value is much lower. Prof. F. S. Dainton (University of Leeds) said The whole tenor of Prof. Hamill’s communication is that ionic processes play the major role in solid glasses which he has investigated whilst our evidence is that in certain crystalline systems efficient energy transfer not involving charge transfer predominates.It is important to point out that there is no fundamental conflict between these apparently contra-dictory conclusions. This can be seen in two ways. From table 1 and fig. 1 of Dr. Hamill’s papers it is evident that in all the systems which he and his collabor-ators have studied the capture of electrons by solutes present in amounts <0.01 mole % is entirely negligible. This is the concentration range with which we are concerned with our solutes and as Dr. Salmon will show Hamill-type experiments using ferric chloride in this concentration in MTHF glass demonstrate that this solute has a negligible effect on the absorption spectrum of the trapped electron.Moreover when these trapped electrons are rendered mobile by warming or by photo-activation they do not react with the dissolved ferric chloride. This suggests either that the electrons are trapped at distances from positive ions derived from the solvent much shorter than the average half-separation of ferric chloride molecules, which in our experiments is generally greater than 100 A or that the reduction of ferric chloride by electrons can only occur when the latter have certain energies > O . If as postulated in the first hypothesis the electron is “ trapped ” at a point where its Coulornbic energy >kT photo- or thermal-detachment of the electron from its hole is likely to cause its return to the group of positive ions from which it originated and the kinetics of the bleaching wiil be given by D = DO exp (-3.t) where D is the optical density.The observed decay law is however D-1 = A + Bt correspond-ing to destruction of electrons predominantly by mutual interaction of one from one track with one from another suggesting that the electroiis are trapped at much 1 Rad. Res. 1962 17 366. 2 Martienssen and Pick 2. Physik 1953 135 709. Bums and Lockyer J. Sci. Instr. 1955 32, 316. 3 Van Dusen and Hamill J. Amer. Chem. Soc. 1962 84 3648 284 GENERAL DISCUSSION greater distances than has been proposed. Since at the doses used [e- trapped]< [FeCl3] such electrons when detached from their “holes” will have higher en-counter rates with FeC13 molecules than with themselves. We therefore prefer the alternative hypothesis that not all electron-FeCl3 encounters lead to reaction.Although a definite choice between these hypotheses for electron trapping in glasses cannot yet be made we can be sure that in perfect crystals there will be no solvent holes and the electron will not be trapped after thermalization but will return to positive centres and on recapture may well cause electronic excitation. Both this excitation and “ primary excitation ” may migrate over long distances by the exciton mechanism to cause chemical change in solutes such as FeC13 or 12. Perhaps I may be allowed to make a comment on terminology inspired by Dr. Burton’s remarks concerning semantics. Solvation is commonly understood to mean the orientation of solvent molecules round a charged or polar solute under the influence of forces of interaction between the solute and solvent molecules.In solid systems forces between the electron considered as a solute and the molecules of the solid certainly exist but it is doubtful whether significant reorientation of solvent molecules round the electron takes place when the latter cease to move through the system. Indeed it seems more likely that in glassy aggregations of solvent molecules there is a distribution of orientation of solvent molecules and that several of these may be capable of attaching an electron with at most minor rearrangement, and with different binding energies. For this reason I think we should refer to the species produced in glasses as a result of irradiation as the “ trapped electron ” and reserve the term “ solvated electron ” for the shorter-lived species produced in irradiated polar liquids.That there is a difference between the two is clearly shown in those solvents such as alcohols in which the absorption spectrum of the solvated electron as measured by pulse radiolysis in the liquid is clearly different from that of the trapped electron produced in the irradiated low-temperature glass. Dr. G. A. Salmon (University of Leeds) said We have investigated the absorption spectrum at 79°K induced in MTHF by C060 y-irradiation at 77°K. This spectrum shown in fig. 1 resembles that presented by Hainill and co-workers except in that we also observe a weak absorption originating at about 20,000 cm-1 and extending to the limit of observation in the u.-v.This spectrum may be compared with those found in irradiated methanol glasses at 77”K.l That the bands A and B are associated with different species may be seen from fig. 2 which shows the dependence on temperature of the absorbance in the two spectral regions. Also in the temperature region 90-95°K it has been possible to ob-serve the rate of disappearance of the two bands and in both cases the decay is found to be a second-order process at the radiation doses used. E.s.r. measurements also confirm the existence of two species since a seven-line spectrum is observed with a superimposed singlet which is more apparent at low microwave powers (fig. 3). The behaviour of the e.s.r. singlet is qualitatively similar to that of band A in the optical spectrum ; i.e.it can be photo-bleached by light of wavelength in the optical band and disappears at a finite rate at 90°K. Similarly, the behaviour of the underlying 7-line spectrum corresponds to that of band B in the optical spectrum. In fig. 4 can be seen the effect of ferric chloride concentration on the size of peak A when measured at 79°K. As Prof. Dainton has already pointed out there is no sig-nificant reduction in the height of this peak at ferric chloride concentrations less than mM. 1 Dainton and Teply this Discussion GENERAL DISCUSSION 285 At smaller concentrations of ferric chloride it is possible to observe the changes in ferric chloride concentration produced by radiation at 77°K. Fig. 5 shows the radiation-induced spectrum in the presence of 1.58 x 10-4 M and fig.6 shows the AA v cm-1 x 10-3 FIG. 1.-y-Radiation induced spectrum in MTHF at 77°K. dose 2.76 x 1018 eV/g fo-7. - - -0 I 1 I 8- - - a -B 79 79 85 9 0 95 100 105 temp. OK FIG. 2.-Temperature profile of the optical spectrum. dose 5.52 x 1019 eV/g ; A 11,000 cm-1; B 40,000 cm-1 dependence of these spectra on temperature. Fig. 6 demonstrates conclusively that the " trapped electrons " disappear without any reduction of ferric chloride occurring 286 GENERAL DISCUSSION Thus these results suggest that whereas secondary electrons can reduce ferric chloride before they are thermalized and " trapped " after " trapping " and sub-sequent release by warming they react either with themselves or with positive ions trapped in the matrix.In addition the result that the e- absorption in pure MTHF -4-2 0 9 - 9dB -15 db - 25db FIG. 3.-E.s.r. spectrum of y-irradiated MTHF at 77°K. dependence on r.f. power 1 10 [FeCl3] M x 103 FIG. 4. disappears according to second-order kinetics means that the distribution of " trapped electrons " has become homogeneous and the electron is not reacting a priori with its parent ion (such a process would obey first-order kinetics). A possible explanation of these results is that the reaction e- + FeC13 -+ FeClz + Cl-involves a significant activation energy and that whereas the electrons during the moderation process will possess this energy after " trapping " and thermalization this reaction becomes impossible but reaction with positive ions can occur GENERAL DISCUSSION 287 v cm-1 x 10-3 FIG.5.-Spectra of FeCl solutions in MTHF at 79°K. x before irradiation ; 0 after irradiation ; dose 2-76 x lo** eV/g I 1 1 1 1 79 79 8 5 90 95 100 105 temp. OK FIG. 6.-Temperature profile of the radiation induced optical absorption. [FeC13] = 1.58 x 10-4 M ; dose 2.76 x 1018 eV/g ; A 8,000 cm-1 ; C 28,000 cm-288 GENERAL DISCUSSION Prof. A. Charlesby (Roy. Military Coll. of Sci. Shrivenham) said Direct evidence for the presence of trapped electrons in polymethyl methacrylate irradiated at liquid N2 temperature can be obtained from electron spin resonance measurements. Thus the spectrum obtained with PMMAfTCNE is similar to that from ionized TCNE (tetracyano-ethylene) prepared chemically at room temperature.Further evidence is provided by thermoluminescence studies in other polymers. In these cases the series of luminescent peaks found can be profoundly modified by the presence of oxygen. However when the oxygen combines with the polymer radicals to give peroxide radical the spectrum reverts to that typical of the anoxic specimen. To be effective oxygen must be present in a molecular form and hence presumably acts as an electron trap. Dr. L. M. Dorfman (Argonne National Laboratory) said The one aspect of the experiments of Dt . Hamill et al. which is most pleasing is that the transient species observed particularly the aromatic anions have well-known electronic absorption spectra ; thus the assignment seems certain. The experiments reported had to be carried out in the solid state.Thus there is left unanswered the important question to which Dr. Hamill has drawn attention, of whether the conclusions reached in his experiments in glasses may be extended to the liquid state. The pulse-radiolysis with microsecond time resolution permits us to answer this question. In the few cases in which our experiments 1 in polar liquids have involved the same systems investigated in the glass by Dr. Hamill, the conclusions may be extended to the liquid state with some important differences. Thus in deaerated ethanol solution of diphenyl we find the adsorption spectrum of the diphenylide ion. In ethanol solution of naphthalene we have identified the naphtlialenide ion. Thus we have unequivocal evidence for the role of ions as transient species in this polar liquid.The diphenylide ion in liquid ethanol has a relatively short lifetime showing a first-order decay with a half-life of 1.7 psec. We believe the lifetime is limited because of the protonic character of the solvent the decay corresponding to the elementary reaction, C12H~o+C2H50H = C12Hll +C2H50-. Our observations extend to free radical species in ethanol. Thus in deaerated ethanol solution of benzyl chloride we observe the benzyl radical and in ethanol solution of triphenylmethylcarbinol the triphenyl methyl radical. In the latter case the observations indicate that the radical is formed by dissociative electron attachment to the carbinol. Dr. E. Hayon (Saclay) said I would like to report some results which in some ways are complementary to those presented in the paper by Prof.Hamill and his collaborators. In recent work Dr. Chachaty and I have carried out we believe we have obtained e.s.r. evidence of radiation-produced electrons trapped in organic glasses at 77°K. On y-irradiation of methanol ethanol and isopropanol as a glass at 77°K they become strongly coloured. Ethanol e.g. turns to a dark purple colour and superimposed on the basic five-line paramagnetic resonance of the CH3CHOH radical is a centre line. This centre line which has a g-value close to the free electron is associated with the purple colour. It is possible to suppress the formation of colour centres and therefore the centre line when certain solutes known to react efficiently with electrons are added initially e.g. acids 12 diphenyl, naphthalene 2-chloroethanol.Addition of alkali somewhat enhances and stabil-izes the purple colour. The colour centres can be bleached on exposure at 77°K to visible light or on warming the ethanol glass to 115°K. In all cases the centre 1 Taub Sauer and Dorfman this Discussion GENERAL DISCUSSION 289 line is attenuated with the disappearance of the colour. A (fig. 1) shows the spectrum of ethanol glass at 77°K ; B 1.0 N NaOH in ethanol ; C 1.0 N HzS04 in ethanol ; D is spectra on warming A to 115°K ; E is after bleaching B with visible light at 77°K; F 0.1 N 12 in ethanol at 77°K; G is spectra after visible bleaching of A ; H is obtained on exposure of G to u.-v. light when a new radical (CH3CHzO) is formed; I spectra of the RO. radical obtained after warming H to 115°K.In conclusion the e.s.r. results obtained on y-irradiation of various alcohol glasses demonstrate quite clearly that radiation-induced electrons are trapped in these low-temperature glasses. A C 0 G e H F 5 0 GAUSS -n FIG. 1 .-0.s.r. evidence of radiation-induced electrons Dr. D. S. Ballantine (Brookhaven Nat. Lab.) trapped in organic glasses at 77°K. (communicated) I should like to report some observations' obtained by R. Ranganathan in our laboratory which seem to indicate " energy transfer " in organic solid systems. The GH was measured for a series of mixed crystals of acrylamide and propionamide which were irradiated at -78". The GH from propionamide is 0-474 while GH from acrylamide is 0.024. All compositions of mixtures show marked negative deviations from the calculated values.Similar indications of " energy transfer " were observed during room temperature irradiations of this same system but the situation is complicated by the fact that at room temperature the acrylamide can polymerize. Most inter-esting are the results with pure acrylamide. The yield of hydrogen against dose is linear to doses of 30 Mrad giving a GH of 0.047. Other experiments indicate that under these conditions acrylamide is " essentially completely " polymerized at less than 5 Mrad. If the polyacrylamide made by irradiation is separated dried 290 GENERAL DISCUSSION and re-irradiated a GH of 0.24 is obtained. One can explain these results on the basis of an " energy transfer" from polymer to isolated acrylamide monomer molecules which remain after acrylamide is irradiated to high doses.Comple-mentary e.s.r. studies on this system are consistent with this picture. Prof. M. Magst (Faculte' des Sciences de Paris) said M. DCrouEde in our laboratory has just completed a reinvestigation of the thermoluminescence of solid cyclohexane after irradiation. Three of his findings may have some bearing on the problems discussed here. (i) Glow curves were observed with maxima at 93 114 150 and 186°K. The first two maxima are not affected by annealing. However in annealed crystals the last maximum which occurs at the transition temperature is much more intense than the maximum located at 150"K while the reverse is true for quenched crystals. Radical recombination in quenched crystals sets in around 140°K.We can conclude that lattice reshuffling in whatever direction the transfmmation occurs sets electrons free. (ii) The thermoluminescence in gas-chromatographically-pure cyclohexane is very weak. An addition of 10 p.p.m. of benzene increases the intensity by a factor of 2.5 an addition of 1000p.p.m. by a factor of 18.5. This effect is not specific to benzene because an addition of 1300 p.p.m. of the saturated compound 2,2,4-trimethylpentane increases the thermo!uminescence yield by a factor of 11. The presence of an impurity seems to be inore important than the nature of the iinpurity. (iii) The spectrum of the thermoIuminescence comprises 3 bands located between 5000 and 6000A. No light emission between 5000 and 2200A was observed.The experiments are now being extended to shorter wavelengths. (iv) A kinetic analysis of the thermoluminescence shows that they are in better agreement with the theory at Bonfiglioli than with the theory of Randall and Wilkins; the activation energy of at least the first three maxima being the same (0- I9 eV & 0.02) within the limits of experimental error. Prof. J. E. Willard (University of Wisconsin) said Recent work in our laboratory, like that discussed in the significant papers presented shows striking differences between the chemical effects produced by gamma irradiation of the glassy state and the polycrystalline state of the same material at the same temperature. This is illustrated by the data in the table. E.s.r. spectra of the irradiated polycrystalline TABLE 1 .-COMPARISON OF E.S.R.SPECTRA OF GAMMA IRRADIATED GLASSY AND POLYCRYSTALLINE ALKYL EODIDES cpd. i-C3H71 n-C3H7I n-C4HgI n-C~H111 state glass cryst. cryst. glass cryst. glass cryst. glass cryst. glass cryst. cryst. cryst. no. of lines 6 30 19 6 7 6 24 6 6 6 15 6 15 total spread gauss 1 60 lo00 500 160 1 60 160 700 160 160 160 350 1 60 350 approx. intensity approx. T. of ratio b rapid anneal OK 1 0-005 0.02 0.2 0.2 0.5 0.03 0.5 0.1 0-2 0-05 0.2 0.05 100 158 145 108 172 106 1 70 118 187 123 250 213 227 m.p. O K 165 182 172 172 1 70 170 187 187 205 205 225 227 a all attempts to obtain the glass were unsuccessful.about 15 min after completion of 20 min 7 x 1019 eV g-1 irradiation GENERAL DISCUSSlON 29 1 alkyl iodides tested differ from the corresponding glasses in that (i} they have many more lines ; (ii} they are much wider exteiiding to 1000 gauss for C2H5I ; (iii) the intensity and hence the implied radiolytic yield of radicals is lower; (iv} they are stable on rapid annealing up to the melting point and they also persist for much longer times at 77". The figure shows the contrast between the 6-line spectrum of glassy C2H51 and the spectrum of some 30 lines observed for the polycrystalline CzHSI at a signal level 160 times more sensitive (the latter also includes the strong central line and hydrogen doublet from the quartz sample tube).The former is attributable to the C2H5 radical. Explanation of the latter seems to require coupling of the unpaired electron to iodine (spin 5/2) as well as to protons possibly in a CzHJ radical. The spectra for the glasses noted in the table and that for polycrystalline i-C3H7I are similar to those reported for these compounds by Ayscough and Thomson1 who did not distinguish between glassy and crystalline samples. The greater yield of C-I bond rupturs in glassy than in polycrystalline C2H5I indicated both by the e.s.r. spectra and by earlier studies of the radiolysis yields of 122 may be due to one or more of the following (i) a greater possibility of cage escape in the glass; (ii) steric effects in the oriented molecular arrangement of the crystal such that the most probable reaction of the activated species does not split C-I bonds; (iii) more C-I bond rupture in the random structure of the glass as a result of greater localization of energy in the C-I bond as contrasted to the de-localization favoured by the long-range order of the crystals ; (iv) ionic reactions in the glasses made possible by the trapping of electrons (the irradiated glasses show an intense blue colour after irradiation while the crystals remain white).Prof. Hamill and co-workers have presented convincing evidence for radical formation by dissociative electron attachment involving solutes in glasses at liquid nitrogen temperature. The implied efficiency of cage escape of the negative ion from the radical (e.g. Ck from C ~ H S C H ~ radical formed from benzyl chloride) is sur-prisingly high in view of photochemical evidence on caging effectiveness in a number of systems.I should be interested in Prof. Hamill's thoughts on this question. 1 Trans. Faraday SOC. 1962,544 1477. 2 J. Anier Chem. SOC. 1957 79 2429 ; 1959 81 761 292 GENERAL DISCUSSION Prof. W. H. HamiU (University of Notre Dame) said In reply to Prof. Willard, in dense media geminate free radicals formed by simple bond rupture are rather likely to experience diffusive re-encounters. Since little or no activation energy i s required recombination is efficient. The denser the medium the smaller the initial separation and the more probable is a re-encounter. In contrast most cases of dissociative electron attachment which have been observed are exothermic with thermal electrons e.g.CBHSCHzCl +e-,C6H&H2* + Cl-. A few diffusive re-encounters of radical and ion cannot reverse the reaction since it would be endo-thermic. Such dissociative attachments are nevertheless more efficient in polar than in non-polar media. It may be that the reaction RX+e-,RX*- is reversed when separation into products is prevented in dense non-polar nedia and that polar media act solvolytically. Prof. F. S . Dainton (University of Lee&) said One of the most notable features in irradiated solid solutions of FeCl3 in aromatic crystals is the rapidly diminishing G(-S) as the dose is increased (see fig. 5 of our paper) which we have attributed to the large lattice dislocation produced when the exciton + FeC13 interaction occurs. We imagined that this was due to the formation of the unsymmetrical and bulky chlorocyclohexadienyl radical C6HGC1.An alternative possibility is that the chlorine atom may be responsible. Consequently I am interested to know from Prof. Magat's paper that Dr. Szwarc has estimated the magnitude of lattice perturbations caused by dissociation of certain bonds. May I ask whether any of these calculations refer to systems of immediate interest to us and if so what are the general conclusions ? Perhaps I may also point out that some of the clearest indications of the im-portance of limited molecular motions in controlling radical isomerizations and disappearance and also thermoluminescence following irradiation is to be found in n-hexadecene-1 which has a well-marked lambda-transition almost 30 deg.below the m.p.1 Prof. F. S . Dainton (Leeds University) said I agree with Dr. Phillips that the data which he presented especially that concerned with solutions of p-nitrophenol, seem to offer further evidence of relatively long-range energy transport in irradi-ated crystalline material as compared with amorphous solids. However his sug-gestion that in freeze-dried crystals the small size of the crystals (lp diam.) limits the migration of the energy is somewhat surprising. If the implication is that the energy can travel to the edge of the crystals before annihilation (and the work of Kallrnan and co-workers shows that this is possible in anthracene) then the linear migration distance can be as large as about 500015. which is much larger than is generally observed and indicates unusually favourable conditions of energy trans-port in carbohydrates.Perhaps the polymeric nature of these substances is also an important factor contributing to the facility of transfer. Dr. Freeman has raised an interesting point. The distances given in our paper are linear distances and it is true that if there is random migration of the energy the actual non-linear distance will be greater than this. Whether this implies that energy can travel further in crystals than electrons in glasses is another question as pointed out in my comment on Prof. Hamill's paper. Perhaps I might add that the fact that as shown in Hamill's paper different solutes have different reaction efficiencies towards electrons suggests that at least some of these solutes can only react with electrons having energy greater than their thermal energy i.e.with those which have migrated a shorter distance. Consequently the migration distances 1 Ayscough McCann Thomson and Walker Trans. Faraday SOC. 1961,57 1487 GENERAL DISCUSSION 293 for electrons which one would calculate from Hamill’s experiments even those involving very efficient electron-scavenging solutes may well be much less than those which an electron may travel before thermalization. The data reported by Prof. Willard seem to indicate that in the polycrystalline alkyl iodides none of the electrons are trapped as such and less react with ethyl iodide molecules than in the glassy iodides. More of the electrons must there-fore return to positive species.If this recapture process results in formation of excited states which correspond to dissociative levels from which radicals would be formed either directly or after migration the broad pattern of these data are explicable. The last point raised by Prof. Willard is not perhaps so surprising. The frag-ments in the glassy cage resulting from dissociative electron attachment are an alkyl radical and an iodide ion. The recombination reaction (1) will have a positive eiithalpy change if the relatively large electron affinity of the iodine atom exceeds ’ the sum of D(R-I) and the trapping energy of the electron at this site. It may well be that this recombination is much less efficient than the recombination efficiency of neutral free radicals to which the photochemical evidence mentioned by Prof.Willard refers. Dr. Cunningham’s remarks are interesting especially in that they indicate that both electron- and energy-migration may occur in the same ionic crystal and that competition leading to protection can be expected on both mechanisms. In the non-ionic crystals which we have studied competition between two energy acceptors may well be regulated by the relative overlap of the absorption spectra of the spectra of each solute with the emission spectrum of the exciton state and in our paper we have suggested that naphthalene may protect FeCl3 by this mechanism. The identification of the excited states responsible for ultimate chemical charge in solid and liquid aromatic systems is difficult. Dr. Siege1 has reminded us of the possibility of triplet trap-to-trap migration in benzene crystals so clearly demon-strated by Robinson and his co-workers.Our reasons for assigning a major role to the singlet exciton rather than to this process are given in the paper. None of these reasons is conclusive but the balance of evidence seems to point to the singlet state as the more significant in our systems. Dr. Kaufmann in his comments on Dr. Cundall and Griffiths’ paper and Dr. Burns have both drawn attention to evidence that G(lB2u state of benzene) in liquid benzene is small and certainly smaller than the value mentioned in ref. (4) of our paper. Since this work is in course of publication I may mention that our most substantial evidence for G(1B2U) is the following. The graph of G(-FeC13)initial plotted against [FeClJ in liquid benzene shows a marked rise from 1.0 to 2.8 in the concentration range 50-200 M then remains constant until [FeC13] = 800 M when a second rise occurs to a value of 11 at the solubility limit of 1.4 mM.The kinetics of the system indicate that these two “ steps ” are due to reaction of FeC13 with each of two excited states of benzene which would otherwise decay by first-order processes. The first step corresponds to a state having a lifetime comparable with that of triplet benzene and hence the second step to a state of shorter life-time formed with a yield +4. We make the tentative identification of this state as so as to maintain a reasonably acceptable energy balance but we have no spectroscopic evidence which ensures certainty in this identification.I agree with the other two comments made by Dr. Burns. Dr. Ballantine’s account of Ranganathan’s interesting work emphasizes a point made above in my comment on Dr. Phillip’s results. Evidently chemical bindin 294 GENERAL DISCUSSION between the monomer units in the polyacrylamide pseudo-morph augments any energy transfer which might occur between acrylamide molecules in crystalline monomer. In this case it will be difficult to specify quantitatively how much intra-macromolecular energy transport is occurring. Prof. S . J. Wyard (University of California) said Using the techniques of electron spin resonance spectroscopy Dr. R. C. Smith and myself have developed methods of making direct measurements of local concentrations of radicals.The methods depend either on an increase of line width in the normal spectrum (g = 2) or on the intensity of a spectrum which appears at half-field strength (g = 4). If the radicals are distributed in pairs the method gives the average separation for each pair. If the radicals are distributed in clusters or in columns the method gives the average concentration within cluster or column. We have applied these methods to frozen solutions of H202 in H20 mainly in the glassy state. With ionizing radiation the local radical concentration increases with the L.E.T. of the radiation as expected but the measured values of con-centration are significantly smaller than the values generally assumed in the radiation chemistry of liquid water. The simplest explanation of this discrepancy is based on a model of local heating, in which part of the energy absorbed from the radiation which is not utilized in producing radicals raises the temperature of the material in the neighbourhood of the radicals and allows them to diffuse a limited distance before the temperature falls again.Measurements on the same radicals produced by u.-v. photolysis give some support to this explanation since with photolysis the radicals are produced in pairs from a single molecule but are trapped about 5 A apart. Other explanations are possible however. Prof. R. F. Firestone (The Ohio State University) said There are two comments which I wish to make regarding the work of Baxendale and Gilbert. The first is a question concerning the accuracy of the dosimetry.How reliable does Dr. Baxendale believe the dosages computed from the dosimeter measurements are ? I would also like to respond to the suggestion that the higher apparent G(-H20) measured with deuterium may be explained by assuming that deuterium reacts eficiently with €€-atoms OH-radicals and other unspecified primary radiolysis products whereas methanol and hydrocarbons react only with H-atoms and OH-radicals. This of course implies that G(H) is equal to approximately 8 and that other reactive species also produce HD with a yield of approximately 4 molecules/ 100eV. It seems more reasonable to propose to the contrary that organic solutes react with ionic intermediates as well as with H-atoms and OH-radicals and that deuterium is incapable of competing for ionic species at mole fractions as small as those employed in the investigations of TOH+HOH+Dz to which Baxendale and Gilbert refer.There is no reason to expect e.g. that methanol is less reactive to-ward H-atoms than is deuterium. If one examines a plot of G(H2) against methanol concentration and a similar plot for water vapour mixtures with deuterium it is apparent that methanol is not less efficient than deuterium at mole fractions of the order of 1/10,000 but that the plateau value for G(H2) is not attained until the methanol concentration is increased to about 1/30 to 1/20 whereas deuterium is completely efficient at concentrations about two orders of magnitude lower. Thus, it appears that methanol reacts very efficiently with H-atoms at all concentrations, but that it begins to compete for ionic intermediates at concentrations below the plateau region.The lower apparent G(-H2O) obtained with organic solutes may be attributed to the fact that they have lower ionization potentials and in some cases higher proton affinities than water. Thus e.g. methanol may accept protons GENERAL DISCUSSION 295 from the H30f ion in competition with H3Of neutralization giving rise to the following reaction sequence : H30++CH30H = H,O+CH,OHz (1) CH,OH;+e= CH,OH+K* (2) or CH,OHz+e = CH,O-+H (3) or CH,OM + e = CH,. +OH. +H (4) Reaction (1) is on the basis of upper and lower limits for the protons affinities of water and methano1,l between 0 and 20 kcal/mole exothermic. It is also energetically possible for methanol to neutralize the water molecule-ion precursor of H3Of, although this appears unlikely in view of the large cross-section for formation of H3O+ in the mass spectrometer.2 Such neutralization if it occurs is not sufficiently energetic to decompose the resulting neutral water molecule but gives rise to the sequence,3 CH30Hf+CH,0H = CH,OHZ +.CH,OH ( 5 ) CH30H++H,0 = CH,OH; +.OH (6) followed by reactions (2) (3) or (4).The net result of reactions of organic additives with either of the major ionic species might then be expected to be simply the forma-tion of one molecule of hydrogen for each water molecule ion formed initially. If, as Hart and Platzman have proposed,4 neutralization of the H30f ion by electrons produces two H-atoms it is clear that any solute which interferes with either the formation or neutralization of this species might be expected to reduce the yield of H-atoms.A possibly important difference between deuterium and the organic solutes is that it is not energetically favourable for deuterium to accept a proton from or to donate an electron to either of the major ionic intermediates. Reference to the curve for NH3 in fig. 3 of Baxendale and Gilbert suggests that ammonia which is not very reactive toward H-atoms contributes to formation of H2 by means of a combination of proton transfer and relatively inefficient H-atom abstraction steps. Dr. J. H. Baxendale (Manchester University) said The higher radical yield ob-served by Firestone using D2 in H20 vapour is paralleled by our most recent observ-ations using H2 and organic compounds in D20.As shown in the figure G(HD) = 10.5 using H2 but appears to approach a lower limit of 7.0 using methanol or propane. Furthermore we have observed a yield of D2 amounting to 0.5 in the presence of hydrogen or propane and to 0.8 with methanol. In the latter case D atoms may abstract in part from the -OD group to some extent and the higher yield may result from this. With H2 and C3Hg it would seem that D2 must be produced en-tirely from D20. The yield moreover is the same as that we report above for G(H2) in the H20 system containing benzene and lends credence to the suggestion that there may be a " molecular" hydrogen yield. This may originate from the ion H- as proposed by Platzman 5 H,O+e-+H-+O€I H-+H20+H2+OH-, 1 Lampe and Field Tetrahedron 1959,7 189.2 Lampe Field and Franklin J. Amer. Chem. SOC. 1957,39 6132. 3 Tal'roze and Frankevich J. Amer. Chem. Soc. 1958,80,2344. 4 Hart and Platman Mechanisms in Radiobiology (Academic Press Inc. 1961). vol. 1 p. 176. 5 Platzman 2nd Int. Congr. Rad. Res. (Harrogate 1962 Abstr.) p. 128 296 GENERAL DISCUSSION but if it is the only source the yield of H- would need to be about twice that which he suggested. 5 3 ' 8 c5 # - - - - - o - o ~ o ~ o A-A-',o/' /a- A 0 0-Mole Fracticn% 0.01 W2,C3H8 0 . 1 1.0 CH3OH 10 FIG. 1-D20 vapour radiofysis. G@2) 0 hydrogen 0.53 f0-06 A methanol 0.80f0-01 0 propane 0 4 8 f0.06 Dr. G. Scholes (University of Newcastle-upon-Tyne) (communicated) There is no doubt as Prof. Schenck has pointed out that one has to consider the role of organic peroxy radicals (R02a) in the oxidation of organic molecules in the presence of oxygen.These radicals play an important part in the radiolysis of oxygenated aquo-organic systems. For alcohols in particular we have shown1 that initial oxidation at a methyl group can lead in the presence of oxygen to the formation of hydroperoxides via RO2- radicals; e.g. the radiolysis of aqueous solutions of t-butanol results in the production of 2-hydroxy-isobutyl hydroperoxide. The effects of ionizing radiations on aqueous oxygenated methanol solutions can be interpreted on the basis of the intermediate formation of the radical -02CH2OH since formic acid is an initial product under certain conditions. Although the experiments reported by Dr.Baxendale were carried out in the gas phase at elevated temperature, one stiZl wonders whether reaction (4) as so written is an actual mechanistic step. Dr. C. B. Amphlett (A.E.R.E. Harwell) (communicated) The mechanism pro-posed by Baxendale and Gilbert for the water vapour+methanol+02 system omits the reaction 2 . CH~LOH+CH~O+CH~OH considered in connection with the oxygen-free system. Without this reaction G(CH20) = 0 when [02] = 0 whereas the minimum value of 2.2 (fig. I) is stated to include only a small contribution from reactions due to possible inleakage of air. -If we consider the following sequence of reactions : Hz-+H + OH H + CHjOH-+H2 + *CH20H OH+CH30H+H2O+*CH20H 2*CH20H-+(CH20H)2 CH20 + CH30H \ 1 Allan Lyon Scholes and Whiston 2nd Int.Congr. Rad. Res. (Harrogate 1962) GENERAL DISCUSSION 297 02+*CH2OH+CH20+ H02 (4) H + 02-+H02 2H02+H202 + 0 2 , the yields of H2 and CH20 are given by the expressions G(CH20) = }{ k3b(CH20H)+2k4(02) }. (ii) k1(CH30H) fIg(OH) +g(H) k,(CH,OH)+ k 5 ( 0 2 ) (k,,+ k3b)(CH20H)+k,(02) At high 0 2 concentrations where CH2OH radicals react predominantly via (4), G(CH2O)-g(OH) = 8.5 ; assuming g(0H) = @) we have in the absence of oxygen G(CH2O) = 3 whence k3a/k3b = 2.9. This compares favourably with other values for the ratio of combination to dis-proportionatiqn in condensed systems e.g. Taub and Dorfman 1 give absolute rates for CH3CHOH radicals which are in the ratio 4 1 ; Scholes Simic and Weiss 2 give data for aqueous solutions of CH30H containing N20 which indicate a ratio of 5-2.When the data in fig. 2 on G(H2) as a function of 0 2 concentration are inserted into eqn. (i) we obtain a mean value of ksfkl FZ 8400 (assuming the irradiation vessel to contain 2 g of water) compared with a ratio of 104 derived from the individual rate constants quoted by Gordon Hart et at. in their paper. Despite this high value the concentrations of methanol and oxygen are such that reaction (1) com-petes effectively with (5) over the range of [02] shown in fig. 2 and an oxygen con-centration of 550 pM would be necessary to reduce G(H2) to 0.01. The shape of the curves in fig. 2 requires that as 0 2 is added initially CH20H radicals continue to be produced by (1) and (2) but disappear principally via (4) with an increase in G(CH20).If we assume that at the maximum of the CHzO curve k5(02) = k1(CH30H) the value of k & ~ derived from the lH2 curve predicts that the maximum will fall at [ 0 2 ] ~ 0 . 6 pM in good agreement with the experimental data. The data are thus quantitatively explicable on the basis of the above mechanism and agree with those of other workers in condensed systems; a inore complete understanding could be obtained if glycol yields were available since these also will be very sensitive to oxygen. Furthermore the competition between (3) and (4) in the presence of traces of oxygen implies that G(CH20) should be dependent upon L.E.T. and it would be interesting to check this point. Prof. M. Magat (Faculte‘ des Sciences de Paris) said Dr. Baxendale has men-tioned in his paper theoretical “ predictions ’’ of GR values for water vapour.Such “ predictions ” made by several authors in particular by Fiquet-Fayard,3 are based on the following argument. We know from mass-spectroscopic measure-ments the appearance potentials of principal ions corresponding to the following : H@+ (normal and excited) H + OH+ + e formations as well as their relative numbers. 1 J. Amer. Chem. Sac. 1961,84,4053. 3 Fiquet-Fayard J. Chim Physique 1960 57 453. 2 this Discussion 298 GENERAL DISCUSSION This enables us to calculate the mean ionization potential. The absorption spectra are also known and a mean excitation energy can be easily deduced. The knowledge of the lowest electronic energy level allows an estimate of the energy carried by the subexcitation electrons ; combining these figures with the W value the number of excitations per ion pair being adjusted it is possible to calculate the average number of the different primary " active " species ions and radicals formed per 100 eV.The quantum yield for the dissociation of excited water molecules is assumed to be I each excited specie decomposing into H and OH. The rate of ion-molecule reactions being large it is assumed that the reactions H20+ + H2O-+H30++ OH@) and OH+ + &0-+&0++ O(b) do occur before neutralization. An assumption is to be made concerning the products of neutral-ization of H30+. Two possibilities are to be considered : H30+ + e H + HzO (1) HzO++ e 2H + OH (II) both being exothermic in the vapour phase. In order to obtain an agreement with Firestone's value of GH = 11.7rf:O.6, it is necessary to assume that reaction I1 occurs exclusively.In that case & (calc.) = 10.9. If reaction (I) is assumed GH(Ca1C.) drops to 7.3-7-6. It is tempting to extend these calculations to liquid water for which a W = 35 eV can be deduced from the G value for solvated electrons. What differences do we have to expect as compared to the vapour phase? The ionization and excitation potentials will be only slightly modified as can be inferred from the work of Vermeil et a2.1 on photo-ionization in liquid phase. Only slight errors will be introduced if water vapour values are used. However (i) reaction (11) between H30Alv. and esOlv. will become endothermic and only reaction (I) is to be considered; (ii) the quantum yield 4 for excited species will become smaller than 1.Let us assume a value of 4 = 0.5 (Matheson determined q5 = 0.44 for the 1850 A line) for all the three excitation bands. We deduce that GH(Ca1C.) = 5-6. This value is to be compared to G(-H2O) = 5.06 given by Hart and Platzman.2 This agree-ment can be-considered as satisfactory. An even better agreement can be reached if one takes into account that the dissociation process H20+*+H+OM+ is a slow one as was shown by mass-spectroscopic studies of Fayard-Fiquet and Guyon4 (isotopic effect in HDO) and Durup 3 (kinetic energy of fragments). This makes it probable that reaction (b) is suppressed and replaced by the sequence H3O'*-l-H20+H30++OH H3 O L . + e&. + H + H2O. In that case G(-H20) (calc.) = 5.1.Of course the excellence of the agreement is fortuitous. Prof. M. Burton (University of Notre Dame) said In general I should like to voice a certain scepticism regarding any conclusions reached in reference to mech-anism based on bookkeeping. By " bookkeeping " I mean the business of adding up the various possible events which are assumed to occur reconciling the result with the observations and deriving conclusions in regard to the mechanism of the processes which have occurred. The danger in such procedure is that one stops the bookkeeping as soon as the arithmetic agrees with the desired results. I have had an experience in which three successive attempts were made to examine the 1 Vermeil Muller Matheson and Leach BUZZ. Soc. Chim. Belge 1962 71 831.2 Hart and Platzman Mechanism of RadiobioZogy (New York 1961) vol. 1 chap. 2. 3 J . Chim. Physique in press GENERAL DISCUSSION 299 results of a vapour-phase radiolysis. The people doing the computation assumed successively an excited molecule mechanism an ion-molecule mechanism and a mixture of the two. With all good intention they were able to reconcile the results to all three mechanisms. I am certain that if another alternative would have been suggested that alternative might have been reconciled with the facts. I suggest to Prof. Magat that in this case had the results of Baxendale and Gilbert been somewhat different from those reported a more detailed computation would have again affected a reconciliation of " theory " with fact. The major advantage as I see it in bookkeeping is that it does force people to examine a variety of possi-bilities.The fact that agreement is finally obtained indicates for the most part, that the man doing the computation knows where to stop. Dr. W. Wild (A.E.R.E. Harwell) said Dr. A R. Anderson and Mr. B. Knight have just completed in our laboratories a study of the gamma radiolysis of H20 vapour with and without the addition of small amounts of NH3 at 120+5"C and approximately 2 atm total pressure.1 The observed decomposition of pure water is extremely low corresponding to G(H2) = 0~0086+0.0016 (mean of 15 deter-minations). This represents the lowest reproducible value which can be readily measured although individual values as low as G(H2) = 5 x 10-4 have been ob-served.The total doses range from 0.36 to 16.6 x 1021 eV g-1 which compare with the dose of 9.2 x 1021 eV g-1 at which Firestone reports a single integral G(H2) = 0.02. Higher yields were obtained in less adequately cleaned irradiation vessels. The low yield does not entirely preclude the possible formation of hydrogen in low yield by a molecular detachment process as a steady state may be established at a very low concentration of hydrogen and their reported G value could cor-respond to slight deviations from the true steady-state value on repeated irradi-ations. The amount of hydrogen from the radiolysis of pure water vapour is negligible in comparison with that obtained in the presence of NH3. With the addition of NH3 (<Om05 mole %) in the absence of 0 2 the only products observed are N2 and H2 and the excellent mass-balance G(H2) = 2.38+_0.23, G(-NH3) = 1.62+0.05 G(N2) = 0.80+0-05 (independent of total dose) pre-cludes the presence of any other products in significant yields.Hydrazine is not detected (<3 x 10-5 mole %) even at low doses and is either not formed by the dimerization of two NH2 radicals or is destroyed immediately by H and NH2 radicals. No significant differences in the initial yields of hydrogen were found at dose rates varying from 0.15 to 3 Mrad h-1 with NH3 concentrations of 0.005 and 0.05 mole % nor with irradiation cells of 100 and 700 ml volume. A tentative radical mechanism is postulated as follows : H20+ H + OH OH+NH3-+NHz+H20 H+NH2+NH3 H+H+H2 NH2 + NH~+N~H.Q NH2 + NH2+NH3 + NH NH+NH+Nz+H2 H + N ~ H ~ + N ~ H ~ + H Z NHz+N~H~+N~H~+NH~ N2H3 + N2H3+2NH3 + Nz (H+NH3+NHz+Hz) 1 to be submitted for publication shortly 300 GENERAL DISCUSSION This stoichiometry indicates that G(-H20) lies between 6.4 and 7.2 which is close to Baxendale and Gilbert's value and indeed approaches the value for gross water decomposition in the liquid state (- 5.6 for acid solutions).The hydrogen abstraction reaction (1 1) is insignrficant at the concentration of NH3 employed. Schiavello and Volpi 1 give an upper limit for kll = 1.6 x 10-17 cm3 molecule-* sec-1 at 150°C which indicates that at 5 x 10-2 mole % NH3 and a dose rate of 0.1 5 Mrad h-1 the maximum rate of reaction (1 1) will be < 10 % of that of reaction (4). However Baxendale and Gilbert's data (fig.3 of their paper) suggest that kll< 10-17 cm3 molecule-1 sec-1 for with the upper limit quoted the hydrogen yield should have reached a constant value at the lowest concentration of NH3 (6 mole %) which they used. The consideration of primary yields using alternative modes of disappearance of the ionic precursors of H and OH radicals has not yet been completed. Dr. A. H. Samuel (Stanford Res. Imt. Calfornia) said One fact that emerges from the paper by Gordon Hart Matheson Rabani and Thomas is that rate con-stants for the reactions leading to " radical " combination which had previously been estimated at about 6 x 1011 M-1 sec-1 are lower by a factor of 2-3 and even more for 2elq-+H2. Diffusion calculations which are still necessary to explain the L.E.T.effect and probably also the gamma-ray molecular yields will be affected. Un-fortunately there are still a number of adjustable parameters including the average number of '' radical " pairs per spur and the diffusion constants; but it seems that it will be necessary to reduce the spur diameter. It would then follow that the electrons are moderated even more rapidly than has been thought to an energy at which they can be captured by suitable constellations of water molecules. This is in complete disagreement with the results presented by Dr. Wyard, showing large inter-radical distances in ice at 77°K. Perhaps this discrepancy may be resolved by considering that more than one-half the radicals observed at 4°K disappear at 77°K; and those that disappear will tend to be those closest together.It also appears that the PI yield observed in liquid water is mostly due to the re-action of radiogenic H30+ and eiq rather than residual electron recapture. In either case one would expect a marked L.E.T. effect on the ratio G(H)/G(eG)-it should increase with increasing L.E.T. The fact that H atoms rather than solvated electrons are observed in ice (pre-sumably they are formed at all temperatures though they are stable enough for ob-servation only below ca. 30°K) seems to indicate that the electron in ice does recombine with the positive ion. Perhaps the constellations favourable to trapping do not exist in the crystal. A test of this view would be to observe the temperature coefficient of the ratio G(H)/G(ei). I would predict that it would grow as the water temperature drops to 0°C and below because non-icelike formations should be getting rarer.Prof. F. S. Dainton (University of Leeds) (communicated) Reference has been made in the paper by Gordon et al. to ionic strength effects (see ref. (6) and (7)) and I would like to draw attention to the possibility of using this effect for a purpose other than that of finding the sign and magnitude of the charge on the solvated electron. The existence of an ionic strength effect depends upon the presence in the chemical potential of eFq of an electrical energy component arising from the potential exerted at the electron by its own ion atmosphere. If the electron reacts with a solute before the equilibrium distribution of ions of opposite sign can be established around it then ionic strength effects should be largely absent.Con-1 SchiaveUo and Volpi J. Chem. Physics 1962,37 1510 GENERAL DISCUSSION 30 I versely in a solution of constant ionic strength diminution of the lifetime of the electron by increasing the concentration of a solute with which it is reacting should cause the apparent rate constant to change progressively towards the value appropriate to zero ionic strength. Dr. Coyle has recently carried out experiments which seein to confirm this prediction. He studied nitrogen yields obtained from deaerated aqueous solutions of nitrous oxide and sodium nitrite at natural pH. In the early stages of this reaction before the nitrate ion coiicentration has reached a sufficient level to interfere the only reactions are : e& + N,O-+N + 0-( -+ OH) e,i+ NO -+NO;-OH+NO,+OH- +NO2 NO +NO$-+2NO, 2NO2 + H2 O+HNOz +HNO, Hence G(N2)-’ = Gi-l{l +k2[N0,]/k,[N20]}.Dr. Coyle found that G; = 2.85 0.1 and loglo(k2/kl) = logl0(O-4 f 0.1) + (I *O8 & 0.1) d i / ( l + &) provided that “201 +0.5[NO,] <2 mM. However if the ionic strength is kept constant at 0.1 5 and the solute concentrations progressively increased the experimental value of k2/kl decreases from 0-75 to 0.4. This effect is shown in the figure where the value of k2/kl is plotted against 0.693/(k1[N20] +k1(k2/k1)[NO;]) taking kl = 1.1 x 1010 M-1 sec-1. If this effect is due to the explanation suggested then the relaxation time z of the ion atmosphere of the electron at p = 0.15 using lithium sulphate as neutral salt is about 3 x 10-9 sec.Two coiiclusions may be drawn from this result. First since T is likely to decrease with increasing ,u it i 302 GENERAL DISCUSSION important not to seek ionic strength effects at high ,u or high solute concentration and generally Ck,[Si] should be less than about 5 x l o 7 sec-1. Secondly it is not surprising that at pH 2 k(e-+ H+)/(e-+ N20) is independent of added neutral salts. Dr. G. Scholes (University of Newcastle-upon-Tyne) said We have determined the relative reactivity of various solutes towards the radiation-produced electrons using the nitrous oxide system discussed in our paper. For many of the solutes investigated there is good agreement with relative rates obtained from the data of Gordon et al.Nitrate ions have a high reactivity towards (H20)-. The reduction of nitrate to nitrite has been studied in the radiolysis of sodium nitrate solutions containing isopropanol (10-1 M). The observed yields were G(N0;) = 2.52 and G(N0;) = 2.86 at sodium nitrate concentrations of 10-3 M and 10-2 M respectively. Addition of other electron acceptors to the nitrate + isopropanol solutions led to a reduction in the radiolytic yield of nitrite. If relative rates compared to nitrate were then calculated assuming simple competition in which nitrite is formed only as a con-sequence of reaction of (H20)- with nitrate it was found that there was a good measure of agreement with the corresponding relative rates obtained by the N2O system. This is shown by the following results (obtained in collaboration with Dr.A. Appleby) where the reactivities are normalized relative to k((H20)-+NzO) = 1-00. The nitrate + isopropanol system therefore suggests itself as a relatively simple one for the determination of reactivity of a solute towards (H20)-. relative reactivity R in the N20 system solute S "201 - 16mM IS] mM R in the NO system [ N o t ] = 1 mM [isopropyl] = 100 mM [Sl mM R thymine 23 1.67 fO.15 0.44 2-08 f0-30 NO 10 1-17 f0-15 10 a 1.19 f0-15 Fe(CN) 2 - 10 0.89 f0.07 1 0.88 f0.08 acetone 10-104) 0.67 &0.07 2 1.01 f0.10 chloroacetate 5-200 0.19 f0.01 5 0- 14 f0-02 solution contained 16 mM N20 Dr. 6. E. Adams Dr. J. M. Baxendale and Dr. J. W. Boag (Mt. Vernon Hospital, Middlesex) (communicated) We have some results which parallel those of Taub, Sauer and Dorfman on the formation of the ion from diphenyl.We used benzo-phenone to capture the electrons. Transient absorption spectra have been obtained by pulsed radiolysis and Bash spectroscopy of de-aerated solutions of benzo-phenone,l These spectra demonstrate the formation of the ketyl radical ion (CsH&C-O- as a primary product resulting from electron attachment to benzo-phenone. In n$utral solution the rapid neutralization of the ketyl ion produces the radical (C&&C-OH and this is shown by the corresponding change in the optical absorption spectrum. Neither of the spectra due to these species is observed in acid solution. In pure de-aerated alkaline methanol a transient absorption is observed with a maximum at 6200A.We attribute this absorption to the solvated electron. In the presence of benzophenone the solvated electron spectrum is absent and is replaced by the spectrum of the ketyl ion. As in the aqueous system the ion is rapidly converted in neutral solutions to the neutral radical. 1 Adams Baxendale and Boag submitted for publication in Pruc. Roy. Suc. A GENERAL DISCUSSION 303 These observations show that in neutral and alkaline solution the solvated electrons react directly with the solute whereas under acid conditions they are first converted to hydrogen atoms which then react differently with the solute. Prof. F. S . Dainton (University of Leeds) (communicated) Mr. 3. Teply in my laboratory has recently irradiated organic glasses of methanol containing 2 % water at - 196°C and the results indicate that the trapped electron in this matrix is related to the solvated electron in liquid methanol in the same way as the trapped and solvated electrons in MTHF are related (see comment on paper by Hamill et al.).On irradiation the glass becomes red and two absorbing species are formed giving rise to the absorption spectrum shown in fig. 1. The band A with a peak I* O.D. A A(& FIG. 1. at 5330A is that of the trapped electron and has associated with it an e.s.r. absorption which exhibits marked r.f. saturation effects. Both the A and B bands increase linearly with dose and Ge-~?32:5D = 3.46 x lo4 cm-l M-' (100 eV)-' whilst G B E2040 - 1950 = 1-24 x lo4 cm-' M-' (100 eV)-'. The A band and the appropriate e.s.r.signal can be photo-bleached by light in the region of the absorption maximum and this causes slight enhancement of the B-band. I @ O.D. A 0-- 190 - 180 - 170 T C ) FIG. 2. Warming causes the rapid fading of the A-band in the temperature interval - 175 to -171" whilst the B peak fades a few degrees higher (see fig. 2). When FeC13 i 304 GENERAL DISCUSSION present the total absorption at -196°C after irradiation is merely the summation of the A B and unchanged Fee13 bands but on warming the A-band fades at about - 178" without any diminution of FeC13 absorption and it is not until - 167°C that this latter band begins to undergo detectable reduction. We tentatively con-clude that the electron in MeOH glass may be trapped closer to positive ions than to the FeCl3 molecules whose average separation was about 70 A in these experiments, and that on thermal or photochemical stimulation the mobilized electron returns to a positive ion.This is supported by the fact that the thermal bleaching between - 172 and - 168°C is first-order in optical density of the A band with k = 2 x 1016 exp (- 8.8 + 2.3 kcallRT) sec-1. Tentatively we write : C H 3 0 H g l ~ + C H 0 H + epl+ aCH20H + CH30H* - 196 Vd -1 or I :;Or CH30H + H -i CHy + *OH or CH2OH + H2 - 167O R + FeC1,-+RCl + FeC12 Dr. J. B. Keene (Christie Hospital and Holt Radium Inst. Manchester) said: Pulse-radiolysis experiments using techniques similar to those of Gordon et al. and Taub et al. are being carried out in our laboratories. The radiation source is a 4 MeV linear accelerator which produces 0.1 A 2 psec electron pulses.The ab-sorption cell has a 16 mm optical path and the region traversed by the light beam can be uniformly irradiated in one pulse with any dose in the range 10-10,000 rads. A variety of light sources is available and the time variation of optical transmission, before during and after the pulse is observed using a photo-multiplier or photo-diode an amplifier and a Tektronix model 545 oscilloscope. Particular attention has been paid to reducing electrical and optical interference and this has made possible the measurement of optical absorption during the pulse itself. Full details of the apparatus have been described.1 Early work with this apparatus2-4 showed the existence of easily detectable transient optical absorptions in a variety of aqueous solutions.In particular the strong absorption in the visible now known to be due to hydrated electrons was observed3 5 The characteristics of this absorption were in excellent agreement with data obtained at the same time by Boag and Hart697 who investigated the effects of a number of solutes and produced good evidence to suggest that the ab-sorption band was due to hydrated electrons. Our original experiments involved only water containing various amounts of dissolved oxygen ; they were however, sufficiently quantitative to make possible an early estimate of the rate constant for the reaction of hydrated electrons with oxygen. The value obtained was 1.5 x 1010 M-1 sec-1.5 1 Keene in preparation.2 British Empire Cancer Campaign 38th Annual Report part 2 1960 p. 498. 3 Radiation Efects in Physics Chemistry and Biolugy (North Holland Publ. Co. 1963) p. 38. 4Keene Nature 1960 188 843. 5 Keene Nature 1963 197 47. 6 Boag and Hart Nature 1%3,197,45. 7 Hart and Boag J. Amr. Chem. SOC. 1962,84,4090 FIG. 1.-Oscilloscope traces obtained with deaerated water for various single 2 psec pulse showing the time variation of hydrated electro 6 0 0 rads c 5430 A t i m e scale l l s e c per large division FIG. 3.-Oscilloscope trace obtained with 50x 10-6N sulphuric acid. The recovery to 100 % after the pulse is much faster than for a similar dose in water and is due to reaction of hydrated electrons with Hf GENERAL DISCUSSION 305 During the past year we have carried out further work on the reactions of the hydrated electron.1 Fig.1 shows typical oscilloscope traces in pure deaerated water for a range of doses. The transmission drops during the pulse due to production of hydrated electrons and recovers back to 100 % soon afterwards. Semi-log plots of data of this kind are shown in fig. 2. With the larger doses the reaction is of mixed order since the electrons react with several of the radiation products. With sufficient low doses the reaction becomes first order with a half-time of about 25 psec. time after end of pulse microseconds hydrated electron concentration of about 1 pM. FIG. 2.-Semi-log plot of data of the type shown in fig. 1. A density of 10-2 corresponds to a Under these conditions the concentration of radiation products is so low that the electrons react only with water.Fig. 3 is for 50 x 10-6 N sulphuric acid at inter-mediate dose. Reaction is prodominantly with H+ the reaction is first order (see fig. 4) and the data can be used to calculate the corresponding rate constant. Curves of this kind were obtained for several different solutes. Measurements were also made with higher concentrations in order to check the independence of rate con-stant on concentration. In these cases the recovery to 100 % after the pulse was too fast to measure. However under these coriditions an equilibrium is established within a few tenths of a micro-second after the start of the pulse such that the rate of removal of electrons by reaction is equal to the rate of production. Measure-ments were made of the equilibrium levels and used to calculate rate constants.Some rate constants obtained in these ways are shown in table 1 where they are corn-pared with the figures of Gordon et al. and Taub et al. presented at this meeting and 1 Keene submitted to Rad. Res 306 GENERAL DtscuSsION TABLE 1 . ~ O M P A R I $ O N OF RATE CONSTANTS REPORTED BY GORDON et d. AT THIS MEETlNG AND BY DORFMAN AND T A U B ~ ~ WITH THOSE OBTAINED IN OUR LABORATORIES. THE VALUES ARE IN UNITS OF 1010 M-1 sec-1 WITH THE EXCEPTION OF k(e,i+H20) WHICH IS IN see-1 reaction Gordon et al. and Dorfman Keene eLq + H+ 2.36 f0.24 2.06 f0.08 C 2.26f0.21 b* 1 eaIq + H202 1.23 -foe14 1.36 20.20 eaq + 0 2 1.88 f0-2 2.16 50.25 elq + N20 0-87 fO.06 0-33 f0.05" ea-q+H20 about 2x 104 (2.7 h0.15) x lo4 less than 4x 104 12 a pH 4.0-4.6.b pH 4.147. C pH 2 1 4 . 3 * A redetermination of the N20 rate constant has now been made. Care was taken to ensure full It is concluded that saturation of the solution; the value obtained was (0.56 i 0-20) X 1010 at 17°C. the original value is in error to lack of saturation. elsewhere. Agreement is good in all cases wth the exception of nitrous oxide. For reaction with H+ no change in rate constant was found over a pH range 2.1-4.3. Measurements at a number of wavelengths were made in order to get the hydrated I I I I 3 I o 2 time after end of pulse microseconds FIG. 4.4emi-log plot of data of the type shown in fig. 3; 50x 10-6 N sulphuric acid; 5430 A; 680 rads. electron absorption spectrum. This is shown in fig.5 in terms of the product of G-value G and extinction coefficient E . It has a maximum at 7200 a. At this wave-length GE is 40,500 (k2800) and if a G-value of 2-5 is assumed the extinction co-efficient becomes 16,200 (-1 1200). 1 Dorfman and Taub J. Amer. Chem. Sm. 1963 88,2370 GENERAL ~ I s c U ~ S I O N 307 The pulse radiolysis technique is now being used in our laboratories in in-vestigations of a wide range of organic and inorganic chemical systems. Among these are (a) the oxidation of ferrous sulphate which has been found to proceed in several distinct stages in tirne,l. 2 (b) reactions of various metal ions with hydrated electrons,3 and (c) reactions in benzene aniline anisole ethyl iodide iodine in cyclohexane and various amino acids.4 1 I I , Dr.J. H. Baxendale (Manchester University) (communicated) If the first-order decay of the electron absorption in water observed by Dr. Keene is the rate of the reaction, H20+e&=+H+OH-, it allows a maximum value to be estimated for the hydration energy of the electron in water. He obtains a half-life of 25 psec for the above reaction. Taking a pre-exponential factor of 1011 would give an activation energy of 12 kcal which sets an upper limit to the endothermicity of the reaction. Using the estimated value of 260 kcal for the absolute hydration energy of the hydrogen ion 5 and 3 kcal for the heat of solution of the H atom the endothermicity of the reaction is calculated to be Se = -44 kcal where Se is the heat of hydration of the electron i.e. Se <2.5 eV.Dr. L. M. Dorfman (Argonne NationaE Laboratory) said I think it is advisable that I present briefly some recent experimental results by Taub and myself 6 which bear on one of the reactions of the hydrated electron. The results which are re-ferred to in the paper by Scholes et al. are concerned with the bimolecular reaction of two hydrated electrons e; +e& the nature of which has been assumed in the paper by Gordon et al. 1 Radiation Eflects in Physics Chemistry and Biology (North Holland Publ. Co. 1963) p. 38. 2 Keene submitted to Rad. Res. 3 Baxendale Fielden and Keene Proc. Chem. Soc. 1963,242. 4 Baxendale Capellos Davies Ebert Fielden Francis Gilbert Keene Land Nosworthy 5 Randles Trans. Farahy Soc. 1956 52 1573. 6 Dorfman and Taub J. Amer. Chem.Soc. 1963,85,2370. and Swallow to be submitted to Nature 308 GENERAL DISCUSSION The experiments were designed to show whether this reaction gives directly a molecule of hydrogen as has been proposed,l or whether freely diffusing hydrogen atoms may be liberated, eai+ea< = H2+20H-, ea,l + eai = 2H(or H) + 20H-. Isotopic experiments with 0.5 M CzH50D in DzO were carried out with a high intensity pulse from our electron linear accelerator. The solutions were strongly basic pH 12-13 to remove the hydrogen ion in the process H++OH-. The hydroxyl radical is removed by the ethanol forming a-ethanol radical. Direct observation in neutral solution indicates this radical does not react readily with the electron. Any hydrogen atom is similarly removed by the ethanol.Under these conditions at high pulse current we find the decay of the electron to be second order, the reaction presumably corresponding to the bimolecular reaction of two hydrated electrons. If a molecule of hydrogen is formed we expect and If freely diffusing hydrogen atoms are formed we expect G(D2) E 0.5 = GD,, G(HD) E 3.5 = GD+ Gelq, The results of three such experiments showing excellent reproducibility are Thus we must conclude that freely diffusing hydrogen atoms are not formed in the bimolecular reaction of two hydrated electrons the product apparently being a mole-cule of hydrogen. Our results also indicate that the rate constant for the reaction is 2 k ~ 1 x 1010 M-1 sec-1 at 23°C. In the strongly basic D20 solutions the data also give Ge;- >2-9 moleculesll00 eV.Dr. J. H. Baxendale (Munchester University) said In connection with the e;,+eLq reaction it would seem to me that although Dorfmann’s work showed that molecular hydrogen ws the product it does not of itself prove that the molec-ular hydrogen yield originates from the same reaction as is assumed for instance, in the paper of Scholes et al. Dr. L. M. Dorfman (Argonne National Laboratory) said Our experiments refer to the total forniation of hydrogen in the reaction eLq+eCq at high concentration of the electron and not separately to the molecular yield. Moreover our fast reaction observations of the second-order decay begin at 0.2 psec after the pulse. In this time range the spatial identity of the spurs has long since vanished and the observations clearly refer to the homogeneous reaction of the electrons in the bulk of the solution.G(D2) = 1.96 G(HD) = 0.63 G(H2) = 0.05. To the extent that the reaction, eai+eai = Hz+20H-may now be considered as a competitive bimolecular reaction the result may be considered pertinent to the molecular yield But I should stress that our experi-ments in no way provide independent support for the validity of a spur model. 1 Hayon and Weiss Pruc. U.N. Int. Cod. Peaceful Uses Afumic Energy (Geneva 1958) 29 80 GENERAL DISCUSSION 309 Dr. G. E. Adam and Dr. J. W. Boag (Mt. Vernon Hospital Middlesex) (com-municated) Using the flash-radiolysis techniques 1 we have also observed a transient absorption spectrum in deaerated ethanol irradiated by a single 2 psec electron pulse.The absorption band which extends throughout the visible region of the spectrum has a maximum at 6700 A and is illustrated in fig. 1. The absorption is very weak in pure deaerated ethanol but is considerably enhanced by an increase in alkalinity (sodium ethoxide). The spectrum is quenched by sulphuric acid and by anthracene a typical electron scavenger. We believe the absorption to be that of the solvated electron similar to that observed in methanolic solutions,2 and analogous to the hydrated electron observed in irradiated water.1 Dr. E. Hayon (SacZay) said In the work mentioned earlier in this discussion on the e.s.r. of ethanol glasses we determined the yields (+ 30%) of total para-magnetic species. G(T) is 8 in unbuffered and alkaline glasses and 10 in acidic glasses.G(e-) is 2 3 and nil in unbuffered alkaline and acidic glasses respectively. In view of the similarity of yields of products obtained from the radiolysis of liquid and glassy ethanol I would like to ask Dr. Dorfman what value of G(CH3CHOH) he obtains on pulse radiolysis of liquid ethanol. Dr. L. M. Dorfman (Argonne National Laboratory) said Our estimate of G(CH3CHOH) in ethanol was briefly discussed by Taub and myself.3 We have not obtained a reliable value for the yield of the a-ethanol radical in ethanol because of the following consideration. We obtained the molar extinction coefficient of the a-ethanol radical in aqueous solution only. The molar extinction coefficient in ethanol may possibly be different because of solvent effects. If we assume the mo.lar extinction coefficient to be the same in ethanol as in water we obtain G(CH~CHOH)E~ to 9 in ethanol.This is not much higher than your data would indicate where I gather the yield is about 6 to 7 on the basis of a yield of 8 for total reducing species. Dr. G. Meshitsuka (Tokyo Metropolitan Isotope Centre) said The results which seem to indicate that the reactions of hydrated electrons are involved as the im-portant processes are being obtained in our laboratory. The yield of 12 from the radiolysis of KI in aqueous solutions shows the minimum in the region of pH2 to 3 in deaerated system. The results can be explained as follows: in lower pH region: eai+H+-+H Hl+I-+H,+I. H+ H+ -+HZ With increasing pH the reaction becomes more important and consequently the 12 yield decreases.In the higher pH region elq reacts with H202 to increase the I2 yield : H+I+H'+I-e& + H202-+ OH' + OH OH+ 1-4 OH- +I. The fact that the amount of the increase of I2 yield in higher pH region is equal to that of the decrease of H202 yield in the same region also can be explained by the proposed mechanism. 1 Hart and Boag J. Amer. Chem. SOC. 1962,84,4090. 2 Adams Baxendale and Boag Proc. Ray. Soc. A in press. 3 Taub and Dorfman J. Amer. Chem. Soc. 1962,84,4053 310 GENERAL DISCUSSION Prof. F. S. Dainton (University of Leeds) (communicated) In solutions con-taining ferrous and cupric sulphates at pH23 to which nitrous oxide is added the participating reactions are : ea~+bi,O-+N,+O-(-,OH) (1) eiq + Cu" +Cd( + Fe'" +Cur' + Fe") H,O,+Fell-+OH-+ OH+Fe"'.(6) Consequently measurements of G(N2) and G(Fe3+) for solutions of different com-position enable k l k2 k3 kq to be determined. These have been made by Dr. Smithies and Mr. Hey and at pH 3.15 and p = 0.01 ; these ratios are 1 5.5 1.4 2-1. When allowance is made for the slightly higher ionic stren@h employed by Scholes et al. the ratio kZ/kl = 5-5 is in good agreement with their value of 4.7. When cupric ion is excluded from the system and ferric sulphate added measure-ments of G(H2) G(N2) G(Fe3+) (at time zero and time 4 m when the initial ferrous salt concentration is low ca. 10-5 M) indicate unambiguously that the hypoferrous species FeI first proposed by Dainton and Jones 1 to account for reactions occurring in acid glasses containing ferrous ions has an independent existence and may either be oxidized by hydrogen ion (reaction (7)) or ferric ion (reaction (8)).Mr. Coyle has found that kslk7 = 45 and decreases with increasing sulphate ion concentration. Fe + HG 3 Fe ' I I + H Fe1+Fe"'+2Fe" The same analysis also leads to relative reactivities of Fe* and Fen towards hydrogen atoms identical with those found by Rothschild and Allen.2 Of greater interest is the fact that whilst these studies and those of Dr. Watt 3 give values of kl k3 k4 = 1 2 2 at p = 0 the value of kllk4 not only differs from that obtained by pulse radiolysis but shows no dependence on ionic strength although we have confirmed that k(H*+e;q)/k(H20;!+e;q) has the expected dependence on p. Dr. Watt also finds that k(eLq +ferricyanide)/k(eLq + H&) at p = 0 has a value almost three times that obtained by pulse radiolysis.These matters are mysterious and underline the need for reactivity ratio measurements even though pulse radiolysis studies may seem to render them redundant. Dr. E. Hayon (SacZay) said In order to elucidate the nature of the primary reducing species formed and the possible effect of 0 2 on their formation in the radiolysis of water neutral air-free aqueous solutions of H202 were recently studied. The following reactions must be considered : OH+Hz02+HOz+HzO (H20)-+H202+OH+HzO+OH-HO + H02-+H202 + 0 2 (HzO)-+ 02-+H20 + OH-H+OZ+HOz 1 Dainton and Jones Rad. Res. 1962 17,388. 2 Allen and Rothschild Rad. Res. 1958 8 101. 3 Dainton and Watt Pruc.Roy. SOC. A 1963 GENERAL DISCUSSION 31 1 Taking G(0H) = 2.25 G(red.) = 2.85 G(H2) = 0.45 and G(H202) = 0-71 if one considers that all the reducing species are formed as electrons (mechanism A) G(-H202) should be equal to 4-77 in lO-4M KBr. Alternatively if H(G(H) = 0.55) as well as (H20)- are formed (mechanism B) G(-H202) = 3-67. It was found experimentally that the initial yield G(-H202) = 3.6+0.1 indicating H atoms and (€320)- are both formed initially on irradiation. However as the de-composition of H202 proceeds sufficient 0 2 is formed to react with the precursors of the H atoms such that all the reducing species are formed as electrons G(H20)- = 2.85. Fig. 1 shows the decomposition of 232pM H202 containing lO-4M KBr I00 -5 0 -I dose 1018 eV/g FIG.1 .-y-Irradiation of neutral air-free solutions of H202. as a function of dose. The full curves are calculated according to mechanism A and B and demonstrate the effect 0 2 has on the decomposition of H202. The H atoms cannot be formed by the reaction (H20)- + H++H + H20 occurring in the spurs since it would be difficult to see why H202 present in concentrations 10-15 times greater than 0 2 cannot compete with Hf whereas 0 2 does when the reaction rates with (H20)- of 0 2 H202 H+ = 1 0.5 1. It seems probable that H20* is the precursor of H atoms H2O*-+H+OH and it is suggested that the role of 0 2 is to quench H20* by a mechanism involving an electron transfer Dr. G. Scholes (University of Newcastle-upon-Tyne) said With regard to the comments of Dr.Hayon I would like to stress that our experiments with aqueous solution of formate and oxygen show that the independently-produced atoms (H") compete between these two solutes. This observation is not compatible with the view that an excited water molecule can react competitively with oxygen and H2O" + 02+(H20)-+ (02)-312 GENERAL DISCUSSION formate but on the other hand in uacuo decompose to give hydrogen atoms prior to dehydrogenation of the organic solute. The experiments of Hummel and Allen using aqueous (ethanol + oxygen) solutions have been recently repeated by Mr. R. L. Willson of our Laboratory. Ethanol solutions of varying concentration (1 0-3 M- 1 M) containing oxygen(2-7 x 10-4 M) were irradiated with COG0 prays. The yields of hydrogen are accounted for by simple competition between oxygen and ethanol for Ha and lead to value of G(Ha) = 0.85+0.05 and to the rate ratio k(H'+02)/k(Ha+ethanol) = 350+35.The yields of hydrogen peroxide varied little over the above solute concentration range (G(H202) = 3.5-3.7) in agreement with the findings of Hummel and Allen. Increasing the ethanol concentration to high values gave larger yields of hydrogen than would be anticipated from the above rate ratio. However the yield of hydrogen peroxide also increased ; e.g. irradiation of solutions containing 5 M ethanol + 6.5 x 10-4 M oxygen gave the values G(H2) = 2.24 and G(H202) = 4.20. Thus under these conditions there is an increase in the extent of scavenging of primary recom-bination processes. The increased hydrogen yield may well be partly associated with scavenging of hydrogen atoms normally undergoing primary back reactions.Dr. E. Hayon (Saclay) (communicated) Early this year A. Reiffsteck deter-mined the yields of H2 by the method of gas chromatography produced in y-irradiation of aerated solutions of ethanol as a function of alcohol concentration. Yields of G(H2) obtained ranged from 0.45 in 2 x 10-3 M ethanol to 1.40 in 1.0 M solutions. Solutes known to react effectively with electrons but poorly with H atoms such as KNO3 and acetone were added separately to aerated 0.4 M ethanol, G(H2) was reduced from 1-20 to 1 .O in 10-2 M KNO3 and 10-2 M acetone and to 0.130 in 10-1 M KNO3 and 10-1 M acetone. From the results in the presence of 0 2 , KN03 and acetone and known rate constants we concluded that H atoms must be the reducing species for which ethanol is competing with these solutes.This is in agreement with the findings of Dr. Scholes. This seems to be contrary to the results on the hydrogen-peroxide system 1 where it is suggested that oxygen reacts with the precursor of the H atom to give a reducing species which is not an H atom. This difference would be explained by saying that the organic solutes (ethanol in this case) competes with 0 2 for the H atom precursor such as to mask the effect 0 2 would have had in its absence. The nature of the H-atom precursor is not known with certainty. It may be an excited water molecule but it is quite clear that it is not formed from the reaction of an electron with a proton in the spurs,2 (H20)-+ H++H + H20.Dr. John T. Allan (Mellon Ins-t. Pittsburgh) said With relation to the paper of Dr. G. Scholes et al. the following results were obtained for aqueous solutions of N20 and 2-propanol irradiated with 2.5 MeV electrons. The yields of Nz Hz, H202 and acetone in 10-2 M 2-propanol solutions were measured as a function of N20 concentration pH and radiation intensity. The ratio of the rate contants for reactions of various solutes with hydrated electrons have also been determined; for the competing reactions (8) and (l) kslkl = 1-65. In neutral solutions containing -3 x 10-3 M N20 the yields of eLq and Ha were respectively G = 2-80$.0-1 and G = 0-60_f0.05. As the N20 concentration was increased to 2 x 10-2 M scavenging of molecular hydrogen precursors occurred to the extent of Gm(H2) = 0.25; reaction (2) is thus a major source of molecular hydrogen.Scavenging of the hydrated electrons which normally enter into reactions (3) and (4) is apparently negligible over this N20 concentration range. The yields of acetone as a function of pH and N2O concentration indicate that the oxidizin GENERAL DISCUSSION 31 3 9 0-g 4 5 -3-species produced in reaction (1) is at least stoichiometrically equivalent to the OH radical. At pH = 13 the chain-reaction (10) was observed and G(N2) was determined at various radiation intensities in solutions of 10-2M N20. The results indicate that reaction 9(a) which occurs in addition to reaction (9) is also a chain-breaking process : 0- 0-I I CH34-CH3 +OH+ CH2-C-CH3 ( 9 4 I H I H 1 .1 1 1 Prof. F. S. Dainton and Mr. D. C . Walker (University of Leeds) (communicated) : A system similar to that reported by Scholes et al. has been studied and also found to exhibit a radiation-induced chain reaction in aqueous solution at high pH. The system studied was an alkaline solution containing N20 and H2 gases. Fig. 1 shows the dependence of G(N2) on pH at constant N2Q and HzO pressures. At pH> 13, G(N2) was independent of pH and proportional to H2 giving a value of 90 at 1 atm pressure. The observations are consistent with the following reaction sequence which is to be compared with reactions (9) and (10) in their paper, e,; + NzO-+Nz + 0-O-+H,+H+OH’ or ea; H+OH-+ea\ (3) At pH> 13 and P N ~ o > 300 mm reaction (2) is the rate-controlling step and it seems likely that the pK for step (3) has a value of 12.75.If the chain-terminating step were OH+OH-+H202 (or O-+O-+Oi- at this pH) the chain length should show an intensity exponent of 0.5. However G(N2) was found to be independent of intensity and since added H2Oz diminished the chain length it was concluded that radiation-produced H202 was involved in the termination process, 0-+H,O,*O2 +H,O. (4 314 GENERAL DISCUSSION Dr. N. Getoff (Reaktorzentrum Seibersdorf and Imtitut fur Radiumforschung und Kerphysik Wien) (partly communicated) Studying the formation of oxalic acid in deaerated aqueous solutions of 0.1 M formic acid and formate saturated with gaseous CO2 up to 1 atm by WO-y-irradiation (dose 1.5 x 1018 eV/ml) a strong pH-dependence was found.Some typical results are presented in table 1. It can be TABLE ~.-~H-DEPENDENCE OF THE CARBOXYLATION OF FORMIC ACID TO OXALIC ACID BY GAMMA IRRADIATION PH 2.8 3.1 4.8 4.9 5-6 6.2 7.8 8.3 10.8 G-value of G-value of oxalic acid oxalic acid (1) (2) - 20 1 -72 -- 2- 50 2-4 1 - 7.6 7-18 -- 8.0 7.57 -- 9.1 The results given in column 2 have been obtained in co-work with Mr. F. Giitlbauer. A newly designed analytical method using ion exchange (Dowex 50 x 8) for quantitative separation of the oxalic acid has been employed in this case. A detailed report on the carboxylation of formic acid will be published elsewhere. seen that in acidic solutions up to pH about 5 the G-value of the oxalic acid formed, is 1.7 to 2.5.However between pH = 5 and 6 it rises very rapidly and reaches a value of about 7.2-7.6. This effect could be explained to a certain extent by the dissociation of the formic acid (pK = 3.75) and by the reaction of CO2 with the hydrated electrons (e;'). As already shown by Czapski et aZ.1 the formate ion (HCOO-) reacts efficiently with the H-atoms yielding the (CO;) radical ion. At higher pH-values this radical ion (Cog) may dimerize to oxalate. A G-value of 1-35 for oxalic acid was found by Hardwick2 when an aqueous solution of 0-1 M sodium formate was irradiated with 3 MeV electrons. On the other hand in neutral and alkaline solutions the reduction of carbon dioxide by (elq) becomes very rapid so that it results in an additional formation of (COY). The (CO;) radical ion in both cases-be it obtained as a result of the dissociation process of the formic acid or as a reaction product of the hydraded electrons with carbon dioxide-leads to the formation of oxalic acid.From experi-ments with evacuated aqueous bicarbonate + formate solutions and others saturated with carbon dioxide up to 1 atm it was found that the hydrated electrons react rather with the C02 than with the bicarbonate ion. This is in agreement to the results obtained by Scholes et aZ. as reported in their paper.3 The reaction mechanism to explain the relatively high yields of the oxalic acid obtained in our experiments in neutral and alkaline solutions is not yet fully under-stood. Experimental results additionally obtained in collaboration with Mr.F. Giitlbauer shows that the carboxylation of aqueous formic acid in the presence of C02 under the influence of 6oCO-y-rays is partly due to a chain reaction. The radical ion (CO;) formed as explained above can react with the formate to form an inter-Further research work on this subject is in progress. 1 Czapski Rabani and Stein Trans. Furaday SOC. 1962,58,2160. 2 Hardwick Rud. Res. 1960 12 5. 3 this Discussion GENERAL DISCUSSION 31 5 mediate probably (COT. HCOQ-). This can then initiate a chain reaction by reducing C02 and leading to the formation of oxalate as follows : COT +HCOO-+(CO . HCQO-) (CO . HCOO-)+C02+CO,+(COO. COOH-) This assumption was supported by experiments studying the yield of oxalic acid as a function of the dose rate as well as of the formate concentration at different pH.The chain reaction is interrupted by the action of the e.g. OH-radicals with (COY . HCOQ-) intermediate. A full report on this subject has been submitted for publication in Radiochimica Acta. Dr. G. Scholes (University of Newcastle-upon-Tyne) (communicated) In some experiments carried out in collaboration with Mr. M. Simic the relative reactivities of solutes towards the independently-produced hydrogen atoms (H") have been obtained. In the radiolysis of deaerated aqueous solutions of sodium deutero-formate containing another organic solute (RH) which can be dehydrogenated, the ratio G(HD)/G(Hz) will be governed by the competing reactions : ka H+DCOO-+HD+CQO-H+RH+H,+R. If the experimental conditions are such that H" is the only dehydrogenating species entering the above competition measurement of the yields of HD and of H2 should lead to the rate ratio k,/kb.These conditions are attained if N20 is also present, since possible complications arising from reactions of (H20)- with the orgallic solutes will be suppressed. In those instances where the solutes do not yield hydrogen on reaction with Ha e.g. Fe(CN)i - the system (NzO + isopropanol) has been used. Here hydrogen arises from the molecular process and from the reaction, k b kc (CHJCHOH + H-+(CH,);OH + HZ, which is in competition with Ha reacting with a solute (X), ka X+H+*XH. From the measured hydrogen yields the rates of reaction of the solutes relative to isopropanol (k,/kd) have been obtained. This latter method is somewhat similar to that used by Rabani and Stein 1 using either ferricyanide or nitrite and several organic solutes which can be dehydrogenated.The following are some representative values obtained by these two methods, showing the range of reactivity of various solutes. solute relative reactivity Fe(CN) 2 -ally1 alcohol HCOO-isopropanol ethanol acetone t-butanol (ka = 1) 180 f40 I04 &23 6.6 f 0.7 2.25 &O-2 0.7 f 0.1 2.8 f0.3 x 10-2 4.4 fo.7 x 10-3 1 Rabani and Stein J. Chem. Physics 1962,37 1865 316 GENERAL DISCUSSION Prof. F. S. Dainton (University of Lee&) (communicated) It is gratifying to find that in those areas where ow results 1 and those of Hughes and Willis overlap they also agree. Since attention is turning increasingly towards alkaline solutions and some disagreements have been reported it will perhaps be useful to outline the present position.Dainton and Watt have given values of radical and molecular yields at pH 6.5 10-8 12-0 12.8 and 13.5 which indicate that G.H’ and GOH increase by about unity as the pH is increased from 10 to 13-5. Haissinsky and his collabor-ators have studied a variety of systems at pH 13 and have proposed entirely different radical and molecular yields. The two sets are shown below and the major point of difference is that Dainton and Watt’s value of GH + ~GH exceeds that of Haissinsky Dainton and Watt p H 12.8 3.55 3-15 0.40 0.60 4-35 Haissinsky 13 2-75 3.15 0.30 0.10 3.35 by an amount equal to twice that by which Dainton and Watt’s value of 2&*02 exceeds that of Haissinsky et al.Whilst this fact is suggestive it is not our purpose to do more than define the area of disagreement. Accepting the authors’ mechanisms for their reactions we agree with Haissinsky that the Dainton and Watt values are uniquely in accord with systems (1) to (6) below that the Haissinsky values agree uniquely with systems (12) (13) and (14) and that the results for systems (7) to (1 1) agree with both sets. (1) Air saturated water ; (2) deaerated soIutions of ferrocyanide and ferricyanide in presence and absence of NzO; (3) deaerated solutions of methanol and ferri-cyanide ; (4) deaerated solutions of nitrous oxide ; ( 5 ) potassium manganate solu-tions ; (6) potassium iodate solutions ; (7) deaerated solutions of potassium tellurite ; (8) deaerated solutions of potassium tellurate ; (9) deaerated solutions of bivalent platinum ; (10) deaerated solutions of hydrogen peroxide ; (1 1) deaerated benzene-saturated solution of tetravalent platinum ; (12) aerated solutions of bivalent plat-inum ; (1 3) aerated solutions of potassium tellurite ; (14) solutions of potassium persulphate.There is no immediately obvious reasons for these discrepancies but there are experimental methods for discovering these and we are presently engaged on this problem. Dr. G. Czapski (Hebrew University Jerusalem) said I would like to present some new data which appear to be important in the interpretation of results in irradiated oxygenated aqueous alkaline solutions. Recently there has been increased interest in strongly alkaline solutions.The paper by Hughes and Willis is one such study. Since we have observed new transient species in oxygenated alkaline solutions there may be some danger of misinterpretation of results from neutral solutions which are extrapolated to basic solution without taking these into account. In pulse radiolysis experiments carried out with Dr. L. M. Dorfman at Argonne National Laboratory we have observed long-lived transient species in oxygen-ated alkaline solutions different from the ones known in the neutral and acid region. We have followed the decay of the transient species by fast spectrophotometry at 2537 A after a 0.4-5 ysec electron pulse at 120 mA and 14 MeV. The results show long-lived intermediates which decay with a first-order rate law with half-lives increasing with pH from 0.01 sec at pH = 9 to about 10 sec at pH = 13.In acid and neutral solutions on the other hand we observe second-order decay of HOz and 0; respectively. Thus in the alkaline solutions it appears that 0; is disappearing rapidly by reaction with OH- and/or OH radical or 0-. f Dainton and Watt Nature 1962 195 1294. GOH GH2 GH202 G-H20 G‘H GENERAL DISCUSSION 317 The new intermediates decaying by first order are probably a basic form of some biradical either of two O, or of 0; and OH-. similar to the H2O3 and H0; which were found 1 in acid solution at pH below 4. Although we have no clear evidence for the identity of this species it appears that in highly alkaline solutions 0; and HOz as such must have only a very short life-time reacting rapidly to give a basic form of the radical (HOZ-).In addition, in order to explain the first-order decay we suggest that some biradicals are formed having very long half-life in alkaline solution. Although these results are of a preliminary nature they should be taken into account in the interpretation of results in strongly alkaline solutions as e.g. in the work of Hughes and Willis. Dr. C. H. Cheek (U.S. NavaE Rex. Lab.) said With regard to the paper of Hughes and Willis it is noted that for oxygenated alkaline ferricyanide solution G(-Fe(CN) -) as expressed by eqn. (9) is well satisfied by G values of the radiolysis products of water at high pH recently obtained from a study of the radiation chem-istry of alkaline hypobromite solutions,2 i.e.G(-Fe(CN)z -) = G(H) + 2G(H202) -G(0H) = 2.8+2(@7)-2.1 = 2.1. This agreement is typical of the results usually obtained by the use of inorganic solutes. When methanol is added however the results no longer agree. The results presented for oxygenated solutions containing ferricyanide and methanol do not exhibit material balance. From the given yields are obtained G(reduction) = G(-Fe(CN&) = 7.8 and G(oxidation) = 2G(CH2O) = 6.4. Ac-cording to the mechanism presented oxygen and hydrogen were produced. Were any gas yields measured? If it can be shown that redox balance is achieved by consideration of the gas yields there will be more assurance that formaldehyde is the only significant organic decomposition product. In general it seems desirable that any work which proposes G values for the radiolysis products of water should also demonstrate material balance.The radical yields obtained in neutral and alkaline solutions by the use of organic solutes are generally higher than those obtained with inorganic solutes, and some explanation for this difference should be sought. The following pro-posal is offered. Let it be assumed that the primary yields Gecq = 2.8 GOH = 2-1, h 2 0 2 = 0-7 and G H ~ = 0.45 are independent of pH. Could it be that there is an additional entity X (possibly excited water molecules) which may contribute up to about 0.8 to the radical yields under certain conditions? The following properties would be required. (a) It would promote additional decomposition of water into H and OH in acid solution thereby accounting for the parallel rise in the radical yields with increasing acidity.(b) It would promote C-H bond dissociation in organic solutes. (c) It would not exhibit measurable effects in neutral or alkaline inorganic Solutions. Dr. G. Hughes (University of Liverpool) said In reply to Dr. Cheek gas yields were not measured. In oxygenated solutions it would be extremely difficult to attempt to measure both hydrogen and oxygen yields. We have recently confirmed our values of the radical yields using formic acid instead of methanol. G(-Fe(CN) 2 -) was found to be 7.8 and independent of formate ion concentration in the range 10-1-1 M. However the higher radical yields observed by us cannot be attributed entirely to the effect of organic scavenger.Thus similar values have been ob-tained 3 from a study of the ferrocyanide + ferricyanide + N20 system. 1 Czapski and Bielski J. Physic. Chem. in press. 2 Cheek and Linnenbom J. Physic. Cknz. in press. 3 Dainton and Watt Nature 1962 195 1294 318 GENERAL DISCUSSION It may be significant that the lower radical yields observed by Br. Cheek and Linnenbom and also by Prof. Haissinsky1 have in general been obtained in systems where the scavenger may undergo changes in oxidation number greater than unity. It would seem preferable to use scavengers where changes in oxidation number are restricted to + I so that the possible number of redox processes is kept to a minimum. Dr. V. J. Linnenbom ( U S . Naval Res. Lab.) said Hughes and Willis find no protective effect of chloride ion on the reduction yield of irradiated ferricyanide solutions when 0.1 M C1- is added to the ferricyanide+methanol system at pH 13.Presumably such an effect could arise because of competition between the alcohol and chloride ions for the hydroxyl radical : OH + CH30H+H20 + CH20H OH + Cl-+OH-+ C1. Should reaction (1) take place unimpeded the CH2OH formed would reduce ferri-cyanide thus increasing the overall reduction yield. However if (2) competes effectively with (I) the C1 atoms would react similarly to OH in oxidizing ferrocyan-ide and the overall reduction yield should be decreased. Since Cl- ion shows no such protective action at pH 13 Hughes and Willis assume that the species of OH present in alkaline solution differs from its counterpart in acid solution and reacts much less readily with chloride ion.Such 8n assumption appears to be unnecessary to explain why (2) docs not compete effectively with (1). If reaction (2) is more correctly considered as an equilibrium reaction with an equilibrium constant K of approximately 10-7 (ref. (1))’ then the relative concentra-tions of chlorine atoms and hydroxyl radicals at any pH is given by [Cl]/[OH] = K[Cl-]/[OH-]. Under the conditions of Hughes’ experiments (0.1 M chloride ion at pH 13), [Cl]/[OH] is of the order of 10-7. In other words the equilibrium reaction (2) is displaced far to left iq contrast to the situation in acid solution where it is dis-placed to the right. This is in accord with experimental evidence that chloride loses its effectiveness as a scavenger for OH at pH > 3.A similar explanation accounts for the effectiveness of bromide ion in protecting molecular H2 from attack by OH at higher pH. The equilibrium constant for the reaction in which bromide replaces chloride in (2) is approximately 102 (ref. (1)). In neutral solutions at dilute Br-concentrations (10-5 M) the [Br]/[OH] ratio is + 1 and complete protection is achieved ; at pH 13 however a concentration of 0.5 M Br- ion has been found necessary before the observed H2 yield reaches its primary G-value yield of 0-45 (unpublished data Linnenbom and Cheek). The possible ionization of OH in alkaline solutions to give the 0- species is a ques-tion still unresolved. As Prof. Hughes himself points out uncertainties in the avail-able thermodynamic data lead to estimates of pK from 8 to > 15.In the light of such an uncertain situation conclusions as to the ionization of OH based on indirect evidence which can be otherwise explained are not justified. Dr. L. S . Myers Jr. (University of Califsrnia Los Angeles) (partly communicated) : Some work on the conversion of OH. to CL which was done in our laboratory by Dr. John Ward bears directly on Dr. Hughes’ discussion of the effects of added chloride ion and is in agreement with the theoretical calculations reported by Dr. Linnenbom. Dilute aqueous solutions containing both thymine and ethanol were 1 Haissinsky and Patigny J. Chim. physique 1963 60 402. 2Uri Chem. Rev. 1952 50 375 GENERAL DISCUSSION 319 irradiated and the thymine-disappearance yield was determined as a function of added chloride-ion concentration.For reasons which will shortly be obvious acid solutions were used. Under such conditions the OH. radical adds to the double bond of thymine and abstracts hydrogen from ethanol. The concentration ratio of the two solutes was arranged so that in the absence of C1- approximately half of the OH. radicals reacted with thymine and half with the alcohol. Chlorine atoms will in general differ from OH= radicals in their relative rates of reaction with a double bond and a saturated structure and thus conversion of OH- to C1. by added Cl-would be expected to cause a change in the thymine decomposition yield. Such was observed. At pH 1 as the C1- ion concentration was increased the thymine decomposition yield increased reaching a maximum at a concentration of about 0.1 M.However at pH4.1 little effect of added C1- ion was noted. A more detailed study of the role of pH with solutions containing 0-1 M Cl- ion showed that between pH 2.5 and 3.5 there is a steady de-crease in the effect of Cl- ion and that at higher pH values there is no effect. The possibilities that the results were caused by salt effects changes in GOH or changes in the ionic state of thymine were eliminated by appropriate experiments. It seems reasonably certain therefore that conversion of OH- to Cl. by Cl- ions does not occur to any appreciable extent above pH 3. This explains the failure of Dr. Hughes to find any effect of added C1- ion in his strongly alkaline solutions. Dr. G.Hughes (University of Liverpool) said Dr. Myer’s remarks on the lack of effect of Cl- are interesting. Although Dr. Linnenboin suggests that the equilibrium OH + Cl-+OH- + C1 would not favour the exchange of OH with chloride ion at pH 13 other factors, e.g. the formation of Cl; in the presence of high concentrations of chloride ion may nevertheless help to promote exchange. The comment of Dr. Thomas is per-tinent to this. However the suggestion that 0- is involved rather than OH was based not only on the fact that it apparently failed to exchange with C1- but also on the marked difference in relative reactivities to methanol and ferrocyanide in acid and alkaline solutions. This is unlikely to be due to similar equilibria. Dr. J. I(. Thomas (Argonne Nut. Lab.) said A point of interest arises in the paper by Hughes and Willis regarding the nature of the reaction of OH radicals with chloride ions.This reaction which is endothermic has in the past been written as OH+Cl-+H++H2O+Cl. At Argonne we have irradiated solutions of sodium chloride with pulses of 13 MeV electrons and looked for the formation of the well-known species Cl; via OH+Cl-+Cl; Cl+Cl-+Cl;. With I mM NaCl at pH7 no transient was observed but on decreasing the pH to 3 and less the Cl; species was identified by its spectrum. At pH 7 it was possible to produce some Cl pro-vided concentrations of NaCl of 0-1 M and greater were used. Probably for the reaction OH+Cl- to occur one of the products must be removed e.g. OH- by H+ or C1 by C1-. It is difficult in the work of Hughes and Willis to decide whether under their conditions (0.1 N NaOH) the reaction would occur due to the equilib-rium being pushed OH-+ Cl- by the high OH-.Dr. L. S. Myers Jr. (University of California Los Angeles) said In considering radiation-induced reactions in aqueous alkaline solutions it is tempting to attribute differences from reactions in acid solutions to changes in the intermediate species (H elq OH etc.) with pH. Dr. Hughes has carefully chosen a system in which this is correct; however this is by no means universally true. Changes in the molecules being attacked may also be of importance. Such is the case with thymine. Similar results were obtained at pH2-5 320 GENERAL DISCUSSION The thymine molecule is a 6-membered heterocyclic ring with two 0x0-groups and a methyl side chain.It ionizes in two stages (pK1 = 9-6 pK2- 13) and has a double bond in the ring which is conjugated with one of the 0x0-groups in the unionized form and with two other ring double bonds in the quasi-aromatic ring of the twice ionized form : 0 II C 0 0-I C /\ I 11 \/ N,‘j ‘C-CH3 /\ I - II \/ H-N t C-CH3 A 1 II \/ H-N C-CH3 Q=C C-H N -0-C C-H N -0-C C-H N I H In air-saturated acid solutions radiolytic attack has been shown (Scholes Ward and Weiss) to be initiated by addition of OH. to the carbon-carbon double bond at the position remote from the methyl group with subsequent formation of saturated ring compounds the hydroxyhydroperoxide and glycol. In air-free solutions the attack is at the same position by either OH- or the reducing species with formation of the glycol and dihydrothymine (Ekert).(The compound formed by addition of both OH. and H- is unstable and spontaneously dissociates to thymine and water.) As the pH is increased above about 7 the yield for the disappearance of the conjugated double bond determined by ultra-violet spectroscopy gradually diminishes until it becomes essentially zero at pH 14. The decrease cannot be related to any known change in the intermediate species. However an excellent correlation with the ionization stage of the thymine can be obtained by assigning double-bond disappearance yields of 2-0 to H2TN (undissociated thymine) 1 -3 to HTN- and 0.0 to TNZ- and computing the expected net yield from the propor-tions of each species present at the various pH values.Although the yields for double-bond disappearance fall with increasing pH, the total yield for destruction of thymine remains essentially constant (2.0) to a pH somewhat greater than 14. Treatment of a neutralized 04-labelled thymine solution irradiated at pH 14 with charcoal to remove NaCl followed by paper chromatography and preparation of radioautographs results in isolation of a single major component which accounts for 80 % of the thymine disappearance. This component has been identified as the unsaturated ring compound hydroxymethyl uracil which is formed by the oxidation of the methyl side chain. Since the rather lengthy analytical procedure might cause secondary reactions it cannot be stated unequivocally that this compound is the true radiation product ; however it is clear that the site of radical attack has changed from addition to the double bond to attack on the methyl group.Thus as the scavenging molecule is ionized and its ring becomes quasi-aromatic the double bonds are inactivated and the hydrogens on the methyl group relatively activated towards free radical reactions. The mechan-ism of attack at present remains in doubt Abstraction of Ha from the methyl group followed by addition of 0 2 should lead ultimately to formation of an aldehyde. An alternative path direct nucleophilic substitution of 0- for H- seems reasonable and would readily account for a hydroxy product. The presence of the methyl side group is not essential for observation of the inactivation of the ring towards free radical addition reactions as the pH is increased.Both uracil and adenine whic GENERAL DISCUSSION 321 lack such a group also show changes in double-bond disappearance yields which are related to their ionization. Dr. 6. Hughes (University of Liverpool) said We agree entirely with Dr. Myers that in any irradiated system it is necessary to distinguish the effects of pH on (a) the primary species produced in the radiolysis and (b) the form of the substrate molecules. The experiments of Dr. Myers clearly indicate the importance of (b). Our sub-strates were chosen intentionally so that they would undergo no change with pH and therefore differences in reactivity could be attributed to (a). Dr. G. Czapski (Hebrew University Jerusalem) said In the paper of Hughes and Willis three important points are discussed (i) there are two different precursors of molecular peroxide one more easily scavengable; (ii) in 0.1 N NaOH the OH radical is in a different form than in acid and neutral solution probably as 0-; (iii) the scavengable precursor of H202 reacts much faster with Fe(CN);4 than with Concerning the first point the conclusion is based on the results presented in table 1 and fig.4rr showing that increasing Fe(CN);4 from 5 x 10-2 to 10-1 decreases CH30H. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 (Fe(CN)-4)3 FIG. 1 .-Hughes and Willis data replotted. G(H202) only by 0.02. Similar conclusions were reached by Anbar Guttmann and Stein1 using H20.18 They measured G(H20218J8) as a function of added H202169 16.In both these systems the yield of H202 drops quickly upon adding low concentra-tion of scavenger (10-3 M Fe(CN);4) and this decrease becomes smaller at higher scavenger concentrations but there is no indication of an asymptotic value of this yield. On the contrary this yield is roughly linear with the cube root of the scavenger concentration with Hughes and Willis G(H202) = 0.7-0-85 ((Fe(CN);4)* and in Anbar Guttmann and Stein 1 paper G(H202l8,l8),~,1.,,-1.~~ = 0.81-0-34 (H202)3. This behaviour is as expected from the Samuel-Magee model. ‘The linearity of the linear plot is not required by theory it is only a good approximation, as found by Sworski over a limited concentration range. At high concentrations it levels off and diverges from the straight line.This argument was used by Mahlman2 and Hayon 3 to show two independent mechanisms for the formation of molecular hydrogen are in doubt as they assume that this linear plot should extend over the whole concentration range an assumption which is in disagreement with calcul-ations of the diffusion model. 1 Anbar Guttmann and Stein J. Chem. Physics 1961,34 703. 2 Mahlman J. Chem. Physics 1960,32,601. 3 Hayon Nature 1862 194 737. 322 GENERAL DISCUSSION Thus it seems there is no proof for these different mechanisms or precursors for either of the molecular products H2 and H202. There seems to be some difficulty to give an alternative mechanism explaining the difference of 2G(OH) between curve I and I1 in Fig. 4. Still accepting the inter-pretation of Hughes and Willis some difficulties appear viz.the identity of the precursor of €3202 (the easily scavengable one according to Hughes and Willis). All possible precursors of a bimolecular process which are postulated sA to centradict other data. The possible species we assume are OH 0- H20+ and €320' (an excited water molecule). I l l 1 L 0 0.2 09 0.6 0.8 1.0 1.2 1-4 W202F FIG. 2.-Anbar Guttmann and Stein data replotted. Hughes and Willis rule out OH and 0-. OH is known to react only about 50 times faster with Fe(CN);4 than with CH3OH; thus in fig. @ one would not expect that -2 x 10-3Fe(CN);2 in the presence of 1 M CH30H would affect so much the yield of G(-Fe(CN),3). As these authors showed the species in the alkaline region (probably 0-) does react with equal rates with CH30H and Fe(CN);4 ; thus the previous arguments rule out this possibility as well.Anbar Guttmann and Stein 1 proved that H20* cannot be the easy scavengable precursor of peroxide. Also it would be difficult to explain most of the reactions of '' OH " in the bulk if this would be an excited water molecule ; thus we are left with &Of as the precursor. In 0.1 N NaOH it is most improbable that an H20+ would survive as such. Summing up it seems that there is no proof for more than one precursor of H202 and if the mechanism of Hughes and Willis is accepted there is some difficulty in suggesting any likely species to be the precursor of H202. Dr. G. Hughes (University of Liverpool) said We agree with Dr. Czapski that our results on the scavenging of molecular peroxide leave some doubt as to the pre-cursors involved.It is unfortunate that in our system we are not able to increase the ferrocyanide concentration further so that this point could be resolved. It is clear however that the precursor of the molecular hydrogen peroxide scavenged by ferrocyanide must be different from that species responsible for the oxidation of ferrocyanide in the bulk of the solution. It is interesting that new transients have been observed in alkaline solution although it is to be feared that these only add to the number of adjustable parameters by which results may be explained. 1 Anbar Guttmann and Stein J. Chem. Physics 1961,34,703 GENERAL DISCUSSION 323 CONCLUDING REMARKS BY F. S. DAINTON It is difficult to sum-up a Discussion of this kind without a brief reference to the history of Radiation Chemistry.Until the mid-thirties the main quantitative work in this subject referred to gases where it is meaningful to describe the efficiency of a radiation chemical reaction in terms of a measurable ionic yield and with only a few exceptions the mechanisms of these reactions were interpreted in ionic terms. Recognition of the high reactivity of atoms and free radicals led to a movement away from ionic mechanisms and the postulation of uncharged odd-electron species as the immediate precursors of the final products; though it was recognized that the free-radicals were in large measure generated from ion-molecule molecular-ion breakdown electron-capture and charge-neutralization processes.Gaseous re-actions seemed to follow the laws of '' homogeneous kinetics '' and liquid-phase reactions to differ from these primarily in that the initial free radical species were localized in tracks or spurs. The digusion model gave a quantitative account of the relative proportions of free-radicals combining or diffusing into the bulk of the liquid so that despite its many adjustable parameters this model seemed to give a moderately adequate account of the factors governing the radiation yields and their dependence on experimental variables. In the last five or six years new techniques most of which are illustrated by contributions to this Discussion have led to new data whih have revolutionized our thinking in radiation chemistry. Mass-spectrometric and vacuum-photo-chemical methods have shown the importance of elimination of molecules especially H2 and CH4 from ions and higher electronic states of molecules.Reinvestigation of gas reactions including isotopic exchanges ((H2 + D2 ; 14CO + CO& oxidation (CO+O2) and decomposition (C02)) has shown that ionic species may dominate the course and speed of a reaction and that excited states of ions and molecules are not to be neglected. (In passing perhaps one may wonder whether Dr. Lind for almost six decades the steadfast protagonist of ionic mechanisms and happily present at this meeting is permitting himself a feeling of satisfaction now that he has seen the wheel come full circle and a restoration to popularity of some of his long-neglected views.) But perhaps the major changes concern our knowledge of condensed systems.Here the techniques of low-temperature ionic polymerization have taught us that both positively and negatively charged primary species can cause major chemical changes in liquids of low and high polarity and in water and the alcohols the solvated electron has not only been spectroscopically identified but the absolute rate constants of many of its reactions measured by means of pulse radiolysis. This latter achieve-ment will surely rank as one of the major contributions of radiation chemistry to our fundamental knowledge of the kinetics of reactions in solution and has stimulated work on other methods of generating these species. Finally experiments with amorphous and crystalline solids and solid solutions have shown that the electron can travel from its point of origin to be trapped as such in glasses and examined spectroscopically at leisure or can bring about chemical change in a solute which may capture it by a dissociative mechanism ; and that excitation is also easily trans-ferred over larger distances through systems either having regular periodicity e.g., crystalline assemblies? or which are chemical aggregates e.g.polymer molecules. In such high density systems there is also the possibility of initial collective excita-tion but the role of this has not yet been delineated with any certainty 324 GENERAL DISCUSSION These many diverse conclusions for particular systems though perhaps at first sight confusing do fall into a relatively simple pattern which at the possible risk of oversimplification it may be worth restating. In all systems whatever the phase, the electronic excitation which leads to some ionization is complete in a time much shorter than a vibration period. In gases extensive charge separation occurs for periods sufficiently long (up to msec in suitable cases) during which ionic reactions can occur and lead to ultimate products or result in generation of uncharged re-active species which may in their turn cause reaction leading to final products. Equally ions or radicals may induce back reactions which reduce the extent of the net decomposition. General rules as to which mechanism will predominate are not easily formulated but suitable experimental tests for the existence of one or other mechanism can be devised and applied and the contributions to this Discussion contain many examples. In non-polar liquids the low dielectric constant polarizability and solvating power ensure that charge separation is rendered both spatially and temporarily minute. Electrons return to a positive ion which is either the original molecular ion or a protonated molecule so that free radicals and excited states determine the pattern of the reactions. These systems are notable in that facile energy transfer may often occur between suitable compounds in binary mixtures. In liquids of high dielectric constant polarizability and solvating power the electron will be in a weaker coulomb field when it is thermalized and may well be solvated as such, surviving sufficiently long to build up its Boltzmann distribution of counterions. In many systems the solvated electron has a half-life with respect to mutual destruc-tion of many psec and it appears that most of its reactions with solutes are diffusion controlled. The initial distribution of these electrons and their protonated solvent counterions is non-uniform to a degree controlled by the L.E.T. and the dose-rate of a particular system so that the diffusion-model is not basically invalidated. In these highly polar solvents the presence of acids or bases will have a very important effect not only in regulating the hydrolysis of solutes but in determining whether solvated electrons are converted by Hf to hydrogen atoms or hydrogen atoms converted by basic anions to solvated electrons before reaction with the solute occurs. If the electrons ‘‘ get away ” to react as e- or H then excited states will be formed in lower yield and this does seem to be the case. Pure crystalline solids contain very few trapping sites for electrons so that although electron mobility may be high chemical change in the solid state is much more likely to take place as a result of the formation of excited states or free-radicals consequent on charge neutralization. Radicals will be trapped in the lattice and radical-radical and radical-solute reactions will generally occur when the temper-ature is raised to confer limited mobility which will be most marked in the region of first- or second-order phase changes. However excitation migration may occur more extensively (linear distances of about 100 A) the more perfect the crystal and the less the lattice vibrations which are excited i.e. it will be favoured the lower the temperature. Migration of this energy to suitable solutes may cause their electronic excitation followed by experimentally detectable chemical dissociation or emission of characteristic radiation. If the solid is amorphous energy transport except through especially suitable molecules will be disfavoured but there may well be very numerous trapping sites for the electron at various distances from its point of origin. If so many of the electrons will be trapped and only escape from these traps by thermal or photo-activation. If electron-capturing solutes are present in concentrations comparable to that of the physical traps in the glass then they may compete for the electron GENERAL DISCUSSION 325 Although the broad pattern of radiation chemical processes is well established much yet remains to be done and past history does not encourage the belief that no surprises are in store. On the contrary there is much to be learnt particularly about energy migration and electron transport about bimolecular interaction of excited states about possibilities of transfer through micro-crystalline domains of liquids about fragmentation of molecules and ions etc. Increasingly we shall learn about the chemical and physical properties of entities of fundamental chemical importance and from the way in which they behave in media of different states of aggregation gain information about the structural nature of these states. As the precision of our knowledge improves we may hope that perhaps the semantic difficulties referred to in the introductory paper will recede

ISSN:0366-9033    

DOI:10.1039/DF9633600232

出版商:RSC    

年代:1963

数据来源: RSC

摘要:

GENERAL DISCUSSION 325 AUTHOR INDEX Adams G. E. 302 309. Allan J. T. 312. Allen A. O. 95 247 253. Amphlett C. B. 296 Ausloos P. 66 239 245 269. Ballantine D. S. 289. Baxendale J. H. 186 295 302 307 308. Beck D. 56,239. Bensasson R. 177. Boag J. W. 302 309. Burns W. G. 124,235 273 283. Burr J. G. 264 270. Burton M. 7 259 298. Busler W. R. 102. Charlesby A. 236 246 268 277 288. Cheek C. H. 317. Clay P. G. 46. Collinson E. 83 153 247. Conlay J. J. 153. Corney N. S. 238. Cundall R. B. 111 261 263 264 266. Cunningham J. 280. Czapski G. 316 321. Dainton F. S. 153 237 239 240 246 247 258, 259,261 283 292 300 303 310 313 316. Denhartog J. 135. Dominey D. A, 35 238. Dorfman L. M. 206,247 288 307 308 309. Durup M. 177. Dworkin A. 177. Dyne P.J. 135 274. Pessenden R. W. 147. Firestone R. F. 294. Freeman G. R. 247 248 250 253 254 269, 278 280. Fmeki K. 19. Futrell J. H. 240 244 278. Getoff N. 314. Gilbert G. P. 186. Golub M. A. 264 276. Gordon S. 193. Griffiths P. A. 111. Guarino J. P. 169. Hummel R. W. 75 245 246. Hamill W. H. 169 292. Hardwick T. J. 240 267 272. Hart E. J. 193. Hauser W. P. 245. Hayon E. 288 309 310 312. Henglein A. 236 272. Hughes G. 223 317 319 321 322. Hummel A. 95. Hummel R. W. 75 245 246. Johnson G. R. A. 46 235. Katsuura K. 267. Kaufman R. G. 262. Keene J. P. 304. Klots C. E. 278. Krauch C. H. 266. Lias S. G. 66. Linnenbom V. J. 318. Lipsky S. 271. Magat M. 177 256 274 290 297. Magee J. L. 19,232 234 235 236 237 247 Martin D.H. 102. Marx R. 177. Matheson M. S. 193. Meshitsuka G. 309. Myers L. S. Jr. 318 319. Palmer T. F. 35 238. Phillips G. O. 281. Rabani J. 193. Ronayne M. R. 161). Rowland F. S. 249. Salmon G. A. 284. Samuel A. H. 257 300. Sandoval I. B. 66. Sauer M. C. Jr. 206. Schuler R. H. 147 269. Scholes G. 214 296 302 311 315. Simic M. 214. Smith D. R. 135 274. Swallow A. J. 234 273 278. Szwarc H. 177. Taub I. A. 206. Thomas J. EL. 193 319. Tiernan T. O. 240 278. Todd J. F. J. 83. Toma S. Z. 268. Ubbelohde A. R. 268. Van Dusen W. 260. Wagner C. D. 273. Walker D. C. 313. Ward J. A. 169. Warman J. M. 46. Weiss J. J. 214. Wild W. 238 257 299. Wilkinson F. 83. Willard J. E. 290, Williams Ff. 102 254 257 259 260. Willis C. 223. Winter J. A. 124. Wyard S. J. 294. * The references in heavy type indicate papers submitted for discussion

ISSN:0366-9033    

DOI:10.1039/DF9633600325

出版商:RSC    

年代:1963

数据来源: RSC