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