OXIDATION AND REDUCTION BY ATOMIC HYDROGEN IN AQUEOUS SOLUTIONS BY GABRIEL STEIN Dept. of Physical Chemistry, Hebrew University, Jerusalem, Israel Received, 4th February, 1960 The oxidizing action of hydrogen atoms (produced by an electrodeless high-frequency discharge) on aqueous solutions of iodide and ferrous ions is examined. It is shown that tho assumption that Hias ions may be the actual oxidizing agents agrees well with the results. Comparison with results in solutions containing ferro-ferri sulphate and ferro-ferri cyanide permit an evaluation of competitive oxidation-reduction reactions due to H atoms. The results are critically compared with evidence from radiation, photo- and electro-chemistry. Atoms and free radicals may, in general, act as either oxidizing or reducing agents, according to whether they acquire an additional electron from a donor or lose their odd electron to an acceptor.For the simplest case, that of a hydrogen atom, this fact is well known. They may act as reducing agents by electron transfer or H atom addition. They may also act as dehydrogenating, i.e. oxidizing agents on molecules from which H may be abstracted whilst the molecular H-H bond is gained. As to electron transfer to H, the electron affinity of H in the gas phase is only of the order of 0.5 eV. In aqueous solution the reaction H-+aq+H,+OH,< (1) would have to occur in addition to provide the maximum gain in energy. This sequence of reactions may be given a pH-dependent form. In acid solutions another possibility arises. The formation of the hydrogen molecule ion H+H+-+H,+ (2) is exoergic by 61 kcal.In aqueous solution the hydration energy of H+ is ap- proximately 260 kcal. If, then, the hydration energy of HZ is around 200 kcal, the species HZas could exist in acid solutions. It would be a strong electron acceptor. The ionization potential of H2 being 359 kcal, the process would be exoergic by some 150-160 kcal. Coulson 1 examined the problem theoretically. The reducing action of hydrogen atoms (produced by electric discharge, and introduced into the aqueous solution) was investigated in great detail. The possible oxidizing action of atomic hydrogen was not considered seriously until recently. Ethier and Haber2 were the first to report briefly experiments which could be best explained by assuming that hydrogen atoms act as oxidizing agents in homogeneous aqueous solutions.The effect of increasing acidity was observed and a termolecular reaction, M,+,+H+H,+,-+M,~,+ +H, was postulated. 235 (4)236 OXIDATION AND REDUCTION B Y ATOMIC HYDROGEN THE EVIDENCE FROM RADIATION CHEMISTRY There was no further work in this field, until the development of radiation chemistry of aqueous solutions. The action of ionizing radiations, e.g. y-rays, on dilute aqueous solutions is best understood according to Weiss,3 if one assumes that the main radical intermediates formed are hydrogen atoms and hydroxyl radicals : HOH--s\w+H+ OH. ( 5 ) In the radiation chemistry of aqueous solutions of Fe2+ ions, in the absence of oxygen, the experimental results could best be explained4 by assuming that H atoms can oxidize ferrous to ferric ions, in acid solution, according to Fe:: +H:aq+Fe,3,$ +H,.The quantitative results of Rigg, Stein and Weiss 5 were confirmed by Barr and King6 and the work was further extended by Rothschild and Allen.7 Im- portant additional evidence was recently obtained by Shubin and Dolin.8 They carried out reaction ( 5 ) in ferrous ion solution in the presence of high pressures of H2. OH radicals are replaced by H atoms according to OH + H2 -+ H20 + H, (7) but the oxidizing equivalents remained practically constant in acid solution. Thus, evidence from radiation chemistry conclusively shows that hydrogen atoms can oxidize ferrous ions to ferric in acid aqueous solution. Regarding the nature of the actual oxidizing species, radiation chemistry could not provide conclusive evidence for or against Weiss’ assumption that Hzaq is the active intermediate.In addition to Haber’s three-body-collision mechanism (4), Uri 9 suggested the possibility of the reaction Fe2+HOH+H--+Fe3+OH- + H,. (8) In radiation chemistry the possibility of the thermalized electrons, eaq, and H atoms, or alternatively H atoms and HZfaq ions being the pair of species dominant at different pH values, was pointed out by Barr and Allen.10 Isotopic experiments using deuterium gave results,ll which did not absolutely contradict the possibility of HZaq being the actual species, but could be reconciled with it only by making special ad hoc assumptions.12 Similarly, the investigation of competition reaction between, e.g.Fez+ and 0 2 , for the reacting species 79 13 was difficult to reconcile with other evidence.12 Thus radiation chemistry could onIy suggest, but not prove, the possible mechanism of oxidation by H atoms in aqueous solution. THE EVIDENCE FROM PHOTOCHEMISTRY The detailed mechanism of the photochemistry of aqueous solutions of ions, e.g. Fez+ or I-, which on u.-v. irradiations yield molecular hydrogen, is still quite obscure. Indeed, even the basic question whether the true initial yields are at all pH-dependent or not, has not yet been conclusively decided. Rigg and Weiss investigated the photochemistry of ferrous 18 and iodide 19 ions. They found both reactions to be pH-dependent and proposed mechanisms, in which the pH- dependence was attributed to the role of Hiaq in the mechanism.However, Lefort and Douzou20 claimed that the ferrous ion reaction at least is not pH- dependent at all. We have therefore recently re-investigated the photochemical oxidation of ferrous 21 and iodide 22 ions, with particular attention to the determination of the pH-dependence of the initial yield. We found that in both cases the initial yield indeed depends on the pH. However, it was proved, as envisaged by Farkas andG . STEIN 237 Farkas 23 and by Platzman and Franck,24 that the primary photochemical formation of the H atoms was itself largely responsible for the pH-dependence observed. Our photochemical results are consistent with comparable results from radiation chemistry, and yield for comparable reactions practically identical rate constants.They prove that oxidation by H atoms occurs. However, whilst the results are entirely consistent with the assumption that H atoms act as oxidants in the actual form of HZ ions, they cannot serve as conclusive proof in this respect. THE EVIDENCE FROM ELECTROCHEMISTRY In investigations of electrode processes resulting in the evolution of H2, it was suggested by Kobosew and Nekrassow,lo that the HZ ion may be the decisive intermediate on high overvoltage electrodes. Horiuti and his co-workers 15 developed in great detail the convincing evidence in favour of this reaction step. Poltorak 16 showed that this active intermediate may penetrate the aqueous solu- tion. Recently, Ives 17 confirmed these results and summed up the electrochemical evidence in favour of the formation of H2f ions on the electrode surface and probably evaporating from there, in the body of the solution. REACTIONS OF ATOMIC HYDROGEN To avoid the ambiguities inherent in photo and radiation chemistry, it is de- sirable that the question whether H atoms can act as oxidizing agents should be investigated using H atoms produced as such and introduced into the solution.Indeed, simultaneously two groups 2 5 ~ 2 6 reported results in agreement on the point. Davis, Gordon and Hart26 produced H atoms by dissociating H2 gas through an internal discharge or by a heated tungsten wire. Czapski and Stein 25 used an electrodeless ring discharge. Both groups reported that H atoms indeed oxidize ferrous ions in acid solution. Their accuracy did not permit the deter- mination of pH-dependence.Using increased accuracy, it was shown recently 12 that the oxidation of ferrous ions by H atoms is pH-dependent. This excludes mechanism (8) as the onZy one operating. However, with ferrous ions the practic- able p H range is limited, The results could not differentiate between the pH- dependent mechanisms (4) or (6), occurring by themselves, or in addition to another mechanism, e.g. (8). The situation is much more favourable with I- ions. Here, the oxidation of I- by H atoms was proved.27 It could be shown that varying the pH from 0.4 to 7.0 and the iodide concentration between 0.006 and 0.6 M, the yield Y of iodine was given by 2A (9) where a = k$/2k, (k2, being the velocity constant of the formation of Hiaq in aqueous solution, and k,, the velocity constant of the recombination reaction, H+H-+H2) ; p = [I-]/([k2,/krJ+ D-]), where kI- is the velocity constant for the oxidation reaction of by H2-aq ; and A g atoms 1.-1 sec-1 is the rate of intro- duction of H atoms into the solution.It was shown that the results disagree with both the pH-independent mechanism and the three-body-collision process (4). They were found to be in excellent agreement with the mechanism involving the specific formation of HZaq which acts as the actual oxidizing species. THE ENERGETICS OF THE H i a q ION From the fact that H;aq can oxidize Iaq in solution an upper limit for SH+aq the energy of hydration of Hiaq can be obtained for the process238 OXIDATION A N D REDUCTION BY ATOMIC HYDROGEN Using the value of (E,+S,-) == 147 kcal, and EH+ = 365 kcal, an upper limit of approximately SHi-= 21 8 kcal is obtained, negIecting energies of hydration of the neutral entities.In our work on the photochemistry of I- soIutions,Zz resuIts were obtained which indicated that the nature of the reducing species changes at about pH = 2. Interpreting this as indicating that the pK of Hzaq is between say 1 and 3, and as- suming the energy of hydration of H+ to be 240 kcal, the energy of hydration of Hiaq has a value of approximately 204 kcal. We thus come to the tentative conclusion that SH; is probably 210 f10 kcal. 2 THE COMPETITION BETWEEN OXIDATION AND REDUCTION PROCESSES BY H ATOMS From the work on the oxidation of I- ions by H i , the velocity constant for the formation of Hiaq from H and H& could be calculated.27 k2 was found to have the value of -102 1.mole-1 sec-1. Thus the formation of HZaq is a relatively slow process. Indeed, it was found that as the pH increases, the reduction of, e.g., ferric to ferrous predominated, since H atoms were captured by the oxidizing agent. The results also indicated that the velocity of reduction by H atoms of the hexahydro Feiz ion was less than, e.g., of the Fe3+0H- complex.12 The role of the complexing group in determining the reduction velocity has been investigated in radiation chemistry22 for the groups OH-, SO:-, C1-, (C1-)2, PO:+, F- and (CN-)6 with Fe3+. The results indicated that in the reduction process, which may in different cases proceed by either group transfer, according to or by conducted electron transfer through the bridging complex, according to the nature of the complexing group is decisive for the reduction velocity. It is of great interest to investigate the detailed mechanism of such processes using H atoms.We investigated first, the reduction of ferricyanide, which prob- ably proceeds by conducted electron transfer. It was shown that when Fe(CN)i- was used,29 the velocity constant of the reaction, H + Fe(CN)i - +Fe(CN)i- + H + , was of the order of 106 1. mole-1 sec-1. The formation of Hi,, was too SIOW ta compete with this process above pH-1, and reduction predominated. Using the large protein containing molecule of ferricytochrome-c, at about pH-7, similarly high velocity constants of the order of 106 1. mole-1 sec-1 were obtained.30 Thus in any particular system, the relation between oxidation and reduction appears to be determined by the pH, oxidation processes by H atoms being favoured in acid slution.The quantitative analysis of the results excludes a purely pH- independent process, such as (8) as the sole reaction path, and excludes the ter- molecular mechanism (4), which cannot account for the quantitative pH-de- pendence. The results are compatible with the assumption of H i being the actual oxidizing species, but some uncertainties remain. It is not known whether a process like (1) may not on detailed analysis yield similar results. The use of H atoms can also yield important results towards our understanding of the detailed mechanism of reduction in ionic solutions, particularly the role of the complexing groups in determining the actual reaction path and the specific velocity constants.Added in Proof.-Since the writing of this paper we have fully investigated the implications of the hydride mechanism suggested in the paper. With Dr. G. Czapski and Dr. J. Jortner we found that it is possible to account for all theG . STElN 239 observed phenomena if we assume that two alternatye pathways of oxidation by H atoms exist. One involves the reaction of H atoms with Hiq to form HZaq. The velocity constant of this reaction, KH=H+ =; 102 1. mole-1 sec-1. H&q formed in this relatively show reaction oxidizes acceptors, c.g., I- as discussed in the paper. This slower reaction is the one also operating in pure aqueous solutions, its slow- ness accounting for the absence of isotope exchange as observed by Friedman and Zeltman. If, however, other acceptors, e.g., metal cations, are present, this may directly form intermediate complexes with H in a reeaction which may have faster rate constants.For example, k g e z + + ~ = 6x 104 1. mole-1 sec-1. When this faster hydride complex pathway is available the slower H i mechanism may not be of kinetic importance. The existence of this faster pathway explains the results of competition reactions between ferrous ions and other acceptors, e.g., 0 2 or alcohols. We could show that the specific expressions derived from the two mechanisms yield different quantitative dependence on the concentration of H+ and the acceptor. We could show that in the case of the iodide ion indeed the HZ mechanism is the only one that fits the results, whilst the hydride mechanism does not agree with it.This latter does, however, account very satisfactorily for the quantitative results in the case of the oxidation of ferrous ions by H atoms. 1 Coulson, J. Chem. SOC., 1956,778. 2 Ethier and Haber, Naturwiss., 1930, 18, 266. 3 Weiss, Nature, 1944, 153, 748. 4 Weiss, Nature, 1950, 165, 728. 5 Rigg, Stein and Weiss, Proc. Roy. SOC. A, 1952, 211, 379. 6 Barr and King, J. Amer. Chem. Soc., 1954, 76, 5565. 7 Rothschild and Allen, Rad. Research, 1958, 8, 101. 8 Shubin and Dolin, Doklady Akad. Nauk S.S.S.R., 1959,125, 1298. 9 Uri, Chem. Rev., 1952, 50, 376. 10 Barr and Allen, J. Physic. Chem., 1959, 63, 928. 11 (a) Gordon and Hart, J. Amer. Chem. Soc., 1955, 77, 3981. (b) Friedman and 12 Czapski and Stein, J. Physic. Chem., 1959, 63, 850. 13 Baxendale and Hughes, 2. physik. Chem., 1958,14, 323. 14Kobosew and Nekrassow, 2. Elektrochern., 1930, 36, 529. Kobosew, Zhur. Fiz. Khfm., 1952,26, 112. 15 Horiuti and Okamoto, Sci. Papers Inst. Physic. Chem. Res. Tokyo, 1936, 28, 321. Horiuti, Keii and Hirota, J. Res. Inst. Cataly8is5 Hokkaido, 1951, 2, 1. Horiuti, Z.physik. Chem., 1958, 15, 162. 16 Poltorak, Zhur. Fiz. Khim., 1953,27, 599. 17 Ives, Can. J. Chem., 1959, 37,231. 18 Rigg and Weiss, J. Chem. Physics, 1952, 20, 1194. 19 Rigg and Weiss, J. Chem. SOC., 1952,4198. 20 Lefort and DOUZOU, J. Chim. Physique, 1956, 53, 536. 21 Jortner and Stein, to be published. 22 Jortner, Levine and Stein, to be published, 23 Farkas and Farkas, Trans. Faraday Soc., 1938,34,1113. 24 Platzrnan and Franck, Farkas Memorial Volume (Jerusalem, 1952), p. 21. 25 Czapski and Stein, Nature, 1958, 182, 598. 26Davis, Gordon and Hart, J. Amer. Chem. SOL, 1958,80,4487. 27 Czapski, Jortner and Stein, J. Physic. Chem., 1959, 63, 1769. 28 Schwartz, J. Amer. Chern. Suc., 1957, 79, 534. Schwartz and Nritz, J. Amer. Chem, 29 Czapski and Stein, J, Physic. Chem., 1460, in press. 30 Czapski, Frohwirth and Stein, to be published. Zeltmann, J. Chem. Physics, 1958, 28, 878. Soc., 11933, 80, 5636.