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.