1982 237The OsO,-catalysed Decomposition of Hydrogen PeroxideBy LBszl6 J. CsBnyi,' ZoltBn M. GalbBcs, and LBsz16 Nagy, Institute of Inorganic and Analytical Chemistry,A. J6zsef University, P.O. Box 440,6701 Szeged, HungaryThe decomposition of H z 0 2 has been investigated in the presence of OsO, as a catalyst. The rate of decompositionis proportional to the first power of the Os04 concentration and to the power 1-1.2 of the hydrogen peroxideconcentration. The pH dependence of the decomposition rate is quite characteristic; a high maximum is attainedat pH 10.6 and a much lower one at about pH 8.3. With the aid of appropriate free-radical reagents and by e.s.r.spectrometry, it is shown that hydroxyl and superoxide radicals are formed during the catalysed decomposition.The rate of formation of the OH radical depends on the concentrations of H202 and OsO,, as well as on the pH.The rate of bleaching of N-dimethyl-p-nitrosoaniline and other dyes by the OH radical as a function of pH exhibits amaximum at pH 8.3.The OH radical is of both primary and secondary origin, and is involved in a chain reaction.The length of the chain is about 80-90 at pH < 8, while at pH > 9 this value drops to ca. 10 and is independent ofthe pH. The apparent activation energy of the reaction route involving the OH radical is 105-1 1 1 kJ mol- atpH < 9, 3540 kJ mol- at pH > 10. Another decomposition route parallels that involving the OH radical, andpredominates at pH > 9.5. In this case the catalysed decomposition can be approximated on the assumption thatperoxo-osmate anion acts as a nucleophile towards the non-dissociated hydrogen peroxide molecule.The tem-perature dependence of this rate component gives an apparent activation energy of 60 kJ mol-1 at pH < 9, ca. 15kJ rnol-l at pH > 10.IN weakly alkaline media osmium tetraoxide is a veryefficient catalyst for the decomposition of hydrogenperoxide. Its catalase-like effectiveness has long beenused for the quantitative removal of H202 in chemicalanalysis. In addition, a H202-OsO, mixture is fre-quently used in preparative organic chemistry as anhydroxylating agent, and consequently it is surprisingthat very little is known of the mechanism of thecatalysed reaction.Chugaev and Bikermanl found that the rate of thecatalysed decomposition varied according to a curve witha maximum when different quantities of sodium hydr-oxide were added to the solution.The maximum wasobserved in the presence of 0.001--0.005 mol dm-3sodium hydroxide. At lower concentrations the decom-position rate was proportional to the amount of OsO,.From a study of the y-radiolysis of an acidic solutioncontaining OsO, and H,O,, Dran2 concluded that acatalysed decomposition reaction takes place, involvingO-containing radicals and osmium(vr1). Domka andMarciniec followed the decomposition of H,02 and theoxidation of indigo carmine catalysed by OsO, in aslightly acidic medium. They put forward alternativesuggestions regarding the mechanism of the reaction.Lunenok-Burmakina 4 found that the dioxygen evolvedduring the cat alysed decomposition of hydrogenperoxide arose solely from H20,, and stated that theOH radical could be detected in the catalysed process atca.pHWe have carried out investigations to try to shed lighton the kinetics of the catalysed decomposition of hydro-gen peroxide and on the role of osmium tetraoxide inthis process. In the present paper some characteristicfeatures of the catalysed reaction are elucidated.EXPERIMENTALAll reagents used were of analytical grade. The stocksolutions of the catalyst were prepared from OsO, (Merck)dissolved in water (10-4 mol dm-3 OsO,) or in 0.02 moldm-3 sodium hydroxide solution (0.02 mol dm-3 OsO,).They were stored in the dark in a refrigerator and werefreshly diluted before use.The hydrogen peroxide solutionwas prepared from stabilizer-free Merck Perhydrol.Ordinary distilled water was redistilled from alkalinepermanganate, and was then passed through columns ofAmberlite IR-120 and Dowex 50 ion-exchange resins inH+ form. Afterwards it was boiled with potassiumperoxodisulphate (1 g dm-3) for ca. 30 min, neutralised withpurified sodium hydroxide solution, and distilled from aPyrex apparatus. Sodium hydroxide and sodium sulphatesolutions were purified according to the method of D'Ansand Mattner.6Methanol, ethanol, propan-2-01, n-butanol, acrylonitrile,and methyl methacrylate were distilled twice before use.The stable free radical of 2,2,5,5-tetramethyl-kphenyl-imidazolin-l-oxyl S-oxide (denoted as tmpio) was preparedas described by Putirskaya and Matus.7In preliminary experiments, attempts were made toremove insoluble impurities by filtering the solutionsthrough a Millipore filter with a pore size of 0.2 pm, butsince no difference was found between the rates of reactionof the filtered and unfiltered mixtures, this step was sub-sequently omitted.Apparutus.-A Radiometer pH 26 pH-meter was usedfor pH measurements, and this meter combined with aRadiometer TTT Titrator and an ABU automatic syringeburette filled with acid was used as pH-stat.Spectro-photometric measurements were made with Unicam SP500, SP 800, and Beckman DB-G spectrophotometers. AJEOL- JES-PE spectrometer was used for e.s.r. measure-ments.Methods.-Concentrations of osmium tetraoxide solutionswere determined spectrophotometrically with the aid ofanthranilic acid * or thiourea @ reagents, and by potentio-metric titration using standard arsenite solution.10The concentration of hydrogen peroxide was determinedeither spectrophotometrically or by titrimetry.In theformer case the sample was added to a 15-50 fold excess ofan acidified iron(I1) solution, and after 30 min the absorbanceof the iron(m) formed was measured a t 304 nm. In othe238cases the sample containing H,O, was added to a knownquantity of an acidified standard arsenite solution, theexcess of which was titrated cerimetrically. The directoxidimetric titration of hydrogen peroxide is not advisablein the presence of higher concentrations of OsO, (> 10-6 moldm-3), since the cerimetric (or permanganometric) titrationis accompanied by an induced decomposition of H,O,.llThe volume of dioxygen evolved during the reaction wasmeasured either with an automatic gas-measuring device l2or with the simple apparatus shown in Figure 1.WhenFIGURE 1 Outline of apparatus used for the determination ofdioxygen formed. (V) = Pyrex reaction vessel; (S) = septum;(M) = manometer; (R) = relay andamplifier; (B) = syringe burette; (P) = plunger; (T) =thermostatted space filled with water at the start ; (G) = glasselectrode; (W) = three-way tapdioxygen is evolved, the level of 1 mol dmU3 H,SO, in themanometer (M) moves downwards and breaks the contactbetween the platinum wire (N) and the sulphuric acid;this actuates the relay (R) of the motor of a slightly modified(reversed-phase) Metrohm Dosimat E 41 5 syringe burette(B).The plunger (P) moves down, sucks water out of thespace (T) to restore the level of the manometer, and themotor stops. The volume of gas evolved can be read offthe meter of the Dosimat as a difference. The sensitivity isf0.02 cmS.Kinetic Measurements.-The measurements were made a t25 f 0.05 OC, unless stated otherwise. The ionic strengthwas maintained at a constant value with the use of Na,-[SOJ or Na[ClO,] (1 mol dm-3 Na+). Decomposition of(N) = platinum wire;J.C.S. Daltonhydrogen peroxide was followed by two methods: (i) bygas volumetry, and (ii) by measuring the concentrationof H,O, in samples taken from the reaction mixture a tgiven time intervals.In the gas-volumetric measurements 42 or 49 om3buffered hydrogen peroxide solution were placed in thereaction vessel [Figure I (V)] and thermostatted for 20 min.The reaction was started by injecting 1.00 cm3 OsO,catalyst solution through the septum (S) and the three-waytap (W) was turned in the manometer direction.To minimize loss of OsO, by evaporation, in some experi-ments a 20-cm3 all-glass syringe (fitted with a thermostatjacket) was used as reaction vessel.The buffered hydrogenperoxide solution was thermostatted in a separate beaker.The reaction was started by introduction of the catalyst,and reaction mixture was sucked into the syringe. Thesyringe was aligned vertically and the gas was displaced bythe plunger. The tip of the syringe was connected by aTeflon tube to a pipette (1 cm3) used for sampling.Somemeasurements were performed in a Teflon reaction vesselfitted with a thermostat jacket.The rate of consumption of the inhibitor dyes was deter-mined spectrophotometrically. To this end the pre-thermostatted components of the reaction mixture [solu-tions (1) and (2)] were mixed by a simple (two-jet) devicemade of glass, connected to the flow-through cell of appro-priate light path in the thermostatted cell housing of aBeckmann DB-G spectrophotometer. The cells weresupplied by two all-glass syringes (5 cm3), the plungers ofwhich were actuated manually via a common shaft. Thesyringes could be filled and emptied separately.The com-position of solution (1) was buffer + OsO, solution, andthat of solution (2) was buffer + H,O, + dye solution.The change in absorbance was recorded a t the appropriatewavelength and a t a chart speed yielding satisfactory timeresolution.Evaluation of Kinetic Results.-Concentration 'us. timecurves were used to determine the initial rates. Initialrates were also obtained numerically. In some cases, whenthe gas volume us. time curves showed short incubationperiods due to the fact that the reaction mixture was notsaturated with 0,, only the sufficiently linear parts of thecurves were used to determine the initial rates.RESULTSDependence on the Concentration of Hydrogen Peroxide.-When the logarithm of the hydrogen peroxide concentiationor the logarithm of the volume of dioxygen evolved wasplotted vs.time the curves obtained were virtually linearup to ca. 30% conversion (Figure 2). Departure fromlinearity indicates that the power of the hydrogen peroxideconcentration in the rate equation is greater than one. ItR = -d[H,O,]/dt = k,b,[H,OJnwas found that n = d(1og R)/d(log [H,O,]) is in the range1-1.2, and increases as the hydrogen peroxide concen-tration decreases during the run.Loss of OsO,.-When the decomposition was followed upto a conversion of a t least goy', and measurements wererepeated after the initial concentration had been readjustedby adding concentrated hydrogen peroxide solution,observable departures between the first and the later runswere not found (see Figure 2).A t concentrations ofosmium tetraoxide higher than mol dm-3 the concen1982 2391.5mIE-0 - 0E"A 1.0 \I520.50 500 1000t l s.57CI)OCI) I I0) 1.0 0 - Ithe larger maximum, appears as an increase of a few percentof the local decomposition rate in the absence of buffers.Otherwise, in the ascending part of the rate vs. pH curve,the rate is proportional to 1/[H+] (Table 1).7 8 9 10 11 12PHFIGURE 3 pH Dependence of the initial rate of catalysed!.5 decomposition. Conditions: 1.8 x mol dm-3 H,O,;FIGURE 2 Kinetic curves for catalysed decomposition. Con-ditions: pH 9.65 maintained by pH-stat; 3 x mol dm-3OsO,; 298 K ; ionic strength: 1 rnol dmP3 Na+ adjusted withNa2[S04]. Curves: (a) concentration vs.time curve, (0) firstrun, (@) second run, the initial concentration having been re-adjusted with concentrated H202; (b) log [H202] vs. timetration of the catalyst was estimated by spectrophotometry.The loss of the catalyst never exceeded 1-1.5%, even a tpH 8.Dependence on the Neutral-salt Concentration.-Whendecompositions were carried out at different ionic strengths(maintained by Na,[SO,] or Na[C10,] in the range 0.05-1 .O mol dm-3), the decomposition curves coincided exactly.Dependence on the pH and Influence of Buflers.-The rateof decomposition depends strongly on the pH. A plot ofthe initial rate us. pH exhibits two maxima; a high maxi-mum a t ca. pH 10.6 and a considerably lower one at ca.pH 8.3 (Figure 3).The height of the lower maximumdepends greatly on the quality (phosphate, borate, carbon-ate, etc.) and the quantity of the buffering substances,while the larger one does not show such a dependence. Thesmaller maximum, superimposed on the ascending arm of3 x lop8 mol dm-3 OsO,; 298 K ; ionic strength: 1 mol dmz3Na+ adjusted with Na,[SO,]; pH maintained by pH-stat byadding H,SO,. (0.) and (U), measured values; (-),calculated via equation (8)Dependence on the Concentration of OsO,.-In the rangepH 9-12 the initial rate was proportional to the concen-tration of the catalyst in the interval 10-9-10-6 mol dmP3.A t lower pH and higher concentrations of OsO, the powerof the catalyst concentration m in the rate equation (i) isR = K[OSO,]~[H,O,]~ (i)less than one.When the concentration of osmium tetra-oxide reaches or exceeds mol dm-3 the rate becomesindependent of the catalyst concentration. In such casesthe reproducibility of the decomposition reaction is poorer.Eflect of Radical Scavengers at pH > 9.5.-The rate is notobservably affected by acrylonitrile ( 10-5-10-3 mol dm-3,methyl methacrylate (10-4-10-2 mol dm-3), ethanol (10-l moldm-3), or H,edta (ethylenediaminetetra-acetic acid, moldm-3) a t pH > 9.5. In the presence of acrylonitrile ormethyl methacrylate no polymer precipitation was observed.It should be noted, however, that polymer precipitationcould not be observed either when an dioxygen-saturatedsolution containing 0.2 mol dm-3 acrylonitrile a t pH 10.6TABLE 1Values of the exponents of the concentrations (mol dm-3) in the rate equation for the catalysed decomposition,Rate = constant .[OSO~]~[H,O,]~[H+]PRange of Exponenth variation Fixed parameters 7 3f------7 m n P[OSO4I [H20210 PH6 x 10-'-8 x lo-' 2 x 10-2 7.7 b 0.81 (9)1 x 10-8-8 x 10-8 2 x 10-2 10.8 0.98 (7)5 x 10-6-1 x 10-3 5 x 10-2 5.4 0.62 (10)r H 2 0 2 1 0 [OSO,l PH9 x 10-3-9 x 3 x 10-8 10.8PH4.3-6.36.2-7.59.0-9.5[OSOJ [Ha021 02.7 x 10-3 5 x 10-21.6 x 10-5 2 x 10-23 x 10-8 2 x 10-21.11 (8)-1.00 (11)-1.05 (10)-1.01 (6)I( The figures in brackets are the numbers of experiments from which the exponents were determined. By using 0.1 mol dm-3Adjusted by adding acid or alkali at the beginning.phosphate buffers. c Maintained with the aid of a pH-stat using dilute H2S04240 J.C.S. Daltonwas irradiated with (cobalt-60) y-rays for 30 min or longerat a dose rate of 2.5 x 1014 eV dm-3.*Detection of the OH Radical.-Because of the highreactivity of the OH radical its direct detection by e.s.r.spectroscopy was not attempted. However, the stableradical tmpio is a selective reagent for the OH radical in thepresence of 02-, H20,, and 02.7 As the adduct formed inthe interaction between tmpio and the OH radical is e.s.r.-inactive, the e.s.r. signal decreased in proportion with thequantity of the OH radical formed. Further evidence forthe presence of the OH radical in the reaction mixture wasobtained with the aid of an a-phenyl-N-t-butylnitrone,PhCH=N(0)But.The hyperfine splitting of the e.s.r.spectrum suggests the simultaneous presence of OH and02- radicals. Both radicals furnished the correspondinge.s.r.-active N-oxyl radical.13 When the OH radical was1- - .0 5 10 15t I minFIGURE 4 Influence of tnm on the bleaching of RNO at differentpH values. Conditions: 298 K ; light path length 0.1 cm,wavelength 420 nm. Curves: (a), pH 8.0, 1.0 x moldm-a OsO,, 3.1 x mol dm-s H,O,, 2.4 x lo-' mol dm-3RNO, 2.3 x lo-, mol dm-a tnm; ( ), pH 8.0, 1.0 xOSO,, 3.1 x lo-, HSO,, 2.4 x lo-' R!!O, 0.0 tnm; (O), pH7.66, 2.20 x OsO,, 2.92 x lo-, H,O,, 1.31 x lo-' RNO,3.02 x lo-, tnm; (a), pH 7.65, 2.26 x lov6 OsO,, 2.92 xlo-* H,O,, 1.31 x lo-' RNO, 0.0 tnmremoved with N-dimethyl-p-nitrosoaniline (denoted RNO) , l4N-oxyl radical production was observed, although in alower yield.The N-chloro- and N-bromo-derivatives of4-hydroxy-2,2,6,6-tetramethylpiperidine l6 could not beused successfully for detection of the 0,- radical.Further, the bleaching of RNO was measured as afunction of the concentration of various competitor mole-cules (comp) . The corresponding relative rate constantswere determined by plotting l/A[RNO] us. [comp]/[RNO],and using the value I"O~+RNO = 1.25 x dm3 mol-ls-l,17 the absolute rate constants were also calculated. Thedata obtained were compared with those in the literatureand furnished further evidence for the participation of theOH radical in the Os0,-catalysed decomposition of hydrogenperoxide.Eflect of Tetranitromethane (tnm) on the Bleaching ofRN0.-In order to learn whether OH or 0,- radicals arethe primary species in the reaction route involving the OH* Throughout this paper.: 1 eV = 1.602 18 x lo-'* J ,radical, the bleaching of RNO was investigated in thepresence of increasing quantities of tnm.It was foundthat the bleaching of RNO can be considerably inhibitedby tnm. When the tnm has been consumed quantitatively,the bleaching of RNO immediately sets in with a higherc cIFIGURE 57.0 7.5 8.0 8.5 9.0PHpH Dependence of the rate components for thedecomposition of hydrogen peroxide. Conditions : 298 K ;1.74 x lo-' mol dm-a OsO,; 8.2 x 10-2 mol dm-3 H,O,.Curves: ( l ) , rate of initiation involving OH radical, &OH;(2), remaining rate; (3), steady-state concentration of OHradical, [OH],bt.; (4), rate of reaction via OH radical, R O H ;( 5 ) , 100 ROH/Rtotal VS.pHEstimation of the Steady-state Concentration of the OHRadical.-An attempt was made to determine thestationary-state concentration of the OH radical bymeasuring the rate of bleaching of RNO. It was assumedthat this concentration is not influenced appreciably whenRNO is added to the reaction mixture in low concen-trations ( % mol dm") ; then, with the known value ofthe bleaching rate of RNO (AA/dt) and accepting thatA[RNO]/A[OH] = 0.5 for the bleaching reaction,17 [OHIstet.can be calculated as in (ii) where AA is the change in theabsorbance at 420 nm, l the path length, [RNO], the initialconcentration of the dye, and E = 3.42 x lo4 dm3 mol-lcm-l,14 the molar absorbancy of RNO at 420 nm.Thevalues obtained can be seen in Figure 5, curve (3).pH Dependence of OH Radical Formation.-BesidesRNO, a series of dyes of different types were applied in thecatalysed decomposition of hydrogen peroxide, such asxylenol orange, p-ethoxychrysoidine, methyl red, methyleneblue, par [4- (2'-pyridylazo) resorcinol] , erythrosine, Congored, neutral red, and erioglaucin-A. Figure 6 shows thatthe dyes behave similarly, i.e. all are bleached as a functio198275 -s00X-50 - \ < TQ,256 7 8 9 I0 11 12PHFIGURE 6 Bleaching of different dyes as a function of pH.Conditions: 298 K; 1.43 x lo-' mol dm-3 OsO,; 1.0 x lo-*mol dm-3 H202; 0.1 mol dm-3 phosphate buffer.Absorbancewas measured a t the appropriate wavelength after 45 min.Curves: ( l ) , 1.43 x mol dm-3 methylene blue a t 665 nm;(2), 3.32 x mol dm-3 par a t 415 nm; (3), 1.43 x lo-,mol dm-3 erythrosine a t 520 nm; (4), 1.85 x mol dm-3RNO a t 440 nm; (5), 1.43 x mol dm-3 Congo red a t 490 nm241of pH according to a curve which exhibits a maximum.Independently of the structures of the dyes, all the maximaare in the pH range 8.2-8.6.Rate of Production of OH Radical.-The rate of decom-position of H,O, decreases and reaches a lowest limitingvalue (remaining rate) when RNO or another OH inhibitoris added in increasing quantities t o the reaction mixture.A t the same time the rate of bleaching of the dye increasesand reaches an upper limiting rate.Twice the limiting ratewas considered to be the rate of initiation ( R i O H ) of the OHradical; R i o H depends on the pH according to a curve witha maximum at pH 8.6 [Figure 5, curve (l)]. The remainingrate of hydrogen peroxide decomposition or the reducedrate of 0, evolution shows clearly that the decompositionhas another independent route, which is not stopped whenall the OH radicals are scavenged [Figure 5, curve (91.The difference between the total rate of decomposition(Rtotal) and the remaining rate in the presence of a sufficientquantity of the appropriate scavenger gives the rate of thatreaction component which proceeds via the OH radical[Figure 5, curve (4)]. The percentage contribution of theOH radical route (100 ROR/Rtotal) is large a t low pH anddecreases considerably as the pH is increased [Figure 5,curve ( 5 ) ] ; at pH > 9.5 this contribution amounts to onlyDependence of the Rate of Bleaching of RNO o n the Con-centrations of OsO, and H,O,.-The relationship (iii) was1-2y0.R i o H = constant.[H,O,]P[OsO,]~ (iii)found, where q was always 1, independent of the pH, whilep = 0.86 a t pH 6.8 and 0.53 at pH 8.3. Similar inform-ation is presented in Figure 7, where the changes in theinitial rate of decomposition of H,O, and in that of theinitiation are plotted as functions of ( a ) the hydrogenperoxide concentration and ( b ) the catalyst concentrationon logarithmic scales.FIGURE 7 Dependence of log rates and of the chain length on the log concentrations of H202 (a) and OsO, ( b ) , respectively.Conditions: 298 K; pH 6.80; 2.2 x lopg mol dm-3 OsO,; and 8.2 x Curves: (l),.log initialoverall rate of decomposition of H202; (2); log initial rate of decomposition of H202 via OH radical; (31, log rate of initiation (logR I O H ) ; (4), length of reaction chain involving OH radicalmol dm-3 H,02, respectively242 J.C.S.DaltonDetermination of the Length of the Reaction Chain involvingthe OH Radical.-If the difference between the number ofmoles of hydrogen peroxide decomposed in the absence andin the presence of a scavenger, applied in a quantity sufficientto remove all the OH radical formed, is calculated anddivided by twice the number of moles of scavenger oxidisedby the OH radical during the same period of time, the lengthof the chain involving the OH radical can be obtained.Thechain length increases as the concentration of H,O, increasesbut decreases with increasing concentration of OsO,[Figure 7(a), (b), curve (4)]. The values obtained by usingdifferent scavengers are in good agreement (Table 2).TABLE 2Length of the chain determined by different inhibitors a tmol 298 K, 8.2 x lo-, mol dm-s H,O,, and 5.80 xdm-s OsO,n n(pH 6.80) (pH 8.10)Inhibitors ,--L--, (-'-,Erioglaucin-A 75 70 69 13.5 13 15 15Safranine T 73 70 70Thymine 78 73 7073RNO 77 75 75 12.5 12 10.5The chain length was dependent on the pH: at pH 6it has a value of 80-90, while at pH > 9 it is ca. 10 andno longer depends on the pH (Figure 8).7.0 8.0 9 .oPH"a410.5FIGURE 8 Dependence of chain length on pH.Conditions:298 K; 1.7 x lo-' mol dm-a OsO,; 8.2 x mol dm-3 H,O,Temperature Dependence of the Decomposition Routes ofHydrogen Peroxide.-The temperature dependence of thecatalysed decomposition of H,O, was determined in theabsence and in the presence of inhibitors (RNO, propan-2-01,tnm, thymine, etc.) by measuring the rate of 0, evolutionand the rates of disappearance of hydrogen peroxide and ofthe dye molecule. The data obtained (Figure 9) indicatethat the apparent energy of activation of the disappearanceof inhibitor, i.e. of the reaction component involving the OHradical, has a value of 105-110 kJ mol-l a t pH < 9,which drops to 35-40 kJ mol-l at pH > 10.The remain-ing rate component in the presence of radical scavenger alsohas a temperature dependence of sigmoidal character : theapparent activation energy changes from 60 to ca. 15 kJmol-l when the pH is altered as above. The overalltemperature dependence can be calculated from the actualtemperature dependences of the reaction components,taking into consideration the actual contributions of thedifferent reaction routes.7 a 9 ?OPHFIGURE 9 pH Dependence of the apparent activation energies.Curves : ( I ) , temperature dependence of bleaching of RNO ;(2), temperature dependence of decomposition of H,O, : (3),temperature dependence of remaining rate component (totalrate minus that of the radical route)Detection of Usmium(vr).-At pH > 10 osmium(vIr1) isreduced by H,O, to an extent depending on the pH.Osmium(v1) was detected by its U.V. absorption spectrumand also by polarography.The formation of differentosmium species in the OsO,-H,O, system will be discussedin a subsequent paper.DISCUSSIONIn the absence of catalyst the decomposition of hydro-gen peroxide was found to be very slow and a homo-geneous process; the extent of the background (un-catalysed) reaction never exceeded 1% of that of thecatalysed one, i.e. the contribution of the uncatalysedprocess can be neglected.The fact that the dependence of the rate on theconcentration of H,O, can be expressed by an exponentof 1-1.2 indicates that the decomposition is a complexprocess. However, it was shown that this fractionalexponent did not arise from the volatility of the OsO, orfrom the purging effect of dioxygen evolution.Investigations using hydrogen peroxide doubly-labelled with oxygen-18 , proved that the dioxygenevolved during the Os0,-catalysed reaction at pH 10.6 isderived exclusively from hydrogen peroxide molecules1982 243and neither scrambling of the oxygens of H2l8O2 noradmixture of oxygen atoms of the water molecules withthe dioxygen was observed.This finding may be usedin support of a two-electron mechanism, however, it mustbe noted that a series of one-electron reactions is possiblewhich also do not require breaking of the 0-0 bond.The Non-radical Mechanism pH > 9.5.-From theobservation that the rate of decomposition of H,O, ishardly influenced by the addition of radical scavengers,it can be concluded that the decomposition does not takeplace (or predominantly not) in one-electron .stepsinvolving the OH radical. Therefore, an attempt ismade below to explain part of the experimental factsassuming two-elect ron react ions.As a plausible explanation one may assume that thecatalysis occurs via a two-equivalent redox cycle inwhich osmium( ~ I I I ) and osmium(v1) species are involvedalternately. If it is accepted that the non-dissociatedhydrogen peroxide is the more oxidizing species and itsanion is the more reducing species, we can write *equations (1)-(3). According to equation (3) the rateOsO, + H02- - OsO, + 0, + OH- (1)OsO, + H202 - Os04 + H,O (2)k2COsOil [H2021 (3)- d[H,O,]anal./dt = ~1[Os04][HO2-] +of decomposition will increase with increasing pH.Anincrease of the peroxide anion concentration is accom-panied by an increase in the first rate component. Onfurther increase of the pH the rate of decomposition willdecrease, because the reoxidation of OsO, will be de-creased. From this it follows that the rate of decom-position should show a maximum, but this mechanismcannot account for the experimentally observed depen-dence of the rate on the concentration of hydrogenperoxide, i.e. n > 1.The observed phenomena can be better explained byassuming the formation of peroxo-osmic acid which,being a strong oxidant, attacks free hydrogen peroxide.There is no reference in the literature to the formation ofperoxo-osmic acid; its existence can be inferred from thepolarographic behaviour of the Os04:H,02 system,'* andthe occurrence of an induced reaction when hydrogenperoxide is oxidized by a one-equivalent oxidant[cerium(rv), permanganate] in the presence of osmiumtetraoxide.ll On the other hand, the formation ofperoxo-acids is a general phenomenon for the transitionmetals, and thus this assumption appears to be reason-able by analogy.In our view, the initial step of the reaction is theformation of peroxo-osmic acid [equationsuitable pH values, this dissociates as an acid(5)].The rate of decomposition of H202* The various rate and equilibrium constants k,,defined by the correspondingly numbered equations.( 4 ~ .~t(4)[equationdoes notkg, etc, aredepend on the ionic strength. In the range pH 1O.P-10.9,however, where the maximum rate of decomposition wasfound, the vast majority of the free hydrogen peroxideis present in the non-dissociated form and this leads usto consider that the peroxo-osmate anion reacts as anucleophile with non-dissociated hydrogen peroxide[equation (6)]. Taking into consideration the dissoci-ation of hydrogen peroxide [equation (7)], rate equation(8) is obtained.O,Os(OH)OO- + H,O, =+= [XI +0, + OsO, + H20 + OH- (6)l- 1It is clear from this equation that the rate of decom-position of hydrogen peroxide is proportional to theanalytical concentration of OsO, and to that of H20,raised to a power somewhat greater than one.Ifthe terms not containing hydrogen peroxide are collectedtogether into suitable constants, a simplified rate equ-ation (9) results. By integration of this, the values ofa and b giving the best fit to the kinetic curves weredetermined with a minimizing program. Satisfactoryagreement can be attained with this approximation upto about 60-70y0 conversion (Figure 3 solid line curve).This mechanism satisfactorily reflects most of theobservations in the ranges pH 9.5-12 and [OsOJ =10+---10+ mol dm-, up to 60-70% conversion. Theremaining experimental facts, e.g. that there is anotherrate maximum at pH 8.3, and the formation of the OHradical, require, however, a more complete explanation.The Free-radical Mechanism, pH < 9.5.-By using thefree-radical reagents tmpio and a-phenyl-N-t-butyl-nitrone, it was demonstrated that the OH radical isformed during the catalysed decomposition. This wassupported by experiments with different dyes.It wasalso found that only restricted formation of the OHradical (RNO bleaching) took place when the superoxideradical was removed with tnm from the reaction mixture.This finding indicates that the OH radical is of bothprimary and secondary origin. It is assumed that thehomolytic fission of the 0-0 bond of the peroxo-osmicacid is the primary source of the OH radical [equation(Wl244 J.C.S. DaltonThe OH radical can be formed, however, in a greaterquantity as a consequence of the appearance of thesuperoxide radical in the reaction mixture. Theappearance of the superoxide radical suggests that inalkaline media the peroxo-anion decomposes in a one-electron redox reaction furnishing osmium(vI1) and thesuperoxide radical [equation (1 l)].Further, the OHradical can be formed in two parallel reactions [equations(12) and (13)].0,- + H,O, 0, + OH + OH- (1.2)(13) 0svI1 + H,O, - 0sV1Ir + OH + OH-The OH radical produced can react either with hydro-gen peroxide, resulting in the superoxide radical, again,or with osmium(v1r) [equations (14) and (15)].OH + H,O, - H,O + 0,- + H+ (14)OH + OsVII + H+ + H,O + OsVIII (15)As regards the reliability of the use of the RNOreagent for the determination of the concentration of OHradicals, the following should be considered.Accordingto Baxendale and Khan1’ RNO is transformed by theOH radical (produced by pulse radiolysis of an RNOsolution saturated with N,O) into an adduct the form-ation of which results in the immediate bleaching of thedye (band at 440 nm) and the appearance of a new bandat 350 nm. This band shows second-order decaykinetics in which half of the lost intensity at 440 nm isrecovered with a rate constant of 7.0 x lo8 dm3 mob1 s-l.Accordingly the equations below can be written.RNO + OH - RNO-OH (loss of band at 440 nm)2 RNO*OH -+ RNO + RNO, + H,ODainton and Wiseall l9 proved by thin-layer chromato-graphy that besides the main product, RNO, (NN-dimethyl-$-nitroaniline) , other substances are formedduring the pulse radiolysis of an RNO solution.Itshould also be added that Holcman and Sehested20pointed out the possibility of the addition of the OHradical to the aromatic ring, and in the case of NN-dimethylaniline showed that the hydroxycyclohexadi-enyl radicals formed not only decay in a second-orderradical-radical process, but also take part in a com-petitive first-order water-elimination reaction. Thereare no data in the literature indicating how RNO andits reaction products with OH are influenced by theconstituents of the reaction mixture investigated here.In this context, the roles of hydrogen peroxide and thesuperoxide radical should be considered. If the adductof RNO were to undergo reactions with H,O,, the radicalyield of the catalysed decomposition of hydrogenperoxide would be altered as shown below.- RNO + H,O + HO,RNo*oH -k H202--(- RNO, + H,O + OHOne has to recollect, however, that the pulse radiolysis ofan RNO solution saturated with dioxygen gives asimpler product distribution and an almost completerecovery of RNO bleached, due to the following re-act ions. - RNO + OH- + 0, - RNO, + H0,-RNO*OH + O,---ISince we do not know enough about the probability of theabove reactions involving hydrogen peroxide, we canstate qnly that the figures obtained for the stationary-state concentration of the OH radical by this method aremerely of informative value and serve as upper estimatesof the true values.It was found that the steady-state concentration of OHpasses through a maximum as the pH is increased.Thismay be interpreted in that the radical species formed instep (11) react with one other before they can diffuse outof the solvent cage [equation (IS)] and OsVII is reduced0,- + OSVII z 0, + OSVI (16)by hydrogen peroxide with an increasing rate when thepH is increased [equation (17)]. Because of the pHOsVII + H0,- --+ OsVI + 0,- + H+dependence of steps (16) and (17) the production of theOH radical decreases as the pH increases. The form-ation of osmium(v1) was observed by different indepen-dent methods at pH > 10.The osmium(v1) formed will be reoxidized by hydrogenperoxide in a presumably two-electron step [equation(ls)]. The rate of regeneration of osmium(vII1) is(17)OsVI + H,O, + 2 H+ + OsVIII + 2 H,O (18)proportional to the square of the hydrogen ion con-centration of the reaction mixture.The length of the chain v is defined as the number ofmoles of H,O, protected from decomposition by removingthe OH radical divided by the number of moles of OHradical captured (=2 [RNOIbleached). It follows fromthis definition that the value of v decreases with increas-ing concentration of osmic acid.The rate of step (19)(19) 0,- + OSVIII - 0, + OSVIIincreases with the increase of the concentration ofosmium tetraoxide, resulting in an increase in the quantityof the end-product, i.e. a decrease in the number ofmoles of hydrogen peroxide protected from decom-position. At the same time osmium(vI1) is obtained inhigher concentration, which gives rise to a higherconcentration of OH via step (13).The latter enhancesthe RNO bleaching, i.e. the value of the denominator ofv is increased. The effect of hydrogen peroxide is justthe opposite. On increasing the concentration of H,O,,reaction (1 1) furnishes 0,- and osmium(vI1) intermediatesin increased quantities, while steps (12) and (13) yield ahigher OH radical concentration, thereby giving back0,- by step (14) at a higher rate, etc. The higheconcentration of hydrogen peroxide enhances the rate ofstep (14) and competes with the reaction between OHand RNO; this results in a decrease of the denominator,i.e. the value of v increases.At higher pH where osmium(vI1) is reduced rapidly by0,- [step (IS)] and by the peroxide anion [step (17)] thelength of the chain is reduced because the re-oxidation ofOsVI by hydrogen peroxide [step (lS)] is slower.The overall apparent activation energy (EapP) andthe temperature dependence of the bleaching of RNO arerather complicated parameters.The parameter Esppcontains the temperature dependence of all stepsinvolved in the decomposition Eapp = El, + Ell +El, + . . . + E19. In the presence of OH scavengers,e.g. RNO or propan-2-01, the apparent energy of activ-ation of the remaining reaction would be determined bythe steps which do not involve the OH radical; Eappmm =E,, - (Elo + El, + E1J. The energy required forbreaking the 0-0 bond of the peroxo-complex is fairlyhigh in comparison with that of steps (14) and (15). Incontrast, the energy requirement of step (1 1) seems to beless than that of (lo), if it is considered that, on the onehand, the transfer of an electron from the peroxideoxygen to osmium(vII1) results in an increase of the 0-0bond strength from ca. 210 to ca.370 kJ mol-l, and, onthe other hand, the bond between osmium(vI1) and thesuperoxide ion formed is weakened because of thedecrease in electronegativity of the core due to thereduction of osmiumlvII1).We thank Dr. G. Putirskaya (Central Research Institutefor Chemistry, Hungarian Academy of Sciences, Budapest)for donating the tmpio stable free radical, and Dr. L.HorvAth (Institute of Biophysics, Biological ResearchCenter of the Hungarian Academy of Sciences, Szeged) fore. s.r. measurements.[1/467 Received, 23rd March, 19811REFERENCES229.Abstr., 1965, 62, 15625~.145.1 L. Chugaev and Bikerman, 2. Anorg. Allg. Chem., 1928,172,J. C. Dran, Commis. Energ. At. Fr., Rapp., 1964,2604; Chem.F. Domka and B. Marciniec, Chem. Anal. Warsaw, 1969, 14,V. A. Lunenok-Burmakina, personal communication.ti V. A. Lunenok-Burmakina, G. G. Lezina, V. B. Emel’yanov,S. K. Rubanik, and L. G. Sevsuk, Zh. Fiz. Khim., 1974, 48, 197.J. D’Ansand J. Mattner, Angew. Chem., Int. Ed. Engl., 1952,64,448.G. V. Putirskaya and J. Matus, Radiochem. Radioanal. Lett.,1978, 35, 227.A. K. Majumdarand J. G. Gupta, Anal. Chim. Acta, 1959,20,532.@ G. H. Ayres and W. N. Wells, Anal. Chem., 1950, 22, 317.lo P. K. Norkus and Yu. Yu. Yankauskas, Zh. Anal. Khim.,1973, 28, 127.l1 L. J. CsAnyi, ‘ Induced Reactions in Chemical Analysis,’Treatise on Analytical Chemistry, eds. I. M. Kolthoff and P. J.Elving, Wiley, New York, 1979, vol. 2, p. 734.la M. 2. GalbicsandL. J. Csinyi, Anal. Chem., 1973, 46, 1784.la J. R. Harbour, V. Chow, and J. R. Bolton, Can. J . Chem.,l4 F. Krajlic and G. N. Trumbore, J . Am. Chem. SOC., 1966,87,l6 A. Rigo, E. Argese, R. Stevanato, E. F. Orsega, and P.16 A. Rigo, E. Argese, E. F. Orsega, and P. Viglino, Inovg.17 J. H. Baxendale and A. A. Khan, Int. J . Radiat. Phys.18 L. J . Csinyi and K. Fulop, Acta Chim. Acad. Sci. Hung.,lo F. S. Dainton and B. Wiseall, Trans. Faraduy Soc., 1968,64,2o S . Holcman and K. Sehested, J . Phys. Chem., 1977,81,1963.1974,62, 3649.2547.Viglino, Inorg. Chim. Acta, 1977, 24, L71.Chim. Ada, 1979, a, 161.Chem., 1969, 1, 11.1963, 88, 193.694