J. CHEM. SOC. FARADAY TRANS., 1994, 90(4), 617-623 617 Electrochemical Study of the Heterogeneously Catalysed Reaction between N,N-Dimethyl-p-phenylenediamine and CO~~~(NH,),CI* at+ Monometallic and Bimetallic Surfaces of Silver and Gold Yao-Hong Chen and Ulrich Nickel" Institute for Physical and Theoretical Chemistry, University of Erlangen-Nurnberg, Egerlandstr. 3, D-91058 Erlangen, Germany Michael Spiro Department of Chemistry, Imperial College of Science, Technology and Medicine , South Kensington, London, UK SW72AY The mechanism of the heterogeneously catalysed reaction between N,N-dimethylp-phenylenediamine and Co"'(NH,)5C12+ at silver, gold and silver-on-gold (Ag/Au) discs has been studied by means of electrochemical methods. Both the mixed (or mixture) potentials and the mixture currents were determined by recording the current-potential curves of the reactants.Silver halide, formed during the reaction, was determined by sub- sequent galvanostatic reduction. The reaction at silver was strongly inhibited by the formation of silver halide whereas the reaction at gold was inhibited by adsorption of the organic compounds as well as by iodide. At Ag/Au discs almost no inhibition occurred. The explanation is that the reduction of the cobalt complex takes place predominantly at the silver surface whereas the simultaneous oxidation of the p-phenylenediamine occurs on the gold. In this way the formation of inhibiting silver halide is suppressed as well as the inhibition caused by the adsorption of p-phenylenediamine.In a heterogeneously catalysed redox reaction between an oxidant (Ox) and a reductant (Red), the metal acts as a con- ductor of electrons from Red to Ox. In contrast to normal heterogeneous reactions, no direct contact of the reactants at the surface of the metal need take Contrary to elec- trochemical reactions no external power supply is necessary to cause the transfer of electrons through the metal/solution interface. The general mechanism of reactions of this kind is Ieschematically displayed in Fig. 1. A suitable model system to study both the kinetics and the mechanism of such catalytic reactions is the oxidation of N,N-dimethyl-p-phenylenediamine (PPD) by Co"(NH3),C12+. This reaction has been thoroughly studied by taking as cata- lyst silver and gold discs as well as the colloids of these noble metals.'-, The reaction occurs irreversibly because of the rapid hydrolysis of the reduced cobalt ~omplex.~ The first Fig.1 Schematic representation of the mechanism by which a nobleoxidation product of PPD is p-semiquinonediimine (SQDI+), metal catalyses a redox reaction an intensely coloured ion radical. SQDI' may give p-quinonediimine (QDI+) either by further oxidation or by dis- chloride could be partially or completely reduced by p-prop~rtionation.~,~The most important redox reactions are phenylenediamines, a reaction which is used to develop pho- summarized in Scheme 1. tographic material. This process, too, depends on the kind Experiments carried out with both noble metal colloids and concentration of halide ions (X-) present in the solution.and large noble metal discs have shown that silver is a better Scheme 2 summarizes the most important reactions catalyst than gold,,** but at the surface of silver the reduction responsible for both the formation and the reduction of silver of Co"'(NH3),C12+ was found to become inhibited because of halide on the surface of silver. the oxidative formation of silver chl~ride.~~~*'~ Chloride rel- The higher the concentration of halide ions the more easily eased from the reduced cobalt complex according to eqn. (4) does silver halide formation OCCU~.~+~ Owing to the different was usually sufficient to slow down the reaction rate con- solubility products, bromide is more effective than chloride siderably.This silver halide formation corresponds to the but less than iodide. On the other hand, the reduction of photographic bleaching process. On the other hand, silver silver halide becomes more difficult with increasing concen- PPD + CO"'(NH,),C~~+ noble metal * SQDI' + Co"(NH,),Cl+ +SQDI' + Co"'(NH,),CI2 noble metal QDI' + Co"(NH,),Cl+ + H' PPD + QDI+ + H+ 2SQDI' c.--Co"(NH,),CI+ CO:~+ 5NH, + C1-Scheme 1 Ago + CO"'(NHJ,C~~~ Ag+ + Co1'(NH3),C1+ (5) Agf + X-*A@ (6) noble metal AgX + PPD b Ago + SQDI' + X-(7) Scheme 2 tration of halide ions. Iodide disturbs more than bromide and much more than chloride. The addition of halide ions there- fore strongly influences the silver-electrocatalysed reaction between N,N-dimet h yl-p- phenylenediamine and Co"*(NH,),C12'.' The gold-catalysed reaction should not be influenced by the addition of halide ions because no insoluble gold salts are formed.However, with iodide an inhibition due to the adsorption of iodide at gold was This effect has also been observed with related catalytic system^.'^^'^ Experiments carried out recently with a mixture of silver and gold colloids have shown a superadditive effect.8 Under certain conditions the inhibition could be almost completely suppressed. In order to obtain more information about the mechanism of reactions of this kind, the oxidation of N,N-dime th y1-p-phen ylenediamine with Co"'(NH ,), C12+ has now been investigated by using metal discs.This paper deals with electrochemical studies on both the chemical and the electro- chemical formation of silver halide at silver and Ag/Au discs. The formation of the silver salt was studied indirectly by sub- sequent electrochemical reduction.' *' The experimental results were compared with the data obtained from the current-potential curves of the reactants and discussed in terms of the theory of mixture potential and mixture ~urrent.~. Experimental Rotating silver and gold discs of 4 mm diameter were taken for the study of both the oxidative formation of silver halide and the recording of the current-potential curves of N,N-dimethyl-p-phenylenediamineand [Con1(NH3),C1]Cl2. The silver halide was formed either electrochemically or chemi-cally with the cobalt complex.The discs were carefully pol- ished for at least l min with 0.25 pm diamond paste before each experiment and then thoroughly washed with water. Usually a rotation speed of 800 rpm was chosen. The zero- current potential was measured against a calomel reference electrode (0.1 mol dm-3 KCl). The Luggin capillary was filled with 0.1 mol dm-, KNO,. The silver halide formed on the disc by chemical oxidation with [Co'n(NH,),Cl]Cl, was determined by subsequent galvanostatic red~ction,'.~ taking the disc as a cathode in an electrochemical cell with a plati- num counter-electrode and a calomel reference electrode (0.1 mol dm-, KC1). Usually 0.1 mol dm-, KNO, + 0.01 mol dmP3 KC1 was chosen as electrolyte in an electrochemical cell.The current-potential curves for both the oxidation of N,N-dimethyl-p-phenylenediamineand the reduction of Co"'(NH3),C12+ were recorded in the same electrochemical cell, the noble metal discs being taken as either anode or cathode, as appropriate. Ag/Au discs were prepared by inserting the gold disc as cathode into a solution of 0.001 mol dm-, AgNO, and applying a current of -100 FA for 0.5, 1, 2, 5 or 10 s. Thus, a total charge of 0.05 mC up to 1.00 mC was passed. A simple calculation shows that 0.10 mC deposits 6.2 x 1014 silver atoms which corresponds to about three silver monolayers on the disc of 4 mm diameter. The thickness of a monolayer is about 0.288 nm according to the diameter of silver." The electrochemical measurements were performed with the potentiostat-galvanostat HEKA PG 284.J. CHEM. SOC. FARADAY TRANS,, 1994,VOL. 90 The N,N-dimethyl-p-phenylenediamine2HC1, the halides and the buffer substances Na,HPO, and KH,P04 were p.a. re- agents from Merck. The cobalt complex [Co1n(NH,),C1]C12 was synthesized according to Schlessinger.'6 All solutions were prepared with doubly distilled water which had been degassed by ultrasound under water jet-pump vacuum and saturated with nitrogen. The phosphate buffers were prepared according to Ssrensen, but diluted 1 :2 with water in order to avoid the formation of silver phosphate. All experiments were carried out under an argon atmosphere at 20 "C. Results and Discussion Experiments with Silver and Gold Discs Fig.2 shows some current-potential curves for the electro- chemical oxidation of N,N-dimethyl-p-phenylenediamine.All curves of this kind were recorded at the gold disc in order to avoid any disturbance because of the formation of silver oxide. For comparison the measurements were also carried out at a glassy carbon electrode. Almost the same results were obtained. The position of the curves depended on the pH. With increasing pH the curves were shifted to more negative potentials until a limiting value was obtained which is near the first protonation constant of N,N-dimethyl-p- phenylenediamine (pK = 6.3)." The insert figure shows the pH dependence of the half-wave potential.'8 Fig. 2 also shows several current-potential curves for the electrochemical reduction of [Co1''(NH,),C1]C12 .The results recorded at the silver disc electrode are indicated by 'S', while those obtained at the gold disc electrode are designated with 'G'.The position of these curves does not depend on the pH but does depend strongly on the kind of electrode material and on the kind and concentration of added halide ions. The higher the concentration of a given halide the more the curves shift cathodically. It is obvious that with silver as electrode the formation of silver halide is responsible for this shift. The electrochemical reactions for both the reversible oxida- tion of PPD and the reduction of Com(NH,),C12+ are sum- marized in Scheme 3.PPDI SQDI' + e-(84 SQDI' QDI' + e-(86) Co1*'(NH3),CIZ'+ e--Co"(NH,),Cl+ (9) Scheme 3 kc -40t --400 -200 0 200 400 ElmVvs. 0.1 rnol dm-3 calomel electrode Fig. 2 Current-potential curves of both N,N-dimethyl-p-phenylene-diamine (5 x rnol drnT3) at three different pHs and CO"'(NH,),CI*+ (5 x rnol dmP3) at pH 6.0 in the presence of the additional halide on a rotating silver disc (Sl-S4)or a rotating gold disc (G144);800 rpm, scan rate 10 mV s-'; (-) no addi-tional halide; (-.-.) 0.1 rnol dm-3 KCI; (---) loP3 rnol dm-3 KBr; (-+ .) lop3 rnol dm-3 KI. (a)pH 4.5, (b) pH 5.6, (c)pH 7.0. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Addition of eqn. (8a)and (9) gives eqn. (1) in Scheme 1, while addition of eqn. (8b)and (9) gives eqn.(2).The reactions (8) and (9) also take place at the surface of the noble metal in the absence of an external power supply if both redox couples are present in the solution together. The noble metal surface then takes up a mixed or mixture potential (Emix)so that it can act simultaneously as an anode for couple (8) and as a cathode for couple (9). An example is shown in Fig. 2. The mixture potential should not be confused with the zero-current poten- tial which is measured for a single redox system. At Emixthe net current is zero, owing to a balance between the anodic current resulting from the oxidation of PPD and the cathodic current produced by the reduction of Co"'. These two currents, termed mixture currents (Imix),are indi- cated by dashed vertical lines in Fig.2. By Faraday's law, Imix is directly proportional to the reaction rate at the surface (vc,Jmol m-2 s-') and is given by eqn. (I):2*14~1g ucat = Imix/nFA (1) where F is the Faraday constant, n the number of electrons and A the surface area of the catalyst (in m2). As long as the initial concentration of PPD exceeds that of the cobalt(r1r) complex the mixture current is likely to be in the region of the limiting current of CO"'(NH,)~CI~+. Under these condi- tions the kinetics were consistent with diffusion control, and both the mixture current and the measured reaction rate were proportional to the [CO"'(NH,),C~]~ concentration but + independent of the concentration of N,N-dimethyl-p-phenyle- nediamine.The heterogeneous reaction rate at both silver and gold discs is then given by eqn. (11):2*5 uCat= khct[Co"'(NH3)5C12t] (11) The curves displayed in Fig. 2 clearly show that the mixture current, and thus the heterogeneous reaction rate constant (k,,Jm s-'), depends on several factors. Most important is the kind of noble metal used as catalyst but also the kind and concentration of halide can strongly influence the position of the mixture potential and the value of the mixture cur- rent. Some values of the mixture potential and the mixture current, calculated from current-potential curves, are listed in Table 1. For the system presented in Fig. 2, the maximum mixture current and therefore the maximum reaction rate were obtained with the gold disc at pH > 4.5 (as can be seen by comparing the lengths of the dashed vertical lines).At the silver disc this reaction rate was obtained only at pH 2 7 and in the absence of additional halide ions. The initial reaction rate, however, was high also at a lower pH. The decrease of the reaction rate, i.e. the amount of inhibition, was found to depend on the relative position of the current-potential curves of PPD, Co"'(NH ,),C12+ and silver halide. The shift of the current-potential curves Sl-S4 to more negative potentials is caused by the formation of silver halide. Inspection of the curve S1 in Fig. 2 shows that in the absence of additional halide ions the formation of silver chloride began at about + 100 mV (us. 0.1 mol dmP3 calomel electrode).The necessary chloride ions were provided by the counter-ions of the salts employed. As at pH > 7 the mixture potential of the p-phenylenediamine and cobalt complex redox couples was considerably more negative, almost no inhibition occurred, but in the presence of additional halide the shift of the current-potential curve of Co"'(NH,),C12 to+ more negative potentials caused a dramatic decrease of the mixture current (cf:Table 1). The more negative the resulting mixture potential, the more difficult was the reduction of silver halide by PPD, because the mixture current for this catalytic process also became very small. In the presence of iodide the mixture current becomes zero. Indeed, the small amount of oxidized p-phenylenediamine present in the solu- tion was then able to oxidize silver! Particularly strong inhi- bition of the silver-catalysed reaction between PPD and Co1"(NH,),C12+ took place at low pH (because of the posi- tive shift of the current-potential curve of the p-phenylene- diamine), in the presence of a high concentration of chloride (shifting the current-potential curve of the cobalt complex to more negative potentials due to the formation of silver halide), and even more so in the presence of bromide and iodide (which shifted the cathodic current-potential curve to still more negative positions).The influence of additional halide on the current-potential curves of CO*~'(NH,),C~~+ recorded at a gold disc electrode is also displayed in Fig.2 (curves G144). In contrast to silver, the mixture current was not decreased by the addition of chloride and bromide, because Emixremained in the region of the limiting current for Co"'(NH3),ClZ+ except when the pH became lower than 4 (see also Table 1). However, as with the silver curve S4, the mixture current decreased strongly in the presence of iodide (curve G4). This was probably caused by inhibition of the electron transfer by adsorption of iodide on the surface of g0ld."*'~9~~ The desorption of iodide only occurred at a rather negative potential. However, not only iodide but also N,N-dimethyl-p-pheny- lenediamine itself and its oxidation products inhibited the catalysis due to an adsorption effect." The current-potential curves G2 and G3 in Fig.3 show the strong shift of curve G1 to more negative positions after a short treatment of the gold disc with the solution of the p-phenylenediamine. The corresponding values of Emixand Zmix obtained from these curves are summarized in Table 2 (left-hand column). The curves SG refer to Ag/Au experiments and will be discussed later. Because of the decrease of the mixture current the rate of the catalytic reaction between N,N-dimethyl-p-phenylene-diamine and Co1"(NH,),C12 should also decrease. In fact, + this decrease was found both at large gold discs and with colloids.2,8 In the absence of additional halide and at pH > 5 the reaction rate between PPD and Co"'(NH3),C12+ was usually considerably slower on the surface of gold than on the surface of silver.2*8 Table 1 Values of Emirand Imix under different pH conditions and with additional halide ([Co"'], = 0.5 mmol dm-3 and [PPD], = 0.5 mmol dm-3 at 800 rpm) Ag disc Au disc PH [KCl]/mol dm -[KBr]/mmol dm -[KI]/mmol dm -Emix/mV 'mixlPA EmixImV ImiJPA 4.5 --95 4 210 28 5.6 --90 15 140 28 7.O --65 28 65 28 6.0 --90 25 90 25 6.0 0.1 --20 5 90 25 6.0 ---50 4 90 25 6.O -1.o none 0 -85 2 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I 0.3f : 'I ................. Q: 1 -20 ...... . i-401 -1 I 1 1 -400 -200 0 200 400 E/mV vs.0.1 rnol dm-3 calomel electrode Fig. 3 Influence of the adsorption of N,N-dimethyl-p-phenylenedia-mine on the mixture current of the redox system N,N-dimethyl-p- phenylenediamine (5 x rnol dm-3)-Co111(NH,),C12+ (5 x mol drn-,) at pH 6.0 on a rotating Au disc (G143) and a rotating Ag/Au disc (SG1-SG3); 800 rpm, scan rate 10 mV s-'.The Ag/Au disc was obtained by inserting a gold disc into a solution of lo-, mol dm-, AgNO, and galvanostatically reducing it for 10 s at -100 PA. Both Au and Ag/Au discs were pretreated with 0.1 mol dm-, PPD for 0 s (full lines, G1 and SGl), 10 s (dashed lines, G2 and SG2) and 30 s (dotted lines G3 and SG3), respectively. Experiments with Ag/Au Discs Fig. 4 shows some current-potential curves for [CO"~(NH,),C~]C~, recorded at several Ag/Au disc elec- trodes. At very negative potentials only the reduction of the cobalt complex took place, but when the potential became more positive, a parallel reaction set in: the oxidative forma- tion of silver chloride.Thus, the cathodic current decreased -200 0 200 400 E/mV vs. 0.1 rnol dm-3 calomel electrode Fig. 4 Influence of both the amount of deposited silver and the addition of chloride ion on the electrochemical reduction of Co1"(NH3),C12+ (5 x lod4 rnol drn-,) at pH 5.6 on Ag/Au discs rotating at 800 rpm, scan rate 10 mV s-'. The amounts of silver deposited on gold discs 1-3 were 0.05,O.lO and 0.20 mC, respectively. (-) No additional halide; (---) 0.1 mol dm-, KC1 (with 0.10 mC silver on the Ag/Au disc). Insert: Dependence of the AgCl formed (QAICI)on the amount of silver (QAI)obtained from curves 1-3. QAga was calculated from the total areas of the corresponding peaks.thin silver chloride layer, the electrocatalytic reaction +between PPD and Co"'(NH3),C12 can therefore remain diffusion-controlled. The limiting current which under certain conditions (e.g. at pH > 5 and with an excess of PPD) equals the mixture current (see Fig. 2) is given by the relation (111) llim nFD, A[CO"'(NHJ,C~~']/a (111)= where D,, (in m2 s-') is the tracer diffusion coefficient of the cobalt complex in the phosphate buffer and or the net current even became anodic. Because of the limited is the concentration of the cobalt complex '3 [Co"'(NH3),C12 amount of silver deposited on the gold disc this oxidative formation of silver chloride stopped when all the silver had been transformed into silver chloride. As the potential rose still further, the only electrode process was again the reduction of the cobalt complex.The more silver had been deposited on the gold disc electrode, the higher was the resulting anodic peak. The amount of AgCl could be deter- mined from the total area of each peak in these curves; it was proportional to the amount of silver deposited on gold (see insert in Fig. 4), but independent of the concentration of chlo- ride present in the solution or released from the cobalt complex. This is shown by the dashed curve, which corre- sponds to the conditions of Fig. 4, curve 2, but was recorded in the presence of 0.1 mol dm-3 KC1. The addition of chlo- ride only shifted the peak cathodically, i.e. the formation of silver chloride began at a lower potential.As the formed silver chloride is porous,1othe deposition of AgCl on the gold disc decreased the limiting current (Imin)for the reduction of Co"'(NH,),Cl2+ only a little. In spite of a Table 2 Values of Emix and Imixobtained from the current-potential curves displayed in Fig. 3 none 90 25 90 25 10 0 6 90 25 30 -40 3 50 20 Discs were pretreated with 0.1 rnol dm -N,N-dimethyl-p-phenyle-nediamine (PPD) at pH 6. t is the pretreatment time of the disc with PPD. in the bulk. At a rotated disc the diffusion layer thickness 6 (in m) is given by the Levich equation:22 6 = 0.643D~.3v'/6f-1/2 (IV) where v is the kinematic viscosity of the solution (in aqueous solution ca.m2 s-') and f is the rotation frequency of the disc (in s-'). Fig. 5 shows some examples for the dependence of the lim- iting current (Ilim)on the square root of the rotation fre- quency (J) for the reduction of Co(NH3),C12+ at an Ag/Au disc. The value of the diffusion coefficient (Dco)was obtained 40 20L4 t I 1 -200 0 200 400 €/mV ws. 0.1 mol dm-3 calomel electrode Fig. 5 Dependence of the limiting current (Ilim)at two potentials on the speed of rotation of a rotating Ag/Au disc for the reduction of Co"'(NH3),C12+ (5 x rnol dm-3) at pH 5.6; scan rate 10 mV s-'.The amount of silver deposited on the gold disc was 0.10 mC. (a) 400, (b) 800 and (c) 2000 rpm. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 from eqn. (HI), taking Zlim at -100 and and +220 mV, respectively (indicated by the dashed line).At -100 mV a value of D,, = 7.16 x lo-'' m2 s-' was obtained, and at +220 mV the slightly lower value of 6.63 x lo-'' m2 s-'. The first value corresponds very well to the values obtained in previous papers in the absence of any phosphate buffer (7.24 x lo-'' m2 s-l lo and 7.30 x lo-'' m2 s-' '). As the diffusion coefficient should not depend on the potential of the electrode in this potential range, it can be assumed that the electron transfer through the interface of Ag/Au and of AgCl/Au is similar to that through the surface of pure silver or gold2,' and that the deviation in the value of D,, is only caused by a slight decrease of the active surface area of the disc.Similar results were obtained when the concentration of [Co"'(NH,),Cl]Cl, was varied between 0.2 and 0.5 mmol drn-,. The amount of electrochemically formed AgCl did not depend on the concentration of [Co"'(NH,),CI]Cl, ,but only on the amount of silver deposited on the surface of gold. In accordance with eqn. (111), the limiting current was pro-portional to the concentration of the cobalt complex. By combining eqn. (I) and (11), the rate constant k,,,/m s-could be calculated: khet = ZmiJ( n FA[Co(NH,) $1 I} (V)+ A value of khet = 5.0 x lo-' m s-' was obtained. This value is somewhat higher than that of 3.6 x lo-, m s-l reported previously.2 The difference can be explained by the slightly higher value of the mixture current obtained at the Ag/Au disc, because it more closely approached the limiting current for the reduction of the cobalt complex. It should be men- tioned, however, that this value is obtained only if the mixture current equals the limiting current of the cobalt complex reduction, i.e.in the presence of an excess of PPD. Fig. 6 shows some current-potential curves for the t i 11 3 -20-5 0 \ -40 --80 0 0.2 0.4 0.6 0 Q,,/mC -400 -200 0 200 400 €/mV vs. 0.1 rnol drn-3 calomel electrode Fig. 6 Influence of the amount of silver on the electrochemical reduction of silver chloride on a rotating Ag/Au disc in a solution of 0.1 rnol dm-, KNO, + 0.01 mol dm-3 KCl; 800 rpm, scan rate 10 mV s-'. Current-potential curves of the reduction of silver chloride started from the zero-current potentials (ca.300 mV) which were measured at the beginning. The amounts of silver deposited on the gold discs were 0.06 mC (curve l), 0.12 mC (curve 2), 0.21 mC (curve 3) and 0.60 mC (curve 7), respectively. (a) pH 4.5, (b)pH 7. Silver chloride was formed by the chemical oxidation of the silver with Co"'(NH,),Cl, (5 x mol dm-,) at pH 6.0 for 15 s. Insert: Dependence of both the zero-current potential (E")and the amount of AgCl formed (QA,,ci) on the amount of silver (QA,) on an Ag/Au disc. QADCl was calculated from the total area of each peak in the figure. 621 reduction of silver chloride deposited on an Ag/Au disc and for comparison the current-potential curves for the oxidation of N,N-dimethyl-p-phenylenediamineat two different pH values.The silver had been deposited on the gold disc gal- vanostatically as described above and the silver chloride was formed by subsequent chemical oxidation of the silver with Co"'(NH,),ClZ +. The more silver had been deposited, the more silver chloride could be formed during the chemical oxi- dation, but as the oxidation time was kept constant at 15 s, any large amount of silver deposited on the gold disc could not be completely converted into silver chloride. In order to avoid any dissolution of the silver chloride formed, the Au/ Ag/AgCl disc was then rinsed with water containing lop3 mol dm -potassium chloride. The current-potential curves were carried out in a cathodic direction, starting from the zero-current potential.It can be seen from the insert figure that the zero-current potential depended on the amount of silver deposited on the surface of gold. The silver chloride formed on the Ag/Au discs was deter- mined from the total area of the cathodic peaks in Fig. 6. When only a small amount of silver had been deposited on the surface of the gold, the silver was almost completely con- verted into silver chloride by the subsequent chemical oxida- tion. Under this condition, the activity of the residual silver (ffAg) was much smaller than unity because of the dominant gold surface. If, however, a relatively large amount of silver had been deposited, so that the silver had not been com- pletely converted into silver chloride, the activity of the silver could approach the value for pure silver, i.e.ffAg % 1. Assuming the validity of the Nernst equation for this gold/ silver/silver chloride electrode, the zero-current potential (EAg/Ag+) is given by RT RT--In aAgo--In ax-(VIb)F F where R is the gas constant, T the temperature in K and Leg, the solubility product of AgX. The zero-current potential, EAgO,Ag+, at the Ag/Au disc can therefore be much more posi- tive than at the surface of pure silver if the amount of silver is very small. In fact, the insert in Fig. 6 shows a strong depen- dence of the zero-current potential on the amount of silver deposited on gold. The points 1-4 correspond to the situ- ation where the chemical oxidation time was sufficient to convert almost all the silver into silver chloride.The points 6-8 refer to the situation where the oxidation time was not sufficient to convert all the silver into silver chloride. Although the zero-current potentials for runs 1-4 are strongly shifted to more positive values, the peaks in the current-potential curves are shifted much less. Nevertheless, the shift can be sufficient to allow reduction of silver chloride by PPD even at a relatively low pH. Fig. 6 shows that at pH 4.5 a mixture current is possible for experiments 1-3, but not for experiment 7. Therefore, at suitable Ag/Au surfaces the formation of silver chloride can be prevented in the presence of PPD as long as the pH is not too low. The influence of the kind and amount of halide ion in solu- tion on the position of the half-wave potentials of the Co1"(NH,),C12+ reduction curves is demonstrated in Fig.7. The full-line curves were recorded at a silver disc and the dashed-line curves at an Ag/Au disc. All sweeps were carried out in the anodic direction, as in Fig. 4. Numbers 2 and 3 refer to experiments in which small concentrations of J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 3 E?12values of 0.5 mmol dm-3 Co"' in the presence of halide at pH 5.6obtained at a silver, a gold and an Ag/Au disc (with 1.00 mC of silver deposited on it) [KCl]/mol dm -[KBr]/mmol dm -[KI]/mmol dm -Ag AU Ag/Au -$202 --40 calomel electrode bromide and iodide were added to the solution. The amount of silver bromide formed depended on the amount of silver deposited on the surface of gold, but it was independent of the concentration of bromide in the solution. The same was true for iodide.Some values of are listed in Table 3. These results can be explained with eqn. (VI) taking uAg< 1 at the Ag/Au disc. The potential was shifted to a more nega- tive value at an Ag/AgBr disc than at an Ag/Au/AgBr disc. Therefore, as with the results obtained with chloride, the reduction of silver bromide occurred more easily on the Ag/Au interface than on the surface of pure silver. As mentioned in the previous section, strong inhibition of the ConI(NH,),Cl2+ reduction occurred at the gold electrode after treating it with a solution of PPD (cf.the current- potential curves G2 and G3 in Fig. 3). At an Ag/Au electrode, however, no inhibition of this kind was observed after pretreating it with N,N-dimethyl-p-phenylenediamine.The curves SG1-SG3 in Fig. 3 are similar to those displayed in Fig. 4 and 5; only the amount of silver chloride formed sub- sequently decreased a little. Inhibition due to the adsorption of PPD was avoided because of the rapid reduction of Co~NH3),C12+ at the silver surface and the simultaneous oxidation of PPD adsorbed at gold. This reaction began immediately after the addition of the cobalt complex and before recording of the current-potential curves. Thus, com- pared with the situation at pure gold discs, a less negative shift of the potential occurred and as a result the mixture current was greater.Some results are summarized in the right-hand column of Table 2. It is clear from the above results that Ag/Au discs display different and catalytically useful properties compared with those of pure silver or gold discs. The introduction of a foreign metal into a metallic catalyst is a well established pro- cedure for modifying surface electronic and geometric struc- tures in order to manipulate catalytic selectivity.' 3*23 Usually the electrocatalytic properties of the bimetallic interfaces are strongly dependent on the relative strengths of the interaction between the reactant and the individual components of the mixed-metal ~urface.~~-~~ Silver and gold have very similar 100 330 110, 150, 350 -10 290 10, 40,250 -50 280 -20,20, 200 -250 150 -240 physical properties (e.g.similar atomic radius and lattice structure).' 5925 The experiments described above have shown that deposition of silver on the surface of gold resulted in a superadditive or synergistic effect. The reduction of Co1yNH3),Cl2+ occurred more rapidly at the surface of silver than at gold while the silver halide, formed according to eqn. (6), could be continuously reduced because PPD could inject electrons into the catalyst at the free surface of gold. The noble metal acting in reactions (1) and (2) is there-fore predominantly silver, whereas the noble metal acting in reaction (7) is gold. The evidence obtained in the present paper thus explains why the inhibition caused by chloride and bromide ions in the catalysis of reactions (1) and (2) by colloidal silver could be overcome by taking a mixture of silver and gold colloids, or better by using Ag/Au particle^.^*^ Conclusion The bimetallic interface consisting of silver islands on a gold substrate possesses unique catalytic properties distinct from those of the pure metals.The formation of silver halide, which is the main autoinhibiting effect in silver-catalysed reactions, can be avoided by electron transfer from the organic reducing agent viu the free gold surface. Thus, even silver halide which has already been formed may be reduced. More- over, the irreversible adsorption of the organic reactant, which is the most important inhibiting effect in gold-catalysed reactions, can be avoided by oxidation of the adsorbed compound via the silver surface.This mutual can- cellation of inhibiting effects has already been used to acceler- ate the colloid-catalysed oxidation of p-phenylenediamines by [Co"'( NH J,Cl] C1 . The authors thank the Deutsche Forschungsgemeinschaft for financial support. References 1 U. Nickel and C-Y. Liu, J. Zmag. Sci., 1990,34,8. 2 R. 0.Farchmin, U. Nickel and M. Spiro, J. Chem. SOC.,Faraday Trans., 1993,89, 229. 3 (a) U. Nickel and C-Y. Liu, J. Photogr. Sci.,1987,35, 191; (b) U. Nickel, C-Y. Liu, P. Lachenmayr and M. Schneider, Bull. SOC. Chim. Fr., 1988,308. 4 M.Spiro, Catal. Today, 1993.17,517. 5 R. 0. Farchmin, Ph.D. Thesis, University of Erlangen-Nurnberg, 1992.6 U. Nickel, K. Kemnitz and W. Jaenicke, J. Chem. SOC., Perkin Trans. 2, 1978, 1188. 7 U.Nickel and W.Jaenicke, J. Chem. SOC.,Perkin Trans. 2, 1980, 1601. 8 Y-H. Chen and U. Nickel, J. Chem. SOC.,Faraday Trans., 1993, 89,2479. 9 U.Nickel, E.Haase and C-Y. Liu, Ber. Bunsenges. Phys. Chem., 1990,94, 726. 10 M.D.Archer and M. Spiro, J. Chem. SOC.A, 1970,82. 11 W.Jaenicke and H. Kobayashi, Elecrrochim. Acta, 1983,28,245. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 623 12 13 14 15 A. M. Creeth and M. Spiro, J. Electroanal. Chem., 1991, 312, 165. (a) C. J. Cali, J. E. Harris, M. E. Bothwell and M. P. Soriaga, Langmuir, 1990 6, 74; (b) M. P. Soriaga, Chem. Rev., 1990, 90, 771. M. Spiro, Chem. SOC. Rev., 1986, 15, 141. (a) J. H. White, M. J. Albarelli and H. D. Abruna, J. Phys. Chem., 1988, 92, 4432; (b) A. Hamelin, J. Chim. Phys., 1991, 88, 1453. 21 22 23 24 M. D. Levi and E. Yu.Pisarevskaya, Electrochim. Acta, 1992,37, 635; G. Horanyi and E. M. Rimayer, J. Electroanal. Chem., 1984,176,339. V. G. Levich, Physicochemical Hydrodynamics, Prentice-Hall, Englewood Cliffs, 1962, p. 69. D. C. Alonzo and B. R. Scharifker, J. Electroanal. Chem., 1989, 274, 167. H. D. Abruna, Electrochemical Interfaces, Modern Techniques for in situ Interface Characterization, VCH Publishers, New 16 17 G. G. Schlessinger, Inorg. Synth., 1967,9, 160. U. Nickel, E. Haase and W. Jaenicke, Ber. Bunsenges. Phys. Chem., 1977,81,849. 25 York, 1991. (a) L-W. H Leung, D. Gosztola and M. J. Weaver, Langmuir, 1987, 3, 45; (b) D. P. Sandoz, R. M. Peekema, H. Freund and 18 19 20 M. Schneider, Ph.D. Thesis, University of Erlangen-Nurnberg, 1989. M. Spiro, J. Chem. Soc., Faruday Trans. I, 1979,75, 1507. M. Noel and K. I. Vasu, Cyclic Voltammetry and the Frontiers of Electrochemistry, Aspect Pub. Ltd., London, 1990. 26 C. F. Morrison Jr., J. Electroanal. Chem., 1970,24, 165. S. Szabo, lnt. Rev. Phys. Chem., 1991,10,207. Paper 3/05888C; Received 30th September, 1993