Effect of Surface Oxidation and Platinization on the Behavior of Platinum Electrodes Reduction of Vanadium(V) and Iodate FRED C. ANSON and DONALD M. KING California lnstitute o f Technology, Pasadena, Calif.
b There has been a difference of opinion in the literature regarding the mechanism by which prior oxidation of platinum electrodes produces increased reversibility in subsequent reactions at the electrode. Experimental evidence is presented for the reduction of iodate and vanadate at pretreated platinum electrodes to support a previous proposal that platinization of the electrode resulting from reduction of the platinum oxides on oxidized platinum electrodes i s the major factor contributing to the observed increases in reversibility.
I
a recent study of the effect of surface oxidation on the behavior of platinum electrodes i t was suggested that reduction of the oxide film on an olidized electrode results in a platinized electrode a t which electrode reactions proceed inore reversibly (1). This hypothesis mas shown t o be consonant with the behavior of the Fe(I1)-Fe(II1) couple a t platinum electrodes ( 1 ) . An earlier study had shonn that iodate ion is reduced a t less reducing potentials a t a previously oyidized electrode ( 2 ) . To account for this behavior an oxide bridge mechanism for electron transfer was invoked. However. on the basis of the later eyeriments n i t h the Fe(I1)-Fe(II1) couple ( I ) , it seemed possible that the behavior of iodate could be explained more satisfactorily by considering that the electrode becomes slightly platinized as a result of the reduction of the oxide film. In addition, it appeared that the rather unusual behavior of vanadium(V) a t oxidized and reduced platinum electrodes reported by Davis ( 6 ) could also be understood in terms of this platinization mechanism. The present study was designed to test these speculations. The behavior of vanadium(T’) and iodate was investigated chronopotentiometricallp M ith platinum electrodes subjected to a variety of pretreatment procedures. The experimental results are in accord with the platinization mechanism and can be adequately accounted for withK
362
ANALYTICAL CHEMISTRY
out resort to a n oxide electron-bridge mechanism. EXPERIMENTAL
Reagents. All chemicals used n-ere of reagent grade quality and were used ITithout further purification. Stock solutions were prepared by weight and measurements were made in oxygen-free solutions. ilpparatus. The chronopotentiometric apparatus has been described (1). The working electrode was a piece of platinum foil 0.9 sq. em. in area that had been spot welded to a short piece of platinum wire which was sealed in glass. Electrode Pretreatment. To bring the electrode into a standard condition before treating it for use in an euperiment, it was immerscd in hot aqua regia for several minutes. This treatment resulted in the dissolution of any finely divided platinum metal on the electrode. Kext the electrode n as conditioned according to one or more of the following procedures : An oxidized electrode was prepared by recording an anodic chronopotentiogram in a solution of the supporting electrolyte (1F H2S04 or pH 3 phosphate buffer). l h e current was interrupted just before the electrode potential reached thc steady background potential. A freshly reduced electrode was prepared by recording a cathodic chronopotentiogram li-ith an oxidized electrode and interrupting the current n hen the electrode potential was approximately 100 mv. less reducing than the steady reduction potential of the supporting electrolyte. l’he resulting electrode was then immediately used to record chronopotentiograms of the cubstances to be investigated. ,4n aged reduced electrode n-as a freshly reduced electrode that had been alloned to sit in a solution of t h e supporting electrolyte for several hours prior to its use in recording chronopotentiograms. A stripped electrode was prepared by dissolving the oxide film from a n oxidized electrode with hot concentrated hydrochloric acid. Previous experiments have shown that no decrease in reversibility results from exposing
unoxidized electrodes to hydrochloric acid (1). A platinized electrode was prepared by passing 1 to 2 ma. per sq. cm. of cathodic current through the reduced electrode for 20 to 50 seconds in a 0.025F K2PtC14 solution. Comparison of transition times for the reduction of Fe(II1) a t both freshly reduced and platinized electrodes showed that this amount of platinization did not change the effective chronopotentiometric area for transition times longer than 5 seconds. The evidence previously obtained to support the contention that a freshly reduced platinum electrode possesses a platinized surface was largely indirect (1). -4clear-cut and direct demonstration of the effect of repeated oxidation and reduction of the electrode surface has now been obtained xyith the follom-ing experiment: h new platinum electrode with a bright lustrous surface v,-as placed in a 1F H2SO4solution with an auxiliary electrode and the pair connected to a source of 60-cycle a x . with the peak to peak voltage amplitude adjusted so that the electrode surface n-as oxidized and reduced during each cycle but no substantial gas evolution occurred. Kithin 2 minutes after the a x . nas applied the electrode surface had lost its luster and after 10 minutes the electrode surface was dark and resembled the surface of electrodes which had been platinized by deposition from a chloroplatiiiite qolution. This experiment leaves 110 doubt that platinization does in fact result from alternate oxidation and reduction of platinum electrodes. RESULTS A N D DISCUSSION
Reduction of Vanadium(V). In Figure 1 are shown cathodic chronopotentiograms for a solution 0.01F in SHaV03 in 1F H2S04. Curves 1 through 4 correspond, respectively, to chronopoteritiograms obtained with the oxidized electrode, the freshly reduced electrode, the stripped electrode, and the platinized electrode. Typically, t n o waves are obtained corresponding to the stepwise reduction of T(V) to T’(1V) and T’(II1) ( 5 ) . Because of the excessive length and lesser importance of the second jvave
1
r
t
i TIME
Figure 1.
TIME
Chronopotentiograms for reduction of 0.01 M V(V) in 1 .OM H2SO4 at current density of 2.2 ma. per sq. cm.
only the first wave is shown in curves 1, 2, and 4. Chronopotentiograms for the oxidation and reduction of the electrode in a solution free of vanadate are not shown but \\ith a current density of 2.2 ma. per sq. em. the transition times involved are negligible compared to those obtained for the reduction of the vanadium. Curve 1 is similar to a chronopotentiograin obtained by Davis with an o\idized electrode. The unusual minimum in the curve was explained by him as resulting from a partial reduction of the platinum oxide film until the remaining oxide could provide electron bridges and allow the reaction to proceed a t the more positive potentials. Curves 2 and 4 shon-, however, that vanadiuni(V) is even more readily reduced a t freshly reduced or platinized c.lectrodes n here no such oxide bridges could be involved. K(.vertheless if the oxide film is chemically stripped from the electrode TI ith hot Concentrated hydrochloric acid the reduction of vanadiuni(V) does proceed much less reversibly, as shonn in curve 3. (Curve 2 gradually deteriorates into curve 3 as successive chronopotentiogranis are recorded.) This behavior is difficult t o account for in terms of a n oxide bridge mechanism but it is readily understood according to the electrode platinization idea as follows: It can be reasonably assumed that electrode reactions mill display their greatest reversibility-Le., least overvoltage-at platinum electrodes that have just been platinized. The s t m d a r d hydrogen electrode is the best known example of this rule. The reason, then, t h a t the reduction of n n a diuni(S7) commences later in curve 1 than in curves 2 and 4 in Figure 1 is because olidation of the electrode before curve 1 n a s recorded converted any finely divided platinum on its surface into platinum oxides thus effectively deplatiiiizing the electrode. The unusual minimum in curve 1 occurs a t the
Figure 2.
Chronopotentiograms for reduction of 0.002M IO3- in phosphate buffer solution of pH 3 at current density of 1.1 ma. per sq. cm.
potential where reduction of the platinum oxide takes place. This reduction produces a fresh layer .of finely divided platinum metal on the electrode 1% hich renders the reduction of vanadium(V) more reversible so the potential climbs back to the value a t which this more reversible reduction proceeds. So minima are observed in curves 2 or 3 because the electrode surface is already platinized and the reduction of vanadium(V) proceeds immediately a t the more oxidizing potentials. T o show that a n initially oxidized electrode behaves identically with a n initially reduced electrode once the oxide film is reduced, curves 5 and 6 were recorded nhile the solution was being stirred. The oxidized electrode (curve 5 ) still exhibits the minimum corresponding to the oxide film reduction but the potential then climbs back t o the value a t n-hich vanadium(T*) is reduced a t a n initially reduced electrode. The small inorphological differences in curves 2 and 4 doubtless result from the difference in the extent of platinization of the electrode in the two cases. The quantity of freshly formed platinum present on the electrode which had been plated in a chloroplatinite solution 11-as about 50 times greater than when it n as merely oxidized and reduced. This is responsible for the slightly earlier potential a t which vanadium(T’) reduction commences on the inore heavily platinized electrode. Curve 3 in Figure 1 corresponds to the reduction of vanadium(V) a t an unplatinized electrode resulting from oxidation of the electrode followed by chemical stripping of the oxide film in hot concentrated hydrochloric acid. The decrease in reversibility is so great t h a t no separate wave for the reduction to V(II1) is observed. Chronopotentiograms similar to curve 3 also result when freshly reduced electrodes are allowed to age before being used. This decrease in rerersibility with time is presumably the result of the freshly formed platinum
metal becoming deactivated either by adsorption of solution impurities or by gradual changes in the structure of the deposit ( 5). The fact that the very irreversible curve 3 can be converted to the much more reversible curve 1 by oxidizing the electrode does not reflect participation by oxide bridges but only the !vel1 linon-n fact that deactivated platinum can be reactivated by oxidation and reduction ( 7 ) . The mechanism of this reactivation is simply t h a t oxidation and reduction of the electrode platinizes its surface. Reduction of Iodate. I n Figure 2 are shown chronopotentiograms for t h e reduction of 0.0028’ KIOl in phosphate buffer solutions a t p H 3. Curves 1 through 4 show, respectively, t h e reduction of iodate a t a n oxidized electrode, at a freshly reduced electrode prepared b y reduction of the oxide film in t h e iodate solution, a t a freshly reduced electrode prepared by reduction of the oxide film in a n iodate-free p H 3 buffer solution, a n d a t a platinized electrode. 1 1 1 of these chronopotentiograms n-ould s l i o ~a second wive corresponding to complete concentration polarization of the iodate a t much longer times but since these second waves are not afferted by electrode pretreatments they were not important for this study. The previously reported ( 2 ) effect of oxidation of the electrode on the reversibility of the reduction of iodate is apparent in the difference between curves 1 and 2 . However, in curves 3 and 4 iodate reduction waves are obtained with unoxidized electrodes a t potentials just as oxidizing as those in curve 1. This proves that the presence of a platinum oxide film on the electrode is not necessary for iodate to commence to be reduced a t these less reducing potentials. It might be argued t h a t platinum oxide is formed on the platinized electrode by chemical reaction of the electrode with iodate. However no VOL. 34, NO. 3, MARCH 1962
363
evidence for such chemical oxidation was obtained in experiments in which the electrode was exposed to a p H 3 iodate solution, removed and washed free of iodate, and a cathodic chronopotentiogram recorded in iodate-free 1F H2S04. No potential pause was observed a t the platinum oxide reduction potential nor a t any other potential before hydrogen evolution commenced. To understand the chronopotentiograms in Figure 2 it is necessary to inquire as to the’ source of the potential inflection in curves 1 and 4. As stated above and shown previously ( 2 ) complete concentration polarization of the iodate has not occured a t the time of the potential inflection in curves 1 and 4. iiccording to the previously invoked oxide bridge mechanism ( 2 ) this potential inflection was assumed to occur when all of the platinum oxide had been reduced and no more electron bridges were available. However, curve 4 displays the same potential inflection at a platinized electrode even though no oxide mas initially on the electrode. The following experiment solved this riddle. A freshly platinized electrode which would have given rise to a curve such as 4 in Figure 2 if used immediately t o record a chronopotentiogram in the iodate solution was instead dipped for about 60 seconds in a 2mM solution of potassium iodide in p H 3 phosphate buffer. This solution corresponds t o the environment experienced by the electrode near the end of a n iodate chronopotentiogram (at all electrodes the reduction of iodate at p H 3 occurs a t potentials more reducing than corresponds to the reduction of iodine to iodide so that iodide is the initial reduction product). When the electrode was Tvashed and used to record a cathodic iodate chronopotentiogram a curve such as curve 2 in Figure 2 resulted instead of curve 4. Thus exposure of the electrode to iodide ion results in the loss of the increased reversibility of the iodate reduction. Iodide is known to be adsorbed on platinum electrodes (6) and it has been previously observed that adsorbed iodide renders the Fe(I1)-Fe(II1) and ferro-ferricyanide couples much less reversible at platinum electrodes (S). It is likely that the adsorption of iodide on the electrode is responsible for the potential inflections in curve 4. When the concentration of iodide a t the electrode surface has increased to the point where significant adsorption on the platinized platinum takes place the platinized surface loses its catalytic properties toward further iodate reduction and the potential inflects to values where iodate is reduced a t unplatinized electrodes. The behavior of the other curves in Figure 2 can now be understood. Iodate is reduced more rerersibly a t a previously oxidized electrode (curve 1) because upon reduction of the oxide a 364
ANALYTICAL CHEMISTRY
I O D A T E C O N C . , M x IO’
Figure 3. Effect of iodate concentration on transition time for first cathodic iodate wave a t an oxidized electrode Prior to recording each cathodic chronopotentiograms electrode was oxidized 10 sec. at 0.9 ma. Current density was 2.4 ma. per sq. cm.
fresh deposit of finely divided platinum is formed a t which iodate reduction proceeds most reversibly. This also explains why the potential a t which iodate reduction commences a t oxidized electrode is identical to the potential a t which the platinum oxide reduction commences and displays the same p H dependence (2). KO iodate reduction can commence before some platinum oxide is reduced to provide a platinized electrode surface. The potential inflection in curve 1 can arise in t n o ways: Iodide can be adsorbed on the freshly formed platinum and interfere with further iodate reduction as in curve 4. I n addition, iodide ion in p H 3 buffer solution is capable of reacting chemically with the unreduced platinum oxides to give soluble iodoplatinum complexes. Experiments shon ed that 90% of the platinum oxide pvas dissolved in 100 seconds in a 2mM iodide solution a t p H 3. I n either case the electrode surface becomes deplatinized and loses its catalytic activity. Curve 2 in Figure 2 results if a second chronopotentiogram is recorded after curve 1 with no pretreatment of the electrode. KO iodate wave is observed because, even though the electrode surface has been freshly oxidized and reduced, the resulting platinized platinum is covered 11-ith adsorbed iodine (resulting from Oxidation of adsorbed iodide by the dissolved iodate). Curve 3 s h o w that a small iodatereduction wave is obtained if the electrode is freshly oxidized and reduced in an iodate-free solution to ensure that no iodide adsorption occurs. Honcver, the large difference in transition times for the waves in curves 1 and 3 is not expected if the oxide film presmt in curve 1 is only effective because i t can be reduced to give a platinized surface. The transition time for curve 1 of Figure 2 is expected to be longer because of the extra time required to reduce the okide film. However, in the
absence of iodate the time required to reduce the oxide at the current density corresponding to curve 1 in Figure 2 is only about one sixth of the observed transition time. The reduction of the oxide will take more time in the presence of iodate because of the lower current efficiency for oxide reduction, but this effect would not be expected to be large enough to account for the difference in transition times in curves 1 and 3. It is also possible that the transition time is larger in curve 1 than in curve 3 because the iodide concentration requires a longer time to increase to the point where significant adsorption occurs due to the consumption of iodide in the chemical reaction with the platinum oxide. A totally unequivocal explanation for this difference in transition times cannot be provided on the basis of the data available. The explanation offered to account for the potential inflection in curve 1 leads to the prediction that, as the iodate concentration is increased, a point should be reached beyond nhich t h e transition time for the m v e should be independent of the iodate concentration. This is to be expected because once the concentration of iodide generated at the electrode surface reaches a value where the electrode is totally deactivated by adsorption or dissolution further increases in the generated iodide concentration should have no effect. Figure 3 shows the results of a set of experiments that confirmed this prediction. Above about O.016N IO3- the transition time does in fact become essentially constant. Conclusions. The present study and a recent similar study (1) have shown that oxide films do not behave as electron bridges in the reduction of iodate, vanadium(V), and iron(II1) nor in the oxidation of Fe(I1). This evidence appears to negate the previous assertions in the literature ( 2 , 6, 8) t h a t platinum oxide films facilitate electrode reactions by acting as electron bridges. Lingane (11, 12) and Sawyer and Interrante (13) maintain that platinum oxide facilitates the electroreduction of oxygen catalytically according t o schemes such as
+ Pt + PtO + 2e1/*02
2H+
-t
-P
Pt
PtO
+ HzO, etc.
(1) (2)
The data of the present and a related earlier study (1) are not sufficient to rule out this mechanism but neither do they support it. The crucial experiment that is required to prove that Reactions 1 and 2 occur is to establish conclusively that oxygen does chemically o-iidize platinum electrodes. Kolthoff and Tanaka (9) performed an experiment designed to detect chemical oxidation of platinum electrodes in air-saturated solutions and concluded that no oxidation of
the electrode occurs. Sawyer and Interrante (IS) claim that oxygen does oxidize platinum electrodes because the electrodes assume more oxidizing potentials in the presence of oxygen than in oxygen-free solutions. Such reasoning, however, merely begs the question. An unambiguous answer to the question as to whether oxygen chemically oxidizes platinum electrodes has yet to be provided. I n any case, however, the demonstration by Laitinen and Enke (IO) that Reaction 2 is inherently extremely irreversible means that mechanisms such as Reactions 1 and 2 could only result in increases in over-all reversibility of electrode reactions in cases where the oxidant is even less reversibly reduced than is platinum oxide. For these reasons it seems likely t h a t the platinization of the electrode result-
ing from reduction of the oxide film on oxidized platinum electrodes is the major factor contributing to increased reversibility of subsequent electrode reactions.
trode Processes,” Chap. 1, Wiley, New York, 1961. (7) Hammett, L. R.. J. Am. Chem. SOC. 46, 7 (1924). (8) Kolthoff, I. M., Nightengale, E. R., Anal. Chim. Acta 17. 329 11957).
ACKNOWLEDGMENT
The alternating current platinization experiment was suggested by C. D. Russell. LITERATURE ‘CITED
( I ) Anson, Fred C., ANAL.CHEM.33, 934 (1961). (2) Anson, Fred C., J . Am. Chem. SOC. 81, 1554 (1959). (3) Anson, Fred C., unpublished experiments. (4) Butler, J. A. V., “Electrical Phenomena at Interfaces,” pp. 206-10, Macmillan, New York, 1951. (5) Davis, D. G., Talanta 3,335 (1960). ( 6 ) Frumkin, A., “Trans. Symp. on Elec-
trochem. SOC.107, 77, (11) Lingane, J. J., J. Electroanal. Chem. 2, 296 (1961). (12) Lingane, J . J., private communica-
tion. (13) Sawyer, D. T., Interrante, L. V., J . Electroanal. Chem. 2 , 310 (1961).
RECEIVED for revien- September 18, 1961. Accepted January 8, 1962. Contribution S o . 2758 from the Gates and Crellin Laboratories of Chemistry California Institute of Technology, Pasadena, Calif. Work supported in part by the U. S. Army Research Office under Grant No. DAORD-31-124-61-G91 and by the E. I. du Pont de Xemours &- Co.
Direct Potentiometric Estimation of the Apparent Activity of the Sodium Ion Use of a Salt Bridge EDWARD S. HYMAN Deparfmenf o f Biochemical Research, louro Infirmary, New Orleans, l a .
b The apparent activity of the sodium ion is estimated potentiometrically using a sodium-sensitive glass and a reference with liquid junction. The Na+ activity of NaCl in water relates to e.m.f. as predicted b y the Nernst equation using concentration and the mean activity coefficient of NaCI. A mixed salt bridge allowed equal representation of a N o + and a CI- in NaCl solutions upon adding NaCl or replacing water with dextrose or ethyl alcohol. The mixed salt bridge lessened the apparent K+ sensitivity of the N a glass, and resulted in representation of a N a + as the N a + to water ratio in dextrose solutions.
T
HE A‘CTIVITY of the sodium ion is probably as important a n aspect of the chemical and biological behavior of sodium as hydrogen ion activity is for hydrogen. The relation of the activity of the sodium ion to total sodium concentration involves the dielectric constant of the solvent, inert solutes making up part of the volume, interionic attraction, nondissociation, and ion association of the dissociated species, Many of these factors cannot be independently assessed. It is important to
estimate sodium activity as a n entity and to study the factors relating the activity to the total concentration of sodium. I n 1922 Neuhausen and Marshall (6) used a sodium amalgam electrode to investigate the state of sodium in blood. Their data showed the estimate of sodium ion to be slightly higher than the estimate of total sodium by gravimetric analysis. It was concluded that all of the sodium was dissociated. Ringer (7‘) came to a similar conclusion with sodium in the presence of egg white, but showed that the sodium amalgam electrode was not specific for sodium. Since that time, particularly in biological sciences, the activity of the sodium ion has been thought of as being almost synonymous with its concentration, with occasional references to coulombic forces that could not be measured. There are reasons for anticipating that the relative importance of factors affecting the behavior of the activity of the sodium ion may be different from those which are familiar in determining the behavior of the hydrogen ion. First, the solvent is usually water’which dissociates to give hydrogen ions.
Upon removal of the hydrogen ion without regard for the activity of that ion, the solvent is almost 56M for hydrogen ion. Second, the range of concentration is different. Hydrogen ion contributes significantly to the ionic strength only below pH 3. This is but a small portion of the estimable range in solutions. I n contrast, the sodium ion contributes to the ionic strength throughout the range in which its activity can be estimated by the glass electrode. The reference electrode must be either a standard electrode with a liquid: liquid interface or another electrode in the same solution as the measuring electrode. The latter is thermodynamically more acceptable because i t avoids the equations for ion transference. Unfortunately, a reference electrode must also relate some parameter of the solution. The hydrogen ion has been studied without transference in aqueous HC1 using a Aga:AgC1:C1reference. The assumption must be made that the activity coefficient of chloride ion equals that of hydrogen ion in all solutions measured. This assumption is inaccurate in the presence of chloride binding by albumin, for VOL. 34, NO. 3, MARCH 1962
0
365
