416 Analyst, June, 1973, Vol. 98, #p. 416-425 Precise Coulometric Determination of Acids in Cells Without Liquid Junction Part 111.” Determination of the Silver Error by Amperostatic Anodic Strippingt BY E. BISHOP AND M. RILEY1 (Chemistry Department, University of Exeler, Stocker Road, Exeter, EX4 4QD) The plating and stripping of silver on platinum electrodes have been examined in the context of the determination of the silver error, which arises from the solubility of silver bromide when the deposition of bromide on a silver anode is used as the auxiliary reaction in the coulometric assay of acids. In order to calibrate the stripping method, it is necessary to plate known amounts of silver quantitatively on to platinum-gauze electrodes. Low recoveries are obtained when the platinum is not fully reduced.The oxidation and reduction of platinum electrode surfaces have been briefly examined, and it is demonstrated that oxide forms on an electrode when its potential is allowed to rise beyond 0.8 V, the termination potential in silver stripping. For calibration purposes, plating and stripping in a 0.1 M solution of silver nitrate in 0.1 M perchloric acid was first investigated. Amperostatic and potentiostatic reduction of the platinum electrode are shown to be ineffective, but chemical reduction leads to excellent plating and recoveries, provided great care is taken completely to remove all traces of reductant. Calibration being satisfactory, stripping in 0.1 M perchloric acid, as in an actual acidimetric assay, has been examined and shown to give excellent recoveries.The anodic stripping curves show an extended second wave, which is identified as arising from reduction of oxygen to hydrogen peroxide at the auxiliary stripping electrode, particularly when the latter becomes plated with silver. The hydrogen peroxide is oxidised at the stripping electrode, and the process is cyclic. IN the coulometric determination of acids, the use of a silver auxiliary anode on which bromide is deposited has been canvassed.1 The slight, but significant, solubility of silver bromide in the electrolyte gives rise to a “silver error” by deposition of silver on the working platinum cathode either by direct electro-reduction or by reduction of silver ion by hydrogen atoms in the compact layer (the combination of which, to give hydrogen molecules, is the rate-determining step of the main cathodic reaction). Additionally, any precipitate of silver bromide in the bulk of the solution is liable to be caught on the cathode and there reduced, although the experimental conditions are chosen so as to avoid the formation of such precipitates. The error incurred is about 1 to 2 C in a total of 5000 C, and a rapid and convenient method is needed for the determination of about 1 mg of silver deposited on a 125-cm2 platinum-gauze electrode.If this amount can be determined with an accuracy of 1 per cent., it will then represent an over-all accuracy of about 2 to 4 p.p.m. in the deter- mination of 0.05 mol of a monobasic acid. Amperostatic anodic stripping proved simple and rapid,l but is of unknown accuracy.In order to assess the accuracy a method must be discovered for quantitatively plating known amounts of silver on to platinum for calibration of the stripping process. Preliminary experiments showed that while the potential rise at completion of the stripping reaction was satisfactorily sharp, the accuracy was poor, recoveries of silver being only 70 per cent. This finding led to the investigation of plating of silver on to platinum, of oxidation of platinum surfaces, and therefore of the pre-conditioning of the platinum electrode. Mechanical stripping, i.e., loss of some of the loose deposit, was suggested by Lord, O’Neill and Rogers2 as the cause of low recoveries. Nisbet and Bard3 obtained 100 per cent. recovery on a platinised platinum electrode, but low recoveries on an oxidised * For particulars of Parts I and I1 of this series, see reference list, p.425. t Presented a t the Second SAC Conference, Nottingham, 1968. 0 SAC and the authors. For Part IV, see p. 426. Present address : Electronic Instruments Limited, Hanworth Lane, Chertsey, Surrey.BISHOP AND RILEY 41 7 electrode. A platinised surface was therefore prepared by a.c. electrolysis in 1 M perchloric acid, and the electrode was finally reduced almost until hydrogen was evolved. Silver was plated on to and then stripped from this electrode amperostatically, and the plating - stripping cycle was repeated several times with anodisation to + l o 6 V each time; progressively shorter stripping times were found for the same plating time.Nisbet and Bard3 reported similar results, which were claimed by them to show that some of the silver is retained on an oxidised platinum surface and can be removed only after reduction of the underlying oxide. Bixler and Bruckenstein4 claimed 100 per cent. recovery when a reduced platinum electrode was used. They confirmed that repetition of the plating - stripping cycle with anodisation to +1-6 V gave short recovery times, but could find no evidence of retention of silver on the electrode. They suggested that the apparent loss of silver arose from partial reduction of the oxidised platinum surface during plating, together with mechanical loss. EXPERIMENTAL The apparatus and reagents used have been described previously.G The cell and circuit used in the present study are shown in Fig.1. The milliammeter shown in series with the working electrodes was used to set the current to the nearest 10pA, but in the later work was replaced with a 10-l2 standard four-terminal resistor and the current set to the nearest 1 pA by using the P3 potentiometer. Titanium(1V) sulphate was prepared by heating 10 g of titanium dioxide with 20 ml of concentrated sulphuric acid for 30 minutes, cooling, diluting the solution with 50 ml of cold water, allowing it to stand for several days and filtering it through a Whatman No. 42 filter-paper. All electrode potentials are given with reference to the standard hydrogen electrode (S.H.E.). Clock r"l I OACS 1- P-3 T I ( t Reversing switch Plat i n urn-wire Fig. 1. Coulometric cell and circuit for the study of anodic stripping of silver from platinum gauze (OACS = operational ampli- fier constant-current source6) RESULTS AND COMMENTS Preliminary plating experiments on untreated platinum-gauze electrodes gave well defined stripping waves, but recoveries were between 50 and 85 per cent.A second, poorly defined wave at +1.0 V is ascribed to oxidation of the platinum. The gauze electrode was reduced at 1 mA for 100 s in nitrogen-purged 0.1 M perchloric acid, and was allowed to stand in this solution for about 10 minutes until its potential became stabilised, before being418 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [ArtdySt, VOl. 98 transferred to the plating - stripping medium consisting of a 0.1 M solution of silver nitrate in 0.1 M perchloric acid.Two cycles of plating and stripping are shown in Fig. 2. The amount of silver plated in each cycle was constant at 400 mC and the respective recoveries were 87 and 73 per cent. The beginning of the second stripping wave can be seen in each instance and, furthermore, in the second plating step the potential takes some time to reach the steady value indicative of silver deposition, which is taken to indicate at least partial reduction of the oxidised surface produced at the tail of the first stripping step. Decrease in stripping time on repetitive cycling has been observed before.8~~ Clearly, oxide formation is involved and requires investigation. Potential - Potential versus S.H.E. - +- 0-6 V i Start - +1*ov \ - + 1.4 v Fig. 2. Consecutive silver plating and strip- Fig.3. Three cycles of oxidation and reduction ping cycles in a 0.1 M solution of silver nitrate in of a 126-cma platinum-gauze electrode at 2 mA in 0.1 M perchloric acid. The short vertical lines oxygen-free 0.1 M perchloric acid indicate points where the current was switched on or off or reversed Oxidation and reduction of platinum after the amperostatic cathodic pre-treatment was examined in nitrogen-purged 0.1 M perchloric acid at total currents of 0.5 to 2 mA. Fig. 3 shows the behaviour of an electrode, previously reduced at 2 mA for 200 s, during three cycles a t 2 mA between 1.4 and 0.5 V. Electrode oxidation and reduction start at +1.0 and + 0.9 V, respectively, although the former is ill defined. The time taken for the potential to rise from 0.5 to 1.4 V decreases with continued cycling although the cathodic step remains essentially constant a t 45 s.Similar behaviour was observed by Feldberg, Enke and BrickerJs who found that the ratio of anodisation to cathodisation times started at 2 : 1 and fell eventu- ally to 1 : 1. In the present work, the initial ratio was about 2: 1 but, after six or seven cycles, became constant at 1.2: 1. Anodisation to potentials greater or less than 1.4 V produced longer or shorter cathodisation times, respectively, and the quantity of electricity required for cathodisation was virtually independent of the current ; for example, halving the current produced an increase in cathodisation time by a factor of about 2.2. A mean value was found for the reduction over a variety of conditions of 0.74 mC cm-2, assuming an electrode area of 125 cm2. Kolthoff and Tanaka' reported a value of 0.92 mC cm-2 for anodic or chemical oxidation.Several chronopotentiometric determinations of oxide have been reported and are summarised in Table I. TABLE I CHRONOPOTENTIOMETRIC OXIDATION AND REDUCTION OF PLATINUM IN OXYGEN-FREE 1.0 M SULPHURIC ACID Transition Electrode Current Quantity of Reference time measured area/cm8 density/mA cm-a electricity/mC cm-8 8 Anodic 2 70-300 0.8 9 Cathodic 6 100 0.98 10 Anodic 0.25 60 0.94 11 Cathodic 0.33 180 0.6 The current densities were much higher than the 8 pA cm-2 used here, but the coverage E. Bishop and B. Cooksey (unpublished work) have values are all of similar magnitude.June, 19731 OF ACIDS I N CELLS WITHOUT LIQUID JUNCTION.PART I11 419 shown that at potentials above 1.0 V the formation of molecular oxygen is in competition with oxide film formation, and attempted to develop growth laws. The oxidation and reduction of platinum are dealt with in detail elsewhere12Js; for the present purpose, it is clear that pre-conditioning of the electrode is necessary for quantitative plating of silver, and that oxidation of the electrode surface occurs when stripping is taken beyond 0.8 V. PLATING AND STRIPPING OF SILVER IN SILVER NITRATE SOLUTIONS- For the efficient quantitative plating of silver, a completely reduced platinum surface is clearly necessary. Three basic types of conditioning, amperostatic, potentiostatic and chemical, have been examined for this purpose.Plating and stripping were carried out in 0.1 M solutions of silver nitrate in 0.1 M perchloric acid. Between experiments, the platinum auxiliary electrode was always cleaned with 1 + 1 nitric acid so as to remove any silver deposited on it during the stripping step, which normally followed the plating step by reversal of the current. A mperostatic cathodic reduction-The pre-treatment was performed in de-oxygenated 0.1 M perchloric acid at 2 to 500 mA. Before reduction, the platinum-gauze electrode was always in an oxidised condition that resulted from a previous anodic stripping run, from prior anodic oxidation to a potential higher than 1.4 V in de-oxygenated 0.1 M perchloric acid, or from immersion, together with the auxiliary electrode, in 1 + 1 nitric acid during the cleaning of the latter.The reduction was followed by treatment to remove any hydrogen remaining on the electrode surface. This treatment was carried out in the same solution and comprised either (a) allowing the electrode to stand for 2 to 20 minutes in the stirred pre-treatment solution, or (b) short-circuiting it to a S.C.E. , connected via an intervening potassium sulphate salt bridge, for 5 to 50 minutes. The electrodes were then rapidly transferred to the silver solution. When the cathodic pre-treatment was of short duration, so that the electrode potential did not become low enough for the evolution of hydrogen, the stripping curves were of the same shape as in Fig. 2, but the over-all recovery was only 85 per cent. When the pre-treatment was more drastic and involved the evolution of hydrogen, and the electrodes were given treatment (a) above, then the stripping curves often showed an additional step just prior to the potential rise, which indicated completion of stripping.Measurement of the stripping time up to a potential of +Om8 V gave recoveries greater than 100 per cent. Fig. 4 shows two such curves at a stripping current of 2 mA for an electrode pre- treated by oxidation and reduction at 2 mA. Curve A was obtained with an electrode re- duced to -0.048 V and allowed to stand for 3 minutes, and curve B with an electrode reduced to -0.072 V and allowed to stand for 2 minutes; the recoveries were 111 and 119 per cent., respectively. Compared with the normal stripping step as shown in Fig. 2, the distortion caused by oxidation of residual hydrogen is apparent, and the hydrogen is evidently underneath the silver plate.Reproducibility was poor, but, in general, the longer the time the electrode was left after pre-treatment the lower was the recovery of silver and the more Potential versus S.H.E. -+ -+ -+ Fig. 4. Distortion of anodic stripping curves due t o the presence of residual hydrogen on the electrode after cathodic pre-treatment and under the silver plate420 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [Analyst, Vol. 98 closely the shape of the curve approached that in Fig. 2. Recoveries of 95 to 105 per cent. were obtained with electrodes left for 20 minutes and the stripping curve still showed a slower potential rise than was normal at the completion of stripping.Short-circuiting the electrode to the S.C.E. appeared effectively to remove the hydrogen and gave normal stripping curves, but the recovery averaged only 90 per cent. A similar recovery was obtained after using Bixler and Bruckenstein’s method of pre-treatment A.* De-oxygenation of the solution for plating and stripping had no detectable influence on the shape of the plating curve or on the recovery of silver. Potentiostatic cathodic reduction-Potentiostatic reductions were performed at command potentials of +0.6, +0.5 and +0-25 V in nitrogen-purged 0.1 M perchloric acid until the current became essentially constant. The platinum gauze was always in an oxidised form before treatment, as in the amperostatic method. The electrodes were then rapidly trans- ferred into the silver nitrate solution.Currents in the reduction started at 1 to 4 mA and decreased rapidly, to one tenth of the initial value in 10 minutes, becoming constant at the two higher command potentials after 30 minutes and decreasing only slowly at +0-25 V; this condition produced the largest residual current, of the order of 0.1 mA. Reduction times from 35 minutes to 3 hours were used, but no significant decrease in current occurred after the first 60 minutes. Silver recoveries as high as 95 per cent. were attained, but the mean re- covery in fourteen runs was only 90 per cent., and there was no appreciable difference between the effects of the three pre-treatment potentials on recoveries. De-oxygenation ofthe plating - stripping solution was without influence.Chemical reduction-Reduction of oxidised platinum by immersing it in a solution of a reductant has been carried out several times. Kolthoff and Tanaka7 used a 5-minute immersion in a 0.01 M solution of iron(I1) sulphate in 0.05 M sulphuric acid, but found that a 0.01 M solution of arsenic(II1) in 1 M hydrochloric acid had no effect in 60 minutes. Ross and Shain14 used a 10-minute immersion in 0.1 M iron(I1) perchlorate solution, Anson and Linganes found an acidic iodide solution to be effective and Anson15 used a 1-minute immersion in a 0.2 M solution of iron(I1) sulphate in 1.0 M sulphuric acid. Trials with immersion of 2 to 10 minutes in 0-2 to 0.5 M solutions of iron(I1) sulphate in 1.0 M sulphuric acid gave recoveries of silver between 90 and 99 per cent.Recoveries improved with increasing im- mersion times, and it became clear that a period of over 15 minutes was required for the platinum-gauze electrode to be reliably reduced a€ ter anodic oxidation or immersion in 1 + 1 nitric acid, and even longer times after cleaning in aqua regia. I t also became clear that the potential displayed by the electrode when immersed in the plating solution gave a good indication of the efficiency of pre-treatments. A fully reduced electrode took up a potential of about 0.75 V, while incompletely reduced electrodes showed potentials between 0.85 and 0.9 V. This was confirmed by immersing two closely similar platinum-wire elec- trodes, one anodically oxidised and the other cathodically reduced, in the plating solutions.The potentials became stabilised at 0.90 V for the oxidised and 0.74 V for the reduced elec- trode, and when a clean silver-wire electrode and a platinum-wire electrode previously reduced for 30 minutes in the iron(I1) solution were placed in the same plating solution, both im- mediately displayed the same stable potential of 0.72 V. This potential was also taken up by platinum-wire electrodes that had been treated with aqua regia, 1 + 1 nitric acid or chromic acid and by immersion, after washing, in the reduction solution for only 1 minute. The platinum-gauze electrode required prolonged immersion because the woven structure presented many hundreds of points of contact in the mesh, which created crevices that were relatively slowly accessible to the solution.Contamination of the silver nitrate solution could arise if the gauze electrode was not completely washed before transfer from the reduction bath. This effect was examined first by adding 0.1 ml of reduction solution that had been diluted 1 + 9 with water to the plating solution just before starting a plating - stripping cycle. This treatment gave a recovery of about 115 per cent. and the stripping curve was much more rounded in the end-point region, while the time required for the potential to rise from +0.8 to +1.4 V was about 30 per cent. longer than usual. Secondly, a reduced electrode that had not been washed was plated at 2 mA for 100 s and immediately stripped at 2 mA. The initial stripping potential was 0.72 V and, after 1900 s, had decreased by only 5 mV.At this stage the current was increased to 5 mA, and after a further 180 s a fairly rapid rise in potential to 0.77 V occurred, followed by a gradual rise over 800 s to 0.8 V. When prolonged washing of the gauze electrodeJune, 19731 OF ACIDS I N CELLS WITHOUT LIQUID JUNCTION. PART 111 421 was performed, by allowing it to stand, after initial thorough rinsing, in three successive batches of 200 ml of vigorously stirred water with thorough washing and draining between each washing, 100 per cent. recoveries of silver were obtained. In addition to the initial potential of the electrode in the plating solution, the shape of the plating curve is also diagnostic of the extent of reduction of the electrodes. This is evident on comparison of Fig.2, which shows incomplete reduction, with Fig. 5 (curve A), which shows a complete plating - stripping cycle at 2 mA in a 0.1 M solution of silver nitrate in 0.1 M perchloric acid with a chemically reduced electrode. The short vertical lines in all of the chronopotentiograms indicate points at which the current was switched on or off. The plating curve in Fig. 5 (curve A) is smooth, while those in Fig. 2 show small negative-going peaks at the beginning of plating. Nisbet and Bard3 observed a similar phenomenon during plating on to incompletely reduced electrodes, Once again de-oxygenation of the plating solution had no influence on recoveries of silver. Fig. 5. Plating and stripping of silver (curve A) and stripping of a silver-free electrode (curve B) by using fully reduced platinum-gauze electrodes PLATING - STRIPPING EFFICIENCIES IN A 0.1 M SOLUTION OF SILVER NITRATE IN 0.1 M As a result of the foregoing investigations, a standard pretreatment procedure for the gauze electrodes was adopted.A minimum immersion time of 20 minutes for oxidised electrodes, or 3 hours, but preferably overnight, for electrodes treated with aqua regia, was used in a stirred 0-2 M solution of iron(I1) sulphate in 1.0 M sulphuric acid. After being thoroughly washed, the electrodes were immersed in clean water for at least 5 minutes. Pre-treated platinum gauze was dried in warm air, weighed, re-treated and plated with about 3 mg of silver. After thoroughly washing it, the electrode was again dried in warm air and re-weighed. The increase in mass was well within 1 per cent.of that calculated from the quantity of electricity passed, thus confirming the quantitativeness of the plating process. Plating and stripping were performed at currents of 2 or 5 mA, the crystal clock was automatically triggered on starting and stopping the plating and on starting the stripping, and was switched off manually when the stripping potential reached 0-8 V. Stripping was started 15 to 30 s after plating so that accurate times could be noted. A series of twenty-two cycles gave the results shown in Table 11. The precision improves as the amount of silver increases. The 95 per cent. confidence limits are &l-6 per cent. for 0.2 mg of silver and 1.0 per cent. for larger amounts. The uncorrected recoveries calculated from the stripping time to 0.8 V exceed 100 per cent.for amounts of 0.4 mg of silver or less, and increase as the amount of silver and the stripping current decrease. The stripping time for a clean, unplated electrode (curve B, Fig. 5) was 3.5 5 0-5 s at 2 mA. If the recoveries are corrected for this effect, they become 99.4 per cent. for 0.2 mg of silver and 99.5 per cent. for 0-4 mg of silver. This “blank” stripping time could be due to a small amount of oxidation of the electrode at 0.75 to 0.8 V, or to the chemical deposition of a partial monolayer of silver on the electrode. Even assuming the extreme value of 100 pF cm-2 for the double-layer capacitance, PERCHLORIC ACID-422 BISHOP AND RILEY : PRECISE COULOMETRIC DETERMINATION [Analyst, VOl. 98 only 0.3 s at 2 mA would be required for charging over the potential range involved.Allen and Hickling16 claimed that platinum was coated with silver on immersing it in 1 M silver nitrate solutions after cathodic reduction. They detected silver colorimetrically after dis- solution from a 100-cm2 platinum foil, and also observed a step in the stripping curve of a pre-reduced 1-cm2 platinum electrode between 0.7 and 0.8 V, corresponding to about 3.5 mC of electricity. On the assumption of a roughness factor of 2, they calculated that a layer of silver seven atoms thick was formed on immersion of platinum in silver nitrate solution, which in the present instance would represent about 0.4 mg of silver, whereas the “blank” stripping time corresponds to 7 pg of silver. It is likely that the deposit, and the stripping step observed by Allen and Hickling, arose from residual hydrogen from the cathodic reduction; no steps taken to remove this deposit were described.TABLE I1 AMPEROSTATIC ANODIC STRIPPING OF SILVER IN A 0.1 M SOLUTION OF SILVER NITRATE I N 0.1 M PERCHLORIC ACID Stripping Approximate mass of Number Mean recovery, Relative standard current/mA silver plated/mg of tests per cent. deviation, per cent. 2 0.2 10 102.9 0.7 2 0.4 6 101.2 0.4 6 0.5 to 1.1 6 100.1 0.4 ANODIC STRIPPING OF SILVER IN PERCHLORIC ACID SOLUTION- Having obtained a satisfactory recovery in silver nitrate solutions, it was then necessary to check the recoveries under the conditions appertaining to “silver error” determinations at the end of an acid assay. After plating in the silver nitrate - perchloric acid solution, the platinum-gauze electrode was removed from the plating solution, drained, carefully and thoroughly washed with water, and allowed to stand in about 100ml of water for several minutes before being transferred to a second coulometric cell, as in Fig.1, containing 0.1 M perchloric acid. It was then stripped at 2 or 5mA, the clock being automatically started and manually stopped at a stripping potential of 0.8 V. The elapsed time between completion of plating and starting of stripping was usually about 10 minutes. Craig, Law and Hamer17 have demonstrated that the rate of dissolution of finely divided silver in perchloric acid solutions is so slow that it can be neglected with confidence. The results of thirteen experiments involving 0-2 to 1.2 mg of silver are summarised in Table 111, and again show an improvement in precision as the amount of silver and the stripping current increase.TABLE IT1 AMPEROSTATIC STRIPPING OF SILVER INTO 0.1 M PERCHLORIC ACID Stripping Approximate mass of Number Mean recovery, Relative standard current/mA silver plated/mg of tests per cent. deviation, per cent. 2 0.2 to 1.2 9 100.9 1.0 5 0.8 to 1.1 4 100.3 0.7 The 95 per cent. confidence limits were 5 2 . 3 per cent. at 2 mA and k 2 . 2 per cent. at 5 mA. Recoveries are again slightly high, but the bias decreases as the amount of silver increases, and is insignificant for amounts in the region of 1 mg. When silver-free, reduced electrodes were stripped, it was found that their initial potential in 0.1 M perchloric acid was about +0.9 V, and so there is no blank stripping time under these conditions.Typical stripping curves for about 0.4 mg of silver and for a reduced electrode at 2 mA are shown in Fig. 6. The potentials displayed by a clean silver-wire electrode and a reduced platinum-wire elec- trode in 0.1 M perchloric acid were approximately 0.43 and 0.98 V, respectively. THE SECOND ANODIC WAVE- It was observed during silver error determinations1 that the second wave in the stripping curve associated with oxidation of platinum was abnormally long. Fig. 7 shows a typical curve for a silver error determination at 2 mA; the rise time from 0.8 to 1.4 V corresponds to about 500mC in this instance, and values ranging from 200 to 700mC are commonly obtained.The silver deposit in these experiments was produced under different conditions and in a different medium from the calibration plating described above. Values of aboutJune, 19731 OF ACIDS I N CELLS WITHOUT LIQUID JUNCTION. PART I11 423 110 mC were found in the preliminary work described in this paper, but pertained to partly reduced electrodes; values for the second wave on stripping in silver nitrate solutions at 2mA were 165 to 170mC for both silver-plated and silver-free electrodes. For stripping into perchloric acid, as described in the preceding section, similar extended second waves appeared, Although variable, the results for the corresponding quantities of electricity were 180 to 380mC, and longer stripping times were associated with longer second waves.Potential versus S.H.E. - + 0-5 V -+ 0-8 -+ 1.1 --+ 1.4 Potential versus S. H . E . + 0.6 V- + 1.ov- + 1.4 V- Fig. 6. Anodic stripping a t 2 mA in 0.1 M perchloric acid of A, silver- plated platinum; and B, silver-free platinum Fig. 7. Extended second wave obtained during a “silver error” determination a t 2 mA These second waves are pertinent to the ensuing pre-treatment of electrodes, and they warranted exploration. It is postulated that they are due to the anodic oxidation of hydrogen peroxide introduced into the solution by the reduction of dissolved oxygen at the auxiliary electrode during the stripping process. Laitinen and Kolthoff l8 claimed that hydrogen peroxide was a product of reduction of dissolved oxygen at a platinum electrode, 0, + 2H+ + 2e + H,O, but later studie~l~-2~ have shown that this is so only under certain conditions of the state of the electrode surface.Peters and Mitchel121 demonstrated spectrophotometrically that significant amounts of hydrogen peroxide appeared in the bulk of the electrolyte when reduction was performed at an “aged” electrode, which was produced by allowing a freshly reduced electrode to stand in oxygen-saturated electrolyte for over 24 hours, but could detect no hydrogen peroxide when the reduction was performed with freshly reduced or pre-oxidised electrodes. They supposed that hydrogen peroxide was also produced at a freshly reduced electrode, but immediately disproportionated to water and oxygen at the “active” platinum surface, and that as the electrode “aged” it lost the ability to catalyse the disproportionation.In the present instance, the pre-reduced auxiliary electrode will be oxidised during plating then at least partially reduced during the first part of the stripping step, and then, concurrently with the reduction of dissolved oxygen, some of the silver stripped from the anode will be deposited on it in increasing amount as the stripping time increases. Kolthoff and Laitinen22 found that dissolved oxygen was reduced at silver at more positive potentials (less charge- transfer overpotential) than at platinum. It did not seem that the condition of the auxiliary electrode was suitable for hydrogen peroxide formation in the present work, and so reduction of dissolved oxygen at silver electrodes was investigated.Deliberate addition of trace amounts of hydrogen peroxide, either during the stripping of silver in 0.1 M perchloric acid, or just prior to stripping a silver- free reduced electrode, produced extended second waves. This observation agrees with that hydrogen peroxide is oxidised at an unoxidised platinum electrode, but424 BISHOP AND RILEY: PRECISE COULOMETRIC DETERMINATION [Analyst, VOl. 98 only qualitative correlation could be made between the amount of hydrogen peroxide added and the increase in duration of the second wave. When oxygen-free 0.1 M perchloric acid was used for silver stripping no extension of the second wave occurred, even with long stripping times, and the recovery of silver was unchanged. When the auxiliary electrode was pre- plated with silver and the solution contained oxygen, significantly extended second waves appeared, even with short stripping times, and hydrogen peroxide was detected spectrophoto- metrically in the cell electrolyte.In two examples with silver stripping times of 120 and 500 s, the second waves were equivalent to 340 and 550 mC, respectively. The peroxodisulphatotitanic(1V) acid method was used to determine hydrogen peroxide. The absorption spectrum of the complex showed a broad peak with A, = 410 nm. A linear Beer’s law graph was used as a calibration graph ; the molar absorptivity of the complex at 410 nm was 40 1 mol-l mm-l. Blanks of freshly prepared 0.1 M perchloric acid electrolyte were used. A 10-ml sample of the cell electrolyte, after about 80 per cent.of the silver had been stripped, was treated with 2 ml of the titanium(1V) reagent and the absorbance at 410 nm was measured. The amount of hydrogen peroxide found in the electrolyte was of the order of low5 mol l-l, but could be only qualitatively correlated with the extension of the second wave. The reaction is probably cyclic. DISCUSSION Clearly, the efficiency of plating silver on incompletely reduced platinum electrodes is significantly below 100 per cent., and the loss is due to the concurrent reduction of remanent oxide, The oxidation and reduction of platinum show behaviour in accord with other work, but the interpretation is still contentious and is dealt with elsewhere. Insufficient attention seems to have been paid to the competitive formation of molecular oxygen in the oxidation half-cycle. Failure of the electro-reduction methods to give fully reduced active electrodes for plating is puzzling, but is tentatively ascribed to electro-sorption of impurities of which minute traces exist in sulphuric acid12 or insufficient prior oxidation, or both.Later work suggests that repeated anodic - cathodic cycling is necessary in preparing a fully active electrode. The iron(I1) reduction method, when correctly applied, gives electrodes at which good plating efficiency is attained, but very thorough washing of the electrode is essential, otherwise cyclic oxidation and reduction of iron(I1) and iron(II1) at the two electrodes occurs and no silver is plated on the electrode. On the basis of plating calibrations involving the use of electrodes pre-conditioned by chemical reduction with iron( II), silver stripping recoveries both in silver nitrate solutions and in pure perchloric acid solutions are excellent for 0.2 to 1.2 mg of silver and are of a precision entirely adequate for the determination of silver errors.The blank found in silver nitrate solutions but absent in perchloric acid shows that about one tenth of a monolayer of silver is deposited chemically on platinum immersed in a silver solution, and this finding is supported by the potentials of silver and platinum electrodes in these media, Hydrogen peroxide is produced at the auxiliary electrode by reduction of dissolved oxygen during the stripping of silver into perchloric acid, and gives rise to an extended second wave in the stripping chrono- potentiogram.The deposition of stripped silver on the auxiliary electrode favours this process because of the lower charge-transfer overpotential of reduction of oxygen on this metal. The phenomenon is absent in de-oxygenated media, and in any event does not affect the recovery of silver; it does, however, add emphasis to the importance of adequate pre- treatment of platinum electrodes on to which silver is to be plated. CONCLUSIONS The present study has shown that silver present on a platinum-gauze electrode can be determined by amperostatic anodic stripping with an accuracy and precision that are adequate for the required “silver error” determination, and that such an electrode can be washed and transferred into the stripping solution without significant loss of silver.The amount of silver present on the working electrode at the end of an acid assay that involves the passage of 5000 C can be determined simply and rapidly by this method. Extended second waves (which do not affect the silver determination) on the stripping curves arise from the oxidation of hydrogen peroxide produced at the auxiliary electrode or oxidation of residual hydrogen on the electrode surface following the acid determination, or both.June, 19731 OF ACIDS I N CELLS WITHOUT LIQUID JUNCTION. PART 111 425 One of us (M.R.) is deeply indebted to the Charitable and Educational Trust of the Worshipful Company of Scientific Instrument Makers for financial support in the form of a Research Studentship. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. REFERENCES Bishop, E., and Riley, M., Analyst, 1973, 98, 313. Lord, S. S., O’Neill, R. C., and Rogers, L. B., Analyt. Chem., 1952, 24, 209. Nisbet, A. R., and Bard, A. J., J . Electroanalyt. Chem., 1963, 6, 332. Bixler, J. W., and Bruckenstein, S., Analyt. Chem., 1965, 37, 791. Bishop, E., and Riley, M., Analyst, 1973, 98, 305. Feldberg, S. W., Enke, C. G., and Bricker, C. E., J . Electrochem. Soc., 1963, 110, 526. Kolthoff, I. M., and Tanaka, N., Analyt. Chem., 1954, 26, 632. Anson, F. C., and Lingane, J. J., J . Amer. Chem. Soc., 1967, 79, 1016. ~- , Ibid., 1957, 79, 4901. Lingane, J. J., J . Electroanalyt. Chem., 1960, 1, 379. Peters, D. G., and Lingane, J. J., Ibid., 1962, 4, 193. Hitchcock, P. H., Ph.D. Thesis, University of Exeter, 1969. Bishop, E., “Coulometric Analysis,” Volume IID of Wilson, C. I,., and Wilson, D. W., Editors, Ross, J. W., and Shain, I., Analyt. Chem., 1956, 28, 648. Anson, F. C., Ibid., 1961, 33, 934. Allen, P. L., and Hickling, A., Analytica Chin$. Acta, 1954, 11, 467. Craig, D. N., Law, C. A., and Hamer, W. J., J . Res. Natn. Bur. Stand., 1960, 64A, 127. Laitinen, H. A., and Kolthoff, I. M., J . Phys. Chem., 1941, 45, 1061. Lingane, J. J., J . Electroanalyt. Chem., 1961, 2, 296. Sawyer, D. V., and Interrante, L. V., Ibid., 1961, 2, 310. Peters, D. G., and Mitchell, R. A., Ibid., 1965, 10, 306. Kolthoff, I. M., and Laitinen, H. A., Science, N.Y., 1940, 92, 160. Hickling, A., and Wilson, W. €I., J . Electrochem. SOC., 1951, 98, 425. Giner, J., 2. Elektrochem., 1960, 64, 491. Lingane, J. J., and Lingane, P. J., J . Electroanalyt. Chem., 1963, 5, 411. Liang, C. C., and Juliard, A. L., Ibid., 1965, 9, 390. NOTE-References 1 and 5 are to Parts I1 and I, respectively, of this series. “Comprehensive Analytical Chemistry,” Elsevier, Amsterdam, 1973, pp. 243-263. Received December 18th, 1972 Accepted January 19th, 1973