Some of the wave forms recorded with the G.C.E. differ from those taken with the P.G.E. The wave for the voltammetric reduction of U(V1) a t the G.C.E. is exceptionally well defined as compared with that obtained with the P.G.E. and even with the D.M.E. The current-voltage curve for the reduction of Cu(I1) in 2M HC1 a t the G.C.E. has two distinct peaks in contrast with that for the reduction a t the P.G.E. which gives only one discernible peak. Both waves are easily measured and increase in direct ratio to an increase in Cu(l1) concentration. Presumably, these peaks represent the reductions Cu(l1) Cu(1) and Cu(1) CuO. The cathodic wave of Fe(II1) in 1M HC10, is well formed; in HC1 it is much less ideal, being extremely broad a t the peak. This nonideal behavior is undoubtedly due to the complexing nature of HCl, which changes the mechanism of reaction at the electrode. The Cr(VI)-H2S04system presents an anomaly. At millimolar or lower concentrations of Cr(V1), the current-voltage curves are fairly well defined and measurable; a t higher Cr(V1) concentrations, they are poorly defined and have a high noise level. The reduction
-
-
of Cr(V1) a t the platinum electrode produces a coating on the electrode that interferes with the reduction of Cr(V1) (6). Possibly, such a coating could also form on the G.C.E. In general, for each of the ions studied the E, values a t the G.C.E. and P.G.E. are close. A more direct comparison is not possible, since the ions were not studied under the same conditions. The precision attainable with the G.C.E., as indicated by the data of Table 11, compares very favorably with that attainable with both the P.G.E. and D.M.E. The relative insensitivity of the G.C.E; to changes in p H is in decided contrast with the behavior of the P.G.E. (8), which responds more quickly to a p H change. This difference in behavior is difficult to explain. It may result from the difference in structure, glassy carbon being completely isotropic, whereas pyrolytic graphite has a well-ordered structure. The G.C.E. has several possible advantages for use in voltammetry. It appears to be inert to strong acids and oxidizing agents. For some currentvoltage curves, especially those for Fe(I1) and U02(II), it provides better definition than does the P.G.E.
LITERATURE CITED
(1) Ada%s, R. N., “Progress in Polarog-
raphy, P. Zuman and I. M. Kolthoff, eds., Chap. 23, Interscience, New York, 1962. (2) Delahay, P., “New Instrumental Methods in Electrochemistry,” p. 131, Interscience, New York, 1954. (3) Ibid., p. 51. (4) Ibid., p. 119. (5) Fisher, D. J., Maddox, W. L..Kelley, M. T., Stelzner, R. W., “ORNL Analytical Chemistry Division Annual Progress Report for the Period Ending November 15, 1963,” ORNL-3537, p. 16.
(6) Kolthoff, I. M., Shams el Din, A. M., J . Phys. Chem. 60, 1564 (1956). (7) Laitinen, H. k.,Rhodes, D. R., J . Electrochem. SOC.109,413 (1962). (8) Miller, F. J., ANAL. CHEM.35, 929 (1963). (9) Miller, F. J., Zittel, H. E., Ibid., 35, 1866 (1963). (10) Miller, F. J., Zittel, H. E., J . Electroanal. Chem. 7,116 (1964). (11) Tokai Electrode Mfg. Co. Ltd., 20 Ahasaka. Tameike-Cho. Minato-Ku. Tokyo, Japan. (12) Yamada, S.,Sato, H., Nature 193, 261 (1962).
RECEIVED for review September 14, 1964. Accepted November 12, 1964. Research s onsored by the U. S. Atomic Energy 8ommission under contract with Union Carbide Corp.
Ligand-Catalyzed Polarographic Reduction of Indium(II1) for Determination of Halides and Certain Organic Sulfur and Nitrogen Compounds A. JAMES ENGEL,1 JOAN LAWSON, and DAVID A. AIKENS Department o f Chemistry, Rensselaer Polytechnic institute, Troy, N. Y .
b Traces of a number of inorganic and organic substances catalyze the polarographic reduction of in(H20)6+3 and can be determined from the resulting catalytic current. Halides, thiocyanate, and a number of organic sulfur and nitrogen compounds can b e determined in this manner. Substances need not b e electroactive in the usual sense to b e detected and the catalytic nature of the reaction ensures high sensitivity. The key factor is the ability of catalytic substances to coordinate with and to labilize the hydration sheath and thus facilitate the electrodeposition reaction. Determination of iodide is given as an example and optimum results are obtained with iodide concentrations from 1 X 1 0 - 5 to 2 X 10-4 M.
M
ligands catalyze polarographic reductions of certain metal ions and greatly enhance the limiting current. The resulting catalytic current ANY
provides a sensitive and selective means for determination of such ligands, even though the ligands may not be electroactive in the usual sense. Analytical application of this effect has, however, received little attention. It has long been known that ligands such as the halides accelerate the normally slow polarographic reduction of many aquated metal ions such as I I I ( H ~ O ) ~and + ~ that addition of high concentrations of these ligands shifts the entire reduction wave to more positive potentials (7, 8, 16, 19). Catalytic waves were not observed by early workers, however, because a t the high ligand concentrations used, the current was limited by the rate of diffusion of metal and not by the rate of metal complex formation. Catalytic waves of metal complexes were first observed by Mark and Reilley (IS)for reduction of Ni(I1) in the presence of low concentrations of certain amines and the waves were used for determina-
tion of catalytically active amines (10, 11, 14). The mechanism was studied by Mark (11), who concluded that the critical step is complexation of Ni(I1) by ligand adsorbed on the electrode surface. The significance of complexation was pointed out by Nelson and Iwamoto (17), who proposed that complexation increases the ease of desolvation and facilitates deposition. This mechanism is supported by the work of Connick and Coppel (g), which shows that introduction of a foreign ligand such as chloride into the coordination sphere of an aquo ion increases the rate of displacement of the remaining water molecules. Hence labilization of the hydration sphere probably plays a key role in all ligandcatalyzed reductions of metal ions. The ligand-catalyzed reduction of In(II1) resembles the ligand-catalyzed Present address, Columbia Union College, Takoma Park, Md. VOL 37, NO. 2, FEBRUARY 1965
203
reduction of Ni(I1) as regards experimental characteristics and probable mechanism. In practice the two catalytic systems are complementary because each is applicable under markedly different pH conditions and exhibits maximum sensitivity toward different classes of ligands. A number of substances that can be determined by catalysis of the In(II1) reduction are listed and the measurement of small concentrations of iodide is described as an example. EXPERIMENTAL
Equipment. Polarograms were obtained using a conventional polarographic H cell. The dropping mercury electrode (D.M.E.) had a drop time in 1 X HC104 of 4.10 seconds and a mass flow rate of 2.09 mg. per second a t a mercury height of 98.5 em. with an applied potential of -0.540 volt. A saturated calomel electrode (S.C.E.) was used as the external reference electrode. The reference electrode compartment of the H cell was filled with 2-X ?;aC104 and connected t o the S.C.E. through an external salt bridge filled with saturated sodium nitrate. The salt bridge ends were closed with inch-diameter plugs of porous Vycor (NO. 7930 glass, Corning Glass Works, Corning, N. Y.). The double salt bridge was necessary to prevent precipitation of potassium perchlorate and plugging of the salt bridge, migration of chloride from the S.C.E. to the D.M.E., and interference with the catalytic In(II1) reduction. The resistance of this cell was 1000 ohms. The salt bridge electrolytes were checked regularly for chloride using silver ion and were changed whenever the chloride test was positive. Polarograms were secured using a Sargent Model XXI recording polarograph without damping. Applied potentials were checked with a potentiometer. S o correction was made for I R drop, which usually was of the order of 5 my. Experiments were performed in a thermostat with a temperature of 22' =t0.2' C. unless noted. Reagents. Snalytical reagent grade materials were used throughout unless noted. h stock solution 0.242M in In(II1) was prepared from indium metal of 99.999% purity obtained from Fairmount Chemical Co., Yewark, N. J. The appropriate amount of metal was dissolved with gentle heating in 55 ml. of 4M H2S04. After dilution t o 250 ml., the solution pH was 1.3. Linde H.P. grade nitrogen gas was used for deaeration. Measurement of Catalytic Current. The catalytic current appears as a small wave with a half-wave potential in the vicinity of -0.50 volt us. S.C.E. Because the shape and the half-wave potential depend on the iodide concentration, the potential for measurement of the catalytic current, denoted iCat, must be selected carefully to achieve optimum sensitivity 204
ANALYTICAL CHEMISTRY
-E,
Figure 1 .
volts
E. S.C.E.
Iodide-catalyzed reduction waves of In(lll)
1M HC104, In(lll) = 10+M Numbers on curves represent [I] X 1 05, M A. HClOa only 8. HCIO4 plus Inllll)
and precision. Typical catalytic waves given in Figure 1 show that -0.54 volt us. S.C.E. is suitable for measurement of the catalytic current with iodide concentrations up to 2 X 10-4 M. At higher iodide concentrations the peak current is shifted to more negative potentials and the current should be recorded a t the peak instead of a t a fixed potential. I n the absence of iodide, In(II1) gives a small background current, ia, which must be subtracted from the total current, it,to obtain itat. Because the catalytic current is proportional to the concentration of I n (111), adjustment of the In(II1) permits setting the catalytic current a t convenient levels. RESULTS AND DISCUSSION
Polarography of In(II1). Appropriate conditions for development of catalytic waves based on formation of In(II1) complexes are readily established from the polarographic behavior of In(II1). I n the absence of complexing anions the rate and extent of electroreduction of III(H,O)~+~ are remarkably sensitive to the proton concentration. The influence of p H on the polarographic limiting current of lX10-31V In(II1) in 1M sodiumhydrogen perchlorate is summarized in Figure 2. At pH 3 the reduction is polarographically reversible with a halfwave potential of -0.52 volt vs. S.C.E. and a diffusion-controlled limiting cur-
rent of 8.3 Ira. The limiting current drops sharply above pH 3 because indium hydroxide precipitates. The limiting current also decreases as the pH is lowered below pH 3 and a t pH 0.5 reaches a lower limit of approximately 2% of the normal diffusioncontrolled value obtained a t pH 3. With decreased pH the current also becomes increasingly kinetic-controlled, and below pH 1.5 the current becomes independent of the mercury height, indicating pure kinetic control (3). Thus In(H,0)6+3is not reduced directly but must first urdergo a chemical transformation, the rate of which is inhibited by protons. These observations suggest that the reduction of In(H20)6+3 proceeds through a species such as InOH(H20)6+2. The influence of foreign ligands on rates of exchange of metal ion hydration sheaths has been summarized by De Maeyer and Kustin (4), who conclude that hydroxide is an extremely effective catalyst. The pK, of I n ( H ~ 0 ) 6 is + ~3.70 (5) and formation of InOH(H20)6+zwould provide a facile path for reduction of In(H20)6+3 and explain the pH dependence of the limiting current. Iodide-Catalyzed Electroreduction of I ~ ( H Z O ) ~ +Suppression ~. of the efficient hydroxide-catalyzed path for In(H20)6+3 reduction limits the current to a very low value in acid solution. Then another ligand which catalyzes In(II1) reduction less efficiently than hydroxide ion can provide an effective pathway for In(II1) reduction, and the
PH
Figure 2. Influence of pH on polarographic limiting current of 1 mM In(lll) in 1 M sodium-hydrogen perchlorate Currents measured a t -0.75
concentration of the catalytic ligand determines the In(II1) reduction current. Iodide is typical of these ligands and provides a good illustration of the analytical application of ligandcatalyzed waves. Typical iodidecatalyzed waves for reduction of In(H20)6+3in 1 M HClO4 are shown in Figure 1 for iodide concentrations from 1 x 10-6M to 2 X 10-4M. These waves represent the current-voltage envelope formed by connecting the maximum currents obtained a t the end of each drop. The catalytic current, i,,,, measured a t -0.540 volt us. S.C.E., is plotted in Figure 3 as influenced by the iodide and In(II1) concentrations. Correction has been made for background currents obtained in the absence of iodide. Iodide evidently catalyzes electroreduction of I ~ ( H * O ) Gthrough +~ formation, a t the electrode surface, of a transient In(II1) - iodide complex such as 1n1(H20)5+2, This complex, like InOH(H20)6+2, is reduced extremely rapidly, regenerating the iodide catalyst. The In(II1) limiting current is therefore determined by the concentration of the iodide complex at the electrode surface, which in turn is set by the concentrations of In(II1) and iodide and the rate constant for formation of the iodide complex. In the absence of significant depletion of In(II1)-Le., when the In(II1) limiting current is less than 1 ~ of 7 the ~ diffusion-controlled valuethe In(II1) limiting current is controlled only by the rate of reaction between In(H20)6+3and iodide. Then the polarographic limiting current be-
volt vs. S.C.E.
comes independent of the height of the mercury column, behavior characteristic of an electrode reaction controlled by the rate of a preceding chemical step (3). The independence of the In(II1) limiting current and the mercury column height is demonstrated by the data in Table I. With 9.68 X lO-3M In(II1) the limiting current is independent of the mercury column height
(between 98.5 and 47.7 cm.) for iodide concentrations between 1.60 X 10-5M and 2.00 x l O - 4 X . With 9.68 x loT4 M In(III), the limiting current is independent of the mercury height within 5% between 8.00 X 10-5M and 2.00 X 10-4M iodide, The 5% variation in limiting current is attributed principally to difficulty in measuring the small currents and in correcting for residual current. Above 2 x lO-4M iodide, a significant fraction of the In(II1) is consumed and the limiting currents show increasing dependence on the rate of diffusion of In(II1). Thus with 1.00 X 10-3M iodide the limiting current is increased by a factor of 1.20 as the mercury column height is increased from 47.7 to 98.5 cm. A limiting current controlled solely by diffusion would be increased by a factor of 1.43 by this increase in mercury column height. Because the limiting current becomes increasingly dependent on the diffusion of In(II1) as the iodide concentration is raised, changes in the iodide Concentration have less effect on the limiting current. Hence the sensitivity toward iodide is greatest a t low iodide concentrations and diminishes as the iodide concentration is increased. As an example, with 9.68 X 10-4X In (111), raising the iodide concentration from 2.0 X l O - * X to 4.0 X lO-4M increases the limiting current from 1.25 pa. to 3.00 pa.; raising the iodide concentration from 2.0 X 10-~~.1'to 4 X IO+ M increases the limiting current from 8.30 to only 8.40 pa. The peaked current-voltage curves shown in Figure 1 suggest that the
Figure 3. Iodide-catalyzed reduction current iodide and In(lll) concentrations
of In(lll) as influenced by
Medium 1 M HClOh currents measured a t -0.54 A. [In(lll)] = 10-M B. [In(lll)) = 10-%4 C. [In(lll)] = 10-*M
volt
VI.
S.C.E.
VOL. 37, NO. 2, FEBRUARY 1965
205
Table 1.
In(III), M 9.68 x 10-4 9.68
x
10-3
Kinetic Nature of Iodide-Catalyzed In(lll) Limiting Current
Iodide, M
x x 1.00 x 1.60 x 8.00 x 2.00 x 1.00 x 8.00
2.00
10-5 10-4 10-3
Limiting current at 9815 cm. Hg Limiting Limiting current, pa. current at 98.5 cm. Hg 47.7 cm. Hg 47.7 cm. Hg 0.29 0.28 1.04 0.72 0.69 1.05 6.20 5.08 1.22
10-5 10-5 10-4
10-8
0.38
2.20 6.88
55.0
0.38
2.24 6.72 44.8
1.00
0.98 1.02 1.22
Limiting current measured at current peak.
II. Determination of Iodide Iodide, -44 X Found, ilv. error Range Taken av. of 4 -0.05 0.22 0.45 0.50 0.00 0.12 1.oo 1.00 0.12 2.96 -0.04 3.00 +0.05 0.08 6.05 6.00 +0.2 0.4 10.2 10.0
Table
15.0 20.0
15.5 19.6
+0.5
-0.4
0.6
0.7
electroreduction of I I I ( H ~ O ) ~ +is~ catalyzed principally by iodide adsorbed on the electrode surface. In a theoretical treatment of catalytic waves Mairanovski: (9) showed that the peaked current-voltage curves result from surface-catalyzed reactions in which the current dropoff is caused by desorption of the catalyst beyond the potential of the current peak. The current dropoff in the iodide-In(II1) reaction corresponds well with the potential region in which iodide is desorbed from a mercury electrode. The dominant role of adsorbed halides as the effective catalyst for reduction of In (111) is confirmed by the relative efficiencies of the halides as catalysts, which fall in the order I- > Br- > C1-. The order of catalytic efficiency parallels the well-known order of adsorption of these ions on mercury but falls in reverse order t o the stabilities of the halide complexes of In(II1) in solution, C1- > Br- > I - (6). Certain substances, especially anions such as sulfate, may complex In(II1) without catalyzing the electroreduction. These substances lower the effective concentration of In(H2O)Bf3 and thus depress the catalytic current. As an example, the catalytic current obtained with 10-3M In(II1) and l O - 3 M iodide is 6.36 pa. in 1 M HC104; in 0.5M H2S04the catalytic current is lowered to only 2.24 pa. Some repression of the catalytic current is apparent in Figure 3, which shows catalytic currents with9.68 X 10-4M, 9.68 X 10-3M, and 206
ANALYTICAL CHEMISTRY
9.68 X 10-2M indium sulfate. Increasing the In(II1) concentration from 9.68 X 10-4M to 9.68 X 10-3M should theoretically cause a tenfold increase in current a t a given iodide concentration. The observed increase in current falls 30% below the predicted value, however, because the fraction of free In(II1) decreases as the concentration of indium sulfate increases. The interference by moderate amounts of sulfate is small and use of indium sulfate avoids the danger of introducing catalytic ligands with the In(III), a severe problem when the In(II1) solution is prepared by dissolving indium metal in perchloric acid, because the latter reaction often yields large amounts of chloride ion. Effect of Temperature. Catalytic currents increase markedly with increase in temperature, with temperature coefficients as high as 80% per "C. reported (15). A temperature coefficient of between 5 and 10% per "C. was found for the pyridine-catalyzed reduction of nickel (22). The iodidecatalyzed current for reduction of In(H20)6+3increases 5% per "C. over the range from 23" to 28" C. These currents were secured using 10-2M In(II1) and l O - 4 M iodide and may be considered representative. Hence temperature control of ~ t 0 . 2 "C. is necessary to obtain catalytic currents within 1% of the correct value. Determination of Iodide. Figure 3 indicates that an In(II1) concentration of 1 X 10-2M should give optimum values of catalytic currents for iodide concentrations from 1 X 10-5M to 2 x l O - 4 M . Results for determination of iodide in this concentration range are summarized in Table 11. Each concentration of iodide was measured on four successive days in a cell thermostated a t 28.8' + 0.1" C., using a single calibration curve. Rest results were obtained for iodide concentrations above 1 X 10-5M. Between 3 x lop5and 2 X 10-4L14iodide, the average error falls between 1.3 and 3.3% and the range falls between 1.3 and 4.0%. S o trend of accuracy or precision with concentration is apparent. With
1 x 10-51+f iodide the range increases to 12% relative although the average error is zero. Below 1 X 10-5M iodide the catalytic current approaches the background current and the precision and accuracy suffer. Thus with 5 X 10+M iodide the average error is -10% and the range is 44%. Determination of Other Catalytic Ligands. d number of inorganic and organic ligands catalyze the electroreduction of In(H20)8+3 and can be determined from their catalytic currents. Catalytic activities of representative ligands are summarized in Table 111, which lists the catalytic current in microamperes for 10-2M In (111) and 10-4-lf ligand in 1M HC104. Of the halides iodide shows the highest and chloride the lowest activity. Fluoride was not studied because it rapidly attacks the D.1LI.E. capillary but probably has little catalytic activity. Of the pseudohalides, thiocyanate shows slightly higher activity than bromide but azide shows no catalytic activity. Azide fails to catalyze the electroreduction of In(II1) because the pK, of hydrazoic acid is 4.77 (1) and in 1M HC104 azide is converted effectively to hydrazoic acid. Although mercaptoacetic acid and 2-mercaptoethanol effectively catalyze electroreduction of In(III), benzyl mercaptan is inactive. The failure of benzyl mercaptan to catalyze the electroreduction is probably caused by the extensive adsorption of this com-
b
Table 111. In(lli) Catalytic Currents of Representative Ligands kat,
pa.,
for 10-4
M ligand aKd
Ell2, volts, 10-2M Ligand types os. S.C.E. In(II1) Halides and pseudohalides Iodide -0.50 .? 24 Bromide -0.50 1.11 Chloride -0.50 0.28 Thiocyanate -0.49 1.45 Organic sulfur compounds Thioacetic acid -0.50 1.63 Mercaptoacetic acid -0.58 2.6a 0.35 2-Mercaptoethanol - 0.49 Thiourea -0.49 0.35 Basic aromatic nitrogens 0-Phenylene-0.50 0.60 diamine 2,2'-Bipyridine - 0.82 1.63b 1,lO-Phenanthi-o-0.57 3.00 line Read at -0.6 volt; catalytic wave appears as shoulder on hydrogen evolution 5
wave.
bLigand concentration 1 x 1 0 - 5 ~ to prevent depletion of In(II1) at electrode surface which would occur a t 10-4.11 ligand.
pound on the electrode surface. Benzyl mercaptan is adsorbed strongly, as shown by the distorted current-time curves of individual mercury drops obtained in the presence of this substance. The adsorbed film could reduce the surface area of the electrode to a very small value and thus inhibit electroreduction of In(III), as has been observed for many other strongly adsorbed organic molecules (18). A number of carboxylic and hydroxylic ligands are inactive, including acetate, tartrate, and acetylacetonate. Oxalate shows slight activity, giving a current of 0.08 pa. Glycine, ethylenediamine, and a number of 0-amino alcohols also fail to catalyze reduction of In(II1). In comparison, aromatic diamines are more active than the aliphatic analogs. Thus o-phenylene diamine shows moderate activity, although aniline, o-aminobenzoic acid, and o-aminophenol are inactive. Ligands bearing heterocyclic nitrogens capable of chelation were among the
most efficient catalysts studied and one of these, 2,2’-bipyridine exceeded the catalytic power of any other ligand by fivefold. LITERATURE CITED
(1) Britton, H. T. S., Robinson, R. A,, Trans. Faraday Soc. 28, 531 (1932). (2) Connick, R. E., Coppel, C. P., J . Am. Chem. SOC.81,6369 (1959). (3) Delahay P., “New Instrumental
Methods ’in Electrochemistry,” pp. 87-114. Interscience. New York. 1954. (4) De &faeyer, L., Kustin, K., Ann. Rev. Phys. Chem. 14, 12-20 (1963). (5) Hattox, E. &I.,De Vries, T., J . Am. Chem. Soc. 58, 2126 (1936). (6) Hepler, L. G., Hugus, Z. Z., Ibid., 74, 6115 (1952). ( 7 ) Kolthoff, I. &I., Lingane, J. J., “Polarography,” 2nd ed., Vol. 11, pp. -51 9-20. _ _ . Interscience. New York. 1952. (8) Lingane, J. J., J . Am. Chem.’Soc. 61, 2099 (1939). (9) RlairanovskiI, S. 0.) J . Electroanal. Chem. 6 , 77 (1963). (10) ;\lark. H. B. Jr., ANAL. CHEM.36, Ok0 (1964).
(11) Mark, H. B., Jr., J . Electroanal. Chem. 7,276 (1964). (12) Mark, H. B., Jr., Reilley, C. N.,
ANAL.CHEM.35,195 (1963). (13) Mark, H. B., Jr., Reilley, C. N., J . Ekctroanal. Chem. 4, 189 (1962). (14) Mark, H. B., Jr., Schwarts, H. G., Jr., Ibdd., 6, 443 (1963). (15) Meites, L., “Polarographic Techniques,” pp. 78-82, Interscience, New York, 1955. (16) Moorhead, E. D., MacNevin, W. M., ANAL.CHEM.34,269 (1962). (17) Nelson, I. U., Iwamoto, R. T., J. Electroanal. Chem. 6 , 234 (1963). (18) Reilley, C. N., Stumm, W., “Progress in Polarography, Vol. I, Chap. V, P. Zuman and I. M. Kolthoff, eds., Interscience, New York, 1962. (19) Schufle, J. A., Stubbs, M. E., Whitman, K. E., J . Am. Chem. SOC.73, 1013 (1951). RECEIVED for review September 16, 1964. Accepted November 30, 1964. Research supported in part by National Science Foundation through Grant GP 1996. A. James Engel was sponsored by National Science Foundation and participated in the 1964 Summer Program in Instrumental Analysis held at Rensselaer Polytechnic Institute.
Electrooxidation of Tetraphenylborate Ion at the Pyrolytic Graphite Electrode W. RICHARD TURNER and PHILIP J. ELVING The University o f Michigan, Ann Arbor, Mich. Tetraphenylborate ion is oxidized at the stationary pyrolytic graphite electrode in aqueous solution to produce two voltammetric waves. The first wave, which occurs as a welldefined peak (E,,z = 0.21 6 volt VS. saturated sodium chloride-calomel electrode) results from a 2-electron, pH-independent process which produces diphenylborinic acid and biphenyl:
B(CE“h--f B(CsH6)2+
B(C&)z+
+
(C6H&
+
2e-
+ H2O -+ B(CGH&OH + H+
The second wave, which is less well defined, represents a 2-electron process, which involves the oxidation of diphenylborinic acid:
+
B(C6HsLOH 3- 2H20 4 NOHIS (CsH& f 2H+ -k 2eThe latter process is linearly pH0.057 dependent: E,lz = 0.92 pH. The slope of -0.057 indicates that one hydrogen ion is produced for each electron liberated in the oxidation.
-
S
its discovery by Wittig (IO), sodium tetraphenylborate (NaTPB) has become a highly useful analytical reagent for the determination INCE
of potassium, certain heavy metal ions, and many basic nitrogen compounds; in fact, its ability to precipitate large cations seems to be very general. While the chemistry of the tetraphenylborate ion in aqueous solution is fairly well understood, a systematic investiga tion of its electrooxidation in water has until now not been reported. As a part of a study of organic oxidation reactions a t the wax-impregnated graphite electrode, Elving and Smith (3) observed two waves for the electrooxidation of the tetraphenylborate ion in aqueous media. This observation led to the development of a method for the direct titration of potassium using amperometric equivalence-point detection ( 7 ) . These results would seem to be somewhat contradictory to those of Geske (4, Tho reported being unable to observe the oxidation of T P B a t the rotating platinum anode in aqueous solution because of the interfering evolution of oxygen. However, Geske was successful in carrying out the oxidation in anhydrous acetonitrile, where a single wave was produced. In one instance, there was an indication of a second wave; however, film formation prevented further examination of this wave. On the basis of his results, Geske proposed the following mechanism for
the electrooxidation of tetraphenylborate ion in acetonitrile:
Subsequently, Geske reported (5) that mass spectrographic analysis of the products of the electrolysis of a mixture of TPB and perdeutero-TPB in acetonitrile revealed the presence of a mixture of biphenyls containing either all hydrogen or all deuterium, thus indicating that both phenyl groups in each biphenyl had come from the same tetraphenylborate ion. I n the present investigation, the voltammetric and coulometric behavior of the tetraphenylborate ion was examined at the pyrolytic graphite electrode in both aqueous and nonaqueous media. EXPERIMENTAL
Sodium tetraphenylborate reagent (J. T. Baker; Fisher Scientific) was recrystallized from chloroform during the earlier phases of the present study; this was abandoned when the commercial material seemed to be of sufficient purity. Chemicals and Reagents.
VOL 37, NO. 2, FEBRUARY 1965
a
207