Hydrogen Overpotential on Electroplated Copper Tin Alloys - The

Hydrogen Overpotential on Electroplated Copper Tin Alloys. I. A. Ammar, and H. Sabry. J. Phys. Chem. , 1958, 62 (7), pp 801–805. DOI: 10.1021/j15056...
0 downloads 0 Views 614KB Size
July, 1958

HYDROGEN OVERPOTENTIAL ON ELECTROPLATED COPPER-TIN ALLOYS

1/T the three points do not give a straight line, but rather a curve with increasing slope a t higher temperatures. The slopes at 25" give AHa2 = 8.2 kcal/mole, and AHA= 5.4 kcal/mole. Equilibria of Species in Solutions of Copper oPhenanthroline (1:1) Nitrate.-As in the case of copper dipyridyl, several species of copper ophenaiithroline exist in aqueous solutions A-o-PhCu(HiO)z + + Bi-o-PhCu(Hz0) (OH) + Bz-0-PhCu (0H)p (BI)Z-(O-P~CUOH) 2 ++

influenced by the basicity of the ligand; as the basicity (or p K a ) of ligands increases, the acidity of the chelates should decrease. Thus in comparing dipyridyl (pKa= 4.4) with o-phenanthroline (pK, 5.0) as ligands for cupric ion, it is found that the o-phenanthroline chelate is the weaker:acid. TABLEIV EQUILIBRIUM CONSTANTS FOR COPPERDIPYRIDYLA N D COPPER0-PHENANTHROLINE IN 0.1 M KNO, Constant

Kaz

The equilibria between these species are as shown previously. The titration curves at 25 and 41.2' with 5 and 10 X lo-* moles/liter of chelate are shown in Fig. 5,9 the values of Y at pH 6.5, 6.75, 7.0, 7.25 and 7.5 are shown in Table IIIj9and the graphs of log Y vs. log ([Ali - Y ) are shown in Fig. 6.9 At 25", % is found to be 2.0 throughout the whole range of pH values and concentrations, and Ka2 = 1.27 X lo-" = 10-10.90. The pH of the crossover is 8.50 so K A = 10-17.00. At 41.2", % is also 2.0 for the whole range of pH values and concentrations, and Kea = 5 X lo-" = 10-10.30. The pH of the cross-over is 8.35; thus, KA = From the values of K,2 and KA at 25 and 41.2", one may calculate AHoz = 8.3 kcal./mole, and AHA = 4.2 kcal./mole. Comparison of Equilibrium Constants.-In Table IV are shown the equilibrium constants at various temperatures for copper dipyridyl (1: 1) nitrate and copper o-phenanthroline (1: 1) nitrate. The tendency to dimerize to species (B& is about the same, but the o-phenanthroline chelate is definitely the weaker acid as evidenced by the lower value of KA. The acidity of 1: 1 metal chelates undoubtedly is

801

KA

Tyw.,

C.

0 25 41.2 0 25 41.2

Copper dipyridyl

0.65 x 1.8 X 4.8 x 3.0 x 5.2 x 12.0 x

10-1' lo-" 10-11 10-17 10-17 10-17

Copper o-phenanthroline

. . . ,......

1.27 x IO-" 5 x 10-11

.... . . . . . . 1.0

2.0

x x

10-17 10-17

Equilibria of Species in Solutions of 1 :1 Cupric Ethylenediamine Nitrate.-In all of the foregoing work the solutions contained 0.1 M sodium nittrate to keep activity coefficients of the metal chelate ions constant over wide ranges of concentration. However, in the case of the ethylenediamine chelate, addition of salt caused formation of a coppercontaining precipitate so in this system salt could not be used and activity coefficients were not constant. Plots of the data as in Figs. 2, 3 and 6 have a very wide scatter of points. Nevertheless, the titration curves appear similar, with a cross-over at pH 8.6 (indicating a ~ K ofA17.2) and the value of Kaz is estimated to be about 10-ll,e. Acknowledgment.-The authors wish t o thank Dr. J. N. Wilson for advice and encouragement and Mrs. Helen B. Heppe and Mr. Sherwood C. Beckley for assistance in obtaining precise titration curves.

HYDROGEN OVERPOTENTIAL ON ELECTROPLATED COPPER-TIN ALLOYS BY I. A. AMMARAND H. SABRY Department of Chenzistry, Faculty of Science, University of Cairo, Cairo, E g y p t Received Auoust SO, 1957

Hydrogen overpotential has been measured on electroplated Cu, Sn and Cu-Sn alloys, in 1.0 N HC1 at 30". Six alloys ranging in composition from 12 to SO$'& Sn have been studied. The overpotential has been measured in the current density range 10+ to 3 x 10-8 amp./cm2. The results have been analyzed statistically and the 95% confidence limits, for the mean overpotential values, have been calculated. Attempts have been made to esplain the overpotential-alloy composition relation through the dependence of overpotential on the heat of adsorption of hydrogen on the alloy.

Introduction Overpoteiitial studies on alloys are numerous. Thus Newbery' observed that the overpotential 7 was higher for Pb-Hg and Zn-Hg alloys than the corresponding values for the components. Fischer2 measured 7 on a number of alloys and found it to be independent of the composition of the alloy and equal to that of the component with the lower overpotential. Harkins and Adams3 found the over(1) E. Newbery, J. Chem. SOC.,109, 1051 (1916). ( 2 ) P. Fischer, Z . p h y d . Chem., 113, 326 (1924).

(3) W. Harkins and H. Adams, THISJOURNAL, 29, 205

(1926).

potential on monel metal t o be lower than the corresponding values for the components. For Cu-Ni alloys, Raeder and co-workers4J observed a gradual change from the overpotential on Cu to that on Ni. Sharp maxima and a flat minimum mere, however, observed for the other alloys studied by the above auth0rs.~#5DeKay Thompson6 studied the relation between the overpotential and composition of brass, and observed a sharp minimum between 15 (4) M. Raeder and J. Brun, 2. physik. Chem., 153, 15 (1928). ( 5 ) M. Raeder a n d D. Efjestad, ibid., 8140, 124 (1928). (G) M. DeKay Thompson, Trans. Electrochem. Soc., 69, 115

(1931).

I. A. AMMAR AND H. SABRY

802

and 18% Cu. He observed no connection between q and the equilibrium diagram of the alloy. DeKay Thompson and Kaye7 observed a maximum between 40 to 80% Fe for the overpotential on Ni-Fe alloys at low polarizing current densities in KOH solutions. For Cr-Ni alloys, a minimum in the overpotential-alloy composition relation was observed between 20 and 30% Cr a t high polarizing current densities.* Croatto and Da Viag observed a rise in q with the increase of Sb and Cd contents in Pb-Sb and Pb-Cd alloys when two phases were present, No general conclusions, as to the relation between q and the alloy composition, could be drawn from the results of the above authors. Furthermore, no close connection between the overpotentialalloy composition relation and the equilibrium diagram or the lattice structure of the alloy could be e ~ t a b l i s h e d . The ~ ~ ~aim of the present investigation is, therefore, to study the relation between overpotential and alloy composition for electroplated CuSn alloys, with careful attention paid to the recent advances in the technique of hydrogen overpotential measurernents,Io and to attempt an explanation for the results obtained. The lack of rigorous experimental techniques in the older work emphasizes the need for the present investigation. Statistical analysis must be used, therefore, in order to establish the overpotential results with certainty. The use of electroplated Cu-Sn alloys is justified by the ease with which they are prepared in a reproducible fashion, and by the occurrence in the literature of reliable data for the overpotential on both C U ' ~ and ~ ' ~Sn.13

Vol. 62

The apparent surface area was used for the calculation of the current density. The cathodes employed were electroplated Sn and Cu-Sn alloys. The overpotential on electroplated Cu was previously investigated.11 Six compositions for Cu-Sn alloys were studied: 12, 28, 42, 55,.67 and 80% Sn. Tin was Deplated from a stannate bathlo operated a t 80 f 1'. position was carried out at a current, density of 2 X 10-2 amp./cm.a for 20 minutes. The compositions (in g./l.) of the various baths used for the deposition of Cu-Sn alloys are given in Table I, together with the appropriate cathodic current density empl~yed.'~The use of potassium salts in place of sodium salts, in some baths, is justified by the fact that potassium salts increase the stability of the hath.17J8 Baths which were kept for some time were, therefore, prepared from potassium salts. For the deposition of the alloys containing 42% Sn, this bath was used17 Sodium stannate, NazSn03.3Hz0 Sodium hydroxide, NaOH Cuprous cyanide, CuCN Sodium cyanide, NaCN 3 Cathodic current density

(7) M.DeKay Thompson and A. Kaye, ibid., 60, 229 (1931). (8) M.DeKay Thompson and G. Sistare, ibid., 78, 259 (1940). (9) U. Croatto and M. Da Via, Ricerca Sei., 12, 1197 (1941). (10) J. O'M. Bockris, Chem. Reus., 43, 525 (1948). (11) S. Wakkad, I. A. Ammar and H . Sabry, J . Chem. Soc., bo20 (19.56). (12) J. O'M. Bockris and N. Pentland, Trans. Faraday Soc., 46, 833 (1952). (13) A. Hickling and F. Salt, ibzd., 36, 1226 (1940); 87, 333 (1941); J. O'M. Bockris and 9. Ignatowicz, ibid., 44, 519 (1948); K. Pecherskaya and V. Stender, J. A p p l . Chem., Russ., 19, 1303 (1946). (14) I. A. Ammar and S. A. Awad, THISJOURNAL, 60, 837 (1956). (15) A. Azzam, J. O'M. Bockris, B . Conway and H. Roaenberg, Trans. Faraday SOC.,46,918 (1950).

x

For the decomposition of all alloys studied, the temperature was kept constant a t 65 f 1' and deposition was carried out for 20 minutes. To emure the constancy of the bath composition during alloy deposition, two copper and tin anodes were used." The composition of the plated alloy was checked by analysis, and when the plating conditions were kept as constant as possible variations only of the order of 1,to 2% in the composition of the alloy were ob-

TABLE I %

SnC14.5Hz0 NaOH

Sn

80 118 118 118 170

12 28 55 67 80

KOH Cu(CN) N a C N K C N

..

80

.

Electrode

N

97

..

..

136 174

, ,

3.3 x 10-3, amp./cm.z

Cu 15 12% Sn 8 8 28% 12 42% 10 55% 67 % 7 80% 7 Sn 15 a Values of mv

.

'I

c

299 429 433 386 396 389 500 633

7 6 9 9 8 7 5 2 u and

11,

36 12 12 12 4

180

,

Experimental The experimental technique was similar to that previously described for electroplated cathodes.14 Electroplating was carried out in a cell fitted with water-sealed ground glass joints. The plating and the overpotential cells were made of arsenic-free glass and were cleaned with a mixture of Analar HNO, and H&Oc acids followed by conductivity water. A platinum wire, sealed to glass, was used as a substrate for electroplating, and was cleaned in the above manner. After plating the electrode was washed several times with conductivity water and was immediately inserted in its position in the previously cleaned overpotentid cell (water-sealed taps and ground glass joints). The cell was completely filled with conductivity water and this was displaced by pure hydrogen before the electrolyte was introduced. The electrolyte was 1.0 N HCl and was purified by pre-electrolysisls on an auxiliary electrode, of the same alloy as the test cathode, at about 4 X amp./cm.2 for 20 hours. A platinized platinum electrode, in the same solution and at the same temperature as the test cathode, was used as a reference. In this manner the overpotential was measured directly. The direct method and the rapid technique were employed.'S The temperature was kept constant at, 30 f 0.5' with the help of an air thermostat.

100 g./l. 10 gJ1. 11.5 gJ1. 2 8 . 5 g./l. 10-2 amp./cm.a

..

56 29 29 29

.. 27

TABLE 11"

3.3 x 10-4, amp./cm.Z

io_-a,amp./cm.X

*L

'I

I

3 X lo-* 2 3 2 2

.. ..

..

*L

15 261 6 13 15 372 7 16 21 384 6 14 20 333 9 20 18 349 7 15 18 337 6 14 11 11 446 5 5 580 2 4 L are approximated

Current density, ampi/ cm.

'I

*.L

233 5 11 328 7 15 347 4 10 288 10 19 311 6 13 293 6 14 396 5 12 530 3 6 to the nearest

TABLE 111 Elec6trode (mv.) Cu" 125 108 12%Sn 107 28% 42% 102 55% 104 105 67% 80% 106 Sn 107

+L 1 3 1 3 2 5 2 4 1 3 2 5 2 4 1 3 c

($0)

3

amp./cm.Z 1.9 X 10-6 3.9 X 10-7 2.3 4.6 3.5 4.0 7.7 X 4.2 X

U

3.7 X 10-6 2.6 X 10-8 3.1 1.9 1.1 1.9

2.3 X 5.3 X

10-B

ItL 8.0 X 10-8 6.2 X 10-8 7.2 4.1 2.4 4.8 5 . 6 X lo-' 1.1

Only the higher Tnfel line slope and the corresponding io value are included (cf. ref. 11). 0

(16) W. Blum and G. Hogaboom, "Principles of Electroplating and Electroforming," McGraw-Hill Book Co., Inc., New Y o r k , N. Y., 1949,p. 328. (17) R. Angles, F. Jones, J. Price and J . Cuthberlson, J . Etectrodeponitor's Tech. SOC.,21, I9 (1946); T i n Itesearch Inst. (London), publications and private communications. (18) F. Lowenheim, Trans. Etectrochem. SOC.,84, 195 (1943).

July, 1958

HYDROGEN OVERPOTENTIAL ON ELECTROPLATED COPPER-TINALLOYS

served. Analar grade reagents were used for the preparation of the various baths.

803

B

0.7

Results Hydrogen overpotential is measured in the current density range to 3 X 10-6 amp./ cme2,in 1.0 N HC1 solution a t 30". The results are statistically analyzed. For this reason, a new electrode and a new solution are used to measure each individual Tafel line. The mean Tafel line is computed from a t least seven individual measurements. To show the accuracy of the results, the mean over-potential +, the corresponding error u, the 95y0 confidence limits fL,19 'and the number of individual measurements N are given in Table I1 for electroplated Cu, Sn and C u S n alloys at three different current densities, ie., and amp./cm.2. The results for Cu plated from a cyanide bath, taken from a previous investigation," are included here for comparison, The mean values for the Tafel line slope b, and for the exchange current (ZO) are given in Table I11 together with their standard errors and confidence limits. The mean Tafel lines for electroplated Cu, Sn and C u S n alloys are shown in Fig. 1. It is clear from this figure that in most cases the overpotential tends to a stationary value at the row current density range,

0.6

0.5

r:

5.= 0.4 0.3

0.2

0.1

log c.d. electrode.20 I n Fig. 2 the mean overpotentia1 is Plotted against the composition- of Fig. 1.-Mean Tafel lines for-electroplated Cu, Sn and Cu-Sn the alloy for 3.3 X IO-', and 3.3 X alloys in 1.0 N HC1 a t 30": I, Cu; 11, Sn; 111, 12% Sn; IV, 28% amp./cm. 2. Sn; V, 42% Sn; VI, 55% Sn; VII, 67% Sn; VIII, SO% Sn. Discussion Previous work on the kinetics of the cathodic hy0.6 drogen evolution reaction on Cu in HC1 solutions has indicated that the slow discharge of Hfions is rate determining.12J1 This conclusion has been based on the value of the Tafel line slope, and on the fact that the capacity of the electrode-solution 0.5 interface is independent of frequency. l 2 This indicates a small coverage of the metal surface with adsorbed H atoms and hence a rate-determining slow discharge has been favored. For Sn cathodes, also in HC1 solutions, the transfer coefficient a is v 6 ,0.4 found to lie between 0.4 and O.45.l3J2 I n the present investigation a value of a = 0.56 is calculated from the mean slope of the Tafel lines on Sn (cf. Table 11Q according to b = 2.303 RT/aF. Since a rate-determining catalytic desorption is charac0.3 terized by a value of a = 2 in the current density range studied in the present investigation, the overpotential results on Sn may be explained by a ratedetermining slow discharge or a slow electrochemical mechanism.2a Distinction between the two al-

L

f

h

(19) 0. Davies, "Statistical Methods in Research and Production," Oliver a n d Boyd, London, 1949. (20) J. Kolotyrkin a n d A. Frumkin, Cornpt. rend. acnd. sci., i3.R.-

S.S., 88, 445 (1941). (21) N. Petland, J. O ' h l . Bockris and E. Sheldon, J . Elechochern. Soc., 104, 182 (1957). (22) J. O'RI. Bockris, "Electrochem. Congtants," Natl. Bur. Standards, No. 524, 1053, p. 243. (23) J. O.'M. Bockris and E. C. Potter, J . Electrochem. Soc., 99, 169 (1952).

50 60 80 70 90 100 Sn, %. Fig. 2.-Relation between overpotential and alloy com0,3.3 x 10-4 position at: 0 , 3.3 X lo+; @, 1.0 X amp./cm.Z 0

10 20

30

40

ternatives could not be made directly since the evaluation of the stoichiometric number or the elec-

I. A. AMMARAND H. SABRY

804

Vol. 62 io = K, exp[-AHo*/RT)

55

-

54

-

AH^*^ - (AHo*),I 53

N

- [(AHa)1 - (AHahI

(2)

and consequently

I 0

(1)

where Kl includes the standard entropy of activation AXo* at the reversible potential, and AHo* is the standard heat of activation a t the same potential. When A&* is independent of electrode material as it has been suggested by Ruetschi and D e l a h a ~ and , ~ ~ when the other terms in K1 are the same for a number of metals, i.e., the rate-determining reaction, the electrolyte concentration and the temperature are the same, io will change with AHo* from one metal to another. I n comparing the overpotential a t two different electrodes, it has been shown that the difference between the values of AHo* is approximately equal to the difference between the heats AHa of adsorption of hydrogen on the two metals when the slow discharge is rate determining on both cases, thusz4

I

20

40

60

80

100

AHo* e const.

- AHa

(3)

From (1)and (3) one gets Sn, %. io KZexp[AH,/RT] (4) Fig. 3.-Relation between AHs (kcal./gram atom) and the a!loy composition for a rate-determining slow discharge mecha- and K z is independent of electrode material. nism. The relation between n and AH, can be deduced tron number23has been hindered by the dissolution when the cathodic current density i is given by of the electrode a t low current densities. i = io exp[-aqF/RT] (5) The results of the present investigation on electroplated Sn are in good agreement with those of and thus from (4) and (5) a t constant i one gets Fecherskaya and Stenderl3 also on an electroplated q K3 (AH./~F) (6) Sn electrode. Although the values of CY and b for electroplated electrodes agree with those obtained and KS is independent of electrode material if a is for electrodes in the form of a wire (Bockris13 and independent of electrode material. Equation 6 can, Hickling13), yet the overpotential results on the in principle, be used to calculate approximate valformer are numerically smaller than those on the ues of q for a metal on which the slow discharge is latter. This may be attributed to a large true sur- rate determining, when K 3 and AH, are estimated. face area for electroplated electrodes as compared Values of AH, for Sn and Cu-Sn alloys, calculated to electrodes in the form of a wire. Thus, a t a con- according to (6) and taking AH, for Cu as 58.5 stant apparent current density (using the apparent kcal./gram atom,24are shown in Fig. 3. However, surface area), the true current density is smaller in Fig. 3 can only indicate that if the slow discharge is the case of electroplated Sn than in the case of a tin rate-determining the heat of adsorption of hydrowire. Hence, the overpotential will be numerically gen decreases with the increase of Sn content in the smaller on the former as compared to the latter. alloy up to about 28% Sn. This is followed by a The dependence of hydrogen overpotential on slight increase of AH, above 28% Sn, and then it the nature of electrode material has been investi- decreases considerably above 67% Sn. It is gated by Ruetschi and DelahayZ4for a rate-de- knownz6that alloys having the composition 39 to termining slow discharge mechanism. They have 60% Sn are mixtures of the superlattices e and q 1 observed a linear relation between 8, for values phases which correspond to the formulas Cu3Sn numerically greater than 0.1 v., and the heat of ad- and Cu&nb, r e s p e c t i ~ e l y . ~From ~ . ~ ~ 0 to 39% sorption of atomic hydrogen on the metal. It is Sn, the alloy is a mixture of the superlattice E and worthwhile t o consider the treatment of the above the solid solution a-phase. 25 For compositions authors in an attempt to explain the relation be- higher than about 61% Sn the alloy is a mixture of tween overpotential and the alloy composition for the superlattice 7' and Sn.25 From this, and on the C u S n alloys studied in the present investigation; assumption that the slow discharge is rate-deterthe assumption being now made that the slow dis- mining, Fig. 3 indicates that AHa decreases with the charge controls the rate of hydrogen evolution for tin content of the alloy except in the region where Cu, Sn and Cu-Sn alloys (cf. the values of b in the mixture of the two superlattices E and ql is Table 111). It is noteworthy that two slopes are formed. The above discussion could have been obtained for Cu plated from a cyanide bath, and this has been attributed to the heterogeneity of the (25) G. Raynor, Inst. Metals Annotated Equilibrium Diagram electrode surface.ll Now, the exchange current for Series, No. 2 (1944). (26) J. Bernal, Nature, 122, 54 (1928). the hydrogen evolution reaction is given byz3 (27) A. Westgren and G. Phragmen, Z . Metallkunde, 18, 279 (1926);

+

(24) P. Riletschi and P. Delshay, J . Chern. Phys., 28, 195 (1956).

2. onorg. Chem., 175, 80 (1928).

July, 1958

HYDROGEN OVERPOTENTIAL ON MERCURY IN PERCHLORIC ACID

805

checked had it been possible to get some direct inThe authors wish to express their thanks to Prof. formation as to the relation between AH, and the J. Horiuti and Prof. J. O’M. Bockris for helpful discussions. alloy structure.

HYDROGEN OVERPOTENTIAL ON MERCURY I N PERCHLORIC ACID SOLUTIONS BY I. A. AMMAR AND M. HASSANEIN Department of Chemistry, Faculty of Science, Cairo University, Cairo, Egypt Received January id, 1068

Hydrogen overpotential has been measured on Hg in 0.1, 1.0, 2.0 and 5.0 N HClOd at 25’. The observed Tafel line slopes are numerically greater than the theoretical value of 0.12 v. expected for a slow proton discharge or a slow electrochemical desorption. Electrocapillary curves have been measured in the same solutions, and the components of the charge of the double layer have been calculated. The Tafel line slopes have been related to the structure of the interface a t the cathodic branch of the electrocapillary curves.

Introduction Criteria for establishing the mechanism of hydrogen evolution on the cathode have been recently formu1ated.l-3 These include the Tafel slope, the electron number and the stoichiometric number. Tafel slopes associated with the three important mechanisms, namely, slow discharge, electrochemical desorption and catalytic combination are ( 2 X 2.303RTIF) for the first, ( 2 X 2.363RT/ 3F) and ( 2 X 2.303RTIF) for th‘e second, (2.303R T / 2 F ) and m for the third. Side reactions, due to the presence of trace impurities, often vitiate the kinetics of hydrogen evolution and the experimental slope is different from what is theoretically expected. However, in some studies, carried out under rigorous experimental conditions, breaks in the Tafel line are observed in the current density range where contributions from resistance overpotential, concentration polarization and ionization of adsorbed atomic hydrogen are diminishingly small and can be n e g l e ~ t e d . ~ -The ~ break in the Tafel line is sometimes attributed to a complex mechanism in which the rate is controlled by more than one step.6 I n other cases, the occurrence of two slopes has been explained on the basis that the potential a t the Gouy-Helmholtz boundary is dependent on the electrode-solution p.d. when the electrode potential is near to that of the electrocapillary maximum.6 Specific adsorption of HsO+ ions on Ag cathodes also has been ~ u g g e s t e d . ~I n this case, the increase of the Tafel slope with increase of cathodic polarization is due to the decrease of the symmetry factor, for the discharge step, as a consequence of the approach of the initial state to the cathode surface. Anomalies in the shape of the Tafel lines for Hg in concentrated solutions of HCI and HBr are at: tributed to the adsorption of anions at the low cur(1) J. O’M. Bockris a n d E. C. Potter, J . Electrochem. Soc., 99, 169 (1952). (2) J. O‘M. Bockris, Ann. Reus. Phys. Chem., 11, 477 (1954). (3) J. O’M. Bockris, “Electrochem. Constants,” Natl. Bur. gtandards, No. 524, 1953, p. 243. (4) J. O’M. Bockris and B. Conway, Trans. Faraday Soc., 48, 724 (1952). (5) I. Ammar and S. Awad, TIXIS JOURNAL, 60, 837 (1956). (6) J. O’M. Bockris, I. A. Ammar a n d A. K. Huq, ibid., 61, 879 (1957).

A.

rent density range.7 This causes a decrease in the overpotential a t low current densities as compared to the values obtained by extrapolating the Tafel line from high to Iow current densities. Adsorption of anions has been confirmed by calculating the surface concentration of anions with the aid of the Gibbs adsorption equation. A marked increase in the surface concentration of anions, with decrease of current density, leads to the appearance of an inflection in the Tafel line. Adsorption of both anions and cations has been studied by measuring electrocapillary curves for Hg in HC1 solutions of different concentrations.* On the cathodic side of the electrocapillary curve, the charge in dilute solutions is due largely to a n excess of hydrogen ions. I n concentrated solutions, however, there is an actual repulsion of hydrogen ions but the charge is due to the much greater repulsion of C1- ions.8 On the anodic side C1- ions are attracted, and in dilute solutions the charge is carried chiefly by the excess of these ions. At the electrocapillary maximum the adsorption of both ions is equal and the net charge is zero. The aim of the present investigation is to extend the studies of adsorption of anions and cations on Hg to solutions of perchloric acid, and to study the overpotential in these solutions with particular attention to the Tafel h e slope. Hence, the relation between the Tafel slope and the extent of adsorption may be elucidated. Perchloric acid is chosen because overpotential studies on Hg in this acid are very few.3

Experimental The electrolytic cell was used to measure hydrogen overpotential as well as electrcicapillary curves. I t was constructed of arsenic-free glass and water-sealed t.ap.s and ground glass joints. The cell had four compartments for the cathode, anode, pre-electrolysis electrode and reference electrode. The reference electrode was a platinized latinum electrode in the same Bolution aa the mercury catEode. A Luggin capillary was used to connect tmhecathode and reference electrode compartments. Overpotential was measured on a mercury cathode in the form of a pool at the bottom of the cathode compartment. This mercury vae introduced from a reservoir attached to the bottom of the cathode compartment. For the measurements of electro-

Z. lofa, Act5 Phyaicochim., U.R.S.S., 10, 903 (1939). (8) M. Devanathan, Ph.D. Thesis, London, 1951.

(7)