J. Electroanal. Chem., 240 (1988) 309-315 Elsevier Sequoia S.A., Lausanne - Printed
in The Netherlands
Short communication ON THE pH DEPENDENCE OF THE REDUCTIVE DESORPTION OF IODINE AT POLYCRYSTALLINE AND SINGLE-CRYSTAL PLATINUM ELECTRODES IN AQUEOUS SOLVENTS
Uniuersrty, College Station,
22nd June 1987; in revised form 31st August
TX 77843 (U.S.A.)
Electrochemical surface science has gained much impetus in recent years from the use of modern surface analytical techniques and the employment of well-defined single-crystal electrodes [l]. While it is now realized that the use of single-crystal surfaces is essential in understanding electrocatalytic phenomena at the atomic level [l], it remains imperative that results obtained from single crystals be compared and/or correlated with those from polycrystalline electrodes. Recently, two independent studies on the cathodic stripping of iodine chemisorbed on platinum electrodes in aqueous solvents have been reported [2,3]; one study was performed on smooth polycrystalline platinum  while the other was carried out at a well-defined Pt (111) single crystal . The comparison of the results from these studies, insofar as the pH dependence of the complete reductive desorption process is concerned, is the subject of this short communication; the establishment of the pH dependence of the iodine-stripping reaction is important in order to appreciate fully the role of hydrogen chemisorption in electrocatalytic reductions carried out at platinum-group metals in aqueous media. EXPERIMENTAL
The study employing smooth polycrystalline platinum electrodes , was based on thin-layer electrochemical techniques; experimental details with respect to preparation of the thin-layer cells, the electrode surfaces, and the electrolytic solutions have been described previously . Thin-layer electrochemical measurements of the absolute surface coverage of chemisorbed iodine were done in 1 M H,SO, by oxidation at 1.2 V [Ag/AgCl (1 M Cl-) reference] for 60 s (which oxidizes the chemisorbed iodine to aqueous iodate) followed by coulometric reduction at 0.7 V 0022-0728/88/$03.50
0 1988 Elsevter Sequoia
(which reduces the aqueous iodate to aqueous iodine) ; the surface concentration of iodine, f?t (mol cm-‘), is obtained from the iodate-to-iodine reduction charge:
r, = (Q -
where A is the electrode surface area determined by underpotential hydrogen deposition , F is Faraday’s constant, and Qb is the background charge measured in the absence of chernisorbed iodine. The study employing well-defined Pt (111) single-crystal electrodes was based on the combined use of electrochemistry with ultra-high vacuum surface spectroscopic methods . Measurement of the iodine packing densities involved quantitative Auger electron spectroscopy, as discussed in detail elsewhere [3,5]; for further experimental details, ref. 3 should be consulted. RESULTS
The present discussion will focus only on the pH dependence of the reductive desorption of iodine at polycrystalline  and single-crystal platinum surfaces  in aqueous solvents. Figures 1 and 2 show graphs of the surface coverage of iodine on platinum as a function of electrode potential near the hydrogen-evolution region at selected pH values; the plots in Fig. 1 were generated using data previously reported for a polycrystalline electrode , while the graphs in Fig. 2 were drawn using data recently published for a Pt (111) single crystal . The data in Fig. 1 clearly demonstrate a pH dependence on the cathodic stripping of iodine at polycrystalline Pt; for example, complete removal of iodine from polycrystalline platinum at pH 7
1.2 \ 1.0 : N I5
% 0 h B
E/V vs. AgCl Fig. 1. Absolute surface packing density of iodme on smooth polycrystalline platinum as a function of electrode potential in (0) aqueous 1 M H,SO, (taken as pH 0), (x) 1 M NaCIO, (buffered at pH 7). (0) 1 M NaClO, (buffered at pH IO), and (+) 1 M NaOH (taken as pH 14) as published by Soriaga and co-workers ; the average standard deviation in the iodine coverages was *6%. The solid lines interconnect the data points and do not represent any theoretical fit. The volume of the thin-layer cell, V = 3.86 81; the area of the electrode, A = 1.04 cm’; temperature. T = 298 K.
0.4 0.3 0.2 0.1
E/V vs. AgCl Fig. 2. Iodine adsorption profile on Pt (111) at negative potentials buffered at pH: (0) 4, (x) 7, and (Cl) 10, as pubhshed by Hubbard
in aqueous 0.1 mM and co-workers .
is attained at -0.65 V but, at pH 10, removal is not complete until -0.78 V . Similar pH dependence is also shown by the data in Fig. 2 for Pt (111); for example, complete removal of iodine at pH 7 is achieved at - 0.63 V but, at pH 10, removal is not complete until -0.70 . It has been suggested (21, on the basis of voltammetric and coulometric evidence, that the pH dependence depicted in Fig. 1 arises because the cathodic stripping of iodine from polycrystalline Pt in aqueous media is coupled with the dissociative chemisorption of hydrogen; it is reasonable to associate the pH dependence displayed in Fig. 2 for Pt (111) also to the participation of hydrogen chemisorption side-reactions. On polycrystalline platinum, the hydrogenative stripping of iodine was found to obey the following reaction : I (adbJ+ H,O + 2 e- = HCadsj+ ILq) + OHwhich may also be written I Cads)
It is important to test reaction (2) in terms of the data in Figs. 1 and 2. This may be accomplished in the following manner. The surface redox reaction in eqn. (2) is quasi-reversible ; hence, one may write its Nernst equation as:
0.0296 PH - 0.0296 l”g[
or: E = K - 0.0296 pH where
(3b) K combines
all the pH-independent
z d >
z 5 c w F
-0.7 -‘I E (l/2) = -0.35 - 0.028 ptl
P" Fig. 3. Plot of E,,,, the potential at which the packing density of iodine chemisorbed on smooth polycrystalline Pt was at half maxlmum, against solution pH. The plot was generated using the data in Fig. 1  as described in the text. The solid line represents the linear least-squares fit (correlation coefficient of 1.00).
of the i th species, and
E&S, = E&s)
E&ads) is for the reaction: H&
+ e- c, HCadsj is for the reaction:
and Ei&,, I (ads)
at which the fractional surface At the half-coverage potential E,,2, the potential ) is one half, eqn. (3) predicts that a plot of E,,, coverage of iodine (0, = Ii/I’i,_ against solution pH would yield a straight line of slope equal to -0.0296 V and an intercept equal to the constant K. At E = E1,2, half of the platinum is covered by iodine and the other half is occupied by hydrogen. Under these conditions, 8u = I’H/IH,max = 0.5 and 8i = 0.5; K then hence, the ratio a H(ads)/aI(ads) may be taken as unity. The intercept simplifies to: K( E = El,z)
= E;;,,,, + E;CadsJ + 0.0296 pI
where p1 = - log a i -Caq)= - log ci-, ci- being the molar concentration of the iodide in solution. Plots of El,* vs. pH are shown in Fig. 3 for polycrystalline platinum and in Fig. 4 for a Pt (111) single crystal. Data for the plot of Fig. 3 were obtained from the graphs in Fig. 1: E,,, was the potential at which I, = 0.5 for a given pH .
’.. n 6 8
P" Fig. 4. Plot of E,,,, the potential at which the surface coverage of iodine on Pt (111) was at half maximum, against solution pH. The plot was generated using the data in Fig. 2  as described in the text. The solid line represents the linear least-squares fit (correlation coefficient of 0.99).
Generation of the data for the plot in Fig. 4 involved drawing a straight line between the last two or three data points in the extreme negative potential region at a given pH; estimates of 0.58, and E1,2 were obtained from this straight line using 0 1,max E I?,,max/rpt = 0.42 iodine atoms per surface platinum atom . The uncertainties associated with the input data for Fig. 4 are thus larger than those of Fig. 3. It is interesting to note that, within experimental error, the E,,* vs. pH plots for polycrystalline and single-crystal Pt are in good agreement with one another. Linear least-squares analysis of the data, represented by the solid lines in Figs. 3 and 4, gave the following results. For polycrystalline platinum: intercept = -0.35 V; slope = -0.028 V/pH; correlation coefficient = 1.00. For Pt (111): intercept = - 0.34 V; slope = - 0.027 V/pH; correlation coefficient = 0.99. Two important points will be stressed from the foregoing results: (i) The slopes for both polycrystalline and single-crystal platinum are consistent with the predictions of eqn. (3); this means that reaction (2) does indeed occur at both polycrystalline and single-crystal platinum. (ii) The intercepts K for both polycrystalline and single-crystal platinum are almost identical. However, in view of eqn. (7) this does not necessarily mean that E”Icadsjis the same for both surfaces. In the Pt (111) study, the concentration of aqueous iodide was maintained at 0.1 mM, which gives p1 = 4.0. The polycrystalline work was performed in the absence of aqueous iodide, but it must be realized that any desorbed iodide is retained inside the thin-layer cavity; hence, at E1,2, the iodide concentration is given by: ci- = o.sr,,,,,A/V where
I’ is the thin-layer
(8) cell volume,
3.86 ~1. For
= 1.1 nmol
A = 1.04 cm2: ci- = 0.15 mM, and p1 = 3.8. EGcadsj is of course dependent upon the crystallographic orientation of the electrode surface. For a Pt (111) surface , E0H(ads) = - 0.20 V [Ag/AgCl (1 M Cl-) reference], but for a polycrystalline surface, there is no unique value for E&&. In this study, Egcadsj has been approximated as the potential at which 8, = 0.5; from studies with roughened  and smooth  polycrystalline platinum electrodes, EGcadsj = -0.1 V. When these numbers are substituted in eqn. (7) EGads, values of -0.36 V and -0.27 V are obtained for the polycrystalline and single-crystal electrode, respectively. It is interesting to note that a small but sharp, pH-independent redox peak was observed at -0.34 V on Pt (111) which has been attributed to the Icads) + I,, reaction involving a minor fraction (- 5%) of the surface iodine; the remaining 95% of the adsorbed iodine remained zero-valent near this potential . This minority reductive desorption reaction is evidently not pH dependent because it is related only to reconstruction of the iodine superlattice and does not involve inclusion of chemisorbed hydrogen into the adlattice; when the I/Pt (111) surface was disordered by anodic oxidation, the spike at -0.34 V was not observed on a subsequent negative scan . The values of E&,,, obtained here indicate that the redox potential of the iodine/iodide couple is shifted in the negative direction by about 0.76 V for polycrystalline Pt, and 0.67 V for Pt (111) when iodine is in the chemisorbed state. As discussed previously , chemisorption-induced redox potential shifts can be used to estimate the relative chemisorption strengths of the oxidized and reduced forms of an electroactive material; for the present study, the difference in chemisorption strengths between surface iodine and iodide, A(AG”), is given by : A( AG’= ) = AG;,,,,
- AG;mcadsj = nF [Eczds, - E&,] - AG,o
of AG,‘;,,,, and AGF-(,ds, are the respective Gibbs energies of adsorption . . for the and iodide, E&, and E&s, are the respective standard potentials couples, and AGdo is the energy involved in the IZcaq)-+ 2 :u.&I,,, and l(ads)/l;ds) (aq) dissociation. The derivation of eqn. (9) does not make any a priori assumptions concerning the individual values of AGG,,,, and AGF-(,,,,; that is, the term AGFmtadsj is retained in eqn. (9) even if, as stated in reaction (6). I,,,, is an unstable species. The chemisorbability of iodide on platinum can be assessed from eqn. (9) only if AG”icads) is known from other experiments such as thermal desorption studies . For estimating A(AG”), the following approximations have been used: Eczd:ds, - EC&:,, = -0.76 V for polycrystalline Pt, = -0.67 V for Pt (ill), and 2AGz = the energy required to break the I-I bond in the gas phase, 150 kJ/mol . Substitution of these values in eqn. (9) yields A(AG o ) = - 150 kJ/mol for polycrystalline Pt and 2: - 140 kJ/mol for Pt (111); the deviation between these two values may in part be due to the approximation that EGcadsj = -0.1 V for polycrystalline Pt. Recent studies on the reductive desorption of iodine at polycrystalline Pt in anhydrous acetonitrile gave A( AC”) = - 160 kJ/mol . The difference in the A(AG”) values obtained in water and acetonitrile is probably only an artifact arising from the neglect of the solvent dependence of AG:.
It will be mentioned that thermal desorption experiments  indicate that AG”tCadsjI - 160 kJ/mol. In view of the present results, this means that the value of to exceedingly weak attractive, if not repulsive, forces. It is AG P-(ads) corresponds therefore not surprising that iodide is expelled from the electrode surface as soon as chemisorbed iodine is reduced to iodide. ACKNOWLEDGEMENT
Acknowledgement this research.
to the Robert
REFERENCES 1 L.R. Faulkner, Charactenzation of Electrochemical Processes, Publication NMAB 438-3, Nattonal Academy Press, Washington, DC, 1987. 2 T. Mebrahtu, J.F. Rodriguez, B.G. Bravo and M.P. Sonaga, J. Electroanal. Chem., 219 (1987) 327. 3 F. Lu, G.N. Salarta, H. Baltruschat and A.T. Hubbard, J. Electroanal. Chem., 222 (1987) 327. 4 J.H White, M.P. Soriaga and A.T. Hubbard, J. Electroanal. Chem., 177 (1984) 89. 5 J.A. Schoeffel and A.T. Hubbard, Anal. Chem., 49 (1977) 2330. 6 A.T. Hubbard, R.M. Ishikawa and J. Katekaru, J. Elcctroanal. Chem., 86 (1978) 271. 7 F.A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry. Wiley, New York, 1980. 8 B.G. Bravo. T. Mebrahtu. J.F. Rodnguez and M.P. Soriaga, J. Electroanal. Chem., 220 (1987) 281. 9 (a) G.A. Garwood and A.T. Hubbard, Surf. Set., 92 (1980) 617; (b) T.E. Felter and A.T. Hubbard, J. Electroanal. Chem., 100 (1979) 473.