Deposition and Stripping Properties of Mercury on Iridium Electrodes S. P. Kounaves *'1 and J. Buffle Department of Inorganic and Analytical Chemistry, University of Geneva, 1211 Geneva 4, Switzerland ABSTRACT The reduction of mercury on a fresh iridium surface is reversible with a half-wave potential of +162 mV (vs. SCE). A Hg(O) monolayer is formed by underpotential deposition (UPD) at + 600 mV during the initial scans of a repetitive cyclic voltammogram. This monolayer acts as bulk Hg(O) during further cycles, and the reduction half-wave potential then shifts towards +410 mV. Anodization of the electrode at a potential > + 1000 mV returns the reduction wave to its original posi- tion. The peak for the oxidation of bulk Hg(O) from the surface occurs at +430 mV followed by a peak for the under- potential stripping (UPS) of the Hg(O) monolayer at + 650 inV. The quantity of Hg(O) for the monolayer corresponds to a 40% coverage of the Ir surface. The underpotential shift for the bulk and monolayer stripping of Hg(O) from Ir was found to be 210 inV. No indications of a Hg-soluble compound formation between Hg and Ir were found. The application of mercury film electrodes to Reagents.--Unless otherwise indicated, all chemicals speciation of trace elements in natural waters is pres- used were analytical reagent grade (Merck). The Hg(II) ently limited either due to the bad stability or reproduci- used for film formation, was prepared by dissolution of bility of the Hg film in varying experimental conditions triply distilled mercury in nitric acid, followed by dilu- (e.g., Hg films on glassy carbon) or to the dissolution of tion with 0.1M HC104 to give 10-3M Hg(II), unless other- the substrate in Hg and the formation of intermetallic wise indicated. compounds with the test metal ion (e.g., films on Pt). A All solutions were prepared with 18 mol/gt water from a detailed study of many substrates (1) showed that these Millipore brand ion exchange system. problems could be overcome by using Ir as a substrate. High-purity (99.95%) nitrogen gas (Carbagas) was used Unfortunately, although the electrochemical behavior of to purge oxygen from solutions and to maintain a nitro- mercury on substrates such as platinum, silver, gold, gen blanket during experimentation. nickel, and carbon is well documented in the literature, the same cannot be said for iridium. During the past 30 Results and Discussion years, only a few studies have dealt with mercury deposi- General considerations.--One of the simpler and more tion on iridium, and these were mainly concerned with direct methods for investigating the interactions of Hg the behavior of mercury as a poison for hydrogen and and the Ir surface is to employ cyclic voltammetry, using oxygen adsorption on the surface of the electrode (2-4). a Hg(II) solution and a "defined" iridium surface. Such a We report here the results of a study of the deposition and stripping properties of mercury at an iridium | 1 i i i i j i i rotating disk electrode (Ir-RDE). The aim of this work -60 B was to quantify the parameters required to optimize the formation and stability of a true mercury film on an -40 iridium substrate. This study was therefore directed to the determination of (i) the reduction characteristics of Hg(II) on Ir, (ii) the oxidation characteristics of Hg(O) - 2O from Ir, and (iii) the interactions of the Ir surface with Hg in terms of its solubility, the underpotential of w deposition/stripping, and work function differences. n. r~ Experimental iO0 Instrumentation.--The electrochemical system has been previously described (1). It consists of a micro- 200 computer controiled potentiostat (Motorola EXORset + C Tacussel PRG5), a custom built medium-exchange/flow- 300 ~CURRENTSCALE through Plexiglas cell, and an inverted polarizing micro- CHANGED scope (Leitz EPIVERT) allowing in-situ observation and = i I I l 1 photography. BOO 600 400 200 0 The working electrode assembly was a Tacussel POTENTIAL,mV vs. SCE EDI-55442 rotating disk electrode (RDE) with a 1 cm Teflon tip into which a 2 mm diam • 10 mm long Ir cylin- der (Heraeus GmhH) was press fitted. The electrode was (c) rotated at 1500 rpm during the experiments. The coun- ~a_-20 terelectrode was a Pt rod. The reference electrode, to JL ...... which all potentials are referenced, was a SCE (238 mV vs. NHE) filled with a saturated NaC1 solution and equipped with a 0.1M NaNO3 bridge. -/ Electrode pretreatment.--The electrode was initially \ polished with diamond spray pastes of decreasing size, E$ 0 l : i , I t I I I I I the smallest of them being 1 ~m until a mirror-like sur- ! 1 face could be seen under the microscope at 250 times. 400 350 300 400 0 -400 The electrode was then washed with deionized water and POTENTIAL, mV vs. SCE cathodized at -2000 mV for 5 rain in degassed 1.0M Fig. 1. (a) Consecutive cyclic voltammograms for reduction and oxi- HNO3. The electrode was kept in this solution until dation of Hg on an Ir substrate. Conditions: 10-3M Hg § 0.1M HCI04, transferred to a test solution. v = 20 mV/s, and ~o = 1500 rpm, Note: Anodic-scan current scale is 5 *Electrochemical Society Active Member. times cathodic-scan current scale. (b) Enlargement of the "bend" at the 1Present address: Department of Chemistry, State University foot of wave-A. (c) Reduction wave for Hg(I) or Hg(ll) at a dropping of New York at Buffalo, Buffalo, New York 14214. mercury electrode. 2495 Downloaded on 2013-10-27 to IP 130.64.26.40 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use 2496 J. Electrochem. Soc.: ELECTROCHEMICAL SCIENCE AND TECHNOLOGY December 1986 curve is shown in Fig. la with 20 consecutive repeating still present (if the change in scan direction occurs at E < cycles. +600 mV) or reformed (when E > +700 mV). On both Pt For the cathodic scan, the most prominent features are (8) and Au (9) a monolayer of Hg(O), which is difficult to the "shift" of the reduction wave-A from -~ 160 mV to 275 oxidize, has been explained as being a result of chemical mV, and peak-B which appears on top of wave-A at ~ 70 interaction between Hg and the substrate material. It inV. Furthermore, as can be seen, the reduction wave is will be shown later that Ir does not form a soluble com- not a clean sigmoidal shape as would be expected for the pound with Hg (there are no amalgam oxidation peaks as diffusion controlled reduction of Hg(II). It is compli- there are for Au and Pt), but that a Hg(O) monolayer is cated by a small bend (shown by the arrow) at the foot of nevertheless formed between 600 and 400 inV. The posi- the wave and a two-step reduction wave for the first six tive shift in the E1/~. value of wave-A and the limited scans. growth of peak-E suggest that the monolayer coverage is The interpretation of the curves in Fig. l(a) requires only partial at first, but increases with the number of taking into account the disproportionation reaction scans, becoming almost complete, as evidenced by the fact that Ell2 approaches that for the reduction of rig(I) or Hg 2+ + Hg ~ ( ~ Hg.o 2+ [1] Hg(II) on the dropping mercury electrode. Its value can In fact, although the bulk of the solution contains only be computed from E~/2 = Eo" + (RT/2F)ln([Hg2+]/2) [see Hg(II) ions, as soon as some Hg(O) droplets are formed at Ref. (7) p. 167] and is +410 mV for the conditions used the electrode surface, Hg(I) can be formed at the elec- here ([Hg 2+] = 0.001M, E'o = 510 mV). trode surface as a result of reaction [1]. The equilibrium Note that the reduction wave can be made to return to constant for reaction [1], K, can be computed from the its original position (cycle No. 1, Fig. la), either by ap- Nernst relationship using the two couples Hg~+/Hg22+: plying a potential > 1000 mV or by washing in concentra- {Eo'}, and Hg22+/Hg~ "} to give ted nitric acid. Both of these treatments destroy the Hg monolayer formed between 600 and 400 mV. E = Eo' + (RT/2F) ln([Hg2+?/[Hg22+]) [2] Determination of E1/2 for Hg(II) --~ Hg(O).--In at- E = E"o + (RT/2F) ln([Hg,2+]) [3] tempting to measure E1/2 for the reduction we must con- sider two limiting cases for the reduction of Hg(II) to where the symbols have their usual meaning. At any po- Hg(O). Case a: Reduction onto Hg(O) already deposited tential, subtraction of Eq. [2] from Eq. [3] gives on Ir, where the corresponding E1/~. is given by wave-A-20 log K = ln(Hg22+]/[Hg"+]) = (E"o - Eo')F/RT [41 (Fig. la), E~/2(Hg) = 275 mV. Case b: Reduction on pure Ir where E~/2(Ir) can be measured either from wave-A-1 The value of K for reaction [1] is ~ 130 at 25~ (5), so that (Fig. la) (lst cycle in cyclic voltammogram, El/2(Ir) = 170 at equilibrium, in a noncomplexing medium such as mV); or by using stripping polarography (Fig. 2). This lat- HClO4, the concentration of Hg(I) will be about 130 times ter measurement, E~/~(SP), is graphically more precise, greater than that of Hg(II). Thus, after the formation of a but less accurate since the surface is no longer purely Ir monolayer of Hg(O) on the Ir electrode, the reduction of due to the longer reduction time.