KINETICS AND MECHANISM OF THE ANODIC DISSOLUTION OF NICKEL IN HCI-DIMETHYLSULPHOXIDE SOLUTIONS

-The anodic dissolution of nickel in HCl-DMSO solutions containing different supporting electrolytes has been studied between 2@45”C. The electrodissolution and electrodeposition are predominantly activated electrode processes. At high anodic potentials and in the presence of perchlorate ions at high concentration passivation sets in. A probable reaction pathway is postulated which explains most of the experimental findings. Passiva-tion corresponds to a salt precipitation-dissolution process at the electrode vicinity.


Complete
quantitative data on metal dissolution processes in non-aqueous electrolytic solutions is rather scarce, since they only exist for iron in HCl-dimcthylsulphoxide (DMSO) solutions[l, 21, and for iron in HCI and HBr solutions in acetonitrile (ACN) [3].These are probably the more extended studies on the subject although the information they provide is still very limited as compared to the number of reports, for instance, on the iron group metals in aqueous solutions.Iron anodes in non-aqueous media dissolve and passivate, the corresponding electrode reactions involving the participation of the solvent.This result encourages the systematic investigation of corrosion and passivity processes of other metals of the iron group, in order to evaluate the corresponding kinetic parameters and to interpret them in terms of a reaction mechanism.
It is interesting to determine for a particular ionic system in non aqueous media whether a common formal reaction pathway is valid for the iron group metals as it apparently occurs in aqueous solutions[4, 51.

EXPERIMENTAL
The electrolysis cell as well as solvent and solution preparations are essentially the same as those already described in previous publicationsC2, 63.The working electrodes consisted either of a nickel rotating disc (2.8 mm dia) or a surface resulting from a nickel rod (2.8 mm dia) coaxially mounted on a Teflon@ rod (15cm dia) cut at 45".Flat

The rest potentials
When the metal is in contact with the HCI-DMSO solution it attains a stable rest potential, E,, which only depends on the HCl concentration according to the following equation already given [9].At 25°C E,(sce)(in V) = (-0298 _t 0.020) -00592 log C,+.
(2) Neither Cl-ion concentration nor Ni(II) ion concentration have any appreciable influence on the rest potential.

Potentiostatic E/l curves
These curves were run with both fixed and rotating disc nickel electrodes.Those obtained with the latter were very reproducible and will be described more extensively.Potentiostatic E/f curves (476-2995 rpm) exhibit the following characteristics (Fig. 1).At a fixed rotation speed, w, the region from the rest potential up to -0-l V (region I), the current increases with the applied potential and it is practically unaffected by stirring conditions.From -@1 V upwards the current increases attaining a maximum (region II), and finally, at about 0.1 V the current decreases with the potential attaining a limiting value at higher potentials (region III).The current around the maximum increases with w.
The E/I curve (region I) approaches a Tafel line between -0-35 and -Oo-20V with a slope comprised between 80 and 90mV (Fig. 2).The limiting current (region III) increases linearly with the 01j2.The occurrence of both the passivating effect and the limiting current regions depends on the presence of ClO, ions in the solution.Thus, if a 0.944 M HCl solution is electiolysed at 25°C under potentiostatic conditions up to l.OV, a steady current is established, without any appreciable passivation.Under these circumstances neither current maxima nor anodic limiting current are observed.This behaviour is reversed when a 0.125 M HCl(1 M KC104) solution is used (Fig. 3).
E/I curves recorded from l.OV downwardly coincide with those recorded from the rest potential upwardly both in regions I and III, although the anodic current maxima is no longer observed.Otherwise the current at the maximum, under potentiostatic conditions, attains a steady value after about 40 minutes.
The anodic current is independent of Ni(I1) ion concentration and apparently it increases linearly with the square root of the H+ ion concentration.Water (up to 10,COO ppm) has no appreciable influence on the dissolution process (Fig. 4).rent with slopes ranging from 70 to 90 mV, at 25°C.At a constant potential the current increases slightly on increasing the H + ion concentration, the probably reaction order is no greater than O-5.

The atwdic limiting current
Region III was studied with the rde under different experimental conditions.At 0.3 V, the anodic limiting current increases linearly with IX"' (Fig. 6) and the corresponding straight lines, at any temperature, intercept the origin ofcoordinates.The limiting current at a constant rotation speed increases with temperature fitting an Arrhenius plot (Fig. 7), with an apparent activation energy equal to 17.8 f 1.0 Gal/mole.This value exceeds the predictions of any simple mass transport process.The height of the limiting current plateau decreases with the concentration of Ni(I1) (Fig. 8), and a reverse effect is produced on increasing the Cl-ion concentration.Neither the H* ion concentration nor the addition of water up to ZOOOppm shows any definite effect in this region of the E/I curve.

Electrode &jj'brcntial capacitance
The electrode differential capacitances were evaluated from the potentiai decay curves at current interruption both for the anodic and cathodic processes.In the @058-0*243 V overpotential range (referred to E,), the apparent differential capacitance is between 20 and 36 @/cm2.In the range -0.8 to -1.2 V its value is about 30 pF/cm'.

Potentiostatic pulses
The current-time profiles (Fig. 9) recorded with various electrolytic solutions in the overpotential range 0.65-155 V exhibit a rapid increase of current up to a maximum value, afterwards decreasing to a very small steady value at tp The charge involved at the onset of passivation, determined after integration of the J/t record, diminishes linearly with the magnitude of the applied potential pulse.The charge, Q, related to the passivation effect corresponds to a thick layer of an insoluble species, such as precipitated Ni(cla&. The charge obtained from the potentiostatic pulses is in close agreement with that evaluated from other techniques, as described further on.A linear Q us t;" dependence is found, although the corresponding straight line does not go through the origin of coordinates.The latter suggests that a constant charge is required before passivation sets in (Fig. 10).

Transition times
Galvanostatic transition times were measured in the range from 0.72 to 19-6 mA using still electrolytic solutions.The error of the galvanostatic pulse up to 1 l-0 mA was less than 2 per cent.Potential/time displays (Fig. 11) exhibit an initial abrupt potential jump, then a region of nearly steady potential and finally a sudden potential increase after a transition time 7.For t < 10 s the product izl" is reasonably constant (Table l), while at t > 10 s it slightly but steadily decreases (Fig. 12).No simple theoretical relationship is adequate to fit the galvanostatic E/t curves.9. Potentiodynumic E/I curves These runs were made with different solutions either still or stirred at potential sweep rates, v, between 5 and 150 mV/s, with automatic compensation of the ohmic drop (Fig. 13).The initial part of the voltammograms is independent of both stirring and v.The voltammograms exhibit an anodic current peak whose height and peak potential increase when D increases.The irreproducibility, however, turns it difficult to establish a reliable quantitative law for both dependences.The charge required for passivation ta set in, increases also with u.Furthermore, the charge required for passivation is larger when the solution is stirred (Table 2).
The voltammograms are more asymmetric in shape as either v or w increase, resembling those of activated processes involving an appreciable ohmic contribution due to the formation of the insoluble passivating salt 1 _--------------a---9--
The E/I curves obtained either with still or stirred solutions (Fig. 14) show the H* ion discharge current[9] from the rest potential down to -0.8 V (portion I).At potential more cathodic than -0.8 V the electroreduction of Ni(I1) is observed @ortions II and III).The Iatter process exhibits a cathodic prewave (portion II).Beyond -l-8 V the simultaneous electroreduction of the solvent takes place.The E/I curve related to Ni(I1) reduction fits at potentials more anodic than -1.4 V a semilogarithmic plot (Fig. 15) involving a cathodic Tare1 slope approaching the ratio 2.3(2 RT/F).

Results indicate that the electrochemical behaviour of polycrystalline Ni in HCl-DMSO
solutions with different supporting electrolytes, exhibits an active dissolution region and a passivation region, the latter being apparently related to the local precipitation of an insoluble salt.In spite of the complexity of the processes involved their explanation in terms of a single reaction pathway may be attempted.

Possible reaction pathways for the active dissolution process
It is obvious, from the above reported results, that Accordingly, two main reaction models can be proposed.Either the molecules or the ions are the predominant species initially adsorbed at the electrode surface.
Mechanism A. Let us assume that the solvent adsorbs preferentially to the anions.The fact that the anodic stationary Tafel slope is lower than 2RT/F suggests that it is rather unlikely that the initial electron transfer is rate determining.Furthermore, the values of the apparent electrode capacitance are low and practically potential independent in the region where the reactions were investigated.Consequently, it is reasonable to admit that the degree of surface coverage by any reaction intermediate is negligible.Therefore, the following first reaction scheme can be discussed: and Ki = ky'k-i Equation ( 5) shows a -1st order dependence with respect to H' ions and a 1st order dependence with respect to HCl.This prediction, however, may be unrealistic since HCl is strongly ionized in DMSO becomes : ia = k~,K,K,K,(~,~)exp(3fE/2). (6) This rate equation involves a Tafel slope equal to 2RT/ 3F if no diffusion effects interfere+ The situation, however, may be rather different if the rate determining step also determines the diffusion rate of species formed at the electrode or participating at the double layer structure.Thus, during the reaction, H+ ions are produced at a rate equal to i, = i,, at the electrochemical interface, then they diffuse out of the electrode region at the same rate: Equation ( 10) implies a cathodic Tafel slope equal to The dependences of the corrosion potential, EC,,, on H+ ion concentration and Cl-ion concentration are obtained from equations ( 6) and ( 11).Thus, at low ca!, p+)X=O =z (CH+)x-cc, when i, = ic,u+ = i,,,, E = dcoTy and it results: After introducing E,,,, as given by equation ( 12) either into equation ( 6) or equation (1 I), one obtains: icorr = (k~K,K,K,)(Cc,-)"4(CH+)3'4 ( 14) din i,,,, 3 a1n i.,,. 1 The predictions of mechanism A are compared to experimental results in Table 3.A partial correspondence between experimental and theoretical results is achieved.The (NiCl+C) concentration is determined by the ionic Equations ( 16) and ( 18) involve respectively the Tafel equilibria involving the different complex ionic species.

Passivution Mechanism
The experimental results obtained in the passivity region indicate that ClO, ions are required for the onset of passivity and the charge required for passivation largely exceeds that for the formation of a nonconducting film of the order of a monolayer thickness.Otherwise, stirring conditions have a definite influence on the passivating current, as revealed by the linear ir us &* plot (Fig. 6).These facts indicate that any solid state mechanism [16] to explain the passivity should be, in principle, discarded.However, a reasonable explanation can be achieved in terms of a precipitation-dissolution mechanism.
Taking into account that passivation occurs when ClO; ion concentration exceeds largely the concentration of other anions, the scheme of reaction, which complements the mechanism of active dissolution, can be put forward as follows: _  Other reaction pathways involving competitive steps, such as the one discussed by various authors[ 18,191 for nickel passivation in aqueous solutions, predict a different i~"~ vs i dependence that the one previously reported in this paper.
The shape of the potentiostatic f/t profile approaches the prediction of a simple precipitationdissolution mechanism (Fig. 9) but the charge required increases as the rate of dissolution decreases.The postulated mechanism implies that a longer time should be required to reach the critical concentration when the rate of complex salt formation (B-5a) diminishes, because its precipitation must occur within a thicker layer adjacent to the electrode.This effect has already been noticed with other passivation processes controlled by a precipitation-dissolution mechanism [Zl,221.The effect of stirring on the passivity current can also be accounted for with a precipitation and dissolution process based upon a nucleation mechanism [23].Accordingly, provided the kinetic conditions imposed upon the reaction scheme are valid, the limiting current density obtained with the rde should depend on the square root of the rotation speed, as found experimentally (Fig. 6).
In conclusion, the results discussed in this paper indicate that the anodic nickel corrosion in HC1-DMSO solutions is an electrochemical activated process which is accompanied by a passivation effect due to a dissolution-precipitation mechanism of a Ni(II] salt.The latter is clearly evidenced when ClO; ions are present in the solution at a relatively high concentration. Fig. I. Stationary potentiostatic E/icurves obtained with the rde; 0,125 M HCLI M KC104; 25°C.(0) 476 rpm; (m) 917 rpm; (A) 917 rpm (run from anodic to cathodic potentials); (0) 2202 rpm; (-t) 2995 rpm.

Ni
inadequate to interpret the reactions of Ni electrodes in HCl-DMSO solutions.Consecutive reaction mechanisms involving the participation of the solvent, such as those earlier proposed for the iron group metals either in aqueous or nonaqueous electrolytes must be considered.A review of these mechanisms is given in[5].Under the present circumstances, taking into account the composition of the metal/electrolyte interface, an interaction between the metal surface and either the solvent molecules or the halide ions must occur.Adsorption of solvent molecules probably occurs preferentially near the potential of zero charge @zc) while anion adsorption should prevail at the anodic side of the pzc.where y + z = 4; 0 < x < 6; 0 < y < 4.Let us assume that steps (A-lHA-3) are at equilibrium, step (A-4) is the rate determining one and reactions (A-5) and (A-6) are equilibria with the participation of the various complexes of Ni with Cl-and DMSO[l3].According to mechanism A, the total ano-2 dissolution reaction of Ni in DMSCHCI solutions Ni + 4 HCl = NiCl:-+ 4 H * + 2 e.(3)If (A-4) is rate determining the anodic cd, i,, can be written as follows:i.= 2Fk, &c, exp(=,.,fEl (4)where eNicr is the degree of surface coverage by the reaction intermediate, k, and 01, are respectively the specific rate constant and the transfer coefficient assist-&g the reaction in the anodic direction of step n; .f= FiRT and E is the electrode ootential measured against the see.Under a quasi-stkdy state for the preceeding steps, equation (4) results: The rate equation for the hydrogen evolution rcaction on Ni in HCl-DMSO is[9]: &It+ = kH+(CG'+) exp(--fW) (11) reaction scheme one considers the preferential adsorption of chloride ion as earlier postulated for the anodic dissolution of iron in acid medium in the presence of halide ions[I5].Anion adsorption predominates at potentials more positive than the pzc.Thus, Ni + Cl-= Ni(Cl-) (B-1) Ni(Cl-) + DMSO = Ni(B)-+ H+ + Cl-Ni(B)-= Ni(B) + e Ni(B)'s-, (NiB)+ + e (NiB)+ + (z + y + 1)HCl + (X -I)DMSO = [NiCl,,(DMS0)X]2-Y + (z + l)clof equation (7) is obtained where X, y, and z are the same numbers already defined after assuming a linear concentration profile extending for Mechanism A and B = CH,SOCH; If steps B-l from the electrode surface up to &, the Nernst diffi-and B-Z were relatively fast processes, the influence of sion layer.If (CH+)x=,, 3 (C,+),,, one obtains: the adsorption process would only appear as a double sNio layer effect in the rate equation.If step (B-4) is rate cyielding now an anodic Tafel line with a slope equal kh = 2Fk_,(l -0,JC,,.(17) to 4KT/3F.After correcting for the HC ion diffusion, equation (15) On the basis of the same rate determining step, becomes: mechanism A predicts the following current density equation for the cathodic discharge of Ni (II) species: i, = i FDH+khK,K& 112 i, = k_-4(Cli(ja+) exp( -n/2) A. E. DELGADO, D. POSADAS and A. J. ARVIA dered as limiting cases of the following cxpresslon: ) obtained from this mechanism is the same as that for mechanism A. but the dependences of E,,,, and L,,., on the H+ ion and Cl-ion concentrations are different:

C
is the concentration of the passivating compound, D its diffusion coefficient and x is a distance normal to the electrode surface.Equation (24) is solved with the following initial and boundary conditions: t) depends on the experimental technique employed.The solution of equation (24) with condiis an irreversible process, as already discussed.The compensa-const) yields for C* the following expression[ 171: tion of Ni(II) charges at the interface occurs more likely through the participation of ClO; ions.When the rate of Ni(II) formation becomes faster than the diffusion of Ni(DMSO), (ClO,), out of the interface, its local concentration attains a value larger than a critical value C* related to the solubility product of the salt, then it precipitates in the electrode vicinity causing a decrease of the anodic dissolution current.The precipitate is in equilibrium with the various Ni(I1) complexes in solution and its dissolution (step B-5a) is determined by Fick's law: the time required to attain passivation (C = C*).For the galvanostatic runs equation (26) predicts it"' = const.Assuming D = 10e5 cm'/s, the estimated value of c* results of the order of 10e4 mole/ cm3, which coincides with the value obtained from Ni(I1) solubility estimations in 1 M KC104.That figure corresponds to an electrolyte of a relatively high solubility.

Table 3
compares the experimental results and the predictions of mechanism B. A closer agreement is now observed.

Table 3 .
Comparison of theoretical