13 ELECTROWINNING AND ELECTROREFINING OF METALS 1 Introduction 3 2 Redox Equilibria 5 24 Electrochemical Potentials 5 22 Single Electrode Potentials 6 2.3 Reference Electrodes ° 2.4 Non Standard Electrode Potentials, 8 2.5 Effect of Complexation on Redox Potentials 10 2.6 © ExipH Diagrams 4 3 Electrochemical Kinet 44 3.4 Reactions at electrodes 14 3.2 Potential Dependence of Electrode Kinetics 18 3.3 Characteristics of the Butler-Volmer Equation 7 4 Mass Transport to Electrodes 24 44 Mass Transport Processes a 42 iffusion Layer Mode! a 4.3 Mass Transport Correlations 23 4.4 Influence of Mass Transport on Electrochemical Kinetics 25 4.5 Mass Transport of tons 26 5 Mass Transfer at Vertical Electrodes 28 5.4 Natural Conveetion 28 5.2 Application te Copper Deposition 30 5.3 Effect of Gas Evolution 30 6 Electrocrystallization 32 64 Influence of Kinetics on Deposit Morphology 33 6.2 Application to Metal Deposits 36 6.3 _Underpotential Deposition 38TA 12 73 14 at 10 4 2 424 12.2 12.3 42.4 425 13 434 413.2 43.3 14 444 14.2 143 14.4 145 18 48.4 48.2 15.3 16 46.4 16.2 7 474 47.2 73 418 475 18 484 18.2 19 Current distribution in cells Primary Current Distribution Secondary Current Distribution Tertiary Current Distribution Current Distribution in 3-Dimensional Electrodes Materials for cells and electrodes ‘Anodes and cathodes Cell design Tankhouse current distribution Energy consumption Copper electrorefining Gathodi Electrolytes Current density and starter sl Anode passivation Deportment of anode impurities Electrowinning of copper Competing Reactions Quality of Copper Cathodes ‘Anodes for Electrowinning Zing electrowinning General Aspects Current efficiency Cathode Purity Effect of Fluoride tons Anode Corrosion Electrowinning of nickel and cobalt Nickel Cobatt Etectrowinning from chloride solutions Electrowinning in Novel Cells ‘Three dimensional Cathodes Operation of Packed Bed Cells Gold and silver Electrowinning Processes Design and operation of gold electrowinning cells Gell Materials and Construction Forced Flow Cells Electrorefining Exerc es Questions Worked solutions Appendix a 43 45 ar 36 56 87 sr 68 ot 4 66 68 348 78 as 26 a7 89 20 1 93 95 96 98 405 39 54 85 58 68 78 87 90 98 12213 ELECTROWINNING AND ELECTROREFINING be METALS “ 1___ introduction his course examines the use of electrical net. metals on a cathode, The power can also be used to oxidize metal to merallic ions _at an anode, The following sections will aspects of the electrodeposition of metals in general with a “Focus ‘on copper, zinc, nickel and gold and also on the more practical conditions und. : metals ate produced, Pee Blectrowinning is the recovery onto a cathode of a metal which will contain a range of impunties which are codeposited fom the impure fecd solution. ‘The feed solution may be coduced_directly leach operstion or mayb yor whol -trowinning. Electrorefining invalves ss sre.anode (often from pyrometallurgical processes ‘or pr vinning)_into-a. solution. from_which.it.is dlectrodeposited as a highly pusified m. often. made from. the pure sneyl. ‘The crude anodes contain material which either dissolves ‘solution, oF luble and Jas anode slime, ‘The build-up of Impurities in solution is controlled by removing patt of the electrolyte (a bleed) and removing netals by precipitation, or other The anode slimes are gencrally treated to yecower valuable metals such as Ag, Au, Te, PGMs, Below ae two tables which summarise the conditions for electrowinning and electrorefining of ange of the more common metals. Table 1.1 Electrowinning of some metals see} ma a eRSeercatee ae Sel : ‘Lathios cathode electrolyte lepietion iaphragm MP! metal anode. trode ‘| Top! “ cece sienhraom | (cRmGae inTRREREE (ERE Ta | sulphate, — Tadd rage date NUT peel yes cathode | | oH72.85 | compartments PoE] + sulphale | go.79 | | suminum [s3 | SRS [070 no 1 Float | stuminum | 052 [ge ti0z00} 870 | no T ——F stiphate | sats 5 | Co25-50 | | | yes, cathode compartments PeisHiSn Ni i starter shoots | | i “aa | suphats ap | Custatter Cu 25-65 | 10-55 | fopper | POICAIEN sreaisistanless | 8 | “H-S0. | lowforSx | | vanable_| LTable 1.2 Electrorefining of some metals Slee Soe SG 5 f, starting |eathode[etectroiyt ltempfpurification etal | anode joyce /fscrapi| “Sheer leyetel | gpl n ipa etectatye. a) crude ‘metal from. | gy rectecro| 12.14 | siphete {yes, cathode : her] 22 fonge| 8, Joo] ase | puts schemas on » b)pHa | anode bags sulphide imate Siete crue cost rine sn 26-30% in | ude | saz] as loosened} gy |82520%) ny 5.40) no organo 5% EH Eee eeee pataly 3 oe sea |e | gig [vanes [eastectnes|*22005 Mosicae | | ineratr ote veined | cad | Pet | PR for conte bullion punt (Fe | Patt ae 2 2rops | sulphate Tberators for ope (cq | bul | 15:20 |Cussziness] per | “Cuss | no foosegcu contol pu rte | varies anode |H:S0.~200 bleed stream Dore crate perodic wor fetal ex| 45. arapnite oF | «or yagg| AG30-125 | anodes fitaon with Wwor [Manode | | 2025 ened ag |1S°'55) “CS45° | bagged aetvated slimes | HNos <3 carbon aig criorde cis | ee | EGei/es toga Consider a cell composed of these evo half-reaetions as shown in Fig. 2.4. Reduction Cathode Oxidation Anode F c i Figure 2.1 An electrochemical cell fees alas Work done in extemal circuit W=LtV = QV F | For conversion of 1 mole of Cu Q = 2q..N >t fe) + (Faraday 96487 coul/njp}) ju} ny * Wotlk done /mol W= nFV where n = stoichiomebric no. of electrons As R increases, V will increase and in the limit _R_— © , V fpproaches the maximum thermodynamic cell voltage, A 22 Single Electrode Potentials _ In the case of the above cell, a voltage (or potential difference) of ITV Would be me the volmmeter if the solutions contained copper and zine ions of concentrations equivalent to unit activity. The copper electzocle would be positive with respect to the zine electrode ‘The cell voltage can be considered to be composed of a difference between the potential of the copper clectrode and that of the zine electrode, e. AE only possible to measure the potential difference (or voltage) bewween_two rodes. A scale of relative potentials can be devised if all potentials are measured seclative to one electrode. For various reasons this standard reference electrode has been chosen as the Standard Hydrogen Electrode (SHE) 2H +2e =H, for which , at unit activity H* and unit fugacity Ho,“Thus for any redox couple, Ox+ne=Red or M" +ne=M the electrode potential(F) is the potential difference of a cell composed of this couple and a SHE, ic, AE=E-E', and a positive AE(or negative AG) implies thar the reaction Ox+H,> Red + 2H" proceeds spontaneously \ For example, the reaction 2Cu(s) + O,(g) + 4H"(aq) = 2Cu™ (ag) + 2H,0. is made up of the difference of the half reactions 0,(g) + 4H"(aq) + 4c = 2H,0 and Cu’*(aq) + 2e = Cu(s) \with a hypothetical cell valtage(under standard conditions) given by AE=V=Eo)- Eo, = 1.229-0.340 = 0.889V ‘Thus, under these conditions, AG! = -nFAE = -496500%0.889 = 343.2K} ‘The large negative value for AG? implies a spontancous reaction as written, \S" liNore that the electrode potential is an intensive property ( independent of the quantity of material) whereas the free energy is an extensive property. ‘Thus, electrode potentials can simply be added or subtracted to give an overall feaction potential difference, but this is not true of addition or subtraction to give another half-reaction, 7 ‘Table 21 gives a selection of the I” values for some couples of importance in the clectrowinning and refining of the more important metals. Note that the half-reactions are wwsitten as reduction ie. electrons are accepted by the species on the leftTable 2.1 Some relevant standard reduction potentials: Half reaction E%, volts Ma" + 22 = Ma “118 Preferred Zn + 2e = Zn 0.76 Reduction or Au(CN); + = Aut 2CN 061 Cathodic Reaction CoAs + 3H" + 3e= Co + AsH, 0.48, Fe" + 2c = Fe Ca¥ +26 = Cd PbSO, + 2e = Pb + SO? Co" + 2e = Co Nit + 2e = Ni HSbO2 + 6H"+ 6e Ph" + 2e = Pb HAsO, + GH’ + 6¢ = AsH(@ + 3H,0 SbHy() + 2H,O 2H? +22 Hy oe oO se=agt+ 0.222 HAsO, + 3H' + Ni* +5e = NiAs + 3H,O. Cu" + 2e = Cu HSbO, + 3H" + Co” + 5e = CoSb +2H,O HAsO, + 3H" + Co* + 5e = CoAs + 3H,O. 039 Ge te=Cu" 052 Fe te 077 Ag’ +e= Ag 0:80, 0, + 4H" + de = 2,0 1.23 MnO, + 4H" + 2e = Mn™ + 2H¢ 1.33 Cl, + 26 = 2Cr 1.36 Preferred Mn" be 149 Oxidation or PbO, + SO? + 4H" + 2c = PbSO,+ 24,0. 1.70 Anodic Reaction Reverse), The equilibrium potentials for the reduction of the more important metal ions are compared in ~~ Figure 2.2 S44 Thus, the only metal which will be co-deposited with copper is silver because its potential ig, more positive than that for the reduction of copper ions. From a thermodynamic paint of view, we Would sot expect lead, iron or nickel to codeposit with copper. On the other hand, ‘one could expect Cd, Fe, Ni, Co, Pb and Cu to co-deposit with zinc while Pb and Cu should codeposit with nickelOrder of Reducibili w 5 A a2 Sereeesshaiibe tat Figure 2.2. Standard reduction potentials for metal ions. 23 Reference Electrodes. ‘The hydrogen electrode is, in most cases, not convenient for use as a reference electrode and therefore other clectrodes with suitable properties are used, the most common of which is the silver/ silver chloride electrode, AgCl(s) + e= Ag(s) + Clq) which has a potential of 0.199V versus SHE at 25°C with saturated KCl electrolyte, 24 Non - Standard Electrode Potentials Most reactions ate carried our undet conditions in which the activities of the species taking part are not unity and one must be able to, quantitatively account for deviations from the standard states, Fora general redox reaction, aA + bB = cC + dD AG = AG’ + RT In fag. ap" / a,*.a,"} where a’s are the activities of the species. Similatly, because of the relationship between AG and K (AG=-nFE), E=E® -RT/nP, Lag ac’. ay! / ay" ay") s is the well-known Nest equation which can be applied equally to half-reactions, such as Ox + ne= Red ESE? - RT/nF. Ln{ a, /a,,} For a general half-reaction, aA + DB + mH’ + ne , més! x10? cy a 931 on 5.26 a 2.03 so? 1.06 Nat 1.33 Co, Ni*, Zn 072 Pet om Fe" 0.60 Osea), 2.10 23000. Note that most metal ions have diffusivities between 5 x 10" m* s" and 10 x 10 m? s" while the proton and the hydroxide ion have unusually high values for the same reason that theit conductivities are high. ‘The effect of concentration on the diffusion coefficients is not as pronounced as in the case of the conductivity and is therefore often ignored. {The thickness of the diffusion layer (and, of course the nfass transfer coefficient) is strongly dependent on the effectiveness of the convective processes in the bulk of the aqueous phase. Some typical values ate shown in Table 4.2. Table 4.2 Some Approximate Diffusion Layer Thicknesses Surface 8mm) | Gas') Freely suspended 100j1m particle of density 4g/em* | 0.07 30x10" Suspended gas bubble 0.02 | 25x 10% Plane surface with longitudinal flow (25em s') on 5x 10° Rotating cylinder (100 em s") 0.04 12.5 x 10° Natural convection on vertical electrode 0.10.2 | 5-10.x 10° Air-sparged vertical electrode 0.04 12.5 x 10% 43 Mass Transport Correlations We have seen that itis possible to estimate the rate of mass transport of a species to a surface if we know the relevant mi t. Unfortunately, as we could expect, the coefficient will depend strongly on the geometry and flow conditions close to the surface. Thus, unlike diffusion coefficients, tables of mass transfer coefficients do not exist. A number been established which enable one to calculate the Sherwood Number, Sh (the dimensionl face concentration gradient averaged over the mass_tansfer susface) for various surface geometties and operating conditions. ‘These correlations use several s0-c yumbers which are given in Table 4.3. s transfer coeffici 23Table 4.3 Some Dimensionless Groups Number(ox Group) | Physical significance Form Sherwood number Gh) | Mass tansfer/Diffusion iD Schmidt number (Se) | Momentum diffusion/Mass diff. | v/D or p/p.D | Reynolds number (Re) | Incttial forces/Viscous forces. | LU/v } Grashof number (Gi) | Buoyancy forces/Viscous forces | 1'gAp/(p.v) Peclet umber (Pe) Flow/Diffusion velocity uyD ‘The symbols and thei dimensions are : D diffusion coefficient (7/1) g_ acceleration due to gravity (L/t’) k,, mass transfer coefficient (L/0) ‘characteristic length (L) 3 fluid velocity (L/) v kinematic 1p/p_ fractional density change H viscosity MI/L.0) It should be remembered that the exact definition of each of these dimensionless groups implies a specific physical system. ‘Thus, for example, the characteristic length I in the Sherwood number will be the beight from the bottom of a planar electrode for mass ttansfer by natural convection but will be the diameter of a spherical particle which is dissolving Many of these correlations have the same general form. They involve the Sherwood number which contains the mass uansfer coefficient. In the case of diffusion, the Shenvood number vaties with the Schmid sumber. The variation of the Shenvood number with flow is more complex because of the different physical origins of flow ie. forced convection due to pumping or other means and natural convection due to density ot thermal gradients. The former involves the Reynolds number while the latter is described by the Grashof number i) a wection Sh = 0,67 (Sc. Gr for 10' < Se. Gr < 10° (Laminar) Sh = 0,31 (Sc. Gr) for 10 < Sc. Gr < 10" (Turbulent) Sh=k,B/D Gr= g(p,- Pp) B/( pV) where p, and p, are bulk and surface solution densities Sc =p/(p D) where jt = solution viscosity and v= kinematic viscosity (j1/p). ii) jas Evolving Surfaces ‘Sh = 0,93 Re®* Sch” where Re = V,.dp/A. and V, is the volumetric gas evolution rate dis the bubble break-off diameter ( 40m for Hy in alkaline solutions and O, in acid solutions ‘and G0jim for Hy in acid and O, in alkaline solutions) 24ed to In many cases, this can be simpli where m=5-10. 10" A” mis” for H, evolution and iy in Am? if) Forced Convection y For laminar flow as shown between two laxge planar surfaces (height>>gap) Sh = 1,85 { Re. Sc. d./L}"* Sh=k,.d,/D where d, =25 Re=U.d./¥ and, for surfaces of finite height, Sh= 147{ Re. Sc. d,/L}" 2/4) }"? 5/B Flow Sh = 0,023 Re™ Sci” iv) Particulate Beds For a packed bed of spheres with Re < 0,1, Sh = 1,09/e. (Re. Se)" Pot a packed bed of spheres with Re > 0,1, Sh = 1,44 Re Sel”? For a fluidized bed (< 25% expansion), Sh = (1-6)", Re”? Sc!” /e. where, Re= Ud p/y and Se = /(P.D) and U = superficial fluid velocity, bed voidage and d = particle diameter vi) Spinning Disk Sh = 0.62 Re, ScX? where Sh = kd/D, Re = da/v, Sc = v/D, dis the diameter of the disk and @ is the angular velocity of the disk. ‘This relationship is the well known Levich equation for mass transport to a rotating disk, an impo ndependent of the position on the .cteristic of which is the local mass transfer coefficient ck tant cha 44 Influence of Mass Transport on Electrochemical Kinetics The current potential curves discussed above were derived and depicted on the assumption that the rates of the reactions at the electrodes are controlled by the rates of the electrochemical 25SS" processes. Thus, for example, the éathodic curve for the reduction of Cu** ions shown ia Fig. 3.3, suggests that the current increases exponentially with decreasing potential. However, the concentration of Cu” ions at the electrode surface will decrease duc to reduction and at a sufficiently high overpotential, this concentration will decrease to effectively zero. Purther increases in the oveipotential cannot inctease the rate Of reduction which has now become contfolled by mass transport and the curve will take on the shape as shown below in Fig. 4.2 ‘The masimum current that can be passed is known as the limiting cuszent density¢j,) Figure 4.2_Current-potential curve for a cathodic reaction involving mass ‘transport. ‘The equation for the current-potential curve including mass transpott in terms of the limiting Current density j,, ean be shown to be of the following form for a cathodic reaction with a high exchange current density, = - (RY /nE). Inf, ~i)/i} in which 9 vp and Ey. is the so-called bi 4,5. Mass Transport of lons In addition to convection and diffusion, the mass transport of ions in aqueous solutions can be eld and it is necessary to briefly review these influenced by the presence of an electrical additional modes of transport. 4) Conductivity Consider two clectrodes of cross-sectional area A placed a distance | apart in a solution of an electrolyte as shown, 26As in the case of any electrical conductor, the resistance and its reciprocal, the conductance or conductivity, is given by( ys aloo Mecluls 4 fe oF) Resistance : R=pl/A where p= resistivity in ohman. Conductivity : K=1/p (chm or Sm! where Sis the Siemen) Unlike metallic conductors, the conductivity of electrolytes is concentration dependent, and therefore we define, Molar Conductivity A=«/c (Smimot) Generally A. decteases with increasing concentration and, approximately, A =A - Ac? where A? is Limiting Molar Conductivity at infinite dilution and is listed in tables, Transport Number B= 4/E4 where 1, is the current carried by species i and © J, is the total current carried by all ions i ‘This enables one to calculate contributions of single ions to the overall conductivity f= GA / EGA, ‘The conductivity of a dilute solution can be calculated ftom the cation and anion values KEEGY, Jn more comples refining or clectrowinning solutions, several empirical xlationships has been, shown to be reasonably accurate. ‘Thus, * For copper eee Resistivity (Stem) = 1.1 {1.0-7310°(L ~ 55) ~ 3x10°(Cyoy~ 200) + 6.6x10°Cg, + 7.6010°C uy, + 82x10°C, +0, 7x0 °C, ai with 'T being the temperature in °C and C, being the concentrations ingh ‘Tables for some of the more important electrolytes are given in the Appendix. b) Combined Diffusion and Migration ‘The total fux/unit area of an ionic species to the electrode surface is given by the sum of that duc to diffusion and migration, J=D@C/d%),. + ti/aF whertis the ‘uansference numberof the ion in question. ‘The limiting current density therefore given by fi FDC, / (1-98 | For example, in a 1 mol/dm* solution of CuSO, fey = 53.6/(53.6 + 80) = 040 “Thus, the limiting current will be some 1.67 times that for diffusion alone. 27On the other hand, a typical copper electzorefining solution also contains about 1.7 mol/ dm’ H,SO, , and in this ease t,, = 0.19 and the limiting current will be only some 1.23 times that for diffusion alone. Why is acid added if this is detrimental to mass wansport of Cu” ions to the cathode? 5. Mass Transfer at Vertical Electrodes m and narrow cathode In industeial cells with vertical electrodes of length up to 1. node gaps, mass transport to the cathode is an important quantity as it limits the current density which can be applied. ‘Thus, most processes operate at current densities of 1/3. to 1/2 of the limiting current density. The flow of electrolyte in the gap between the electrodes is generally relatively Jow with values of 1 to 10 cm/min being used. 5.1 Natural Convection In the absence of gas evolution at the eathode (euch as in the electrowinning of copper), natural or free-convective flow is generally the most effective means whereby the electroactive ions are wansported from the bulk of the electrolyte to the surface of the cathode, Natural convection is caused by density of temperature differences in the electrolyte which results in a concentration gradient from the cathode. For a vertical cathode, the electrolyte is partially depleted of relevant metal ion at the cathode surface. This results in a lowet density of the electrolyte in the vicinity of the cathode which, in turn, induces vertical flow of electrolyte close to the cathode as shown in Figure 5.1. Obviously, in a refining cell in which a metal is being dissolved at the anode, the electrolyte becomes denser close £0 the anode surface and the low generated by natural convection is in the opposite direction, ie. downwards. In combination with the upward flow at the cathode, this sets up a circulating flow between the electrocles which can be quite effective in enhancing ‘mass transport. jo --Welocity Profiles at Half-Height __ 28 | “ — Wining g 20 4 = | Eis #10 Fo 4 i “10 ° 5 10 1S 202830 Distance from Cathode, mm Figure 5.1. Variation of flow as a function of distance from cathode in copper fefining and winning cells. For constant concentration in the bulk and at the clectrode surface, ie. constant cursent density, the following correlation has been found to describe mass transport due to natural gonvection induced by density differences, 28Sh = C.(Se.Gr)* a constant and the Shenvood number Sh /D, the Schmidt number hof number Gr = {g(,- PIL"} / (p,) where C = g{a(C,-CJL} /v? Se= v/D and the G and k, is the mass transfer coefficient, L the length of the electrode, D the diffusion coefficient of the electroaetive species, v the kinematic viscosity of the electrolyte and @,,0, the density of the electrolyte in the bulk and at the cathode surface respectively. © is the so-called densification factor which relates the den concentration. sity of an electrolyte to the For most electrolytes, Sc>10" and C = 0.67 ) For a simple binary electrolyte such as copper sulfate, the above can be rearranged to give an ple binary PP weed to gi expression for the average mass transfer coefficient over the electroxle, ky, = 119 {a(C,— CJ}! DY!" m/s Note that mass transport is a relatively wenk fonction of al the relevant parameters and that the average value decreases as the length of the electrode increases. The difference C, ~ C, is obviously related to the current density at the electrode, In the case of a misture of electrolytes, such as CuSO, and H,SO,, the situation is somewhat thore complicated because of the presence of the acid. ‘Thus, as we have previously seen, in solutions typical of copper refining or winning, most of the electrical current is cartied by the protons in the electrolyte, ie. the transference number tj, ~ 1 and t, ~ 0. Thus, protons migrate to the cathode but ate not reduced (in the case of copper deposition) and they therefore accumulate in the vicinity of the cathode. At steady-state, the rate of migration to the clectrode is balanced by a diffusion flux in the opposite ditection. ‘This increase in acidity at the cathode is shown is shown in Figure 5.2 ‘This increase in acidity affects the density of the electrolyte at the cathode surface and therefore the density gradient. It tends to reduce the density gradient and thereby also the mass transport of copper ions to the cathode. c cut Distance from Cathode Figure 5.2 Concentration profiles near the cathode in a refining cell 29Under these conditions, the calculation of the mass transfer coefficient is mote complex and only the final result will be shown, expressed as the limiting curtent density, ic for i, = 114835 z (Iete.)” Do", (@/VL)"* A/m? = Hyfty/ = ted} Dea/Dy* where 0,,i8 the value of the densification factor for CuSO, and 64, the value for H,SO, z's the no of electrons involved and & = 5.2 Application to Copper Deposition The above can be used to estimate the limiting current density during clectzowinning ox refining of copper from an electrolyte containing 40 g/! of copper and 150 g/' sulfuric acid in cells containing vertical electrodes of height Im at 60°C . ‘The electrolyte is normally circulated through the cells at such a slow flow rate that the influence of the forced flow on mass transport to the eathode can be neglected with the major component in mass transport beisig natural convec ‘The relevant values of the required parameters are De, = 12x 10" m/s, D,, = 4.3 x 10? m*/s at 60°C 64x10" m/s 0.07 4) = 06, -5 x 10% m'/mol for 630 mol/m’ of CuSO, y= 0, = 5.2.x 10° m'/mol for 1530 mol/m of HSO, from which, = 1.37 x 10" m'/mol and i, = 830 A/m?* ‘Operating plants generally use current densities berween 250 and 350 A/m? which is in the range of about 1/3 to 1/2 of the limiting current density one can calculate the 4 to give In the absence of the supporting electrolyte, ie. the acid in this case, limiting current density us sve equation with 6 = Gand ty 1000 A/m? Thus, although the use of acid in the electrolyte to increase the conductivity is desirable, it has a negative impact on the mass transport of copper ions to the cathode surfaces Gates Berkb $ 4.24) ng the a 5.3 Effect of Gas Evolution Mass transfer to the cathode can be considerably enhanced as a result of the flow generated by the rising oxygen bubbles from the anode. This effect is shown schematically in Figure 5.3(a) and the consequences are that i) mass transfer at the anode is greater than that at the cathode. ‘This has consequences for the speciation of, for example, iron in the electrolyte as we shall see later. ii) mass transport is greater towards the top of both anode and cathode.} a = @ ) Figure 5.3. Effect of gas evolution at the anode (a) and cathode (b) on mass. transport ‘This effect is even more pronounced in the case of hydrogen evolution at the cathode which is applicable in zinc and some extent nickel but not copper electrowinning As shown schematically in Figuce 5.3(b), the rapid flow of electrolyte to the sutface of the cathode when a bubble of hydrogen detaches from the sutface is very effective in increasing mass transport 10 the electrode. ‘The following empirical relationship has been found to apply in this case, kx (cm/s) = 5x 105 in “Thus, for jy in A/m2 Thus, for i = 50.A/m?, which is equivalent to 90% current efficiency at 500A /m* k= 75x 10° cm/s compared to 7 x 10% cm/s for natural convection, On the other hand, evolution of gas results in a finction of the electrolyte volume between the anode and cathode to be taken up by gas bubbles. ‘This decreases the conductivity of the solution, particularly in the upper part of the cell where the entrained gas volume is the greatest. The data in Figure 5.4 shows how this effect can give rise to an uneven vertical cuuent disttibution which is obviously more noticeable at higher current densities | Mass Trane Mass Transfer Coeff, emis | 2 Gas Holdup 200 25030035000 0 02 04 08 os 1| Cument deat, Vind Fraction of Height from Bottom Figure 5.4 Effect of current density on gas hold-up and on the current distribution in a typical Cu electrowinning cell. The ratio ili, is that of the local current density to the mean value for the overall cathode. 34An effective method of enhancing mass tsnsport to vertical cathodes is by sparging. air into the cell below the electrodes. ‘The results in Figure 5.5 demonstrate the inerease in the average mass transfer coefficient and the uniformity of mass transport across the cathode as a result of ouch sparging, i Aaitation by Air Sparging | a8 ce | 3 io tL | 0s vot comectr | | o Seat Air Flow, Lin" dm? Figure 5.5 Effect of air sparging on the average mass transfer coefficient and the uniformity of mass transfer to a vertical cathode. The disadvantages of this are obviously increased complexity of the cells, the potential to increase the concentration of suspended solids and the increased generation of acid mist. 8 Electr The cathodic deposition ofa hydrated metal ion M(H,O),." onto a metal electrode M occurs in several steps as shown schematically in Figure 6.1 ’) Charge transfer to the hydrated metal ion in the double layer with the formation of a neutral or partially charged “ad-atomt” or “ad-ion” which may still be partially hydrated. i) Migration of the ad-atom by surface diffusion to a gsowth site (such as an edge or kink) at whieh further cordination to metal atoms in the lattice occurs, iil) Growth of the crystal latice uccessive transfer and diffusion steps. This sequence of steps, any one of which may be rate-determining, is shown schematically below. lattice is that the c “The essential difference between an aqueous metal ion and a metal ion in & gidination sphere of the former is composed of water molecules while that of the latter is composed of other metal ions i.c. metal - H,O bonds have to be broken and replaced by metal metal bonds. As could be expected, the rate of this process is influenced by a number of factors such as the nature and concentration of the metal ion and the anions present, the presence of inhibitors and the nature of the substrate. In the case of electrocrystallization, an important additional parameter is the potential difference between the solution and the substrate 32@ a Figure 6.1. Steps in the electrocrystallisation process 6.1 _ Influence of Kinetics on Deposit Morphology 4) Charge Transfer ‘The exchange current density, of metal-metal ion electrodes varies benween wide limits depending on the nature of the metal ion and its state in solution. For simple ions, the following table gives some idea of the range of reversibility nent of water molecules by other molecules or anions in the metal co-ordination sphere can either accclerate or inhibit charge transfer e.g. chloride tends to increase the i, for the deposition of the iton-group metals while cyanide can prevent deposition completely. These effects ate complex combinations of thermodynamic and kinetic factors, 33Table 6.1 Classification of Metal Deposition Reactions Normal i (Aem) Inert ] [10 10* 10°- 10° 10°- 10" | Lead) Copper(t) Tron (iD) Ting) Zinc(tl) Cobalt(tt) | Mercury (11) Manganese(I) Nickel({1) Cadmium(!) Chromium(ID) Silver() Platioum(11) Gold) fi) Crystallization ‘The formation and moxphology of electro-deposits is dependent on several factors the most important of which is the relative rates of nucleation and growth. In crystallization in general, the rate of nucleation is a strong fanction of the degree of supersaturation of the solution. In the case of electrocrystallization, the degree of supersaturation is determined in addition by the ‘overpotential which therefore determines the nature of the deposit as follows: Rate of Nucleation o¢ Cone. of ad-atoms o Overpotential A high nucleation rate will create a large number of small crystallites ie. a fine-grained smooth deposit. This is generally observed and, in the extreme, powders are formed as the current ity approaches its limiting value. SEM mictogsaphs of copper nuclei formed on a platinum surface are shown on Figure 6.2. In this case, the overpotential for nucleation on a foreign surface exceeds that for growth of the nuclei and therefore the density of nucle is not great. This has obvious implications in regard to the adhesion of copper deposits to foreign substrates, @ a Figure 6.2 Nucleation of copper on stainless ste! The use of complexed metal ions, such as Ag(CN), also tends to promote the formation of dense fine-grained deposits. Furthermore, growth can be inhibited by the addition of various levelling or brightening agents which absorb preferentially on the crystal defects which are the growth sites. Higher periodic current densities produced in superimposed pulsed or A.C current also act to promote nucleation and produce fince deposits at the same mean cusrent