Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Fundamental aspects of nickel electrowinning from chloride electrolytes Ji, Jinxing 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1994-894024.pdf [ 8.66MB ]
Metadata
JSON: 831-1.0078500.json
JSON-LD: 831-1.0078500-ld.json
RDF/XML (Pretty): 831-1.0078500-rdf.xml
RDF/JSON: 831-1.0078500-rdf.json
Turtle: 831-1.0078500-turtle.txt
N-Triples: 831-1.0078500-rdf-ntriples.txt
Original Record: 831-1.0078500-source.json
Full Text
831-1.0078500-fulltext.txt
Citation
831-1.0078500.ris

Full Text

FUNDAMENTAL ASPECTS OF NICKEL ELECTROWINNING FROM CHLORIDE ELECTROLYTES by JINXING JI B.Eng., Shanghai University of Technology, 1982 M.Eng., Shanghai University of Technology, 1985 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES The Department of Metals and Materials Engineering We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA February 1994 ©JinxingJi, 1994  In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. Ifurther agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives.  It is understood that copying or  publication of this thesis for financial gain shall not be allowed without my written permission.  (Signature):  Department of Metals and Materials Engineering  The University of British Columbia Vancouver, B.C., Canada Date: February 14,  1994  Abstract  ii  Abstract Nickel electrowinning from chloride electrolytes is an innovative and efficient process developed and commercialized mainly by Falconbridge Ltd. Several fundamental aspects related to this process have been addressed in this thesis, including the thermodynamic study of nickel electrolytes, the measurement and modelling of the cathode surface pH during nickel electrowinning and the kinetic study of nickel reduction and hydrogen evolution. The major apparatus and equipment used include a surface pH measuring device, an EG&G rotating disc electrode, a SOLARTRON 1286 Electrochemical Interface and a RADIOMETER titrator system. All of the experiments were carried out via computer control. The thermodynamic study includes the activity coefficient of the hydrogen ion and the spe ciation of nickel electrolytes to obtain a better understanding of the properties of nickel electrolytes. The activity coefficient of the hydrogen ion (y +) was measured using a combination glass pH 11 electrode. It was found that was greater than 1 in concentrated NiCl 2 solutions and increased significantly with increasing NiCl 2 concentration. The addition of NaC1 increases ‘y ÷ whereas the 11 ,  addition of 4 SO decreases it. Theoretically, several useful equations were derived based on 2 Na Meissner’ sand Stokes-Robinson’s theories to calculate the single-ion activity coefficients including These equations are the two-parameter (q and h) functions, capable of predicting with rea sonable accuracy single-ion activity coefficients in any concentrated pure electrolytes and in mixed electrolytes of the type 1:1 + 1:1, 2:1 + 1:1 and 2:1 + 1:1 + 1:1. The accuracy of the calculations may be further improved when the Meissner parameter q is adjusted properly and the effect of ionic strength on the hydration parameter h is taken into account. A series of speciation diagrams for nickel species was plotted with and the effect of the ionic strength on the equilibrium constants being taken into account. It was discovered that the predominant nickel species in the acidic region are Ni 2 and NiCl in concentrated pure NiC1 2 solutions and Ni , NiCl and NiSO 2 4 in concentrated sulfate-containing NiCl 2 solutions. The traditionally accepted electroactive species NiOH is negligible until the NiCl 2 concentration is lowered to the order of 106 M. When the pH increases, the formation of insoluble Ni(OH)$) should be expected if the NiCl 2 concentration is higher than 106 M. The pH where Ni(OH)S) starts to form decreases with increasing NiCl 2 concentration and temperature. A limited number of electrowinning tests were carried out under conditions similar to those employed in the industrial process in order to obtain information concerning the current efficiency of nickel deposition. It was found that higher nickel concentration, higher pH and the addition of NaCl, 3 B0 and NH H C1 improved the current efficiency of nickel deposition. However, the addition 4  Abstract  iii  of sulfate decreased the current efficiency of nickel. In 0.937 M NiC1 2 at 60°C, the pH may go as low as 1.5 for a current efficiency above 96 %. Nickel deposition was also found to be a steady-state process since the amount of acid added to the electrolyte at a constant pH increased linearly with time. To acquire data on the cathode pH behaviour during nickel deposition, the cathode surface pH was measured using a flat-bottom combination glass pH electrode and a fine mesh gold gauze as cathode. Nickel was deposited on the front side of the gold gauze and the pH electrode was positioned in the back and in direct contact with the nickel-plated gold gauze. The cathode surface pH was always found to be higher than the pH in the bulk electrolyte, and if the current density was suf ficiently large, it would eventually reach a level causing precipitation of insoluble Ni(OH)S) on the cathode surface. Lower bulk pH, higher nickel concentration, higher temperature and the addition of 3 C1 effectively depress the rise of the cathode surface pH. Additions of NaCl and 4 B0 and NH H 4 S 2 Na O also depress the rise of the cathode surface pH but to a much smaller degree. Also, agitation of the electrolyte decreases the cathode surface pH. In order to predict the cathode surface pH, mathematical modelling in the case of 0.937 M NiCl 2 and 2 M NiCl 2 was carried out. The model was in reasonably good agreement with the experimental data. Nickel deposition and hydrogen evolution were studied using a rotating disc electrode. The hydrogen evolution was found to be affected strongly by the RPM. The rate of nickel deposition was first order with respect to the activity of nickel ion and zero order with respect to the activities of chloride and hydrogen ions. The rate of hydrogen evolution was found to be first order with respect to the activity of hydrogen ion and to be zero order with respect to the activities of nickel and chloride ions. These findings indicate that nickel deposition and hydrogen evolution proceed independently. The Tafel slopes obtained from the partial polarization curves were 94 mV/decade for nickel deposition and 112 mV/decade for hydrogen evolution. Hydrogen evolution was also studied using a rotating nickel-coated Pt disc electrode in 2.5 M NaCl solution in the absence of nickel ions. The rate of hydrogen evolution was first order with respect to the activity of hydrogen ion and zero order with respect to the activity of chloride ion. According to the relationship between the limiting current density and the square root of rotational speed, hydrogen evolution was mass transfer controlled under the limiting conditions and the buffering actions of 3 B0 and NH H C1 were negligible. The magnitude of the limiting current 4 density at a given pH or a given acidity in the presence of sulfate can be well explained considering the activity coefficient of the hydrogen ion. Further studies of nickel electrowinning should be directed towards hydrogen evolution on the nickel substrate in nickel-containing electrolytes, focusing on the hydrogen bubble’s nucleation,  Abstract  iv  growth, coalescence and detachment. The use of addition agents affecting hydrogen evolution by way of adsorption, change in interfaciai tension or destruction of atomic hydrogen is worth investigating. The identity of intermediate species during nickel reduction is not clear. The identification of these species would be quite rewarding in clarifying the mechanism of nickel reduction. The nucleation of nickel and crystal growth in the initial stages of deposition on various substrates including titanium, stainless steel, copper and nickel are other important aspects of nickel electrowinning which should be investigated.  Table of Contents  v  Table of Contents  Abstract Table of Contents  v  List of Tables  ix  List of Figures  xii  Acknowledgements  xix  Nomenclature  xx  Introduction  1  Chapter 1 Literature Review on Nickel Electrodeposition  7  1.1 Nickel matte chlorine leaching process  7  1.2 Plant practice of nickel electrowinning  10  1.3 Nickel electrodeposition in chloride and chloride-sulfate electrolytes  15  1.4 Kinetics and mechanism of nickel electrodeposition  20  1.5 Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media  26  Chapter 2 Thermodynamics of Nickel Chloride Solutions 2.1 Activity coefficients in multicomponent nickel chloride solutions  34 34  2.1.1 Measurement of activity coefficient of hydrogen ion  35  2.1.2 Calculation of mean activity coefficients and water activity  47  2.1.3 Calculation of single-ion activity coefficients  55  2.1.3.1 Single-ion activity coefficients in aqueous solutions ofpure electrolytes  56  2.1.3.2 Single-ion activity coefficients in aqueous solution of mixed NiCI -HC12 NaCZ  61  2.2 The pH for the formation of insoluble nickel hydroxide  67  2.3 Distribution of nickel species in aqueous solutions as a function of pH  72  Table of Contents  vi  Chapter 3 Electrodeposition of Nickel in Various Electrolytes  84  3.1 Experimental apparatus and set-up for nickel electrodeposition  84  3.2 Electrodeposition of nickel at 25°C  85  3.3 Electrodeposition of nickel at 60°C  87  3.4 Electrodeposition of nickel in 2 M NiC1 2 + 6 M HC1  95  3.5 Measurement of current efficiency of nickel from the acid volume  98  Chapter 4 Surface pH Measurement during Nickel Electrodeposition  103  4.1 Experimental apparatus and set-up for surface pH measurement  104  4.2 Characterization of gold gauze  106  4.2.1 Electrochemical properties of gold in chloride solution  106  4.2.2 Investigation of new 500-mesh gold gauze  107  4.2.3 Investigation of nickel-coated 500-mesh gold gauze  111  4.3 Effect of nickel concentration on the surface pH in pure NiCl 2 solutions at 25°C  116  4.4 Effect of sulfate on the surface pH in 2 4 NiCl S Na O solutions at 25°C  118  4.5 Effect of sodium chloride on the surface pH in NiCl -NaCl solution at 25°C 2  ..  119  4.6 Effect of boric acid on the surface pH in 3 -H 2 NiCl B 0 solution at 25°C  120  4.7 Effect of ammonium chloride on the surface pH in 4 -NH 2 NiCl C 1 solution at 25°C  126  4.8 Effect of temperature on the surface pH in pure nickel chloride solution  127  4.9 Effect of ultrasound on the surface pH  132  4.10 Surface pH measurements at 60°C  133  Chapter 5 Modelling of Surface pH during Nickel Electrodeposition  135  5.1 Modelling of surface pH for the solution 2 -HC1-H NiCl 0  136  5.2 Modelling of surface pH in 0.937 M NiC1 2 at bulk pH 2.5 and 25°C  142  5.3 Modelling of surface pH in 2 M NiCl 2 at bulk pH 2.5 and 25°C  149  Table of Contents  Chapter 6 Rotating Disc Electrode Study of Nickel Electrodeposition  vii 152  6.1 Fundamentals of the rotating disc electrode technique  152  6.2 Experimental apparatus, procedures and conditions for the RDE tests  156  6.3 Reaction orders of the rates of nickel reduction and hydrogen evolution with respect to the concentrations of electrolyte components  160  6.4 Effect of RPM on the hydrogen evolution and electrode potential during nickel electrodeposition  164  6.5 Polarization curves of nickel reduction and hydrogen evolution  167  6.6 Nickel electrowinning at high current density  177  6.7 Hydrogen evolution on the nickel cathode in electrolytes without NiC1 2  178  6.8 Probable mechanisms for nickel electroreduction and hydrogen evolution  187  Chapter 7 Conclusions  193  Chapter 8 Recommendations for Further Work  195  Bibliography  196  Appendix 1 Correction for Liquid Junction Potential in the pH Determination 206 Appendix 2 Single-ion Activity Coefficients in Pure Electrolytes  211  Appendix 3 Single-ion Activity Coefficients in Mixed Chloride Electrolytes  214  Appendix 4 Computer Programs for the RADIOMETER Titrator  221  (1) pH titration  221  (2) RED OX titration  224  (3) pH-stat tests  226  Appendix 5 Computer programs for the SOLARTRON 1286 Electrochemi cal Interface  228  (1) Recovery of lost experimental data from the SOLARTRON’s data file  229  (2) Galvanostatic experiments  229  (3) Potentiostatic experiments  231  (4) Linear potential sweep experiments  232  Table of Contents  viii  (5) Cyclic voltammetry experiments  233  (6) Galvanostatic anodic dissolution  235  (7) Potentiostatic anodic dissolution  236  Appendix 6 Computer program for the SOLARTRON 1286 Electrochemical Interface together with the RADIOMETER titrator  239  Biographical Data  242  List of Tables  ix  List of Tables  -  Table 1  Reactions taking place during nickel electrowinning  10  Table 2  Operating conditions for direct nickel matte electrowinning  12  Table 3  Operating conditions for electrowinning from nickel sulfate electrolyte  13  Table 4  Operating conditions for electrowinning from nickel chloride electrolyte  14  Table 5  Temperature coefficients of the overpotentials of nickel cathodic deposition and anodic dissolution in 1 M NiC1 2 and 1 M NiSO 4 at pH 1.5  25  Table 6  Properties of pH responsive electrodes  37  Table 7  Activity coefficients of hydrogen ion in aqueous solutions of pure and sulfatecontaining nickel chloride in the pH range 1-4 at 25,40 and 60°C  38  Equilibrium quotients for the reaction SO + H = HSO at 25°C based on equation (90)  45  Activity coefficients of hydrogen ion in aqueous solutions of sulfate-containing nickel chloride in the pH range 1-4 at 25°C  45  Table 8  Table 9  Table 10 Characteristic parameter q for pure electrolytes at 25°C Table 11  49  Mean activity coefficient of NiCl 2 and activity of water in aqueous solutions of nickel chloride at 25°C  51  Table 12 Mean activity coefficient of NiSO 4 and activity of water in aqueous solutions of nickel sulfate at 25°C  52  Table 13 Mean activity coefficient of HC1 in aqueous solutions of hydrochloric acid at 25°C  53  Table 14 Mean activity coefficient of HC1 in mixed aqueous solutions of NiC1 -HC1 at 25°C 2  54  Table 15 Activity of water in mixed aqueous solutions of NiCl -HC1 at 25°C 2  54  Table 16 Parameters for Stokes-Robinson’s hydration theory equation  56  Table 17 Activity coefficients of hydrogen and chloride ions in aqueous solutions of HC1-NaC1 at 25°C  65  Table 18 Comparison between calculated and experimental activity coefficients of hydro gen ion in aqueous solution of NiCl -NaCl-HC1 at 25, 40 and 60°C 2  65  Table 19 Comparison of activity coefficient of hydrogen ion in electrolytes of sodium chloride and calcium chloride at 25°C  66  List of Tables  x  Table 20 Dissociation quotient of water at 25°C  69  Table 21  Equilibrium quotients of nickel hydrolysis at 25°C  69  Table 22 The pH’s for the formation of Ni(OH)S) in different solutions  71  Table 23 Equilibrium quotients in solutions of pure nickel chloride at 25°C  75  Table 24 Equilibrium quotients in solutions of mixed nickel chloride and sulfate at 25°C  75  Table 25 Current efficiencies of nickel deposition in 3 -H 2 NiC1 B 0 and 4 -NH 2 NiCl C 1 at pH 2.5 and 25°C (two hours for each run)  87  Table 26 Current efficiencies of nickel deposition in various solutions at 60°C and pH 1.1  88  Table 27 Current efficiencies of nickel deposition in various solutions at 60°C and pH 1.5  89  Table 28 Current efficiencies of nickel deposition in various solutions at 60°C and pH 2  89  Table 29 Current efficiencies of nickel deposition in various solutions at 60°C and pH 2.5  89  Table 30 Current efficiency of nickel deposition in 2 M NiCl 2 + 6 M HC1 at 25 and 60°C  96  Table 31  96  Current efficiency of nickel deposition in 2 M NiC1 2 + 6 M HC1 at 95°C  Table 32 Calculated activity coefficients, activities and electrode potential shifts in 2 M 2 + 6 M HC1 at 25, 60 and 95°C NiC1  97  Table 33 Errors in current efficiency due to ±0.01 pH shift in 200 mL 0.937 M NiCl 2 at 2 A/m 300 (0.09 A) for 2 hours  102  Table 34 Errors in current efficiency due to ±0.01 pH shift in 200 mL 0.572 M NiCl 2 0.365 M NiSO 4 at 300 A/m 2 (0.09 A) for 2 hours  102  +  Table 35 Dimensions of gold gauzes  105  Table 36 The coefficients of the 8 x 8 multilinear equations for the surface pH modelling of the aqueous solution of 2 -HC1-H NiC1 0  141  Table 37 Density and viscosity of aqueous solutions of NiC1 2 + HC1 at 25°C  144  Table 38 Diffusion coefficients in 0.937 M NiCl 2 at 25°C  145  Table 39 Equilibrium quotients in 0.937 M NiC1 2 at 25°C  146  Table 40 Diffusion coefficients in 2 M NiCl 2 at 25°C  150  Table 41  150  Equilibrium quotients in 2 M NiC1 2 at 25°C  Table 42 Tafel slopes determined from the partial polarization curves  176  List of Tables  xi  Table 43 Calculated Tafel slope and reaction order for the rate of nickel reduction when the effect of the coverage of the cathode with the adsorbed nickel species is taken into account  188  Table 44 Calculated Tafel slope and reaction order for the rate of the reduction of nickel ions for different mechanisms  189  Table 45 Tafel slope and reaction order for the rate of hydrogen evolution with respect to the concentration of hydrogen ion  192  Table 46 Liquid junction potentials and the corresponding pH shifts in nickel chloride solutions at 25°C  209  Table 47 Liquid junction potentials and the corresponding pH shifts in mixed sulfate containing nickel chloride solutions at 25°C  210  List of Figures  Xii  List of Figures  Figure 1  The electrolyte conductivity and viscosity of 2 M (NiC1 2  Figure 2  Flowsheet of the Falconbridge nickel matte chlorine leaching process  Figure 3  The electrolyte conductivity and viscosity of NiC1 2 solutions at various tem peratures  16  The partial current density vs. potential for the deposition and dissolution of nickel on a platinum RDE at 1,000 rpm (1 M NiC1 2 -2 M NaCl 0.01 M HC1 at 21°C)  21  The overpotentials of nickel cathodic deposition and anodic dissolution in 1 M 2 and 1 M NiSO NiC1 4 at 100 Aim 2 and pH 1.5  25  Concentration of hydrogen ion as a function of its activity in nickel chloride solutions  40  Concentrations of hydrogen plus bisulfate ions as a function of hydrogen ion activity in sulfate-containing nickel chloride solutions at 25°C  42  Figure 8  Sub-section distribution curve of nickel species in 3.92 M NiC1 2 at 25°C  42  Figure 9  Concentration of hydrogen ion as a function of its activity in sulfate-containing nickel chloride solutions at 25°C  46  The activity of water in aqueous solutions of nickel chloride as a function of ionic strength (I = 2 mNC1 3 )  53  Figure 4  +  ) at 60°C 4 NiSO  4 8  -  Figure 5 Figure 6 Figure 7  Figure 10  Figure 11  Calculated activity of nickel ion as a function of its concentration at different temperatures  Figure 12  Concentration of hydrogen ion as a function of its activity in aqueous solutions of sodium chloride and calcium chloride at 25°C (HC1 added continuously)  66  Dependence of the pH of nickel hydroxide formation on the nickel concentration and temperature in nickel chloride solutions  68  Figure 14  Distribution curves of nickel species in nickel chloride solutions at 25°C  76  Figure 15  Distribution curves of nickel species in sulfate-containing nickel chloride solutions at 25°C  80  pH titration curve of dilute solution of nickel chloride (13.6 g/L 2 •6H NiC1 0 , 150 mL sample, 25°C and 2 mL/min. speed)  82  Figure 13  Figure 16  59  List of Figures  xiii  Figure 17  Schematic drawing of the apparatus for nickel electrodeposition tests  84  Figure 18  SEM photomicrograph of the cross-section of nickel deposit obtained from 0.937 M NiC1 2 at 750 Aim , bulk pH 2.5 and 60°C 2  90  The potential of nickel electrode as a function of time in 0.937 M NiC1 2 at 750 A/m , bulk pH 2.5 and 60°C 2  91  Sub-section potential and nature ofnickel cathode as a function of time in 0.937 M 2 at 750 A/m NiCI , bulk pH 2.5 and 60°C 2  91  The potential of nickel electrode as a function of time in 0.937 M NiC1 2 at 1,000 Aim , bulk pH 2 and 60°C 2  92  SEM photomicrograph of the cross-section of nickel deposit obtained from 0.937 M NiCl 2 at 1,000 2 A/m bulk pH 2 and 60°C ,  93  SEM photomicrograph of the morphology in the black zone of nickel deposit obtained from 0.937 M NiC1 2 at 1,000 AIm , bulk pH 2 and 60°C 2  93  Figure 19 Figure 20 Figure 21 Figure 22 Figure 23 Figure 24  The acid volume added to the electrolyte as a function of time during nickel electrodeposition from 0.937 M NiCl 2 (55 g/L Ni ) and 0.572 M NiC1 2 2 0.365 M 4 (55 gIL Ni NiSO 2 and 35 g/L SO) at 300 AIm , different pH’s and tempera 2 -  tures  100  Schematic drawing of the apparatus for surface pH measurement and associated equipment  104  Figure 26  SEM photomicrograph of 500-mesh gold gauze (dull side) (20 kV, 500X)  108  Figure 27  SEM photomicrograph of 500-mesh gold gauze (shiny side) (20kV, 500X)  108  Figure 28  SEM photomicrograph of 500-mesh gold gauze (dull side) (20 kV, 2,000 X)  109  Figure 29  SEM photomicrograph of 500-mesh gold gauze (shiny side) (20kV, 2,000 X)  109  Figure 30  SEM photomicrograph of 500-mesh gold gauze (cross-section) (20 kV, 4,000 X)  110  Figure 31  Schematic drawing of the 500-mesh gold gauze  111  Figure 32  EDX diagrams of 500-mesh gold gauze coated with nickel layer of different thicknesses (0.05-1 jim) and after anodic dissolution (20 kV, 7,000 X)  112  SEM photomicrograph of 500-mesh gold gauze coated with -0.05 jim (nominal) thick nickel film (20 kV, 2,000 X) (morphology)  114  SEM photomicrograph of 500-mesh gold gauze coated with --0.5 jim (nominal) thick nickel film (20 kV, 2,000 X) (morphology)  114  SEM photomicrograph of 500-mesh gold gauze coated with —1 pm (nominal) thick nickel film (20 kV, 2,000 X) (morphology)  115  Figure 25  Figure 33 Figure 34 Figure 35  List of Figures Figure 36 Figure 37 Figure 38 Figure 39 Figure 40 Figure 41  Figure 42 Figure 43 Figure 44  Figure 45  Figure 46 Figure 47 Figure 48 Figure 49 Figure 50 Figure 51  xiv  SEM photomicrograph of 500-mesh gold gauze coated with -1 Im (nominal) thick nickel film (20 kV, 2,000 X) (cross-section)  115  Surface pH as a function of time at 50 A/m 2 (500-mesh gold gauze, 0.937 M , bulk pH 2.5, 25°C) 2 NiCl  116  pH titration curves for different NiCl 2 concentrations at 25°C (150 mL sample and 0.5 mLJmin speed)  117  The surface pH as a function of current density for different NiCI 2 concentrations at 25°C (500-mesh gold gauze and bulk pH 2.5)  117  The surface pH as a function of current density for different sulfate concentrations at 25°C (500-mesh gold gauze and bulk pH 2.5)  118  pH titration curves for different sulfate concentrations at 25 and 60°C (0.937 M NiC1 0.937 M NiC1 , 2 2 + 0.365 M 4 SO 0.572 M NiCl 2 Na , 2 + 0.365 M NiSO 4 and 0.572 M NiCl 2 + 0.365 M NiSO 4 + 0.365 M 4 SO 150 mL sample and 2 Na , 0.5 mLlmin speed)  118  pH titration curve for 0.937 M NiC1 2 2 M NaCl at 25°C (150 mL sample and 0.5 mLfmin speed)  120  The surface pH as a function of current density in 0.937 M NiCl 2 -2 M NaCl at 25°C (500-mesh gold gauze and bulk pH 2.5)  120  Distribution curves of boric acid species in aqueous solutions containing 5 and 40 g/L 3 B0 at 25°C H  122  -  Distribution curves of boric acid species at 25°C (considering 3 B0 and B(OH) H only)  123  pH titration curve for free boric acid at 25°C (0.485 M 3 B0 30 mL sample, H , and 0.5 mL/min speed)  123  Volume difference of 5 M NaOH between 2nd and 1st peaks as a function of boric acid concentration at 25°C (30 mL sample)  123  pH titration curve for the boric acid in 2 M NaC1 0.485 M 3 B0 at 25°C (30 mL H sample, and 0.5 mLJmin speed)  124  Volume difference of 5 M NaOH between 2nd and 1st peaks as a function of boric acid concentration in solutions containing 2 M NaC1 (30 mL sample)  124  pH titration curve for 0.937 M NiCl 2 0.485 M 3 B0 at 25°C (150 mL sample H and 0.5 mL/min speed)  125  The surface pH as a function of current density in 0.937 M NiCl 2 0.485 M 3 B0 H at 25°C (500-mesh gold gauze and bulk pH 2.5)  125  -  -  -  List of Figures Figure 52 Figure 53 Figure 54 Figure 55 Figure 56 Figure 57  Figure 58 Figure 59  Figure 60  Figure 61  Figure 62 Figure 63 Figure 64 Figure 65 Figure 66  XV  pH titration curve for the free ammonium chloride solution at 25°C (1.31 M C1, 30 mL sample, and 0.5 mlJmin speed) 4 NfT  126  pH titration curve for 0.937 M NiC1 2 1.31 M NH C1 at 25°C (150 mL sample 4 and 0.5 mlJmin speed)  126  The surface pH as a function of current density in 0.937 M NiC1 2 1.31 M NH C1 4 at 25°C (500-mesh gold gauze and bulk pH 2.5)  127  pH titration curves for 0.937 M NiC1 2 at different temperatures (150 mL sample and 0.5 mL/min speed)  127  The surface pH as a function of current density in 0.937 M NiC1 2 at different temperatures (500-mesh gold gauze and bulk pH 2.5)  127  The potential of nickel electrode vs. time in deaerated 0.937 M NiCl 2 at bulk pH 2.5 and 60°C (50 Nm , with N 2 2 bubbling and under agitation except where marked)  129  The potential of nickel electrode vs. time in non-deaerated 0.937 M NiCl 2 at bulk pH 2.5 and 60°C (50 Nm , under agitation except where marked) 2  129  The potential of nickel electrode vs. time in 0.937 M NiCl 2 at bulk pH 2.5 and 60°C (50 Nm , under agitation except where marked, with prior air bubbling for 2 10 minutes)  129  The potential of nickel electrode vs. time in the non-deaerated 0.937 M NiCl 2 at bulk pH 2 and 25°C (50 A/m , without prior deaeration, under agitation except 2 where marked)  129  The potential of nickel electrode vs. time in deaerated 0.937 M NiC1 2 at bulk pH 2 and 25°C (50 Nm , with 10 minutes prior N 2 2 bubbling, under agitation except where marked)  130  The potential of nickel electrode vs. time in 0.937 M NiCl 2 at bulk pH 2 and 25°C , with 10 minutes air bubbling, under agitation except where marked) 2 (50 Nm  130  The effect of ultrasound on the surface pH in 0.937 M NiCl 2 at bulk pH 2.5,25°C and c.d. 80 and 180 Nm 2  132  The surface pH as a function of current density at 60°C without agitation in various electrolytes (500-mesh gold gauze)  133  pH titration curves for highly concentrated solutions at 60°C (3.92 M NiC1 2 and 3.555 M NiCl 2 + 0.365 M NiSO , 150 mL sample and 0.5 mLfmin speed) 4  133  The potential of nickel electrode vs. time in deaerated 3.92 M NiCl 2 at bulk pH 2 and 60°C (50 Nm , with 10 minutes prior N 2 2 bubbling and under agitation except where marked)  134  -  -  List of Figures  xvi  Figure 67  Interactions between species in the solution 2 -HC1-H NiCl 0  136  Figure 68  Definition of X-coordinate for the surface pH modelling  136  Figure 69  The viscosity and density of aqueous NiC1 -HC1 solution at 25°C (dashed lines 2 contain no HC1, solid lines contain 0.1 M HCI, density times a factor of i0 3 kglm , 3 absolute viscosity times i0 kg/msec, kinematic viscosity times 106 m /sec) 2  143  Figure 70  Modelled surface pH in 0.937 M NiC1 2 at bulk pH 2.5 and 25°C  148  Figure 71  Sub-section distribution curve of nickel species in 0.937 M NiC1 2 at 25°C  149  Figure 72  Modelled surface pH in 2 M NiCl 2 at bulk pH 2.5 and 25°C  151  Figure 73  Schematic drawing of the rotating disc electrode  152  Figure 74  Dimensions of the surface of the rotating disc electrode  156  Figure 75  Schematic drawing of the apparatus for the rotating disc electrode study  156  Figure 76  The effect of ohmic drop on the polarization curve (0.937 M NiC1 , pH 2, 25°C, 2 1,000 rpm, 5 mV/sec and bare Pt)  158  The current density vs. time for potentiostatic operation (0.3 M NiCl 2 , 0.005 M HC1 <pH 0.90>, 25°C, 2,000 rpm, Ni-coated Pt) 2 CaCl  159  ...  Figure 77 Figure 78 Figure 79  Figure 80  Figure 81  Figure 82 Figure 83 Figure 84  +  2.7 M  The current density vs. time for linear potentiostatic anodic dissolution (0.937 M 2 + 0.485 M 3 NiC1 B0 pH 2, 25°C and 2,000 rpm) H ,  160  The current densities of nickel reduction and hydrogen evolution as a function of nickel concentration (NiC1 2 + CaC1 2 = 3 M, pH 1.1, 25°C, 2,000 rpm and Ni-coated Pt disc)  161  The current densities of nickel reduction and hydrogen evolution as a function ofHC1 concentration (0.3 M NiC1 2 + 2.7 M CaC1 , 25°C, 2,000 rpm and Ni-coated 2 Pt disc)  161  The current densities of nickel reduction and hydrogen evolution as a function of chloride concentration [0.5 M Ni(C10 2 + 3 M (NaC1 + NaC1O ) 4 ) + 0.005 M 4 HC1, 25°C, 2,000 rpm and Ni-coated Pt disc]  162  The effect of rotational speed on the current efficiency in various electrolytes and at different pH’s (25°C, -0.850 volt vs. SCE and Ni-coated Pt disc)  165  The effect of rotational speed on the electrode potential in electrolytes of pure nickel chloride (started with Pt substrate at 25°C)  166  Polarization curve at a sweep rate of 2 mV/sec (0.3 M NiC1 2 + 2.7 M CaCl , 2 0.005 M HC1 <pH —0.9 >, 25°C, 2,000 rpm and Ni-coated Pt disc)  168  List of Figures Figure 85 Figure 86 Figure 87 Figure 88 Figure 89 Figure 90  Figure 91  Figure 92  Figure 93  Figure 94  Figure 95  Figure 96 Figure 97 Figure 98  Figure 99  xvii  Polarization curves of combined nickel reduction and hydrogen evolution in different electrolytes (2,000 rpm, pH 2, 25°C, 2 mV/sec, Ni-coated Pt disc)  168  Polarization curves of combined nickel reduction and hydrogen evolution in 0.937 M NiC1 2 at different pH’s (2,000 rpm, 25°C, 2 mV/sec, Ni-coated Pt disc)  172  Current efficiency of nickel over the potential range covering the whole polarization curve (0.937 M NiC1 , pH 2, 25°C, 2,000 rpm, Ni-coated Pt disc) 2  173  Partial polarization curves of nickel reduction and hydrogen evolution in 0.937 M 2 at pH 2, 25°C and 2,000 rpm (Ni-coated Pt disc) NiC1  174  Tafel plots of the partial polarization for nickel reduction and hydrogen evolution 2 at pH 2, 25°C and 2,000 rpm (Ni-coated Pt disc) in 0.937 M NiCl  176  Polarization curves for hydrogen evolution on Ni-coated Pt electrode in 2.5 M NaCl, 2.5 M NaCl + 0.365 M 4 SO and 2.5 M NaC1 + 0.485 M 3 2 Na B0 at H different RPM’s (25°C, 2 mV/sec. —2 pm Ni-coated Pt disc)  179  Limiting current density for hydrogen evolution as a function of the square root ofRPM in electrolyte containing no nickel ions at different pH’s (25°C and —2 pm Ni-co ated Pt disc)  181  Reaction order for the rate of hydrogen evolution with respect to hydrogen ion concentration in the electrolytes containing no nickel ions (25°C, 2,000 rpm and —2 pm Ni-coated Pt disc)  182  The limiting current density for hydrogen evolution in different electrolytes vs. the concentration of hydrogenion (25°C, 2,000rpm and —2 pm Ni-coated Pt disc)  183  The slope of (L vs. ,IRPM) as a function of hydrogen ion activity in different electrolytes (25°C and —2 pm Ni-coated Pt disc)  183  The current density of hydrogen evolution as a function of chloride concentration in 3 M (NaC1 + NaC1O ) + 0.01 M HC1 (25°C, 2,000 rpm and —2 p.m Ni-coated 4 Pt disc)  184  Tafel plot of hydrogen evolution in 2.5 M NaCl at pH 2, 25°C and 2,000 rpm (2 mV/sec and —2 p.m Ni-coated Pt disc)  184  Polarization curves of hydrogen evolution in electrolytes without nickel ions at different pH’s (25°C, 2,000 rpm, 2 mV/sec and 2 pm Ni-coated Pt disc)  185  Polarization curves of the hydrogen evolution in electrolytes without nickel ions at different acid concentrations (25°C, 2,000 rpm, 2 mV/sec and —2 pm Ni-coated Pt disc)  186  The possible routes for hydrogen evolution  190  Figure 100 Separated view of a combination glass pH electrode  206  List of Figures  xviii  Figure 101 The equivalent conductivities of electrolytes (KC1, NaC1, NiCl , Na 2 4 S 2 O and )at25°C 4 NiSO  206  Figure 102 In-situ screen output during pH titration  222  Figure 103 dpHJdV vs. volume for pH titration  222  Figure 104 pH vs. volume for pH titration  223  Figure 105 dpHJdV vs. pH for pH titration  223  Figure 106 In-situ screen output during REDOX titration  224  Figure 107 dPOTENTIALIdV vs. volume for REDOX titration  225  Figure 108 POTENTIAL vs. volume for REDOX titration  225  Figure 109 dPOTENTIAL/dV vs. POTENTIAL for REDOX titration  226  Figure 110 In-situ screen output for pH-stat test  226  Figure 111 Volume vs. time for pH-stat test  227  Figure 112 In-situ screen output for reading data from SOLARTRON’s data file  229  Figure 113 Potential vs. time for galvanostatic experiment  230  Figure 114 Potential vs. time for galvanostatic experiment  230  Figure 115 In-situ screen output for potentiostatic experiment  231  Figure 116 Current density vs. time for potentiostatic experiment  232  Figure 117 In-situ screen output for linear potential sweep  232  Figure 118 Current density vs. potential for linear potential sweep  233  Figure 119 In-situ screen output for cyclic voltammetry  234  Figure 120 Current density vs. potential for cyclic voltammetry  235  Figure 121 In-situ screen output for galvanostatic anodic dissolution  235  Figure 122 Potential vs. time for galvanostatic anodic dissolution  236  Figure 123 In-situ screen output for potentiostatic anodic dissolution  237  Figure 124 Current density vs. time for potentiostatic anodic dissolution  237  Figure 125 In-situ screen output for galvanostatic electrolysis with pH measurement  240  Figure 126 Surface pH vs. time for galvanostatic electrolysis with pH measurement  240  Figure 127 Potential vs. time for galvanostatic electrolysis with pH measurement  241  Acknowledgements  xix  Acknowledgements I would like to express my sincere gratitude to Dr. W. Charles Cooper for his thoughtful supervision and numerous constructive discussions in the progress of this thesis work, for his conscientious reviewing and editing of this thesis, and for his fatherly care of my life since I worked for him.  I would like also to express my appreciation to Dr. David B. Dreisinger and Dr. Ernest Peters for their acceptance of myself to study in this department, for all the conveniences and instruments provided for my experiments, and for their consistent interest in and enlightening advice to my thesis work. Many thanks are due to my fellow graduate students and research engineers in the hydrome tallurgy laboratory for their great assistance in various ways in the years of working together. Special thanks are extended to Ms. Mary Mager for her technical help in using the scanning electron microscope and to the technicians in the machine shop and the secretaries in the department for their support. The generous financial support of this research provided by Falconbridge Limited is greatly appreciated. In addition, I wish to thank the personnel in the Metallurgy Technology Centre at Falconbridge Limited for their many valuable comments during project review meetings.  Nomenclature  Nomenclature ROMAN SYMBOLS AND ABBREVIATIONS a  activity of an ion (cation or anion) or compound  a,..  activity of water an ion-size-related parameter,  (A)  A  constant of Debye-Huckel equation for the activity coefficient, which is equal to 0.509 (molefkg)” 2 for water at 25°C in equation (93); predominant species  abs  absorbed  AC  alternating current  A.C.S.  American Chemical Society  ads  adsorbed  aq  aquated; aqueous medium; solvated; hydrated  b  the bulk of electrolyte  B  coefficient of ion-size term of Debye-Huckel equation for the activity coefficient, which is equal to 0.329 x 1010 m (molefkg)’ for water at 25°C 1 in equation (93); B = 0.75 0.065q in Meissner’s equation (99) -  C  C = 1 + 0.055 q exp(-0.0231 ) in Meissner’s equation (100); molarity, i.e., 3 moles of solute per litre of solution, (molefL)  Cb  molar concentration in the bulk of electrolyte, (moleIL)  C  molar concentration at the electrode surface, (molelL)  Calcd.  calculated  c.d. or C.D.  current density, (Aim ) 2  CE  current efficiency, (%)  d  electrode gap, (m); wire diameter of gold gauze, (jim)  D  diffusion coefficient, 2 /(m sec)  D÷  cation diffusion coefficient, 2 /(m sec)  Nomenclature  xxi  D  anion diffusion coefficient, 2 /(m sec)  Dsait  diffusion coefficient of a salt, 2 l(m sec)  DC  direct current  DH  Debye-HUckel  Diff.  difference  DSA  dimensionally stable anode, often made of the titanium substrate with a noble metal oxide coating, such as Ti-Ru0 2 for chlorine evolution  e  charge of an electron  E  electrode potential, (volt) corrosion or mixed potential, (volt) standard electrode potential, (volt)  EDX  energy dispersive X-ray  EMF  electromotive force  eq  equilibrium  EW  electrowinning  Exptl.  experimental  f  rational (or mole-fraction scale) activity coefficient  f±  mean-ion rational activity coefficient  F  Faraday constant, (96,500 C/equiv.)  FRP  fiberglass reinforced polyester  g  gaseous phase; standard acceleration, (9.81 mlsec ) 2  AG  Gibbs free energy, (kJ/mole)  AG°  standard Gibbs free energy, (U/mole)  h  hydration parameter; electrode height, (m)  [Itj  molar concentration of the hydrogen ion, (mole/L)  Q 2 H  hydroquinone 6 (HOC O 4 H H)  i  current density, 2 (A/rn ) ; species i  4orr.  corrosion current density, (Aim ) 2  Nomenclature  xxii  d t  diffusion current density, (Aim ) 2  L 1  limiting current density, (A/rn ) 2  0 i  exchange current density, (A/rn ) 2  I  current, (A); ionic strength, I  IHP  inner Helmholtz plane  j  speciesj  J  flux of matter, 2 (kmo1Im s ec)  k  Boltzmann’s constant, (1.3807 x 1023 J/°K); rate constant of a reaction; Sievert’s law constant  =  -  0.5 m z 1  rate constant of a backward reaction kf  rate constant of a forward reaction  K  thermodynamic equilibrium constant  K,  solubility product of an insoluble compound  K,,,  ionization constant of water  L  electrode length, (m)  m  molality, i.e., moles of solute per kilogram of water, 2 (molelkg•H 0 )  M  molecular weight, (g/mole); unit for molar concentration, (moleIL); metal  n  the number of electrons transferred; the number of total species in the solution  N  mole fraction of the solvent  ] 2 [Nj  molar concentration of the nickel ion, (molefL)  OHP  outer Helmholtz plane  Ox  oxidant  P  pressure, (atm)  PGM  platinum group metals  pH  negative logarithm to base 10 of the activity of hydrogen ion, pH  ppm  parts per million  =  —  log(a+)  Nomenclature  xxiii  PRC  periodic reverse current  PTFE  polytetrafluoroethylene  PVC  polyvinyl chloride  Q  equilibrium quotient; the number of coulombs, (C); benzoquinone 0) 4 H 6 (0C  Q  solubility quotient of an insoluble compound ionization quotient of water,  Q = [H9.[OH1  r  the radial coordinate in the poiar coordinate system  r  radius of an ion, (m)  R  molar gas constant, 8.3 14 J/mole•°K; regression coefficient (IRI  R,  ohmic resistance, (2)  RDE  rotating disc electrode  r.d.s.  rate-determining step  Re  Reynolds number, Re  Red  reductant  rpm, or RPM  revolution per minute, RPM =60 cx I (2it)  RRDE  rotating ring-disc electrode  s  space between wires of gold gauze, (urn); electrode surface; solid  Sc  Schmidt number, Sc = v I D  SCE  saturated calomel electrode  SEM  scanning electron microscopy  SHE  standard hydrogen electrode  S.S.  stainless steel  t  transference number (t  T  absolute temperature, (°K)  TBP  tributyl phosphate, P (CH [CH 0 3 ) 2 (O) ]  TIOA  triisooctylamine, ) 2 ( [CH N 3 ] 7 CH  V  volume, (mL); flow velocity of the solution, (mlsec)  1)  rIV w• 2  1); electrolysis time, (mm.); temperature, (°C);  Nomenclature water  w  x  xxiv  -  distance from the electrode surface, (m); unknown molar concentration of a species, (molelL)  X  mole fraction of the solute; ligand; anion  y  molar activity coefficient mean-ion molar activity coefficient  z  charge on an ionic species; the vertical distance from the disc surface (m)  z  charge on a cation  z  charge on an anion  GREEK SYMBOLS cathodic charge transfer coefficient (x anodic charge transfer coefficient (f3  1) 1); stability constant  ö  thickness of the diffusion layer, (m)  &ff  effective thickness of the diffusion layer of a binary electrolyte thickness of hydrodynamic boundary layer, ö 0  TI  =  3.6’J  overpotential, (volt) molal activity coefficient mean-ion molal activity coefficient logf*  1 q 7 —O.510 ( 1 l + cfij in Meissner’s equation (101)  logf’+ = logy+I I z z_ I in Meissner’s equation (97) K  conductivity, (mho/m) equivalent conductivity, (mho•m /equiv.) 2  Ii  mobility, (m /sec•volt); viscosity, (kglm.sec) 2  v  kinematic viscosity of electrolyte, v  =  iJp, (m /sec) 2  the number of moles of cation per mole of solute the number of moles of anion per mole of solute  Nomenclature 12 V  xxv =V++V_ 1 V 2 angular velocity of the disc, (radian/sec) azimuthal coordinate of the poiar coordination system osmotic coefficient, for aqueous solution, electrical potential of the solution, (volt)  0  —1000 lna/(l8 v ); m 1  Wi  electrical potential at the outer Helmholtz plane with respect to the bulk solution, (volt)  p  density, (kg/rn ) 3  0  coverage of electrode surface with adsorbed species, (0  1)  Introduction  1  Introduction Nickel is used primarily in the production of metal alloys. The production of stainless steel accounts for 64 % of total nickel consumption, and nickel-based and copper-based alloys account for another 12  %[1]•  Other important applications include a base deposit for chrome plating, powders  for coinage and catalysts, and oxides for anodes in rechargeable batteries. According to the 1988 statistics, 175,000 tons of nickel were produced electrolytically, amounting to 31 % of the total nickel production in the non-Communist wor1d . The major electrolytic nickel producers are INCO 21 Ltd. in Canada, Falconbridge Ltd. in Norway, Sumitomo Metal Mining Co. in Japan, Outokumpu in Finland, Jinchuan Non-Ferrous Metals Corp. in China, Rustenburg Base Metals Refmers in South Africa, and Société Le Nickel in France. Three electrolyte systems are in use today in the nickel industry. Direct nickel matte elec trowinning in mixed chloride-sulfate electrolyte, whose overall cell reaction is Ni 2 = 3Ni+ 2S, is S 3 in operation at INCO’s Thompson, Manitoba nickel refmery, Sumitomo’s Niihama nickel refinery and Jinchuan’ s nickel refinery. Nickel sulfate electrowinning, whose overall cell reaction is 2NiSO 4 + 2H 0 = 2Ni + 4 2 S0 + 02, is practiced at Outokumpu’s nickel refinery and by Rustenburg Base 2 2H Metals Refiners. Nickel chloride electrowinning, the most innovative and efficient electrolytic process, whose overall cell reaction is NiCl 2 = Ni + Cl , is employed by Falconbridge’s Nikkelverk 2 A/S. Sumitomo Metal Mining Co. and Société Le Nickel. The copper content of the nickel matte affects the process of choice. Nickel mattes with a lower copper content (< 7 %) are usually suitable for direct matte electrowinning, pressure ammonia leaching or 3 -FeCl leaching. Nickel mattes with a high copper content (>25 %) are often pro 2 C1 cessed by atmospheric or 4 S0 pressure leaching, or 2 2 H -CuCl leaching Cl . Except for direct matte 31 electrowinning, the nickel matte is first subjected to leaching. Following the leaching or matte electrowinning, the electrolyte must be purified of impurities including copper (Cu ), cobalt (Co), 2 lead (Pb ), arsenic (AsO), iron (Fe 2 ), manganese (Mn 2 ) before being pumped to the cathode 2 compartments in the electrowinning tankhouse. The steps involved in each process, that is, leaching, purification, hydrogen reduction and electrowinning of nickel were recently well reviewed respectively by Conard , Kerfoot and Weir 21 , Burkin 131 41 and Hofirek and Kerfoot . 51 The desired electrowon nickel should be a dense coherent deposit with a smooth surface to minimize occlusion of electrolyte. This goal is not difficult to achieve even in the absence of any addition agents due to the high overpotential of nickel itself. As a result of its low exchange current density, nickel reduction is normally under activation control and the resulting cathode deposit consists of fine grains. An undesirable feature ofnickel electrowinning is the inevitable simultaneous  Introduction  2  hydrogen evolution. Hydrogen evolution during nickel electrowinning is favored both from the thermodynamic and kinetic points of view. Whether nickel reduction or hydrogen evolution takes precedence in the cathodic discharge depends entirely on their respective electrode potentials as expressed by equations (1)-(2). +  =  =  where:  -  (T —298) +  —0.257 + 0.93 x 4 10 ( T —298)  jjj- (T 4  =  E+,H,c) +  =  0+0.90 x 3 1W ( T  =  0.90 x 3 l(T ( T  —  —  —  298) +  298)+  in  + JR 1 +1 lNi  00591T log aN.2+ + JR 1 + flNj + 2 x 298  -  (1)  in aH+ + JR 1 +r  298  1 flH loga++JR 2  (2)  O.OS 1 TH 298) 91 +JR 2 ThI  R is the ohmic resistance between the reference and working electrodes, (2). I is the cathodic current, (A). lNi 1 TIH2  is the total overpotential of nickel deposition, (volt) is the total overpotential of hydrogen evolution, (volt)  Thermodynamically, in order to make nickel reduction become the leading cathodic reaction, the pH should be greater than 0.257/0.059 1  =  4.3 under conditions of unit activity of the nickel ion  and 25°C. Such a high pH level is hard to maintain in practice without pH buffers. The use of pH buffers in electrowinning cells may not be welcome on account of the complexity of the chemistry and the possibility of contamination in the leaching, purification and solvent extraction circuits. The major reason why a pH higher than 4.3 is not practical in nickel electrowinning is the possible formation ofinsoluble colloidal nickel hydroxide Ni(OH)) on the cathode surface. Once Ni(OH)S) is formed on the cathode surface, the nickel reduction will be hindered greatly and a coherent high quality nickel deposit can no longer be obtained. The structure, surface appearance and properties of cathode nickel will all be affected significantly by the formation of insoluble Ni(OH)S). As the nickel ion concentration and temperature increase, the pH at which insoluble Ni(OH)S) forms will decrease further. As can be seen from equations (l)-(2), the temperature affects the equilibrium electrode potentials very little. The temperature, however, affects the overpotential of nickel reduction significantly.  Introduction  3  Kinetically, nickel reduction is not favored either in view of the high overpotential of nickel and the low overpotential of hydrogen evolution on the nickel substrate. In spite of the fact that hydrogen evolution can be alleviated to a less significant extent by optimizing the operating con ditions, the hydrogen evolution is unlikely to be surmountable completely during nickel electro winning. The factors influencing both nickel reduction and hydrogen evolution are the pH and temperature of the electrolyte, nickel ion concentration, current density, agitation, electrolyte composition, addition agents and the nature of the cathode surface. Lower electrolyte pH in nickel electrowinning is not feasible. At low pH, copious hydrogen gas will evolve on the cathode, consuming energy unnecessarily and hence raising the operating costs. The other disadvantage of the abundant hydrogen evolution is the absorption of hydrogen into the cathode nickel which causes harmful stress in the nickel. The only advantage of hydrogen evolution is the enhancement of the mass transport near the cathode. Theoretically, only in sulfate electrowinning will the pH of the electrolyte decrease as the electrowinning proceeds due to the anodic generation of acid. This is why a diaphragm cell is always used in nickel sulfate electrowinning to control the catholyte pH. Another reason for using a diaphragm in nickel sulfate electrowinning is to avoid contamination from the scaling of the deposit from the lead anodes. In nickel matte and nickel chloride electrowinning, however, dia phragm cells are used for the purpose of collecting anode slimes and chlorine gas, respectively. Nickel matte chlorine leaching and nickel electrowinning in chloride electrolytes which were mainly developed and commercialized by Falconbridge in Norway represent the future direction of nickel hydrometallurgy. In chloride solutions, it is likely that the nickel ion forms a nickel chloro complex (NiCl). Contrary to most of the complex species where complexation causes a substantial negative shift in the equilibrium potential and cathodic overpotential, the formation of a nickel chioro complex actually promotes the reduction of nickel ions. One argument as regards the function of chloride ions is that the promotion of nickel reduction comes from the interfacial phenomenon represented in particular by “chloride ion bridge” theory rather than from the bulk solution chemistry . This 6 argument might be true since it has been found in this thesis work that the reaction orders of the rates of nickel reduction and hydrogen evolution with respect to the chloride activity are both equal to zero when the chloride concentration is above 0.4 M. The advantages of using chloride electrolytes compared with sulfate electrolytes can be summarized as higher electrical conductivity and lower viscosity (Figure 1) of the electrolyte, lower cathodic nickel and anodic chlorine overpotentials, higher solubiity of NiC1 2 than NiSO , higher 4 activity coefficient of nickel ion and easier nickel-cobalt (Ni-Co) separation by solvent extraction in the preceding purification stage. These properties result in a higher current efficiency of nickel reduction, lower cell voltage and thus a lower energy consumption. The lower JR drop across the  4  Introduction  electrolyte means that less heat is generated in the cells. This lower heat generation together with 2 suggests the possibility of using high current density nickel electro the higher solubiity of NiC1 winning. Still another advantage is realized in the chlorine leaching stage where the leaching rate 2 is also favorable from the viewpoint of of nickel matte is high. The higher solubility of NiC1 leaching. The nickel matte chlorine leaching and electrowinning processes avoid the extensive anode handling encountered in matte electrowinning and the contamination by anode scale and the shortened life expectancy of the lead anodes. Past problems such as the corrosive nature of acidic 2.8  1.7  2.6  1.6 C’)  C’)  jj 2.4  1.5th ‘C  1.4  2.2  E  1.3  1.6  1.1  1.4  1.0  0’  0  0.9  °1.2 1.0  0  0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  1.8  0.8  8  chloride electrolytes and the toxicity of chlorine gas no longer exist. Advances in materials engineering have made available suitable materials for the construction of the cell and the anodes. Anodic chlorine gas is recovered completely in the tankhouse and recycled to the leaching stage. Figure 1 The electrolyte conductivity and viscosity 1 ) at 60C 4 2 + NiSO of 2 M (NiCl  NI2, (M)  Although nickel electrowinning from nickel chloride electrolytes has been in commercial application since 1980, many fundamental aspects remain to be understood. As far as hydrogen evolution is concerned, very little definitive knowledge is available in the literature. One popular 21 and the nickel hydroxy belief is that hydrogen evolution results from the decomposition of water . As to 81 2 exist as a buffer in the stabilization of the electrolyte pH complexes including Ni(OH) the reduction of nickel ions, many contradictory mechanisms prevail in the literature. Different conclusions reached by different authors can in most cases be traced to the different experimental conditions, such as electrolyte composition, electrode material and JR drop compensation. Basic questions such as how the pH is related to the acidity of the electrolyte, how to estimate the activity coefficient of hydrogen ions in the concentrated electrolytes, what is the magnitude of the liquid 19 in junction potential need to be considered. These questions were recently explored by Peters some detail for the highly acidic and concentrated nickel chloride electrolytes. One important point ) on the pH 4 which has not been mentioned heretofore is the effect of sodium perchiorate (NaC1O and electrode potential readings. Sodium perchlorate is an inert electrolyte which is often used in maintaining an electrolyte of constant ionic strength. Certain precautions, such as using a double liquid junction, have to be exercised in using sodium perchlorate at a high concentration level.  Introduction  5  The use of high current density electrowinning of nickel in chloride electrolytes is an attractive alternative. Before going far in this direction, a more reliable understanding ofthe physical chemistry and electrochemistry of the process is needed. The operating conditions for achieving both an acceptable current efficiency and a high quality of cathode nickel need to be determined. The present thesis addresses a number of the above-mentioned problems and focuses on the fundamental aspects of nickel electrowinning in chloride electrolytes. Thermodynamically, a series of distri bution curves of nickel species is plotted using the available equilibrium constants (quotients) with the effects of the activity coefficients and ionic strength being taken into account. The activity coefficients of hydrogen ions, nickel ions and chloride ions have been studied theoretically. The activity coefficient of hydrogen ion has been measured experimentally using a combination glass pH electrode, even though such measurements are not perfectly rigid from the viewpoint of ther modynamics. The error in such measurements has been addressed with respect to the liquid junction potential. The applied experiments embrace simple nickel electrodeposition under various conditions, the measurement of cathodic surface pH during nickel electrodeposition and the study of electrode kinetics using a rotating disc electrode. As to the electrodeposition tests, pure nickel instead of a dimensionally stable anode (DSA) has been used as the anode in order to simplify the cell con struction and test procedures. These experiments are selective and not extensive, considering the successful industrial operating conditions and the work done by other 121 investigators’° The . measurement of cathode surface pH has been carried out using a self-designed apparatus mainly at  25CC for electrolytes of major importance. An effort has been made to understand the change in the cathode surface pH and to model it mathematically for the electrolytes with a simpler compo sition. In the study of electrode kinetics in the presence of concentration polarization, the rotating disc electrode is the electrode of choice, as a uniform and known diffusion layer near the working electrode surface can be established with satisfactory precision. The reaction orders of the rates of nickel reduction and hydrogen evolution have been determined with respect to the three most important electrolyte components, i.e., nickel ion activity, chloride ion activity and pH. In addition, a series of polarization curves has been constructed using the technique of linear potential sweep and attempts have been made to separate the combined polarization curves into partial polarization curves for some important electrolytes. Although not exactly the same as in nickel electrowinning, hydrogen evolution has been studied using a nickel-coated platinum substrate in sodium chloride solution without nickel chloride. The results of the above-mentioned studies should permit the establishment of some guidelines for choosing electrolyte compositions and operating conditions for nickel electrowinning from chloride solutions. They should also assist in understanding the various phenomena involved in  Introduction  6  the physical chemistry of nickel chloride electrolytes and in the electrochemistry of cathode nickel reduction and hydrogen evolution. Finally they should provide directions for further studies and for the improvement of current industrial operations.  Nickel matte chlorine leaching process  7  Chapter 1 Literature Review on Nickel Electrodeposition 1.1 Nickel matte chlorine leaching process The nickel matte chlorine leaching process was developed and commercialized primariiy by Falconbridge Ltd. in Norway ’ ‘. It was studied on a laboratory scale between 1966 and 1969, 10 and was tested further on a pilot-scale between 1970 and 1972. Industrial scale chlorine leaching was started in 1975 and by 1981 the changeover to the new chlorine process including the purification stage was completed. However, improvements continued to be effected until 1987. The nickel-copper matte, comprised mainly of Ni , Cu 2 S 3 S and a Ni-Cu alloy with a ratio of 2 7Ni: 3Cu contained 40-45 % Ni, 25-30 % Cu, 20-22 % S, 2-3 % Fe and 1.0-1.5 % Co. The leaching proceeded in two stages. In the first stage of leaching, the matte was leached in CuCI 2 in the presence of Cl 2 gas. The selective dissolution of nickel was made possible by controlling the redox potential of the slurry at a predetermined value through an appropriate ratio of matte to chlorine. It was found advantageous to keep the liquid:solid ratio low enough in order to have a highly concentrated NiCl 2 solution and to make cuprous ions soluble in the leach slurry. The temperature was controlled at around 110°C, the boiling point of the slurry at atmospheric pressure to take advantage of the agitation effect of boiling. The boiling temperature was also beneficial for concentrating the slurry, as each tonne of chlorine afided would produce one tonne of steam. The leaching was quite fast and the heat generated was adequate to maintain the slurry at the boil. The chlorine reacted almost completely during the leach. The sulfur contained in the matte was transformed mainly to elemental sulfur with less than 1 % oxidized to sulfate. The principal chemical reactions taking place during the leaching were: 2Cu(I) 2 S 3 Ni NiS  +  +  2 Cl  =  2Cu(II)  +  2C1  2NiS  +  Ni(II)  2Cu(II)  2Cu(II)  +  =  Ni(II)  +  S  +  (3) +  2Cu(I)  2Cu(I)  (4) (5)  Cu S 2 +S=2CuS  (6)  The solution resulting from the first stage leaching contained about 200 g/L Ni 2 and 50-70 g/L . The second stage leaching was aimed to precipitate Cu 2 Cu 2 as CuS by adding fresh matte. The redox potential was also controlled. The major chemical reactions were: 2 S 3 Ni Ni  +  +  S  S +  +  2Cu(I)  2Cu(I)  =  Cu S 2 +S=2CuS  =  2NiS  Ni(II)  +  +  Ni(II)  S 2 Cu  +  S 2 Cu  (7) (8) (9)  Nickel matte chlorine leaching process  8  CHLORINE LEACH PLANT  NICKEL FOR SALE  Pb-PURIFICATIQ_N rALArMENt PLANT  REFINERY  COBALT FOR SALE  ROASTING H5O PLANT  PROD. PLANT -  COPPER FOR SALE  -  ,,  PM-RECOVERY PLANTS  PM- MATTE -FURNACE  PU- MATTE-LEACHING  —3 0’  Figure 2 Flowsheet of the Falconbridge nickel matte chlorine leaching process 1101  Nickel matte chlorine leaching process S  +  2Cu(I)  =  Cu(II)  +  CuS  9 (10)  To reduce the input of matte, the process was modified later on. Instead of adding more fresh matte, the slurry from the first stage of chlorine leaching was processed in an autoclave at 140-145°C. At such a high temperature, the following two reactions would occur: NiS  +  2Cu(I)  =  Ni(II)  +  S 2 Cu  Cu S 2 +S=2CuS  (11) (12)  The solution from the second-stage leach contained around 230 g/L Ni 2 and 0.2 g/L Cu . The 2 leach residue was filtered and washed. The filter cake, containing 15 % Ni and 50 % Cu, was then roasted in a fluidized bed furnace to transform the sulfides to oxides. The calcine was subsequently leached selectively in the spent copper electrolyte 4 S0 + CuSO 2 (H ) with a 90 % copper recovery 4 and resulting in a solution containing 95 g/L 4 S0 and 50 g/L Cu 2 H . 24 The residue from the calcine leaching stage, containing 55 % Ni, 18 % Cu and all the PGM metals, was subjected to dilute HC1 (20-30 g/L) leaching at 95°C. Most of the nickel and copper could be leached Out and the dissolution of the PGM metals could be minimized by adding a small amount of matte during leaching. The filtrate, after the removal of its iron via precipitation, was pumped to the first-stage chlorine leach. The resulting residue contained primarily PGM metals. The purification process consisted of three stages, (1) precipitation of Fe and As, (2) solvent extraction of Co and other minor elements, and (3) precipitation of Pb and the remaining impurities. In the first-stage of purification, the very concentrated pregnant solution from the chlorine leach stage was neutralized using NICO 3 under oxidizing conditions of Cl 2 gas to precipitate Fe as Fe(OH) 3 and As as arsenate. In the second-stage of purification, triisooctylarnine (TIOA) (15 vol. % in an aromatic solvent) was used as the extractant for cobalt removal, reducing the Co concentration from 5 g/L to 1 mg/L. The resulting raffmate contained 230 g/L Ni , <0.001 g/L Co 24 , 0.15 g/L Pb 2 , 24 fl 4 g/L HC1 and 0.01 g/L organic. 4 0.15 g/L 2 The organic was then removed by passing the solution through an activated carbon column. To remove lead (Pb ) and manganese (hi 24 ) the solution was diluted to about 85 g/L Ni 2 24 using the anolyte from the tankhouse. The Pb 24 and Mn 24 were removed as precipitates formed by using NiCO 3 under the atmosphere of Cl 2 gas. After this stage of purification, the solution contained only < 0.02 mg/L Pb 24 and < 0.05 mg/L Mn . The other trace impurities, Co, Fe, Cu and As, were 2 also further removed. The purified NiC1 2 solution was then pumped to the electrowinning tankhouse. Nickel matte chlorine leaching was also investigated separately by Société Le Nickel (SLN) at Le Havre-Sandouville in France . SLN switched to the chlorine leaching of nickel matte fol 112 lowed by electrowinning from the chloride solution in 1980 after a tragic fire. The nickel matte,  Plant practice of nickel electrowinning  10  imported from SLN’s facility in Doniambo, New Caledonia, containing 75 % Ni and small amounts of S, Fe, Co and virtually no Cu, was leached in a ferric chloride solution in the presence of chlorine gas. The ferric chloride came from the iron removal stage. The resulting pregnant solution after leaching contained NiC1 2 (200 g/L Ni ), CoC1 2 2 (5 g/L Co 3 (10 g/L Fe) and elemental ), FeC! 2 sulfur. After filtering the slurry to remove the residues, the solution went to the purification stage where ferric ions were separated by solvent extraction using iributyl phosphate (TBP) as the extractant. The FeCI 3 recovered from scrubbing the ferric loaded organic phase with fresh water was recycled in part to the chlorine leach stage, and the rest was sold after being concentrated by evaporation. The removal of cobalt from the nickel chloride solution was achieved using TIOA solvent extraction. The cobalt-free nickel chloride solution was subsequently subjected to selective electrolysis to remove the small amount of lead and then passed through an activated carbon colunm to remove any remaining impurities including trace organics. Finally, the purified nickel chloride solution was pumped to the electrowinning tankhouse. 1.2 Plant practice of nickel electrowinning Commercial cathode nickel is produced by three distinct processes, viz, direct nickel matte electrowinning in mixed chloride-sulfate electrolyte, electrowinning from nickel sulfate electrolyte and electrowinning from nickel chloride electrolyte. The major reactions which take place in each process are listed in Table 1. Table 1 Reactions taking place during nickel electrowinning Process  Matte EW 2 S 3 Ni  Anode reactions  =  Ni Cu  Cathode reactions  2 3Ni =  =  N?  4 EW NiSO  2S  +  2 Ni  +  2e  2 Cu  +  2e  2e  =  Ni  +  +  2 2H+2e—H  Desired cell reaction  2 S 3 Ni  =  3Ni  +  2S  6e 2H/)  =  4Lt  2 Ni  +  0.35  + °2 +  2e  =  4e  Ni  2 2H+2e=H 4 2N1SO  +  0 2 2H  4 S 2 2H 0 E:eu at25°C, (volt)  2 EW NiC1  =  2N1  2Ct  =  2 Cl  +  2e  2 Ni  +  2e  =  Ni  2 + 4 2H 2e=H +  2 N1CI  =  Ni  +  2 Cl  + °2  1.48  1.61  : The other impurities As, Co and Fe will also be dissolved The cell voltage across the anode and cathode is composed of several terms. As expressed in equation (13), these terms are the equilibrium cell voltage, ohmic drops across the electrolyte and the contact zones, and the anodic and cathodic overpotentials.  Plant practice of nickel electrowinning  ,  Ece=Ece+  dE RT (T—298)+—ln dl nF  11  ,.  ....  reactart  )  (13)  As to the direct nickel matte electrowinning, the matte contains mostly Ni 2 and a significant S 3 amount ofNi-Cu alloy. The impurity content especially copper in the matte should be low. However, 2-3 % copper content was found to be beneficial as the resulting anode slime was porous and thus the voltage drop across this slime was decreased. The electrolyte is basically a mixed chloride-sulfate in the presence of boric acid. Among the three processes listed in Table 1, the (absolute) equilibrium cell voltage is the smallest for direct matte electrowinning. However, there is large voltage drop across the anode slime, and this voltage drop increases as the anode slime becomes thicker. During matte electrowinning, a small percentage of the anodic current is wasted in dissolving some impurities and in oxidizing sulfur and water to sulfate and oxygen gas. Thus more nickel is deposited on the cathode than is dissolved at the anode. As a result, it is necessary to replenish the electrolyte continuously. This is done by leaching a portion of ground anode residue in an air-agitated reactor. The Ni 2 in the anode residue remains practically unleached S 3 : 21 2N1 + 2H 4 S 2 0 +02= 2NiSO 4 + 2H 0 2  (14)  The major disadvantages with direct nickel matte electrowinning are the high-grade nickel matte required, extensive handling of matte anodes and residual anodes, high residual anode, large voltage drop across the voluminous sulfur anode slime, and extensive purification of impurities. Direct nickel matte elecirowinning is currently in operation at INCO’s Thompson Nickel Refinery in Manitoba, 16 Canada at Sumitomo’s Niihama Nickel Refinery in Japan 91 and at the Refinery ’ 17 of Jinchuan Non-Ferrous Metals Corp. in China° . Their operating conditions are listed in Table 2. 1 ,  Nickel electrowinning from pure nickel sulfate electrolyte is being practised at Outokumpu’s nickel refinery in Finland 1211 and at Rustenburg Base Metals Refiners Ltd. in South Africa . 1 Their operating conditions are listed in Table 3. In nickel sulfate electrowinning, since an equivalent amount of sulfuric acid is generated in the anode compartment, the anode compartment must be separated from the cathode compartment. This is usually done by using a cloth diaphragm and circulating electrolyte from the cathode compartments into the anode compartments. The acid rich spent anolyte is recycled to the leach stage. The advantages of the sulfate electrolyte are the insignificant corrosion and the inexpensive lead or lead alloy which serves as the anode. The most efficient nickel electrowinning is carried out in chloride electrolyte. The process was developed and commercialized mainly by Falconbridge Ltd. in Norway. Falconbridge Nik kelverk A/S in Kristiansand-S in Norway , Sumitomo Metal Mining (SMM) in Japan 101 ’ and 1 Société Le Nickel (SLN) at Le Havre-Sandouville in France 112 are using this process. Their operating  Plant practice of nickel electrowin fling  12  Table 2 Operating conditions for direct nickel matte electrowinning Company  61 INCO’  9 ’ 7 Sumitomo’  Jinchuan°  Production, tlyear  45,000  22,000  24,000  Anode matte, (%) No. of cells  73 Ni, 2.5-3.0 Cu, 0.8 72.3 Ni, 4.9 Cu, 20.8 68 Ni, 6 Cu, 1.8 Fe, 1.0 [31 Co, 0.6 Fe, 0.2 As & Co and, 23 S 20 S  608  /  207  precast concrete  concrete  concrete  FRP  FRP  plastic  636 LJh  /  /  0.9 x 1.6 x 5.8  /  1.15 x 1.45 x 7.43  Anode dimensions, (m)  1.1 x 0.7 x 0.063  0.97 x 0.77 x 0.05  0.8 x 0.37 x 0.05  Cathode dimensions, (m)  1.0 x 0.7 x 0.013  1 cm thick  0.88 x 0.86  woven polypropylene  /  I  modacrylic cloth  Tetron membrane  /  wooden (spruce) box  /  /  Anode spacing, (cm)  21  15  /  Anodes/cell  27  39  74  Cathodes/cell  26  38  36  Anode cycle, (day)  15  20  /  Cathode cycle, (day)  10  10  4  Residual anode, (%)  25  /  25  Mother blank  S.S.  316L S.S.  Ti  Initial anode weight, (kg)  238  220  /  Final cathode weight, (kg)  88.5  /  /  75 Ni , 51 C1, 2 28 Na, 8 3 B0 H ,  70-80 Ni , 80 Cl; 2 40 Na, 8 3 B0 H ,  75 Ni , 70 Cl, 35 Nat, 2 B0 & balance SO H 63  120 SO  120 SO  9,000-10,000  /  13,500  240  200  240  -.3-4  /  /  50  /  /  96  /  3-6  /  / 3.5  351  /  /  Cell construction Cell liner  Electrolyte flow/cell Cell interior dimensions, (m)  Anode diaphragm Cathode diaphragm Cathode frame  Catholyte, (g/L)  Current per cell, (A) C.D., (A/rn ) 2  pH Temperature, (‘C) CE, (%) Cell voltage, (volt) Energy consumption, (kwh/kg-Ni)  Plant practice of nickel electrowinning  13  Table 3 Operating conditions for electrowinning from nickel sulfate electrolyte Company  21 Outokumpu’  Rustenburg  18,000  20,000  126  152  I  precast concrete  PVC  GRP Atlac 4010(6 mm thick)  /  5501db  1.2 x 1.22 x 6.6  1.15 x 1.17 x 6.56  Anode dimensions, (m)  8 cm thick  I  Cathode dimensions, (m)  0.97 x 0.89  /  Pb  Pb-Sr-Sn  polyester cloth  /  Cathode diaphragm  /  woven terylene  Cathode frame  /  wooden (Oregon pine) box  Cathode spacing, (cm)  13  16  Anodes/cell  49  41  Cathodes/cell  48  40  Anode cycle, (year)  5-6  /  Cathode cycle, (day)  7  6  Starter sheet cycle, (day)  2  2  acid-proof steel (AISI 316)  Ti  75  /  Catholyte, (gIL)  BO H 4 1 97 Ni, no 3  , 120 4 2 80 Ni SO & 2 Na B0 6H 3  Spent anolyte, (gIL)  70 Ni, 45 4 4 S 2 H O 1  , 50 4 2 50 Ni S0 2 H  Current per cell, (A)  20,000  14,000-15,000  ) 2 C.D., (A/rn  200-230  205-230  pH  3.5  3.5  Temperature, (°C)  60  60-65  CE, (%)  96-97  96-98  Cell voltage, (volt)  3.6 3•71  3.6-3.9  Production, 1/year No. of cells Cell construction Cell liner Electrolyte flow/cell Cell interior dimensions, (m)  Anode Anode diaphragm  Mother blank Final cathode weight, (kg)  Energy consumption, (kwh/kg-Ni) §:  30 cells were devoted to the preparation of starter sheets.  /  Plant practice of nickel electrowinning  14  Table 4 Operating conditions for electrowinning from nickel chloride electrolyte Company  Falconbridge’°’  1111 Sumitomo  . 1121 SLN  54,000  2,500  16,000  3281  40  80’  reinforced concrete  precast concrete  /  FRP  FRP  /  4,000 LJh  1,500 LJh  /  0.8 x 1.6 x 7  /  /  /  0.79 x 0.9 x 0.01  I  DSA  DSA  graphite  41 polyester dynel cloth 12  polyester fibre  plastic  /  FRP  /  Anode spacing, (cm)  14.5  15  /  Anodes/cell  46  39  31  Cathodes/cell  45  38  30  Anode cycle, (year)  /  I  /  Cathode cycle, (day)  /  8  3-4  Starter sheet cycle, (day)  /  2  I  Mother blank  /  Ti  Ti  Final cathode weight, (kg)  /  75  80(12 mm thick)  Catholyte, (gIL)  2 60 Ni  50-45 Ni 2  I  Spent anolyte, (gIL)  54 Ni 2  /  I  Current per cell, (A)  24,000  14,000  I  220  233  500’  /  1.0-1.2  /  60  55-60  /  98-99  92  /  Cell voltage, (volt)  /  3.0  I  Energy consumption, (kwh/kg-Ni)  /  3.0  /  Production, i/year No. of cells Cell construction Cell liner Electrolyte flow/cell Cell interior dimensions, (m) Cathode dimensions, (m) Anode Anode diaphragm Anode frame  ) 2 C.D.,(A/m pH Temperature, (‘C) CE, (%)  §: 24 cells were devoted to the preparation of starter sheets. *: 15 cells were used to prepare starter sheets. ¶: Calculated on the basis of 4 days and 12mm thick nickel cathode assuming 100 % current efficiency  Nickel electrodeposition in chloride and chloride-sulfate electrolytes  15  conditions are listed in Table 4. The dimensionally stable anode (DSA) used in nickel chloride electrowinning is chemically inert, stable in dimensions and has a long life expectancy. The use of DSA avoids the extensive anode and scrap handling encountered in direct matte electrowinning and also prevents the contamination of the anode scale experienced in nickel sulfate electrowinning. The chlorine gas evolved on the anode is collected completely and recycled to the leach stage. The cathode nickel, after being washed, is heated at 700°C to reduce the hydrogen content to less than ’. The purity of the cathode nickel can reach as high as 99.97 % Ni’. 12 5 ppm  1.3 Nickel electrodeposition in chloride and chloride-sulfate electrolytes As early as 1977, Falconbridge conducted extensive experiments on nickel electrowinning from pure nickel chloride electrolytes on a pilot scale . The two cells used were actually of 1 industhalsize,viz,0.8 x 1.6 x 7 m . Thecathodehadthedimensions 1.14 x 0.63 m 3 andthegraphite 2 anode had the dimensions 1.3 x 0.62 x 0.06 m . The distance between the two anodes was 18.9 cm. 3 The number of cathodes was 31 in one cell and 26 in the other cell. The maximum conductivity of electrolyte was found at a nickel concentration of around 130 g/L Ni 2 at temperatures 40, 60 and 80°C. The test conditions were 130 g/L Ni , 60-65°C, pH —l and 200-250 A/m 2 2 with a current efficiency of 97-98 %. Sodium chloride and boric acid were not added to the electrolyte as no beneficial effects were found with their additions. At pH 1, the cathode deposit obtained had a very good quality from the viewpoint of purity and surface appearance. The results obtained at pH 1 were even better than at pH 2 in that the occasional pitting on the cathode surface could be avoided. The current density was found to have the potential to be raised further, as good quality nickel cathode could still be attained at current density up to 400-500 A/m . The restricting factor in the 2 use of a high current density was found to be the overheating at the cathode contacts. The impurities Pb, Fe, Cu, Co. Mn, As and Zn were also studied. Lead was found to deposit completely with nickel, accounting for 3 ppm in the cathode when the electrolyte contained 0.4 mg/L Pb . The lead 2 content of the cathode was almost the same as in the electrolyte 0.4 x iO /130 x 106 3 ppm. As for iron, its content in the cathode was somehow higher than that in the electrolyte, 87 ppm versus 5 mg/L, as 5 x iO / 130 x 106 38 ppm. Although it was not found that iron had any adverse effect on the cathode nickel, the cathode surface pitting observed at pH2 might be associated with the iron in the electrolyte. The presence of copper had a deleterious effect on the cathode nickel. Due to its higher electrode potential, copper would probably be deposited at the limiting rate and cause harmful dendrite formation on the cathode. When the electrolyte contained less than 1 mg/L , dendrites did not form. However, the cathode nickel contained 27 ppm Cu. The tolerable 2 Cu concentrations of other impurities were 1-4 mg/L Co , <2 mgfL Mn 2 2 and <2 mg/L Zn . 2 Gong et al 1 recently investigated the electrowinning of nickel from nickel chloride electro lytes. The cathode was a pure nickel sheet, the anode was a platinum foil or Mn0 -coated titanium 2  Nickel electrodeposition in chloride and chloride-sulfate electrolytes  16  sheet, and the space between the anode and the cathode was 6 cm. The test was run typically for 3—4 hours. Whether a diaphragm was used in their tests was not indicated. The nickel ion con centration for electrowinning was selected on the basis of their measurements of the electrolyte conductivity and viscosity at 30—192 g/L Ni 2 and 25—80°C (see Figure 3). Higher conductivity and lower viscosity of the electrolyte were preferred for the electrowinning of nickel. The rec ommended operating conditions were 120 g/L Ni , 65°C, pH 1, 150—250 A/rn 2 , under which the 2 current efficiency was around 96.5 %. 3.5  4  3  3.5  Temp., (t)  —. E o  w  25  53  0  50  A  60  0  70 • 80*  2  1.5  —  C.)  0  1  C  8  0.5  Temp., (°C)  80 A  0  0  70 •  60  50  25  C  A  0  20406080100120140160180200 [N12+J, (gil.)  0.5 0  I  0  I  I  I  I  20406080100120140160160200 LNI2+1, (gil.)  Figure 3 The electrolyte conductivity and viscosity of NiC1 2 solutions at various temperatures The quantitative relationships between the impurity (Pb , Zn 2 , 2+ 2  (2+)  contents in cathode nickel and the electrolyte were established experimentally in the electrolyte containing 81.8 g/L , 44.5 g/L Ct, 141.5 g/L SO and 31 gIL Na at a current density of 300-400 A/rn 2 Ni , pH 2.2 and 2 temperature 68-70°C. The examined impurity contents in the electrolyte were in the ranges of , 0.8 ppm Zn 2 0.8 ppm Pb , 14 ppm Cu 2 2 and 35 ppm Co . It was found that a linear 2 relationship existed between the impurity contents in the cathode nickel and the electrolyte. These linear relationships indicate that the impurities are reduced on the cathode probably at a limiting rate. The effects of the impurities Mg , Mn 2 , Zn 2 2 and Al in the range of 5-2,000 ppm on the current efficiency, deposit quality and purity, surface morphology and crystallographic orientation were examined by Gogia and Das when nickel was electrowon from sulfate electrolyte (60 gIL , 12 g/L 4 2 Ni SO and 12 g/L 3 2 Na B0 at a current density of 400 Nm H ) , temperature 30°C and 2 pH 2.5. It was found that the current efficiency was affected very little by these impurities. However, it was observed that the cracking, curling or peeling of the nickel deposit would occur in the presence of these impurities, especially at the higher level of their concentrations. Concerning the con tamination of the cathode deposit, the least occurred with Mg 2 and the greatest with Zn . Based 2 on the acceptable quality of cathode nickel, the tolerable limits of impurity concentrations in the  Nickel electrodeposition in chloride and chloride-sulfate electrolytes  17  electrolyte were  500 ppm Mg , 250 ppm Mn 2 2 and 5 ppm Al. The reasons , 100 ppm Zn 2 behind the effects of these impurities were not stated clearly by Gogia and Das. It is believed that  the effects of these impurities may not be purely electrochemical in nature, as their equilibrium potentials are all well below that of nickel. As regards zinc, it may be deposited with nickel at a potential much more positive than the Zn /Zn equilibrium potential, due to the so-called under2 potential deposition phenomenon’ . Underpotential deposition is mainly due to the formation of 291 a solid solution of Zn-Ni. Therefore, the activity of zinc in the metallic (cathode) phase is greatly reduced and the electrode potential of Zn 7Zn is shifted in the positive direction. 2 The purpose of nickel electrodeposition is to obtain a coherent and compact cathode deposit in most cases. However, nickel powder can also be produced via electrodeposition. Ostanina et t301 studied the production of nickel powder from a mixed nickel chloride-sulfate electrolyte a1 containing 47.8-57.4 g/L 2 •7H 4 NiSO 0 , 200 g/L NaCl and 50 g/L NRC1 at pH 4.5-4.8 at a tem perature of 50°C. In order to maintain a large portion of fine particles, the (nominal) current density was increased linearly during electrolysis to account for the increase in the actual cathode area. The current density was increased linearly at a rate of 600-10,000 A/m h from the initial limiting c.d. 2 up to 2,200-10,000 A/m . When the current density reached the predetermined maximum, the 2 powder on the cathode was shaken off. Then the electrolysis was continued again in the same cycle. This technique gave not only a uniform nickel powder but also a lower energy consumption. On the basis of the experimental results, mathematical models of the mean size of the powder and the current efficiency were developed as functions of the rate of current density increase, the maximum c.d. and the time interval of powder removal. A new technique was developed by Teschke and Galembeck 311 to produce large nickel particles from an electrolyte containing 300 g/L 2 •6H 4 NiSO O , 45 g/L 2 •6H and 30 g/L 3 NiC1 0 B0 at 30°C H on a PTFE-covered nickel cathode. The typical pinhole diameters on the PTFE film were around 5 im. The electrons could transfer only through the paths of pinholes, cracks and protrusions in the PTFE layer. Due to a very small fraction of active cathode area, an apparent current density of 100 AJm 2 resulted in a very negative cathode potential of -2 volts vs. SCE. The nickel particles had the following characteristics: (1) The particles had little direct contact with the substrate metal. A slight disturbance of the electrolyte would make them fall off. (2) The particles were approxi mately hemispherical in shape. (3) The bases of the particles were parallel to the PTFE layer. The problem with Teschke and Galembeck’s method lies in the fact that it is quite difficult to control the real active cathode area. Philip and Nicol 32 conducted a limited number of tests on nickel electrodeposition from pure nickel chloride solutions. The highest current efficiency they obtained was 94.3 % under the conditions of 1 M NiC1 , 0.1 M HC1, 53°C and 225 A/rn 2 2 in an electrolysis of 6-hours duration. The  Nickel electrodeposition in chloride and chloride-sulfate electrolytes  18  authors believed that chloride ions play a catalytic role in the nickel cathodic reduction, reducing the overpotential of nickel deposition and thus resulting in little or no simultaneous hydrogen evolution. The process, without the need of a pH buffer, was also relatively insensitive to the pH fluctuations in the feed solution and less demanding in the requirements for diaphragm materials. The impurities, 2 Fe 2 , Zn and 2+ were also investigated in the electrolyte of 1 M NiCl , 2M 2 NaCl and 0.01 M HC1. It was found that for ferrous ion, its effect was negligible when its con centration was between i0 3 and 10.2 M. However, when its concentration reached i0’ M, it was reduced as well with nickel, accounting for 13 % of iron in the cathode deposit at 250 A/m 2 and 20 % at 1,200 A/rn , compared with 60 % and 52 % at 250 and 1,200 A/rn 2 2 respectively in the sulfate electrolyte. While for zinc, 25 % Zn was found in the cathode nickel deposit at 600 A/rn 2 when the electrolyte contained 10.1 M Zn , compared with 42 % in the sulfate electrolyte. 2 It was pointed out by Finkelstein et a1 33 that a high cathode current efficiency and a satisfactory nickel deposit should be obtained from chloride media. For the optimization ofthe current efficiency and improvement in the nature of the deposits, the following conditions should be met: higher concentrations of nickel and chloride ions, higher temperature and 0.01 M acid. They found that vigorous agitation was advantageous since it reduced the problems associated with the mass transfer. Fujimori et al 1 stated that one disadvantage of the all chloride electrolyte was the high internal stress of the nickel deposit. The presence of sulfate in the electrolyte would decrease the magnitude of this stress. The causes of the internal stress were not indicated. The high stress would increase the tendency for short circuits, leading to an inefficient operation. The deposit stress was found to rise with increasing chloride concentration. Therefore, although high chloride concentration lowers the cell voltage due to the increase in the electrolyte conductivity, a compromised chloride con centration must be chosen. They found that the deposit stress became less severe as the temperature and pH rose. The development of the stress inside the deposit may be attributed to the interaction of atomic hydrogen with the nickel cathode and the adsorption followed by dissolution of hydrogen into the body of the cathode. This is in keeping with Fujimori et al’s” observation that the stress could be increased by lowering the temperature and pH. The specific adsorption of chloride ions on the cathode surface and the occlusion of electrolyte might also be responsible for the internal stress inside the cathode deposit. The electrowinning of nickel from nickel chloride electrolytes was also investigated in a three-compartment diaphragm cell . The cathode reaction was the reduction of nickel ions, 1 whereas the anodic reaction was the decomposition of water with oxygen evolution. The purpose ofelectrolysis was to produce hydrochloric acid in the centre compartment. The cathode diaphragm, made from woven terelyne, was used to control the pH of the catholyte (NiC1 ). Without a cathode 2 diaphragm, the pH of the catholyte would become low enough so as to result in a deteriorating  Nickel electrodeposition in chloride and chloride-sulfate electrolytes  19  current efficiency for nickel deposition. The anode diaphragm, made from a porous ceramic membrane, was more critical and was employed to prevent chlorine evolution on the anode. The leakage of chloride ions into the anode compartment was found to be around 0.4 % according to the amount of acid generated. The anolyte was a 34 % 4 S0 solution. The catholyte contained 2 H 50-74 gIL Ni 2 (as NiCl ) at 55°C, pH 2,200 A/m 2 2 and 2.9 volts. The current efficiency of nickel was over 97 % even when the nickel ion was stripped down to 55 g/L. The overflow electrolyte from the centre compartment contained around 30 g/L HQ and 1.5 g/L 4 S0 This process seems 2 H . to be successful technically; however, it may not be economically viable. The demanding properties of the anode and cathode diaphragms and the difficulty in controlling the appropriate flow rates of the feeds (NiCl 2 and H 0) are likely to discourage would-be users. 2 An interesting work was carried out by Sabot et al 3 on nickel electrorefining in 5 M Cad 2 electrolyte at 98°C. Calcium chloride was believed to provide the best electrolyte system for nickel electrorefming considering its electrochemical properties and price. The conductivity of the electrolyte increased with temperature. However, interestingly the conductivity of the electrolyte did not increase continuously with CaC1 2 concentration. The maximum conductivity was found at a CaC1 2 concentration around 2.5-3 M. The reason for choosing 5 M CaCl , which was not optimum 2 as regards the conductivity of the electrolyte, was based mainly on the electrochemical reversibility of the nickel electrode Ni fNi. The purification of 5 M CaC1 2 2 could be achieved by anionic ion exchange or by TBP solvent extraction. Ferronickel, which contained 94.4 % Ni and the rest Fe, Co and Cu, was tested and a cathode product with 99.7 % Ni was obtained after electrorefming in 2 + 0.5 M NiCl 5 M CaC1 . The current efficiency, pH of the electrolyte and the effect of pH on the 2 current efficiency are not disclosed. One important method which has been used widely in nickel electroplating to improve the plating quality is to use periodic reverse current (PRC) . Although the current efficiency in 1 electroplating is not as serious a concern as in electrowinning, the benefits achieved sometimes outweigh the cost due to the loss of current efficiency. The example given by Teschke and Soares is the electroplating of nickel in an electrolyte containing 300 g/L 2 6H 4 NiSO O , 45 g/L 2 6H NiCl O , 30 g/L 3 B0 at 80°C. The PRC parameters studied were the ratios of current amplitudes and pulse H widths in each cycle of deposition and dissolution. Using the PRC technique, not only can the current density be raised, but also the growth morphology can actually be controlled by changing the periods and amplitudes of each cycle. Boric acid is widely used in nickel electrowinning and electrorefming. The purpose of using boric acid, besides improving the cathode quality, is to control and stabilize the electrolyte pH near the cathode surface. However, the interpretation of the function of boric acid is quite controversial, mainly between two opinions whether it is a buffer or a surface catalyst. In a dilute solution of  Kinetics and mechanism of nickel electrodeposition  20  boric acid, the buffering point pH is around In the presence of nickel ions, does the buffering point pH remain the same? Using the pH titration technique, Tilak and co-workers 38 investigated the behavior of boric acid in an electrolyte of 4 NiSO S 2 NaC1 Na O at 55°C. Two series of titrations were carried out, one on the effect of boric acid concentration (0.1-0.5 M) in 0.97 M NiSO 4 0.33 M SO 1.33 M NaCI and the other on the effect of nickel sulfate (0.1-0.5 M) in 0.3 M 3 2 Na 4 B0 H -  -  -  0.33 M 4 SO 1.33 M NaC1. They found that the buffering capacity of the electrolyte increased 2 Na both with increasing concentrations of boric acid and NiSO . To account for this fact, a weak 4 -  complex of nickel with boric acid was assumed to exist. Their thermodynamic calculations and mass balance were based on the following four reactions (15)-(18): 1 K  B0 3 H  =  (15)  H+H B 2 O 2 K  +2H 2 Ni B O  =  (16) 3 B Ni(H 2 ) 0  4 K  Ni + 2 SO  (17)  4 NISO  =  5 K  +H 2 Ni O  =  (18)  NiOH+H  The equilibrium constant for the reaction (19) was calculated to be log K = -12.2 K  +2H 2 Ni B 3 O  =  3 B Ni(H + 2 ) 2H O  —  -11.1. (19)  Although the complex of nickel with boric acid is likely to exist, the accuracy of the authors’ calculations is questionable. Besides the question of the reliability of the equilibrium constants used in the calculations, Tilak et al did not consider the nickel chloro complex and the formation of bisulfate. As for bisulfate, however, such an omission may create only a marginal difference as the initial pH’s for their pH titrations were around 3.5. Hoar& 391 believed that the nickel-boric acid complex could be reduced much more favorably.  1.4 Kinetics and mechanism of nickel electrodeposition Most of the mechanisms proposed so far are related to nickel electrodeposition from the Watts bath 3 -NiC1 4 (NiSO 2 B ) H 0 which was developed for nickel refining and especially for nickel electroplating. The addition of chloride is to promote the dissolution of the nickel anode. It is normally believed that the nickel ion is reduced in two consecutive one-electron charge transfer steps. Two linear regions are observed on the Tafel plot (Figure 4). Considerable controversy exists regarding the mechanism and the rate-determining step in the electroreduction of nickel ion.  Kinetics and mechanism of nickel electrodeposition  21  The most difficult problem is how to identify the intermediate species involved in the electron transfer. Previous researchers have attempted to resolve this problem relying mainly on certain assumptions without appropriate experimental measurements and/or thermodynamic calculations. Most of the investigators assume that monovalent nickel hydroxide (NiOH.) is the intermediate species involved in nickel electrodeposition from both sulfate and chloride solutions. Others believe that monovalent nickel chloride (NiC1) is the intermediate species for nickel reduction in chloride solutions.  0.1 0 -0.1 -  o  ° -0.3 -OA  : 0.5 -0.6 -0.7 -0.8  Figure 4 The partial current density vs. potential for the deposition and dissolution of nickel on a platinum RDE at 1,000 rpm (1 M NiCI 2 -2 M NaC1 0.01 M 321 woo HCI at 2lC) -  -0.9  0.1  1  10  C.D., (A/m2)  100  Nickel electrodeposition is further complicated by simultaneous hydrogen evolution, where atomic hydrogen may be adsorbed on the cathode surface and further be absorbed into the body of the cathode. The Tafel slope and the exchange current density for nickel reduction should be obtained from its partial polarization curve which does not contain that part of the current due to hydrogen evolution. How to subtract the effect ofhydrogen evolution may bring another inaccuracy into the determination of the kinetic parameters of nickel reduction. The following are three popular mechanisms’ proposed in the literature for the Watts bath and/or chloride solutions. The first mechanism, as expressed in reactions (20)-(22), has two one-electron transfer steps and assumes NiOH as an intermediate species involved in the electron transfer process. This mechanism has been recommended for electrolytes of chloride° , Watts bath 1 , and chloride and perchlorate 411 . 42 +H —NiOH+H 2 Ni O  (20)  r4..  (21)  NiOH+e  -*  (NiOH)  (NiOH)+H+e =Ni+H 0 2  (22)  This mechanism, despite being the most popular one, is very doubtful, as the concentration of NiOH is negligible according to thermodynamic calculations. The second mechanism, as represented  1 One should keep in mind that coordinated water molecules are usually omitted when writing the reactions. This may be misleading. For instance, the reaction N? + 1120 = Ni0H + H should be written more correctly as 6 0) 5 [Ni(H 2 ] 0 2 [Ni(H O H] ) + J[4.• The hydrated hydrogen ion H O is nomially written as 3 1i omitting the H 0. 2  Kinetics and mechanism of nickel electrodeposition  22  in reactions (23)-(25), also has two one-electron transfer steps; however, it assumes NiCl as an intermediate species associated with the electron transfer during nickel ion reduction. This 33 43j and the Watts bath. The rateZ 3 mechanism has been proposed for chloride electrolytes( determining step may shift with the magnitude of the overpotential. (23)  + C1 =NiCl 2 Ni  (24)  r4..  NiCl+e  (NiCl)  —* at highzT  r4s.  (NiCl) + e  -, at Iowufl  Ni  +  (25)  cr  The third mechanism, as described in reactions (26)-(27), seems to be the simplest one, involving neither the nickel hydroxy complex nor the nickel chloro complex. This mechanism is the most likely one when considered together with the results of the present investigation. Ni.2+ + e  r4.  —*  (26  Nz  (27)  Ni+e=Ni  The third mechanism was proposed by Ovari and Rotinyan t451 based on their studies on the cathodic reduction of nickel ions in pure nickel chloride electrolyte at 55°C, and by Ragauskas and Leuk 1 in view of their test results obtained in the electrolyte 3 M KC1 + 0.1—0.6 M NiC1 sminas 2 at : 451 25°C. The kinetic equation for the third mechanism can be expressed ast Ni = 1  (28)  (ctFE”I 2Fk [Ni.2 ] exp— RT  3  where x is approximately equal to 0.5. Equation (28) was also found to be true by Vilche and 471 for expressing the kinetic rate of nickel reduction in the electrolyte of 1 M NiC1 Arvia 2 + HC1 + NaC1 at temperatures 25-75°C. The chloride ion was found not to be involved in the electrode reactions. The slight increase in the electrode reaction rate resulting from the addition of NaC1 was attributed entirely to the enhanced activity of the nickel ion. Ovari and Rotinyan t451 defended their mechanism expressed in reactions (26)-(27) using their other supportive results on the measurements of the exchange current density, corrosion potential and corrosion current. =  (29)  k [ 0 Ni]’  corr. =  kcorr.I-  =  (30)  -I  Const. +  RT ’H (1+ aNI + )F 112 1 1  Const.  0.0591 — 2  at 25°C  (31)  Kinetics and mechanism of nickel electrodeposition  23  Vilche and Arvia t471 derived similar equations for the corrosion potential and current at 25°C. (kC,R [ 2 Hi = (1+ aM + 0.5)F lfllL ,N, 4 k RT  =  J  0.059 loI 2  j  —  0.059 pH 2  (  F F(l+aNJF 1 [Hj 1 k 1 01 exp—E = kN exPIL ,.) Ecori.] RT  (32)  (33)  The electrochemical impedance technique was found to be useful in identifying the existence of intermediate species involved in nickel reduction . It seems certain from these studies that 451 the monovalent nickel aciion is involved; however, it is still very difficult to distinguish between Ni, NiOH and NiCl.  Wruck° using the rotating disc electrode technique, studied the effect of electrolyte com , 1 position, temperature and rotational speed on the current efficiency of nickel electroreduction under the conditions of 0.16-0.5 M NiCl 2 plus 2 M NaCl at pH 1.2-2.4, temperature 25,40 and 60°C, and RPM 600, 1,600 and 3,000. Some of his findings are consistent with the results of the present research. For instance, Wruck found that the current efficiency of nickel increased with pH, tem perature and overpotential but decreased with increasing mass transfer rate. He also observed that the sharp drop in current density would occur when the potential became sufficiently negative. This phenomenon was called cathodic passivation. It was believed that the cathodic passivation was caused by the formation of insoluble nickel hydroxide on the cathode surface due to the hydrogen evolution reaching a limiting condition. Wruck’s contributions to understanding nickel deposition are outstanding. However, his modelling and related calculations are subject to question. Based on the reactions (34)-(36): (34)  fast  0 = NiOH + H 2 2+H Ni rds NiOH+e  -*  (35)  (NiOH)  (36)  fast  (NiOH)+H+e = Ni+H 0 2  Wruck derived, without providing any details, the following equation (37) for the rate of nickel ion reduction: Ni  =i [NiOH 1  [H4]{exP[  (2 —a)F T1NI]  —  exP(_c-nNi)}  For the rate of hydrogen evolution, he derived the equation (38):  (37)  Kinetics and mechanism of nickel electrodeposition  .0  112 t 1112 = [’fl  1  r(1—a)F 1 lH 1 j RT 2  24  (  —  czF ‘ii )1 2 exPL%_nH  (38)  The concentrations of nickel and hydrogen ions in equations (37) and (38) are those at the electrode surface. The problem in deriving equations (37) and (38) is that Wruck confused the overpotential with the electrode potential. Furthermore, he studied neither of the effects of acid and chloride concentrations, and in his calculations, only Ni 2 and NiOW were considered for the nickel species. Consequently, the validity of equations (37) and (38) is suspect. Platinum was used as the anode; however, how the contamination by chlorine gas was eliminated is not indicated. Whether the ohmic drop between the working and reference electrodes was compensated or not is not mentioned either. Another question concerns the omission ofthe nickel chioro complex NiCP in the equilibrium calculations. A work somewhat similar to one of Wruck’s studies was carried out by Hurlen . Using a 521 stationary nickel wire as cathode, he studied the kinetics of nickel reduction at 25CC in the electrolyte 0.03—4 m NiC1 2  . He found that the Tafel slope was equal to 21 mV/decade at the 2 0—7 m CaC1 lower overpotential and 118 mV/decade at the higher overpotential. Based on his thermodynamic and kinetics studies, he believed that the reduction of nickel ion proceeds by a two-step mechanism +  with Ni(I) as the intermediate species. The first electron transfer step, Ni(ll) parallel reactions and is independent of kink sites on the cathode surface: Ni(H O 2 )+e 2 =Ni(H 0 ) 0) 2 NiC1(H  +e  =  NiC1 ( o) 2  —*  Ni(I), involves two  (39)  (40)  Reaction (39) was dominant at low and moderate chloride activity and high water activity, while reaction (40) predominated at high chloride activity and low water activity. The second electron transfer step, Ni(I) — Ni, occurred directly on the kink sites. Chloride ion was found not to be involved in this second electron transfer. However, Hurlen could not confirm the existence of Ni(I) using the rotating ring-disc electrode technique. Nevertheless, he argued that Ni(I) species was soluble in the electrolyte and its life was too short or its concentration was too low to be detected on the ring. The problem with Hurlen’s studies is mainly the lack of the investigation of the effect of the mass transfer rate on the nickel reduction. In addition, his belief that Ni(I) is soluble may be incorrect. The rest potential of nickel in a solution containing 2 M NaC1 or 2 M NaC1O 4 was found to be not equal to either the potential of HfH /Ni; however, it was dependent on the pH and the 2 2 or Ni anion in the electrolyte . It was also discovered that at low pH the initial portion of the cathodic 42 Tafel plot was related mainly to the hydrogen evolution, whereas at higher pH and higher current density the hydrogen evolution and nickel reduction occurred simultaneously. However, the method  Kinetics and mechanism of nickel electrodeposition  25  used to determine the parameters for hydrogen evolution is not quite correct. Piatti et a1 t42 determined the parameters for hydrogen evolution from the initial portion of the polarization curve at a low pH. The problem is that even at a lower pH and lower current density, the nickel reduction is still there. That part of the current which is due to the reduction of nickel ions must be deducted in order to study the hydrogen evolution. Piatti et al believed that the mechanism for nickel reduction is expressed in reactions (20)-(22). However, they did not study the effect of chloride concentration and mass transfer rate. The convection of the electrolyte was achieved by bubbling hydrogen gas through it.  The effect of the temperature on the equilibrium and polarization behavior of the nickel electrode in chloride electrolyte (1 M NiC1 ) was studied by Vagramyan et al at pH 1.5 and tem 2 perature 25275°Ct531. For the tests at temperatures above 100°C, the cell together with the reference electrode was placed in an autoclave. It was discovered that the cathodic overpotential decreased and the exchange current density increased with increasing temperature (Figure 5 and Table 5). The authors ascribed the major reason for the acceleration of nickel reduction to the increase in the  activity of the nickel ion. However, there are many doubts about their research. As has been found in the present experiments, dissolved oxygen and agitation affect greatly the electrode potential.  .  -0.1  C  a) -0.3  a a,  -  -0.4  0.2  a,  Vagramyan et al did not study the effect of these two factors. Furthermore, they even did not indicate whether or not the electrolyte was deaerated before the tests. Finally they did not mention if the partial current density of hydro gen evolution was subtracted from the total current density.  C  0.0.5  0  25  50  75  100  125  150  175  Figure 5 The overpotentials of nickel cathodic deposition and anodic dissolution in 1 M NiCI 2 and 1 M NiSO 4 at 100 A/rn 2 and pH  200  Temperature, (‘C)  Table 5 Temperature coefficients of the overpotentials of nickel cathodic deposition and anodic dissolution in 1 M NiC1 2 and 1 M NiSO 4 at pH 1 •153] Electrolyte  JrlIJT, (mV/°C)  aria/aT, (mV/°C)  25-75°C  175-200°C  25-75°C  175-200°C  1 M NiC1 2  2.4  0.2  -3.5  -0.4  1 M NiSO 4  3.3  0.4  -4.8  -0.7  Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media  26  1.5 Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media Nickel electrowinning is always accompanied by simultaneous hydrogen evolution under normal operating conditions due to nickel being electrochemically negative to hydrogen, high nickel overpotential and low hydrogen overpotential. The hydrogen evolution in nickel electrowinning wastes electricity and the adsorbed hydrogen is responsible for some defects in the cathode nickel deposit. The absorption of atomic hydrogen and adsorption of nickel hydroxide and basic salts might increase the undesirable high internal stress and hardness, and reduce the ductility of the nickel cathode deposit. In the case of high current density electrowinning, the prevention or at least a lessening of hydrogen evolution is necessary; otherwise, the surface pH will increase markedly, leading to the precipitation of insoluble nickel hydroxide. This precipitate will in time depress greatly the reduction of nickel ion and make impossible the achievement of a coherent cathode nickel deposit. Boric acid is commonly used in the Watts bath to stabilize the electrolyte pH. Yeager and co-workers 55 believed that the simultaneous hydrogen evolution during nickel electrowinning might well be compared with the hydrogen evolution on the nickel substrate in the acidic media without nickel ions, as they found that hydrogen evolution and nickel reduction did not have any evident interactions. In Yeager and co-workers’ tests, the cathode was positioned horizontally and the electrolyte was injected parallel to the electrode surface at a velocity of 45 cm/sec. The electrolyte was saturated with hydrogen gas. The rate of reduction of nickel ion was found to be independent of pH, while the rate of hydrogen evolution was independent of nickel ion concentration. From the experimental data, the Tafel slopes were calculated to be 123 mV/decade for hydrogen evolution and 103 mV/decade for nickel reduction in 0.5 M NiC1 2 at pH 2,25°C and in the current density range of 10-1,000 A/m . However, the reaction order with respect to chloride 2 ion concentration was not determined. Furthermore, the ionic strength was not held constant in their tests. On the basis of their experimental studies on hydrogen evolution on a nickel cathode in 0.005-1 M 4 S0 and 0.01-0.5 M HC1, Tamm et al derived equation (41) to describe the over2 H potential of hydrogen evolution: —i  =  E, +  1—a  1—aRT —  —i- ln[H] + RT ln i  .  (41)  where Wi is the potential at the outer Helmholtz plane. It has to be pointed out that equation (41) does not consider the concentration polarization. When the concentration is taken into account, equation (41) becomes:  Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media  —=E  +  1—a  i—  1—aRT I ( i RT [H1bl—iJ -ln + 1 lnz  27  (42)  Equation (41) was successful in explaining the experimental results for the Tafel slope and the effect of acid concentration. It was also successful in elucidating the (around 60 mV) lower overpotential in HC1 solution than in 4 S0 solution on account of the specific adsorption of chloride ions on the 2 H nickel cathode surface leading to a decreased One of the problems associated with hydrogen evolution during nickel electrowinning is the formation of pits on the cathode. To eliminate these pits, the adsorbed hydrogen bubbles must be removed either via chemical oxidation or mechanical agitation. Chemical oxidation was found to be effective using hydrogen peroxide. The stability of hydrogen peroxide in a nickel electrolyte was studied by Chen et a1 . Their electrolyte contained 65 gIL Ni 1 , 30 g/L Na, 40 g/L C1, 6 g/L 2 B0 at 60°C. The mechanism for hydrogen bubble destruction was supposed to be due to the H 3 reaction H 0 2  2H,  21120. However, the true mechanism is not well understood. Chen et al formulated an equation from their experimental results to represent the decomposition rate of peroxide as: +  [H 0 2 ,J where  =  =  J 0 2 [H  exp(—5.5 x 3 10 t )  (43)  is the initial peroxide concentration and r is the time in seconds. The decomposition reaction was 2H 0 = 2H 2 0 + °2• It can be calculated from equation (43) that almost no peroxide 2 will be left in the electrolyte after one hour. On the other hand, considering the standard electrode [11202]  potential E°  1.776 volt vs. SHE (at 25°C) for the cathodic reduction of peroxide, 11202 + 2H + 2e =21120, it is understood that peroxide should be reduced preferentially to water on the cathode =  rather than nickel reduction and hydrogen evolution. Therefore, the behavior of H O in suppressing 2 2 pits on the cathode remains a mystery. In practice, it may be necessary to add peroxide con H tinuously to the electrolyte during nickel electrowinning. Considering the higher decomposition rate and higher consumption of hydrogen peroxide (100 mLlkg-Ni) in depressing hydrogen bubbles, sodium hypochlorite (NaC1O) was tried and was found to be more effective than hydrogen peroxide . Thermodynamically, hypochiorite may still 58 be reduced on the cathode, as the standard potential is 0.81 volt vs. SHE at 25°C for the reaction Cl()  0 2 H  2e  Ci + 20H. One alternative to avoid hydrogen pits on the cathode nickel is to add some surfactants which can increase the surface tension between nickel and the hydrogen bubbles. However, such an addition may cause some undesirable side effects. +  +  =  Ovari and Rotinyan 451 investigated hydrogen evolution during nickel deposition from pure nickel chloride electrolyte under the conditions of 55°C, 0.25-2 M NiC1 2 and pH 0.75-2.5. They  Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media  28  discovered that the rate of hydrogen evolution obeyed a simple kinetic expression: 2= ZN  Fk[H’] exp(_)  where a is approximately equal to 0.5. Ovari and Rotinyan also measured the surface pH on the cathode using a micro glass pH electrode. They found that the surface pH was always higher than the bulk pH if the bulk pH was above 1.5. They claimed that the surface pH increased initially with increasing current density; however, it did not increase further when the current density reached a certain level. This experimental finding could be an artifact. The range of current density studied was quite limited, only up to 150 A/m . It is believed that the surface pH would increase continuously 2 with the current density. The major reason for this conclusion is that the change in current density was not large enough. The simultaneous nickel reduction and hydrogen evolution was modelled by Diard and Le 59 in chloride media (0.5 M NiC1 Gorrec 2 + 1 M KC1) at pH 2. A rotating disc electrode was used for their experimental tests. After realizing the difficulties in identifying the intermediate species involved in electron transfer during nickel ion reduction, Diard and Le Gorrec came up with the idea of proposing a theoretical mechanism first and comparing the theoretical response with the experimental results. Their modelling was based on the mechanism that the nickel ion was chemically reduced by molecular hydrogen produced electrochemically at the cathode. H+s+e=H  (45)  H  (46)  +  H  +  e  =  (g) 2 H ( 2 )=H H ( 2 )  +  2 Ni  (47) =  Ni  +  2H  (48)  where s represents the cathode substrate. However, their mechanism is difficult to accept, as it is unlikely that the reduction ofnickel ions is caused by hydrogen gas at ambient pressure. Furthermore, their mathematical modelling is too complicated to be understood. Using the rotating disc electrode technique, Dorsch° studied the simultaneous nickel depo sition and hydrogen evolution on a gold substrate in a nickel sulfate electrolyte (1 M NiSO 4 + 20 g/L B. 3 H ) 0 It was observed that nickel was not deposited and only hydrogen gas evolved until a minimum current density was exceeded. In addition, it was found that at higher total current density the partial current density of hydrogen evolution remained constant and was equal to its limiting current density. It was concluded that the minimum current density was equal to the limiting reduction rate of W ions at pH below 3. At a pH above 3, the minimum current density was limited  Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media  29  by the diffusion of hydroxyl ions or complex ions such as NiOW. However, such conclusions axe not reasonable. The question is how to explain under the limiting condition of hydrogen reduction why insoluble nickel hydroxide Ni(OH)S) does not form and why the nickel deposition can still proceed smoothly. Interestingly, a membrane cell was used by Ragauskas and Leuksminas in their study of simultaneous hydrogen evolution during nickel electrodeposition from chloride electrolytes . On 611 one side of a palladium membrane (16.2 p.m thick), cathodic reactions (nickel reduction plus hydrogen evolution) took place in the electrolyte 3 M KC1 + NiC1 . On the other side of the 2 membrane, the permeated atomic hydrogen was completely oxidized anodically in 0.1 M NaOH at a potential of 0.700 volt vs. SHE. The collection efficiency of the atomic hydrogen permeation is unbelievably high, amounting to 97 % ofthe hydrogen evolved on the cathodic side of the membrane. The permeation time through the Pd membrane plus 1 pm thick nickel layer was only 2 seconds. Based on the current determined from the anodic oxidation of permeated atomic hydrogen, the partial current densities for nickel reduction and hydrogen evolution could be separated. There was a peak on the nickel partial polarization curve. The hydrogen evolution after the peak was believed to be due not to direct decomposition of water but to the reduction of water by univalent nickel ions. The presence of univalent nickel ions was claimed to be confirmed by the rotating ring-disk electrode studies. The hydrogen evolution and hydrogen content in the cathode nickel can be affected by superimposing a sinusoidal alternating current (AC) on a direct current (DC) electrolysis process . 2 The studied conditions were: AC frequency 20-5,000 Hz, electrolyte containing 210 g/L 7H 4 NiSO O 2 , 25 g/L 3 B0 and 9 gfL NaC1, temperature 24°C, pH 4.2, DC current density H 100 AIm , and the amplitude ratio between AC and DC 1:1, 5:1 and 10:1. The hydrogen content 2 in the cathode nickel was determined by vacuum extraction while heating the samples to 500°C and the hydrogen gas was determined volumetrically. Higher hydrogen evolution and lower hydrogen content in the cathode nickel were found when using combined AC and DC electrolysis rather than DC alone. Also it was found that the minimum hydrogen content in the cathode nickel occurred at an AC frequency of 50 Hz. The effects resulting from superimposing a sinusoidal AC were due not to the anodic dissolution of nickel, but to the decrease in the thickness of the diffusion layer. The reduction of nickel ion proceeded with little concentration polarization. However, the simultaneous hydrogen evolution took place near the limiting rate. Therefore, the decrease in the thickness of the diffusion layer led to a substantial increase in the mass transfer rate towards the cathode. The lower hydrogen content in the cathode nickel may also be due to hydrogen ionization on the nickel cathode. However, one problem with such a technique is that the current efficiency  Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media  30  of nickel will be sacrificed to a certain degree. Current reversal and straight DC electrolysis have been investigated in the case of nickel. It was found that the technique did not generate any differences in the appearance of cathode nickel if soluble organic additives were absent . However, in the presence of soluble organics, such as 631 pitting additive caprylic acid 2 (CH 3 CH C 6 ) OOH, hydrogen pitting on the cathode was found to be more serious than when using DC alone. The tests were conducted in an electrolyte containing 105 g/L Ni , 85 g/L Cl-, 120 g/L SO and a small amount (< i0 M) of caprylic acid at pH 2.5, 2 65°C and 350 A/rn . When using the current reversal electrolysis, the addition of sodium chlorate 2 (NaC1O or tetramethylammoniurn sulfate 4 ) 3 N0 ) 3 [(CH S 2 1 did not prevent the pit formation on the cathode. The reasons for the formation of hydrogen pits under current reversal are not given. Temperature was found to have a significant effect on hydrogen evolution and the hydrogen content of the cathode nickel in an electrolyte containing 0.43 M 2 •6H and 0.5 M 3 4 NiSO 0 B0 at H c.d. 50 A/rn 2 and pH 1.5. The temperature increase from 20 to 50°C brought down the hydrogen content from —21 ppm to —7 ppm. A further increase in temperature affected the hydrogen content very little. However, the current efficiency of nickel increased continuously with increasing temperature, —42 % at 20°C, —50 % at 50°C and —93 % at 100°C, respectively. It was believed that atomic hydrogen in cathode nickel existed in the form of nonstoichiometric nickel hydride NiH. (n = 0.1-0.9). This nickel hydride was not stable even at ambient temperature. The subsequent decomposition of this nickel hydride caused dislocations in the structure of the nickel cathode. The adsorbed atomic hydrogen can be absorbed into the nickel cathode. The solubility of atomic hydrogen in nickel is quite low, in the order of 10.6 mole/cm . However, nickel is a 3 hydride-forming metal and the absorption of atomic hydrogen into nickel is an exothermic pro . The solubility of atomic hydrogen in most metals obeys Sievert’s law, that is: 651 cess [H] = 0 k-fl  (49)  where P 112 is the partial pressure of hydrogen gas, and k is the Sievert’s law constant. Actually, Sievert’s law was derived from the following equilibrium, H)=2M—H=2M—H  (50)  The two commonly used addition agents in nickel electroplating, the leveller 2-butyne 1 ,4-diol OH-CC-CH (CH O 2 H) and the brightener sodium benzene sulfonate 3 -SO 5 (CJ{ N a) were found to affect the hydrogen evolution as well during nickel deposition in a Watts electrolyte at 50°C. It was found that 2-butyne 1 ,4-diol increased the hydrogen evolution while sodium benzene sulfonate depressed the hydrogen evolution.  Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media  31  Another issue related to the hydrogen evolution during nickel electrodeposition is the cathode surface pH which is usually higher than the bulk pH. The higher surface pH may cause numerous undesirable consequences. The formation of ferrous hydroxide was found to be responsible for the abnormal electrodeposition of an iron-nickel alloy . The theoretical prediction of surface pH was 67 attempted by Dahms and &oll[67] although their model was over simplified. In order to make the prediction, the partial current density for hydrogen evolution must be known in advance. The problem with Dahms and Croll’s modelling is that they did not take into account the electrical migration of ions, such as, bisulfate, soluble nickel and ferrous hydroxides and the ion pair NiSO . 4 The activity coefficient of the hydrogen ion and a proper correction for the equilibrium constants due to the ionic strength were not considered either. In a generalized case of divalent metal electrodeposition in an electrolyte without any pH buffers and ligands, the distribution of ionic species in the diffusion layer and the surface pH were analyzed theoretically by Harris. The increase in surface pH was believed due to the reduction of hydrogen ions. The electrolyte adjacent to the cathode surface would become basic as soon as the reduction of hydrogen ions reached its limiting condition: 1112  >  FDH+[H]b  /6  (51)  The formation of soluble metal hydroxy complexes would suppress the rise of the surface pH. The drawbacks of Harris analysis lies in the fact that the selection of parameters such as diffusion coefficients and equilibrium constants is arbitrary and the specific details of the calculations are not given in his paper. Of course, equation (51) must be modified if there are any pH buffers in the electrolyte. One of the methods of measuring the surface pH involves the use of a micro pH-sensing electrode. Pt/H 2 and antimony micro electrodes have been used to measure the cathode surface pH during the electrolysis of acidic sodium chloride solution. The change in the surface pH was found to be quite significant under the conditions of 25°C and 0.5-100 A/m 2 even though the tip of pH-sensing electrode was positioned 150 iim away from the cathode surface. The electrolyte was not agitated mechanically in the tests. However, the 112 bubbles evolving on the cathode affected the flow of electrolyte nearby. Controversial results prevail in the literature concerning the surface pH during nickel elec trodeposition. One paper published by Berezina et al° 1 deals with the surface pH measurement  during nickel electrodeposition in 0.89 M NiSO 4 without and with 3 BO (0.5 M) at 20°C. It was H found there was a maximum in the surface pH at a bulk pH of around 3 on the curve of surface pH versus the bulk pH. It is difficult to understand why the surface pH decreased as the bulk pH  Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media  32  increased. Besides, the authors’ surface pH values are too high, reaching above 9 at 20 AIm 2 and 20°C. Such an ermneously high surface pH value may be ascribed to the method used for the measurements. They determined the surface pH based on the electrode potential of the cathode immediately after the current was switched off. In the presence of boric acid, they found that the maximum in the surface pH disappeared mysteriously. However, according to the results of the present investigation, the surface pH is again too high in the presence of boric acid, being above 8.5 at2O A/rn , bulk pH —2.6 and 20° C. The problem is that the authors did not consider the formation 2 of insoluble nickel hydroxide Ni(OH) 2 at these high surface pH’s. In their subsequent studies in chloride electrolyte under the conditions of 0.3-1 M NiCl 2 and bulk pH 0.5-5, the effect of the surface pH on the mechanism of nickel reduction was determined at 23°C’: (52)  ( aFE [Ni ] 2 expl RT [0H]  (53)  N1 = 1  2Fk[Ni.2 ]  When the surface pH > 6.5,  ‘N =  2Fk  <  (  aFE” RT  6.5,  When the surface pH  2 .  .  exp—  —____  Kuhn and 12 Chan reviewed the reliability of surface pH measurements during nickel electrodeposition using different techniques. Each technique, as they pointed out, had certain drawbacks, associated with the effect of gas bubbles and the current flow. The convenient technique was to use a pH-sensing electrode, such as black Pt/H , Pt-quinhydrone, Sb 2 /Sb and glass pH 3 0 2 electrodes, among which the glass pH electrode was the most widely used. The use of the glass pH electrode with a flat bottom was not mentioned in their review. They proposed new techniques such as using an optically-transparent electrode together with a UV-visible indicator which are not easily feasible. An innovative technique for measuring the surface pH is to use a flat-bottom glass pH electrode together with a very fine gold gauze cathode . In a study carried out in an unstirred dilute 87375 (<0.2 M NiCl ) nickel chloride electrolyte Deligianni and Romanldw 2 41 used as the cathode a 2,000 mesh gold gauze, having an aperture diameter of 7 .tm and a thickness of 5 Im. Their results appear to be reasonable qualitatively. They found that a higher nickel concentration and the presence of boric acid resulted in a lower surface pH. Using a rotating pH electrode, they discovered that the surface pH decreased continuously with increasing RPM . Nevertheless, there are certain dis 31 crepancies in their study. Their tests were based on a potentiostatic step or liner potential sweep. It is not clear whether or not the ohmic drop between the working and reference electrodes was compensated. Furthermore, their curves of the surface pH vs. potential and of the current vs. potential are not well defined. The problems might arise from the nature of the cathode substrate. It may be unlikely that they precoated the gold gauze with nickel and deaerated the electrolyte  Hydrogen evolution during nickel electrodeposition and on a nickel substrate in acidic media  33  before any tests. They studied hydrogen evolution on gold and platinum substrates in the absence of nickel ions whereas a nickel substrate should have been used. In addition, a platinum anode was used. How to prevent the contamination by chlorine gas was not indicated in their paper. An odd explanation was given to account for the plateau observed in the curve of the surface pH vs. potential. They attributed the plateau to the formation of the nickel monohydroxy species (NiOW). In one test conducted in an electrolyte containing 15 gIL 2 •6H at bulk pH 3.5 and c.d. —52 A/m 4 NiSO 0 , 2 the surface pH was observed to rise to an unbelievably high value of 11.7 within 14 seconds. In summary, there are three possible mechanisms for hydrogen evolution during nickel elec trodeposition: (1) Hydrogen evolution via electrochemical reduction of hydrogen ions which is unimportant when pH >3 due to the low concentration of hydrogen ions. (54)  r43.  1130k + e  + Ni -*  Ni  -  H +H 0 2  Ni—H+Ni—H=H)+2Ni  or  2 0 3 H 0 +Ni—H+e=H)+H +Ni  (55)  (2) Hydrogen evolution via homogeneous chemical reaction between water molecules and uni valent nickel ions occurring at the cathode surface after the nickel reduction reaches a peak. Ni(!) +1120 +Ni  —  Ni —H +Ni (Il)+ 0H  Ni-H+Ni-tç,,--H)+mi  or  0+Ni-H+e=H)+OW+Ni 2 H  (56)  (57)  (3) Hydrogen evolution via electrochemical reduction of water molecules which occurs at a more negative potential than in the case of mechanism (2). 0 2 H  +  e  + Ni -*  Ni  -  (58)  H, + OW  ,+Ni -H =H)+mi 41 Ni -H  or  0 +Ni -H+e =H)+0Fr+Ni 2 H  (59)  Activity coefficients in multicomponent nickel chloride solutions  34  Chapter 2 Thermodynamics of Nickel Chloride Solutions 2.1 Activity coefficients in multicomponent nickel chloride solutions The significance of the activity coefficients of the hydrogen and nickel ions has been well described by Peters 91 for highly acidic and concentrated nickel chloride solutions. The importance of activity coefficients has also been increasingly realized in the present experimental work on the pH titrations and surface pH measurements. As the activity coefficient of the hydrogen ion is required to interpret more accurately the experimental results for pH titrations and surface pH behavior, and in the speciation study of nickel species and mathematical modelling of the surface pH, an effort was made to deal properly with this subject. In addition, the pH for the formation of insoluble nickel hydroxide was estimated to provide an upper limit for the surface pH. The dis tribution of nickel species with pH is also important in understanding the solution behavior of nickel chloride and in the selection of nickel species for the surface pH modelling. The single-ion activities in the electrolyte NiC1 -HC1-NaCl can be expressed by the following 2 equations (60)-(6 1): aH+ = YII+ mJ,+  = “3’Ni2+ mN.2+ a+  mN+  =  acr = cr •mcr  (60) (61)  where y is the single-ion activity coefficient and mis the molal concentration. As the activity and the activity coefficient are usually considered as dimensionless parameters, sthctly speaking, the activity should be expressed as: (62)  m a =y—=n  where m jd is the molal concentration at the standard state, which is equal to unit molality and is 3 usually omitted in writing. The activity of nickel chloride can be represented as: aMI =  •  2= (a)  a = (m•  (63)  Therefore, the cube of the mean activity coefficient of nickel chloride is equal to the product of nickel ion activity coefficient times the square of the activity coefficient of the chloride ion. ()3  ()2  =  (YNI2i’  (64)  The important points to be remembered are that the single-ion activity coefficient of neither the nickel ion nor the chloride ion is equal to the mean activity coefficient of NiCl , and the activity of 2 2 is not equal to the sum of Ni NiCl 2 activity plus Cl activity.  Measurement of activity coefficient of hydrogen ion  35  (÷ + y)/2  Ni 7  + acr  (65)  The pH, one of the most important parameters in hydrometallurgy and electrometallurgy, is defined strictly on the basis of the activity of the hydrogen ion, rather than on the concentration of the hydrogen ion.  pH = —loga÷ = —log(y+. mH+)  —logm+  —logç+  (66)  In concentrated electrolytes of nickel chloride, the substitution of the concentration for the activity of the hydrogen ion creates a very serious error. There are three different concentration scales, that is, mole fraction, molality and molarity, among which the molarity is most commonly used in hydrometallurgy. The molality and molarity have the following relationship:  ci  —  -  (67)  p-0.0Ol £C M 1  where C is the molarity and Mis the molecular weight. Corresponding to three concentration scales, there are three different kinds of activity coefficients, i.e., rational activity coefficientfon the mole fraction scale, molal activity coefficient y on the molality scale and molar activity coefficient y on : 61 the molarity scale. For the mean activity coefficients, their relationships are as follows (68)  f=y(l +0.0l8Xv ) m 1  C 1 p + 0.001(18 v f±=Y±  —  ) 1 YCIM  0 p  p—0.00lECM 1 C =y±— 0 mp p° 0 p  = y(1 + 0.001 EmM )— = 4 where:  0 mp  0 p  is the density of pure water  p M  is the density of electrolyte  v m C  (69)  (70)  (71)  is the molecular weight of solute is the number of moles of ions formed by the ionization of one mole of solute is the moles of solute per kilogram of water is the moles of solute per litre of electrolyte  2.1.1 Measurement of activity coefficient of hydrogen ion The principle which the electrode technique uses to measure the pH of the unknown solution X is based on the cell voltage established in the following cell.  Measurement of activity coefficient of hydrogen ion Electrode reversible to hydrogen ion, Soln. X  36 Reference electrode  Salt bridge  The common pH measuring electrodes are hydrogen, glass, quinhydrone and antimony. Their reactions and potential expressions can be represented as follows: (1) Hydrogen electrode Hydrogen electrode is the primary pH electrode and the ultimate standard for the determination of pH. H  +  +e  =  1 H)  (on Pt black),  EH+/H  = —  2.303RT pH F  (72)  (2) Glass electrode Glass electrode is a secondary pH electrode. There is no electrode reaction for the glass pH electrode. The measurement is based on the liquid junction across the glass membrane. EG=EG—  2.303RT pH F  (73)  (3) Quinhydrone electrode Quinhydrone electrode is also a secondary pH electrode. Quinhydrone is an equimolecular compound of benzoquinone (0CJ1 0) denoted as Q, and hydroquinone 4 4 H 6 (HOC 0 H) , sym bolized as H Q. The mixture is slightly soluble in water, approximately 4 g/L at 25°C. 2  Q + 2H + 2e  =  Q 2 H  (on Pt or Au),  Q = E, 11 EQ, Q 112  —  2.303R TpH  (74)  (4) Antimony electrode Antimony electrode is a secondary pH electrode too. The pH measurement is based on the potential established between metallic antimony and antimony oxide. 3 + 6H + 6e 0 2 Sb  =  2Sb  +  0 2 3H  —  ,  Esb , 3 0 2 sb = E:bO,Sb  2.303R TPH  (75)  A comparison of these four pH electrodes is summarized in Table 6. Even though the hydrogen electrode is the ultimate standard thermodynamically for the determination of pH, to set up a reliable hydrogen electrode presents many technical difficulties. How to make an accurate pH measurement in nickel chloride solutions with this electrode is not an easy task. Considering the combined factors of accuracy and convenience, a simple and straight forward measurement as carried out in the present thesis work is to use a glass pH electrode. Using  Measurement of activity coefficient of hydrogen ion  37  Table 6 Properties of pH responsive electrodes 7 Property  Hydrogen electrode  Glass Electrode  Quinhydrone Electrode  Antimony Electrode  pH range  unlimited  0-14  0-8  0-11  pH response  Nemstian  nearly Nernstian  Nemstian  variable  ± 0.001  ± 0.002  ± 0.002  ± 0.1  Precision (pH) Temp., (C)  unlimited  Convenience of measurement  80  low  Measurement time (mm.)  30  high  30-60  <  1  unlimited  medium  high  5  3  Versatility  low  high  medium  medium  Electrical resistance  low  high  low  low  strong reducing  E° drift, variable limited pH asymmetry range, salt error potential, high  action, air must be removed  Disadvantages  resistance, sodium error  poisons such as CN, S, oxidizing 2 ,H 2 SO agents, reducible organic substances, noble metal ions, e.g., Ag  Interference  dehydrating solutions, some  proteins, some amines  colloids, fluorides, surface deposits on the electrode  defective response, not completely reversible some oxidizing agents, Cu ion, anions of hydroxy acids, e.g., oxalates, citrates, tartrates  a solid-state pH meter and reliable pH calibration buffers, the problems associated with the pH electrode can be overcome to a satisfactory extent. Although the theory behind these measurements may not be very rigorous thermodynamically, it appears that the data obtained are quite compatible with the experimental observations in pH titrations and surface pH measurements, and in good agreement with theoretical calculations when the parameters are properly chosen. The principle of the measurement is quite simple. Based on the definition of pH, pH  =  —log a+  and  = YJf*h1fl+  YH÷CH+  (76)  It should be noted here that as an approximation the molality is substituted by molarity for the convenience of calculations. Such an approximation will not produce a serious error. Using the previous equation (67) and the data in Table 37 (in the following section 5.2), it can be calculated thatm  =  1.01 Cfor0.937 2 MNiCl a ndm  =  1.03 Cfor2MNiCl . 2  Measurement of activity coefficient of hydrogen ion  38  Table 7 Activity coefficients of hydrogen ion in aqueous solutions of pure and sulfate—containing nickel chloride in the pH range 1-4 at 25,40 and 60°C Solutions  Temp. (C)  2 +2 M NaC1 0.937 M NiCI  From [HC1] vs.  When corrected for liquid junction potential  Error  (%)  25  7.35  6.25  15  25  2.69  229  15  40  2.35  I  /  60  2.22  /  /  2 2MNiC1  25  8.01  6.51  19  3MNiC1,  25  33.3  27.1  19  3.92 M NiC1 2  25  96.4k  802  17  60  48.8k  2 + 0.365 M 4 0.937 M NiC1 SO 2 Na  25  1.68  1.50  11  4 0.572MNiCl+0.365MNiSO  25  1.34  121  9.7  0.572 M NiC1, + 0.365 M NiSO 4+ 0.365 M 4 SO 2 Na  25  0.935  0.846  9.5  2 0.937 M N1C1  §: Linear fitting was restricted to the linear portion on the right-hand side of the graph in Figure 6-D The activity coefficient of the hydrogen ion was assumed to be constant in the nickel-containing solutions over the pH range to be studied. Starting from a higher pH level, a certain amount of concentrated hydrochloric acid’ was added and the corresponding pH was measured using a combination glass pH electrode . This step was repeated to obtain a series of sets of data points. 2 If the initial concentration of hydrogen ion before adding any hydrochloric acid is assigned the value C , and the concentration of hydrogen ion resulting from the addition of HC1 0 3 the value C , 1 the following equation will hold: 1+C 0=! 114 a CH+ = C  = 1O”/y,+  (77)  1 Hydrochloric acid should be used as highly concentrated as possible in order to keep the volume increase of the system to a minimum. 2 The combination glass pH electrode was purchased from Baxter/Canlab. 3 HC1 was assumed to be fully dissociated and the buffering action from nickel ions was assumed to be negligible.  Measurement of activity coefficient of hydrogen ion  39  By plotting the line of C, versus the activity of hydrogen ion, the activity coefficient of the hydrogen ion can be determined from the reciprocal of the slope. A series of lines of this type is shown in Figures 6-7. These graphs are surprisingly linear except for 3.92 M NiCI 2 at pH above 2 (Figure 6-D) and when the total sulfate concentration reaches 70 g/L SO (Figure 7). The activity coefficients of the hydrogen ion extracted from the slopes by linear fitting are indicated on the graphs and listed in the third column from the right in Table 7. When the solutions contain no sulfate ion, the calculations are straightforward. For the sulfate-containing nickel chloride solutions, the activity coefficients ofthe hydrogen ion were obtained via a certain conversion, which will be shown shortly. The purpose of such a conversion is to deduct the amount of bisulfate. The data in the second column from the right in Table 7 were obtained when the effect of liquid junction potential was taken into account, the calculation method for which is well documented as detailed in Appendix 1. Due to the lack of equivalent conductivities at 40 and 60°C, corrections were not made at these two temperatures. The error in the far right-hand column simply means that the liquid junction potential, if it exists, may give rise to such a discrepancy in the determination of the activity coefficient of the hydrogen ion when using a combination glass pH electrode. A review of the activity coefficients of hydrogen ion in Table 7 leads to the following obser vations. Firstly, ‘H is larger than one in the concentrated pure nickel chloride solutions and increases dramatically with the increase in NiCl 2 concentration. Secondly, the addition of 2 M NaC1 increases value. Thirdly, the addition of sulfate decreases the the value. When 3.92 M NiC1 2 at 25 and 60°C (Figure 6-D) is considered, the lines bend somewhere around pH 2. The reason for the occurrence of this curvature is not well understood, as it was not observed even for 3 M NiCl . If 2 the pH electrode is assumed to perform well in this solution and the liquid junction potential, if it exists, is considered to be constant, the only reason for this curvature might be related to the existence of soluble nickel hydroxy complexes. To confirm this speculation, the distribution curve was calculated and one portion towards the soluble nickel hydroxy complexes was amplified (Figure 8). This graph seems to support the speculation. Three soluble nickel hydroxy complexes, i.e., Ni OH, 2 NiOW and Ni (OH), become gradually important at pH above 2. Due to their very small per 4 centages, one may question that this may result from the inaccuracy of the calculation. In preparing this graph, the error in the calculation itself was controlled on the basis of the mass balance of the total nickel concentration under the condition of Iz[Nz]I/[Ni]T x 100 < 106. Consequently, the calculation error is negligible and the calculated results reflect the real situation if the equilibrium quotients employed are accurate. In the presence of sulfate ions, some hydrogen ions are combined with sulfate ions to form bisulfate ions. Thus, the “activity coefficients” extracted from the slopes of the lines in Figure 7 are not the real activity coefficients of the hydrogen ion. The amount of bisulfate ions must be de  Measurement of activity coefficient of hydrogen ion  40  0.040 0.036 0.032 0.028 0.024 0  0.020  (A  5 I  0.016 0.012 0.008 0.004 0.000  0  0.01  0.02  0.03  0.04  0.05  0.06  0.07  0.08  0.09  0.1  0.07  0.08  0.09  0.1  activity of H+ 0.050 0.045 0.040 0.035 0.030 0  0.025  (B) 0.020 0.015 0.010 0.005 0.000 0  0.01  0.02  0.03  0.04  0.05  0.06  activity of H+ Figure 6 Concentration of hydrogen ion as a function of its activity in nickel chloride solutions  Measurement of activity coefficient of hydrogen ion  41  0.016  0.937 M N1CI2  +  2 M NaCI at 25°C  0.014 0.012  0  0.010  -  -  w  -D 0  (C)  0.008  -  0 0 0  0  0.006 0.004 0.002  -  -  -  0.000 0.00  0.01  0.02  I  I  I  I  0.03  0.04  0.05  0.06  0.07  0.08  0.09  0.10  0.07  0.08  0.09  0.10  activity of H+ 0.0022  0.0020 0.0018 0.0016 0.0014 G)  g (D)  0.0012  o.ooio 0 0  5  0.0008  I  0.0006 0.0004 0.0002 0.0000 0.00  0.01  0.02  0.03  0.04  0.05  0.06  activity of H+ Figure 6 Concentration of hydrogen ion as a function of its activity in nickel chloride solutions (concluded)  Measurement of activity coefficient of hydrogen ion  42  0.16 0.14 0.12 0.10 0.08 ()  =  0.06 0.04 0.02 0.00 0  0.01  0.02  0.04  0.03  0.05  activity of  0.06  0.07  0.08  0.09  0.1  H+  Figure 7 Concentrations of hydmgen plus bisulfate ions as a function of hydrogen ion activity in sulfate-containing nickel chloride solutions at 25C 0.008 [NiOH+]  p  0.007  [Ni(OH)2(aq)]  p  0.006  [Ni(OH)3<-,.] [Ni(OH)4<2->.J  0.005  [NI2OH<3i-,.] [N14(OH)4<4+>]  0.004  A  0.003 0.002 0.001 0.000 -0.001 0  1  2  3  4  5  6  7  8  9  10  11  pH  Figure 8 Sub-section distribution curve of nickel species in 3.92 M NiC1 2 at 25CC  12  13  Measurement of activity coefficient of hydrogen ion  43  ducted for the accurate calculations. The accuracy of the calculated activity coefficients of the hydrogen ion, of course, depends on the reliability of the equilibrium quotients used in the calcu lations. It is obvious from Figure 7 that the following linear relationship holds: aJ,+  (78)  10”=y[HCl]  wherey is the reciprocal ofthe slope. As the concentration ofadded HC1 is equal to the concentrations of hydrogen plus bisulfate ions, equation (78) is equivalent to the following equation (79): aH+ =  10” =  +  [HSO 1 4 )  (79)  In the calculations, only seven species are considered, that is, H, SO, HSO, NiSO , Ni 4 , NiCl 2 and C1. Therefore, it is necessary to find seven equations and to solve for the concentration of the seven species. Besides equation (79), there are three chemical equilibria and three mass balance  equations (80)-(85): Qi  H+SO  =  HSO 2 Q  4 + 2 Ni SO  =  (80)  (81) 4 NISO  3 Q  Ni + 2 cr =NiCl  (82)  ] = [SO4]T 4 [SO1 + [HSO 1 + [NiSO 4  (83)  [Cr] + [NiCfl = [Cl]T + [Cl ifrd nct = [Clip + 10”/y = [Cl] +  (84)  2 + [NiCfl + [NiSO [Ni ] = [NuT 4  (85)  From equation (79) and equilibrium (80) the following relation (86) can be derived:  At a given pH, i.e., a is known, the above equation is a function only of the concentration of the hydrogen ion. From equation (83) and equilibria (80)-(81), equation (87) can be obtained:  Measurement of activity coefficient of hydrogen ion  44  Since [SO] is a function of [114], [Ni 4] will also be a function of [if 2 ] as [SO4]T is known. It 4 follows from equation (84) and equilibñum (82) that: (88)  [Cl]T+a,Jy  [Cu  + 24 [Ni 3 Q ]  =  Combining equations (85) and (88), equation (89) can be obtained:  [Ni24](1  (89)  [CuT + 14 2 [ 3 +Q Ni ]  +Q [ 2 SO]]_[Ni]T=0  As [Ni] and [CuT are known, y can be obtained from the slope of the lines in Figure 7, and [Ni J 2  and [SO] are both a function only of [114], [114] can be solved definitely based on equation (89) using a simple bisection calculation method. The required equilibrium quotients Q , 81 3 at 25°C can be found from the literature ,Q 1 2 and Q that is, log Q ), log Q 4 ) and log Q = -0.17 (2 M NaC1O 4 ). 4 1 = 0.95 (2 M NaC1O 2 = 0.57(1 M NaCIO Alternatively, there is an equation for Q : 19 1 at 25°C  [HSO]  1 logQ =  [H] [SO1  =  1.99—  36 . 2 q1 O  (90)  0.4’Ji  where I is the real ionic strength of solution. For the solution 4 -NiSO NiCl S 2 , Na O the real ionic strength is equal to: 1 z = (4[Ni 1 C 4] + [NiCfl + [114] + [Na4] + [Cu + 4[S0 2 j + [HSO]) 4  I =  (91)  while the formal ionic strength can be expressed as: I  =  +4 CNjSO +  (92)  Thus calculated ionic strengths and equilibrium quotient Q 1 for three sulfate-containing nickel chloride solutions are listed in Table 8. Using four sets of Q 1 values for each solution, i.e., calculated at formal ionic strength and at real ionic strength (Table 8), log Q ), and log Q 4 1 = 0.95 (2 M NaC1O 1 1.99 (at I = 0), the activity coefficient of the hydrogen ion was calculated to see which values of 1 would generate reasonable data. The calculated concentration of the hydrogen ion is plotted Q against its activity in Figure 9. =  Measurement of activity coefficient of hydrogen ion  45  Table 8 Equilibrium quotients for the reaction SO+H=HSO at 25°C based on equation (90)  2+ 0.937 M NiCI  0.572 M NiC1 2+  0.365 M NaSO 4  4 0.365 M NiSO  2 + 0.365 M 0.572 M NiC1 4 + 0.365 M 4 NiSO SO 2 Na  [H], (M)  —0.016  —0.020  —0.029  [SOt], (M)  —0.155  —0.136  —0.323  [Ni], (M)  —0.360  —0.443  —0.328  [HSO], (M)  —0.004  —0.005  —0.014  ], (M) 4 [NiSO  —0.207  —0.224  —0.393  [Ct], (M)  —1.523  —0.899  —0.971  [NiCfl, (M)  —0.370  —0.269  —0.215  Formal!  3.906  3.176  4.271  RealI  —1.98  —1.76  —1.91  log Q (atformall) 1  -0.257  -0.128  -0.313  (atreall) 1 logQ  0.157  0.225  0.178  Solution  §:  The concentrations of individual species are the mean values in the pH range from 4 to 1.  Table 9 Activity coefficients of hydrogen ion in aqueous solutions of sulfate-containing nickel chloride in the pH range 1-4 at 25°C Solution  y  1,1+  log Q 1  log Q 1  (fonnal 1)  (real 1)  (2 M MaC1O ’ 9 ) 4  (1=0)  1.38  1.50  1.68  3.10  18.7  1.23k  /  1.50  I  /  1.09  1.20  1.34  2.27  12.5  0.990*  /  1.21  /  /  0.572 M NiC1 2 + 0.365 M N1SO 4+  0.634  0.734  0.935  2.20  16.7  0.365 M 4 SO 2 Na  0.578  I  0.846  /  /  0.937 M NiC1 2 + 0.365 M 4 SO 2 Na  0.572 M NiCl 2 + 0.365 M NiSO 4  §:  Corrected for the effect of liquid junction potential  log Q 1  =  0.95  log Q 1  =  1.99  Measurement of activity coefficient of hydrogen ion  46  0.07  0.06  0.05  .  x  0.04  0  (A)  C) C 0  0  0.02  0.01  0.00 0  0.01  0.02  0.03  0.04  0.05  0.06  0.07  0.08  0.09  0.1  0.07  0.08  0.09  0.1  activity of H÷ o.o 0.08 0.07 0.06  .  (B)  0.05  0  d (  0.04 0.03 0.02 0.01 0.00 0  0.01  0.02  0.03  0.04  0.05  0.06  activity of H+ Figure 9 Concentration of hydrogen ion as a function of its activity in sulfate-containing nickel chloride solutions at 25CC  Calculation of mean activity coefficients and water activity  47  0.14  0.12  0.10  0.08  (C)  :: 0.02  0.00 0  0.01  0.02  0.03  0.04  0.05  0.06  0.07  0.08  0.09  0.1  activity of H+ Figure 9 Concentration of hydrogen ion as a function of its activity in sulfate-containing nickel chloride solutions at 25CC (concluded) It can be seen that all of the lines in these three graphs in Figure 9 are quite linear. The activity coefficients of the hydrogen ion, which are marked on these graphs, were calculated from the inverse slopes of these lines. For convenient comparison, they are summarized in Table 9. As has been determined, the activity coefficient of the hydrogen ion in 0.937 M NiCl 2 is 2.69. When the sulfate ions exist, the activity coefficient of the hydrogen ion should be less than 2.69, if the total nickel and chloride concentrations are kept constant. Considering this fact, the data in the third column from the right in Table 9 look reasonable. 2.1.2 Calculation of mean activity coefficients and water activity When the single-ion activity coefficient is not available, one normally uses the mean activity coefficient instead. As will be described shortly, the values of the mean activity coefficient and the activity of water are still required to calculate the single-ion activity coefficients. There are a few equations available in the literature for the calculation of mean activity coefficient subject to the limitation of different concentration levels. (1) Debye-Huckel equation 61  Calculation of mean activity coefficients and water activity  AIz+•zj’jT  logf =  0  —  I  0.1 m  (93)  I  1m  (94)  1+Ba’Ji  AIz+•z41i  logf =  + bI  —  ,  48  1+Ba,IT  where A, B are Debye-HUckel constants, which are equal to 0.509 (mole/kg)’ 2 and 0.329 x  1010  1 ( 4 m mole/kg) for water at 25°C, 1 is the ionic strength (mole/kg), d is an ion-size-related parameter (m), and b is a constant adjustable to suit the experimental data. (2) Guggenheim equation 6 logf=—  AIz+.z_l’Ji +bI  i+.qi  (95) ,  Ilm  where b is an adjustable parameter which is equal to Blzzi (3) Stokes-Robinson equation’ 61  AIz+z_Ii h  logy =  ——loga,, —log[1 +0.018(v —h)m]  —  1+Baqi  V  (96)  where h is the hydration parameter of the solute, v is the number of moles of ions for each mole of solute, and m is the molality of the solute. The mean activity coefficients of the electrolyte can be determined experimentally and can be calculated based on certain empirical equations. ’ developed an easy and practical method to calculate the mean activity coefficient 08 Meissner which was claimed to be quite successful in chloride media. The only parameter for Meissner’s theory is a parameter q which is available at 25°C for the pure aqueous solutions of electrolytes, derived by Meissner himself. For convenience, all of the necessary equations are summarized as follows: logy=Iz.z_flogT’  (97)  logf= log[1 +B(1 +0•1J) _B]+logr*  (98)  where:  B  =  0.75  C  =  1+0.055 q ) 3 exp(—0.023 1  logf  *  =  —  0.065q  qi 5 . 0 — 107 i+cqi  (99) (100) (101)  Calculation of mean activity coefficients and water activity  49  The symbol I in equations (98), (100) and (101) is the total formal ionic strength of the electrolyte. For those electrolytes important to nickel electrodeposition, their parameter q values at 25°C are listed in Table 10. Table 10 Characteristic parameter q for pure electrolytes at 25°d ’ 81 Electrolytes  2 NiC1  4 NiSO  HQ  NaC1  2 CaCl  NH C 4 1  q° (25°C)  2.33  0.025  6.69  2.23  2.40  0.82  15  9  4.5 —6  3 —4  15  4.5 —6  Applicable L, (m)  When the calculations are to be undertaken at temperatures other than 25°C, Meissner also supplied equation (102) to correct for the effect of temperature on the q value.  1 F  —  Q”C) 1  —  —  (102)  0.0027(t —25)]  ]  I z+• z_ I  “(25°C)[  Even for solutions of mixed electrolytes, the q value can still be calculated from equation (103) on the basis of the fraction of the ion strength’. qiz—  —  I  I  (jf’i II  IIOL  (103)  I1i,2 m rllj I i=1,3,...iI) j=2,4...I) —  —  1 also derived from the Gibbs-Duhem relationship the following equation (104) to Meissner° calculate the activity of water in a pure solution (only one cation and one anion) of electrolyte : 2  —55.5 ln[a=  I z+.z_I  +2J1.d(lnr±)  (104)  1.0  The first term on the right-hand side of equation (104) can be calculated readily, while the second term is somehow difficult to calculate. Equation (104) can be rewritten as: —55.5 ln[a,J =  —  F  I z+. z_  (105)  +2F(I)  361  00 l 1 at2(W)_ex1 IZZI J  36F(I)] 1000 j  (106)  1 Odd numbers denote calions and even numbers denote anions. 2 There is an error in Meissner’s original equation. The base 10 logarithm should be changed to natural logarithm.  Calculation of mean activity coefficients and water activity  where: F  d(lnr)=  50  d{ln[l +B(1  =  d{ln[1 +B(1 +o1I) _B]+2.303()}  =  —  ‘([0. lIqB (1+0. 1±B(10.1J)_B  —  ‘Ji+0.007590qI 3 exp(—0.0231 )  [  2 1.700(1+CJij  (107)  dl  Equation (107) can be solved numerically. The activity of water in a mixed (more than one cation, or more than one anion, or both) solution of electrolytes is expressed as° : 1 — —  r  i 1 2  Lal ( 2 W  where: = 1 x 2  rLa (W)J 23  r[a (W) 34  C 1 2 12 C +C+C+•  (109) =  (110)  c  x= 2  +C+C+. 1 C 2  (108)  =  +rn+m+•• 1 m 2  where C is in units of mole/L and m in 2 molelkg-H 0 . A few exercises will be carried out to show how good or how poor these calculations are. For aqueous solutions of pure nickel chloride at 25°C, it is shown in Table 11 that the maximum error is less than 1 % for the activity of water and 7 % for the mean activity coefficient of NiCl 2 over the nickel chloride concentration range 0.2-5.0 m. These errors are quite acceptable in practice. For aqueous solutions of pure nickel sulfate, the experimental and calculated mean activity coefficients of the NiSO 4 and the activity of water are listed in Table 12. It can be seen that when the NiSO 4 concentration is less than or equal to 2 m, the errors are quite small, less than 1 % for the activity of water and 6 % for the mean activity coefficient of NiSO . For aqueous solutions of pure 4 hydrochloric acid, the experimental and calculated mean activity coefficients of HC1 are listed in Table 13. It is shown that the error is also small, the maximum being less than 4 %. These three examples for solutions of pure NiC1 , NiSO 2 4 and HC1 demonstrate that Meissner’s method will generate acceptable results for the mean activity coefficients and the activity of water in aqueous solutions of pure electrolytes. As nickel chloride is one of the most important electrolytes in nickel electrodeposition, the activity of water was calculated and plotted in Figure 10 as a function of ionic strength at temperatures 25, 60 and 90°C.  Calculation of mean activity coefficients and water activity  51  Table 11 Mean activity coefficient of NiC1 2 and activity of water in aqueous solutions of nickel chloride at 25C  2 NiCl  1761 Exptl.  (mole/kg)  Calcd. (this work)  a  ) 2 Y±YicL  a  Diff. (%)  ) 2 Y±QiCl  Diff. (%)  0.2  0.868  0.991  0.479  0.991  0.00  0.447  -6.68  0.4  0.907  0.981  0.460  0.980  -0.10  0.439  -4.57  0.6  0.960  0.969  0.471  0.968  -0.10  0.463  -1.70  0.8  1.016  0.957  0.496  0.955  -0.21  0.499  0.60  1.0  1.082  0.943  0.536  0.941  -0.21  0.542  1.12  1.2  1.150  0.928  0.586  0.926  -0.22  0.592  1.02  1.4  1.221  0.912  0.647  0.910  -0.22  0.651  0.62  1.6  1.293  0.894  0.720  0.893  -0.11  0.721  0.14  1.8  1.366  0.876  0.805  0.874  -0.23  0.806  0.12  2.0  1.442  0.856  0.906  0.855  -0.12  0.904  -0.22  2.5  1.633  0.802  1.236  0.803  0.12  1.213  -1.86  3.0  1.816  0.745  1.692  0.748  0.40  1.617  -4.43  3.5  1.969  0.689  2.26  0.694  0.73  2.14  -5.31  4.0  2.100  0.635  2.96  0.640  0.79  2.79  -5.74  4.5  2.202  0.586  3.76  0.587  0.17  3.60  -4.26  5.0  2.292  0.539  4.69  0.536  -0.56  4.60  -1.92  §:  The symbol, 4), is the osmotic coefficient. For NiC1 , 4) = —1000 1na/(18 v 2 ) = —1000 Ina/(54 mMc,) m 1  For precise calculations, attention should be paid to the units of concentration. From the thermodynamic point of view, it is more convenient to use molality, which is usually denoted by the symbol m in the units mole/kg-H 0, as it is independent of temperature and pressure. However, 2 in practical applications, it is more convenient to use molarity, which is normally signified by the symbol C in units mole/L. The conversion between these two units is given in equation (111): Ci  p-0.OOl  M 1 C 1=1  where p is the density of solution (kg/L) and M. is the atomic weight of species i.  (111)  Calculation of mean activity coefficients and water activity  52  Using 0.937 M NiC1 2 (55 g/L Ni ) solution at 25CC as an example, the density of this solution 2 is around 1.107 kgf1i . Therefore, 83  =  p  —  0.001 x (58.7  x  CN+ + 36.45  x  (112) —  0.937 1.107—0.001 x(58.7 xO.937+36.45 x 1.874)  Thus the formal ionic strength is  =  3 x 0.953  =  0953  2.86 m. It can be determined from Figure 10 that  at this ionic strength the activity of water is around 0.94. For the highly concentrated 3.918 M NiC1 2 (230 g/L Ni ), the density of solution is around 1.447 kgfL at 25C 2 . 831  =  1.447  —  (113)  3.918 = 4.207 0.001 x (58.7 x 3.9 18 + 36.45 x 7.836)  It can also be determined from Figure 10 that the activity of water is --0.62 at the formal ionic strength 3 x 4.207 = 12.62 m. Table 12 Mean activity coefficient of NiSO 4 and activity of water in aqueous solutions of nickel sulfate at 25C 4 NiSO  t761 Exptl.  (mole/kg)  Calcd. (this work)  a  a  Diff. (%)  Duff. (%)  0.2  0.533  0.996  0.105  0.997  0.10  0.109  3.81  0.4  0.488  0.993  0.0713  0.994  0.10  0.0732  2.66  0.6  0.465  0.990  0.0562  0.990  0.00  0.0584  3.91  0.8  0.456  0.987  0.0478  0.987  0.00  0.0500  4.60  1.0  0.459  0.984  0.0425  0.984  0.00  0.0446  4.94  1.2  0.472  0.980  0.0390  0.980  0.00  0.0408  4.62  1.4  0.492  0.976  0.0368  0.976  0.00  0.0379  2.99  1.6  0.517  0.971  0.0353  0.973  0.21  0.0356  0.85  1.8  0.551  0.965  0.0345  0.969  0.41  0.0338  -2.03  2.0  0.589  0.958  0.0343  0.965  0.73  0.0324  -5.54  2.5  0.708  0.938  0.0357  0.954  1.71  0.0296  -17.09  §:  For NiSO , .p —1000 1na/(1 8 Ev 4 m) = —1000 1na/(36m, 1 ) 4  Calculation of mean activity coefficients and water activity  53  Table 13 Mean activity coefficient of HC1 in aqueous solutions of hydrochionc acid at 25C Concentration (m)  ExpU.  Calcd. (this work)  Duff. (%)  0.01  0.9048  0.9036  -0.13  0.02  0.8755  0.8736  -0.22  0.05  0.8404  0.8320  -1.00  0.10  0.7964  0.7881  -1.04  0.20  0.7667  0.7538  -1.68  0.50  0.757 1  0.7349  -2.93  1.00  0.8090  0.7783  -3.79  1.50  0.8962  0.8624  -3.77  2.00  1.009  0.977  -3.18  3.00  1.316  1.293  -1.75  4.00  1.762  1.751  -0.62  1.1  1.0  0.9  a) as  0.8  0.7  0.6  0.5  0.4 0  2  4  6  8  10  12  14  16  Ionic strength, (m) Figure 10 The activity of water in aqueous solutions of nickel chloride as a function of ionic strength (I = 2 mNC, 3 )  Calculation of mean activity coefficients and water activity  54  Table 14 Mean activity coefficient of HC1 in mixed aqueous solutions of NiC1 -HC1 at 25C 2 (I = ‘NiC + ‘HCl = 3 mNjc + m,JC, 3 moles/kg) 2 NiC1  HQ  (mole/kg)  (mole/kg)  Expt1.  Calcd. (this work)  Diff. (%)  Exptl.  Calcd. (this work)  Duff. (%)  —0.000  3.00  0.935  1.05  12.3  1.32  1.29  -2.3  0.133  2.61  0.875  0.972  11.1  1.26  1.26  0.0  0.395  1.82  0.761  0.825  8.4  1.16  1.16  0.0  0.632  1.10  0.666  0.703  5.6  1.07  1.07  0.0  0.795  0.616  0.606  0.629  3.8  1.01  1.01  0.0  0.897  0.308  0.570  0.584  2.5  0.979  0.965  -1.4  ‘Y±qlcr)  Table 15 Activity of water in mixed aqueous solutions of 2 NiC1 HC1 at 25C 2 NiC1  HQ  (mole/kg)  (mole/kg)  ExptL  Calcd. (this work)  Duff. (%)  0.801  0.401  0.9379  0.924  -1.48  1.00  0.501  0.9166  0.901  -1.70  1.20  0.602  0.8958  0.875  -2.32  1.51  0.757  0.8577  0.832  -3.00  1.92  0.959  0.7984  0.769  -3.68  2.32  1.16  0.7407  0.704  -4.95  0.802  0.20 1  0.9456  0.939  -0.70  1.00  0.251  0.9301  0.920  -1.09  1.20  0.301  0.9125  0.899  -1.48  1.40  0.351  0.8913  0.877  -1.60  1.82  0.455  0.8499  0.826  -2.81  2.22  0.555  0.7973  0.773  -3.05  2.62  0.656  0.7409  0.7 18  -3.09  Calculation of single-ion activity coefficients  55  For the mixed solutions, calculations of mean activity coefficient and water activity become much more complicated and less reliable. Complete sets of experimental data have not been collected so far. In the following, only limited experimental data will be presented. Khoo et al used the following electrochemical cell to measure the mean activity coefficient of HC1 in the mixed aqueous solution of NiC1 -HC1 at 25CC at five different total ionic strengths, i.e., 0.1,0.5, 1,2 and 2 3 moles/kg. Pt, H 2  (g, 1 atm)  2 (mB) I HC1 (mA), NIC1  I AgC1 I Ag  Under the condition oftotal ionic strength of3 moles/kg, their experimental results and the calculated data based on Meissner’ s method are listed together in Table 14. Examination of the data in Table 14 indicates that the difference between the calculated y> and the experimental values is somewhat large especially when the ratio [NiC1 J/[HC1J is small. There are some reservations regarding Khoo 2 et al’s Ycl) values. It appears that these values are too good to be true in the case of Y±HC1). For the activity of water, Awakura et al made some measurements using a transpiration method. Their experimental results and the calculated water activity based on Meissner’s equation are summarized in Table 15. Although the differences in Table 15 look quite acceptable, there is a question as to Awakura et al’s experimental procedure, since they mentioned that the hydrochloric acid concentration was determined by pH measurement. In such strongly acidic solutions, a glass pH electrode will certainly not perform well. Even for a hydrogen electrode, the reliability of conversion from the pH measurement to the acid concentration is still questionable in such a high level of acid. 2.1.3 Calculation of single-ion activity coefficients The importance of single-ion activity coefficients has been recognized for some time. However, due to many difficulties in determining these coefficients whether experimentally or theoretically, a traditional approximation is to use the available mean activity coefficients instead, or in the worst cases, an assumption of unity has to be made. For nickel chloride solutions, in particular, these two traditional approximations would result in a serious error. As shown recently by Peters , the 91 solution of NiCl -HC1 demonstrated some unusual behavior as regards the activity coefficients of 2 hydrogen and nickel ions, especially in highly concentrated nickel and HC1 solutions. A further theoretical exploration of this system is detailed in the following section, and several useful equations have been worked out.  Single-ion activity coefficients in aqueous solutions of pure electrolytes  56  2.1.3.1 Single-ion activity coefficients in aqueous solutions ofpure electrolytes In a book edited by Pytkowicz, Stokes-Robinson’s hydration theory is introduced. This theory relates the molal mean activity coefficient ofelectrolytes at high concentration to the lowering of water activity and the degree of hydration of ions. (114)  h ln’±=iz÷.z_IlnfDH;lnaW1n[1+0.018(vh)m] where: v h  ---  ---  v. (i.e., number of moles of ions produced by one mole of solute) hydration parameter, proportional to the number of moles of water bound to one  v  +  mole of solute (Ii m z. JDH  ---  ---  ---  =  v.. h  +  v.. hj  concentration of electrolyte, 2 (mole/kg-H 0 ) valence of cation valence of anion the electrostatic contribution (Debye-Huckel equation) —AJT  fDH 1  1 1+Ba4 where: €1 is an ion size parameter (m); I is the ionic strength (mole/kg), and A, B are Debye-Huckel constants, 0.509 (mole/kg)” 2 and 0.329 x 1010 2 (mole/kg)” for water at 25°C, respectively. The 4 m actual values of parameters v, h, and d for electrolytes of interest are listed in Table 16. Table 16 Parameters for Stokes-Robinson’s hydration theory equation 6  ‘  v  h  a, (A)  Range fitted, (m)  3  13  4.86  0.1-1.4  2 Cod  3  13  4.81  0.1-1.0  HC1  2  8  4.47  0.01-1.0  NaC1  2  3.5  3.97  0.1-5.0  2 CaCI  3  12  4.73  0.01-1.4  NH C 4 1  2  1.6  3.75  /  Electrolyte 2 NiCI  A caution should be exercised here for the concept of the hydration parameter h introduced in Stokes-Robinson’s hydration theory. It can be seen from the h numbers listed in Table 16 that they are not equal to the real primary hydration number. For instance, NiCl 2 has six coordinated water molecules in dilute and moderately concentrated solutions, and has only four coordinated water  Single-ion activity coefficients in aqueous solutions of pure electrolytes  57  molecules when highly concentrated HC1 is added. However, h is directly proportional to the hydration number of the solute. The values of the parameter h in Table 16 were derived from curve fitting based on the experimental data. There is a reasonable speculation that the values of the hydration parameter h should decrease as the electrolyte becomes more concentrated. For a general formula of an electrolyte with complete dissociation: (116)  MX =v÷M+v_X  where M denotes a cation and X represents an anion. The symbol V 12 is defined as equal to v + V.. On the basis of the Gibbs-Duhem relationship and Stokes-Robinson’s hydration theory, a general equation has been developed to calculate the single-ion activity coefficients in aqueous solutions of pure electrolyte. z. 12 z_ V ln=1ny—  VZ  —Iz_I  h 1na  ln[1 +0.018(v —vh)m] 12  (117)  12  The detailed derivation of equation (117) is documented in Appendix 2. Three assumptions were used in developing equation (117). (1) Anion (such as chloride ion) is assumed not to be hydrated. (2) Water bound to one or both ionic species is no longer part of the bulk solvent. (3) The Debye-HUckel theory gives correct values for the activity coefficients of hydrated ions on the mole-fraction scale. The specific equations for individual electrolytes can be derived from the general equation (117). For the pure electrolytes of 1:1 univalent chlorides, such as HC1, NaCl, KC1 or NH C1 etc. z = 1, 4 12 = v÷ + V. = 2. When these numbers are placed into equation (117), it Izi = 1, v = 1, V. = 1, and V follows that: ><.  h lna+  1  .  1n[1 +0.018(2—1 xh)m]=lny±—lna 1  h log = logy—1oga h logy_ = 2 x logy—logy = logy+loga In terms of the osmotic coefficient, 4, for 1:1 pure electrolyte:  (118)  (119)  Single-ion activity coefficients in aqueous solutions of pure electrolytes —100Olna rn 182v 1 ...  —1000lna m 12 18v  —  58 (120)  l000xloga, , —2.303 x 4 36m  (121)  loga=—O.01563m  Replacing equation (121)into equations (118) and (119), it follows that: h  logy÷  =  log y  log a  =  logy + 0.00782hm  logy_  =  h logy+loga  =  logy—0.00782hm  —  (122)  (123)  Equations (122) and (123) are exactly the same as those developed by Bates et al. Bates et a1 showed that these two equations were quite successful for solutions of HC1, LiC1, NaCl, KC1, RbC1, C1. For example, 2 m HC1 has y = 1.009, y÷ = 1.42 1, = 0.7 17 while 3 m HC1 has 4 CsC1 and NH = 1.316, ‘y+ = 2.357, ‘y = 0.735. The parameter h in equations (122) and (123) is the hydration parameter of the cation or electrolyte, as the anion is assumed not to be hydrated. According to ‘  Robinson and Bates , when the hydration of the anion is taken into account, equations (122) and 881 (123) can be simply rewritten as equations (124) and (125):  logy  =  log ‘‘± —  logy_  =  logy+  (h—h_) log a, , 4 2  =  1ogy + 0.00782(h h_)m  (h÷—h_) loga 2  =  logy—0.00782(h—h_)mØ  (124)  —  (125)  For the pure electrolytes of 2:1 divalent chlorides, such as NiCl , CoCl 2 , MnCl 2 , MgC1 2 2 and . z =2, Izj = 1, v. = 1, v =2, and v 2 CaC1 12 v÷ + v. = 3. Again, when these numbers are put into the general equation (117), the following equation is obtained: ln + . 2 < 3 l 1 n[1+0.018(3— h.lna =1ny± 1 1xh)m] — +  (126)  h =2xlny±—-lna+ln[1 +0.018(3—h)m] h . = 2 x logy—--loga, 24 logy , + log[1 4  + 0.018(3  —  h)m]  (127)  Single-ion activity coefficients in aqueous solutions of pure electrolytes  2 2x1ny + =lny±+lna--1 _=3x1 n[1+0.018(3— ny±—iny h)m] 2 x logy_= 1ogy+-loga —log[1 + 0.018(3 —h)m] In accordance with the osmotic coefficient, —  ..  —1000 in a, , 4 m 1 18v log a, , 4  =  —  —  —1000 in a, , 4 m 12 18v  —  —  ,  59 (128)  (129)  for 2:1 pure electrolytes:  —2.303 x 1000 x log a, , 4 54m  (130)  —O.02345m  (131)  Substitution of equation (131) into equations (127) and (129) leads to:  2 = 2 x log’y—loga + log[1 logy =  + 0.018(3  2 x logy+ 0.00782hm + log[1  + 0.018(3  2 x logy_ = logy+-loga —log[1 + 0.018(3 =  logy— 0.00782hm  —  log[1  —h)m]  —  —  (132)  h)mj  h)m]  + 0.018(3  —  (133)  h)m]  Figure 11 Calculated activity of nickel ion as a function of its concentration at dif ferent temperatures  160 140 120 100 +  c’J  z  80 60 40 20 0  0  0.5  1  1.5  2  2.5 3 N1CI2, (m)  3.5  4  4.5  5  Figure 11 shows the calculated activity of nickel ion as a function of the con centration of nickel ion based on the equations of (97)-(102), (104) and (132). Bates et a1 have derived the same equations as (132) and (133). For other types of pure electrolytes, no equations have as yet been published in  the literature. Following the same steps as for 1:1 and 2:1 electrolytes, the equations for 3:1, 1:2 and 2:2, etc, electrolytes can be easily obtained.  Single-ion activity coefficients in aqueous solutions of pure electrolytes  60  (1) 3:1 electrolytes, such as Aid 3  3 =3 x 1ogy—1oga + 2 x log[1 + 0.018(4 logy  —  h)m]  (134)  =3x1ogy+0.00782hm+2x1og[1+0.018(4—h)m]  3x1og=1ogy+1oga—2x1og[1+O.O18(4—h)m] =  (135)  1ogy— O.00782hm —2 x log[1 + 0.018(4— h)m]  , 4 —1000 ma) m 1 18v  —  —  —1000 1na m 12 18v  —  —  (136)  —2.303 x 1000 x 1oga 72m  (2) 1:2 electrolytes, such as 4 SO 2 Na  2 x logy÷ = 1ogy—h 1oga =  —  log[1 + 0.018(3 —2h)mJ  1ogy— 4 x 0.00782hm  —  log[1 + 0.018(3  _ = 2 x 1ogy+h 1oga +log[1 + 0.018(3 2 logy =  —  —  —  —10001na m 12 18v  2h)mJ  2h)mJ  2 x 1ogy+ 4 x 0.00782hm4 + log[1 + 0.018(3  —10001na 1 18vm  (137)  —  —2.303 x 1000x1oga  (138) 2h)m] (139)  54m  —  (3) 2.2 electrolytes, such as NiSO 4  ÷ 2 logy  =  h 1ogy—1oga  =  1ogy+0.00782hm  1ogy_  =  h 1ogy+1oga  =  1ogy—0.00782hm  —10O01na m 1 18v  —  —100OIna m 12 18v  —2.303 x 1000x1oga —  36m  (140)  (141)  (142)  Single-ion activity coefficients in aqueous solution of mixed 2 -N1CI HCI-NaCI  61  2.1.3.2 Single-ion activity coefficients in aqueous solution of mixed NiC1 -HC1-NaC1 2 By definition, the mixed solutions contain more than one cation, or more than one anion, or both. Here to be considered is a mixed chloride solution of NiCl 2 + HC1 + NaCl with a common chloride anion. The following symbols have been assigned: mHc,  ---  molality of HC1  hHc,  ---  molality of NiC1 2  hN  ---  molality of NaCI  hNj  ---  hydration parameter of HC1 hydration parameter of NiC1 2 hydration parameter of NaC1  (143)  m =mHC,+mNQ+mN1  XHCI = /m 111 m h  =  ,  (144)  and XN, = mN,/m  mN/m  (145)  +XN. hNIc +XN,.  = XHcI  11 logHc,) + XN log Y±(NicL + XN, log’y±(NCl) X XN + XN,) log 1 (XHC1 + 2 cr = ,  XHCI  —  2  + XN— +XN, —L!  3  2  — X 2 NC log[1 + 0.018(3  —  hN)mNj  log a, , XHCI log{1 4 —  + 0.018(2  —  hHCl)mHC,J  (146) —  XN, log[1 + 0.018(2  —  hNl)mN,1  +log{1 + 0.018[(2 hHCl)mHC, + (3— hN)mN + (2— —  Based on the same Gibbs-Duhem relationship and Stokes-Robinson’s hydration theory and applying the same assumptions as in developing equation (117) for pure electrolytes, the equation (146) has been developed to calculate the activity coefficient of the chloride ion. The details for developing equation (146) are documented in Appendix 3. In terms of the osmotic coefficient, 4), (XHC, + NiCL 2 + XNI) log 7 cr = XHC,  log Y±(HCI) + XN log YiN + XN, log Y±(Nacl)  + X—— + XNZ’  11 _0.00782c1{ 1 c X  —XHC, log[1 + 0.018(2 hHCl)mHCI] —  —XN, log[1 + 0.018(2 + log[ 1+0.01 8[(2  where:  4)=  —  —  3 m N+2 mNI)  2 X log[l + 0.0 18(3  (147) —  hNIC, ) 2 mNC]  hN,)mN,]  hHC,)mHCZ +  —1000lna 18 2v m 1  —  J  +  (3— hN)mN + (2— hNl)mN,] }  —2.303 x 1000 xloga, , 4 =  l m 2 ( 8 mNjcL 2 3 + Hcl + 2 mNo.c,)  (148)  Single-ion activity coefficients in aqueous solution of mixed NiCI -HCI-NaCl 2 Once Ycr is known,  YN2+  and  62  can be easily calculated as follows:  1og+ = 2 x log’±(llcl)— logy  (149)  + =3 x logy— 2 xlogy 2 logy.  (150)  logy+ = 2 xlogyN,)_logy  (151)  The above equations can be applied as well to solutions of pure electrolytes, such as, HC1, NiCl , 2 NaCl, or of any two-component combinations, such as, NiC1 -NaC1 and HC1-NaCl. In 2 -HC1, NiC1 2 the case of mixed solution of HC1-NaCl, XN =0, can be simplified as:  mNQ  hHc,  —i— +X,,  log’y = X 111 log YHc,) + XN, log yq,) + c, 1 X 1 log[1  + 0.0 18(2  + log[ 1+0.01 8[(2  —  —  =0 and XJJJ + XNQJ1 = 1, equation (146)  hHC,)mffci] —X, log{1  hHC,)mHC, +  In accordance with the osmotic coefficient,  —1000 in a = 18 Yv m 1  Once log cr is known,  H 1  —  hN,c,)mN,1  (152)  ,  —  where:  + 0.018(2  log a  (2— hNI)mN,)] }  logy = XHC, log Y±(Hc,) + XN, logyj(NQ) 0.00782hm —XHc, log[1 + 0.0 18(2  J  + log[1 + 0.018(2  —  h )m] (154)  —  hHC,)mHC,l —XNl log[1 + 0.0 18(2  —  —2.303 x 1000 x ioga =  (155)  mHc, + 2 2 ( 18 mN1)  can be readily solved.  and  log ‘y ÷ =2 x log’y±(c1) log 11  (156)  logy+ = 2 x logy±(N1)—logy  (157)  —  Robinson and Bates 881 developed a somewhat different equation shown as follows: log cr = XHCI log Y±(Hci) +XN, log Y±(N,) + log a, , 4 = XHC,  logy(Hc,) +XN, logy<J) 0.00782hm —  (158)  Single-ion activity coefficients in aqueous solution of mixed NiCI -HCI-NaCl 2  63  Equation (158) looks quite different from equation (153); however, the difference between them is not very significant. As XHC, —*0 or XHc, -4 1, these two equations are close enough to each other. The largest difference occurs when XHc, = 0.5. Several numbers are shown in the following to elucidate this point. C1(thLç work) 1  1 +0.018(2—h)m  =  [1+0.018(2  CF(Robj,ison &Bales) 7 ‘  —  —  —  hHc,)mHc,}X  [1+0.018(2  1+ 0.018[2 —(0.5 x 8 + 0.5 x 3.5)] (mild  —  hN,)mN,]’  (159)  + mN.,Q)  1 +0.018(2—8)m ,g1 +O.Ol 11 .S)mN, 3 — 2 ( 8  The value of this ratio depends on the concentrations of HC1 and NaC1. mHc,=mN1  Ycr( work) ‘C1 (Robüison  0.1  0.5  1  1.5  2  2.5  3  0.993  0.963  0.928  0.889  0.848  0.803  0.755  For the mixed solution of NiCl -HC1, mN, =0 and XNd, =0, equation (146) can be simplified to: 2 + 0) log  ,+ 11 (X  +(  11 log Y±(ilc,) +XN 1og, + 0 x log Y±(NQ1) X 1 c r=,  hHd,  XHdj——+XN----+ 0  hN, x__Jloaw —X,log[l + 0.018(2 ‘  —  hHC,)mHC,]  (160) XNjc, log[1 + 0.018(3— hN)mN] 2 —  +log{1 + 0.018[(2  —  hild,)mild, +  —0 x log[1  + 0.018(2  —  hN,) X 0]  (3— hN)mN + (2— hN,) x 0]}  , + Nid1 11 (X ) log 7 2 cr = XHcj logyflC,) +XN  +(  11 h ,  __,J1oaw 2 XHd,—-- +xNICI  —X, log[1 + 0.018(2  —  c,)m 1 h ,] 11 1 (161)  2 X NjC log[1  + 0.018(3— hN)mN]  +log{ 1+ 0.018[(2 hHCI)mHC, + (3— —  In terms of the osmotic coefficient,  ,  Single-ion activity coefficients in aqueous solution of mixed NiCl -HCI-NaCI 2 (XHCI + 2KNIcL,) logy  =  0 + XNc log YjQiC) XHC, log’y  _O.00782(xHC! + XNIC,_f!  J  (2m,c, + 3mNiclj  NiC1, iog[ 1+0.018(3— hNicljrnNc,j +  where:  =  64  —  X, Iog[1 + 0.018(2  Iog{ 1+0.01 8[(2  —  —  hHC,)mHC,}  (162)  hHC,)?nHC, + (3_ hNic1l)nM,]}  (163)  , 4 —10O0lna —2.3O3xl000xloga, = 18 Evgmj mnc, + 2 2 ( 18 mNic, 3 )  891 once developed an equation similar to equation (161). When Jansz  cr  is known, + 11 and y  can be calculated. Thus, log. =2 x log Y±(HCI) log  (164)  + =3 xlogy— 2 x logy 2 logy.  (165)  —  For the solution of mixed NiCl -NaC1, equations similar to (161)-(165) would be derived readily, 2 just replacing HC1 with NaC1. There have been very few reports in the literature concerning experimental measurements of single-ion activity coefficients in mixed solutions. Majima and Awakura° 1 determined the activity of hydrogen and chloride ions in solutions of NaC1-HCI at 25°C via measurement of the electro motive force of a cell composed of a black Pt working electrode and a Ag/AgC1 reference electrode. Their measured activities of hydrogen and chloride ions have been converted to the corresponding activity coefficients and listed in Table 17 together with the calculated data based on our equation for mixed 1:1+1:1 solutions. As can be seen from the data in Table 17, experimental and calculated activity coefficients of hydrogen and chloride ions have the same consistent trend. Although they do not match exactly, the differences between them are not exceedingly large. Also it can be seen clearly that the addition of NaCI raises the magnitude of the activity coefficient of the hydrogen ion. One comment needs to be made in the case of Majima and Awakura’s measurements° . There is a question regarding 1 their treatment of the liquid junction potential. The sign in Henderson’s equation for the liquid junction potential is incorrect as equation (166).  ± uI(CI,d—C,O)  Ejp —Eo E  ‘ —  dF  n  C 1 ZU  1  ‘  z 1 C u 1 0  ) 10 zuI(CId—C i=1  1=1  (166)  Single-ion activity coefficients in aqueous solution of mixed N1CI -HCI-NaCI 2  65  Table 17 Activity coefficients of hydrogen and chloride ions in aqueous solutions of HC1-NaCI at 25CC NaC1  HC1  (M)  (M)  0  0.5  0.808  0.42  0.788  0.69  0.735  0.983  2  0.5  1.48  0.49  1.58  0.60  0.974  0.907  2  1  2.92  0.66  2.08  0.60  1.116  0.882  2  1.5  3.54  0.68  2.77  0.60  1.284  0.855  2  2  4.29  0.78  3.73  0.59  1.486  0.826  3  0.5  2.22  0.60  2.33  0.59  1.170  0.867  2.5  0.5  1.75  0.56  1.92  0.59  1.067  0.887  1.5  1.5  3.03  0.58  2.21  0.61  1.164  0.877  1  2  3.22  0.56  2.31  0.63  1.210  0.872  0  3  3.13  0.77  2.33  0.72  1.293  0.863  1 ExptL°  Calcd. (this work) Y±tpcz)  Table 18 Comparison between calculated and experimental activity coefficients of hydrogen ion in aqueous solution of NiC1 -NaC1-HC1 at 25,40 and 60C 2 hHc, =8,  hMc = 13,  Solution  =  Temp.,(C)  2.33,  hN,  qHc,(2sc)  3.5, 6.69  2.23 10  Exptl. y+ 0.937 M NiC1 2 +2 M NaC1 +  11  11.5  12  Calcd.  25  7.35  3.10  5.69  6.73  7.30  7.90  2 + 0.0374 M HC1 0.937 M NiCI  25  2.69  1.18  1.56  1.68  1.74  1.81  2 + 0.0426 M HC1 0.937 M NiC1  40  2.35  1.15  1.52  1.64  1.70  1.75  0.937 M NiC1 2 + 0.0457 M HC1  60  2.22  1.11  1.46  1.57  1.62  1.68  2 M NiC1 2 + 0.0125 M HC1  25  8.01  2.73  5.36  6.52  7.18  7.90  2 + 0.00297 M HCI 3 M NiCI  25  33.3  6.63  19.0  26.0  30.4  35.4  3.92MNiC1 + 2 0.OO1O5MHC1  25  96.4  13.9  54.4  81.7  100  122  2 + 0.00208 M HC1 3.92 M NiC1  60  48.8  10.1  35.1  50.9  61.2  73.6  0.O138MHC1  Single-ion activity coefficients in aqueous solution of mixed NiCI -HCI-NaCI 2  66  In equation (166), the minus sign in front of RTIF should be a plus sign. This may be only a printing error. Secondly, Majima and Awakura used the equivalent conductivities at infmite dilution to calculate the liquid junction potential. This is questionable in such concentrated solutions. Table 19 Comparison of activity coefficient of hydrogen ion in electrolytes of sodium chloride and calcium chloride at 25°C =  2.23,  qcacL s 2 oc) =  2.40,  qczc,.c) =  11.5,  hN,  =  3.5,  h,qj = 12,  hHcz  =  8  Electrolyte Exptl.  Calcd.  1 MNaC1  1.18  0.98  2 M NaCl  1.85  1.61  —  1.66 (at 10  —  0.06 M HG)  3 M NaCl  3.05  2.89  —  2.97 (at l0  —  0.04 M HG)  4 M NaC1  5.35  5.45  —  5.53 (at l0  —  0.02 M HG)  2 0.937 M CaCl  2.03  1.69  —  1.76 (at l0  —  0.05 M HG)  2 M CaC1 2  7.79  7.14  2 3MCaC1  31.0  1.01 (at l0— 0.O9MHC1)  —  7.27 (at iO  0.0 15 M HG)  31.1 4 (atlO — 0.OO4MHCI)  0.09 0.08 0.07  .0.05  0 a 0 0  .0.04 0  0  z  0.02 0.01 0.00 0  0.01  0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09  activity of H+  0.1  0  0.01  0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09  0.1  activity of H+  Figure 12 Concentration of hydrogen ion as a function of its activity in aqueous solutions of sodium chloride and calcium chloride at 25°C (HG added continuously) For mixed aqueous solutions of NiC1 -HC1 and NiC1 2 -HC1-NaCl, unfortunately, there are no 2 published experimental results for the activity coefficients of the hydrogen ion. As presented in  The pH for the formation of insoluble nickel hydroxide  67  Section 2.1.1, the activity coefficients of the hydrogen ion have been estimated experimentally in the present work. Some of the activity coefficients are listed in Table 18 together with those values calculated using the equations developed in this thesis. In the course of the calculations, it was found that the q value supplied by Meissner t811 for HC1 which is 6.69 did not generate compatible results. One possible reason for this is that the q values given by Meissner are not universal, as they were derived only from pure electrolytes and their validity was never seriously checked for mixed nickel chloride electrolytes. Considering this fact, the q value for HCI was changed sequentially in the calculations to see which one would produce compatible results. As shown in Table 18, when q forHCl is equal to 11.5, the calculated data are in general consistent with the experimental results under the conditions with or without NaCl and a nickel concentration from 0.937 up to 3.92 M. This q value of 11.5 for HC1 produces coincidentally satisfactoiy results for the activity coefficient of the hydrogen ion in solutions of sodium chloride and calcium chloride. The combination glass pH electrode was used for the experimental tests in the solutions of sodium chloride and calcium chloride. The activity coefficients of hydrogen ion were determined from the inverse slope of the linear fitted lines in Figure 12. The calculated results in Table 19 were obtained from the previous equations calculated at two levels of acidity. The comparisons between the experimental and calculated data are quite favourable. 2.2 The pH for the formation of insoluble nickel hydroxide As far as the surface pH during nickel electrodeposition is concerned, it is important to know at what pH insoluble nickel hydroxide starts to form. Although there is a concentration polarization near the cathode surface during nickel electrodeposition, the precipitation pH estimated from the bulk nickel concentration will give a safer upper limit where the surface pH can ultimately go  without the risk of the formation of insoluble nickel hydroxide. The solubility product, K,,,, of . This value does not 911 nickel hydroxide at 25C is cited as 5.47 x 10b6 by the CRC Handbookt account for the effect of ionic strength. It was found in the present calculations that this value was applicable for up to 1 M NiC1 . However, it resulted in some serious errors at higher nickel con 2 centrations. As early as 1962, Ovchinnikova et a1 measured the precipitation pH for the formation of nickel hydroxide in solutions of nickel chloride at temperatures of 25 and 5YC. Their data are reproduced in Figure 13. As can be seen from Figure 13, the pH for the formation ofnickel hydroxide decreases both with increasing the nickel concentration and temperature. As compared with the present results, the trends are actually the same and the differences are only around 0.5 pH unit. Ovchinnikova et al claimed that the addition of 2.05 M NaCl caused the pH for the nickel hydroxide  The pH for the formation of insoluble nickel hydroxide  68  formation to decrease by 0.2 pH unit. The present calculations have virtually confirmed this and the decrease in pH is on the same order of magnitude when 2 M NaC1 is added to 0.937 M NiC1 2 solution.  Figure 13 Dependence of the pH of nickel hydmxide formation on the nickel con centration and temperature in nickel chlo ride solutions 1  0  (Note: The horizontal axis was NiCI 2 (gIL) in the originalpaper. It is believed that this should beNi (gIL) based on our knowledge) 2  0  40  20  60  80  100  120  140  160  Ni, (g/L)  Mesmert S ] summarized the dissociation constant of water in different media Baes and 9 (Table 20) and the equilibrium constants of nickel hydrolysis (Table 21) at 25°C. These equilibrium constants are better named as equilibrium quotients as they are a function of ionic strength. The advantages in using equilibrium quotients instead of equilibrium constants in calculations lie in the fact that the consideration of activity coefficients can be avoided.  Here a few exercises will be carried out to show how to calculate the precipitation pH for the formation of nickel hydroxide in solutions of 0.937 M NiC1 2 (55 g/L Ni) and 0.937 NiC1 2+2M NaC1 at 25°C. If only three species, i.e., Ni , Ct and NiCl are considered in solutions of nickel 2 chloride, the concentrations of free nickel and chloride ions can be calculated as follows: [Ni ] 2 =  [Cu  ([Cl] 3 —{ 1 + Q  —  ([Cl]. — [Ni]T)} [Ni]T)} + { 1 + Q 3 2 + 4Q [NIIT 3 3 2Q  (168)  [Cl]T =  (167)  2 [ 3 1+Q ] Ni  where Q 2 3 is the equilibrium quotient of the reaction Ni  NiCl, [Ni]T and [ClJ are the total nickel and chloride concentrations, respectively. For 0.937 M NiCl , [Ni] = 0.937 M, [Cl]T = 2 2 x 0.937 = 1.874 M. When log Q 8 is used, it can be calculated that [Ni ) 4 J 2 3 = -0.17(2 M NaClO +  Ci  =  0.479 M, [Cl] = 1.4 16 M, [NiCfl = 0.458 M, and the real ionic strength is equal to 0.5 x (0.479 x 4+ 1.416+0.458) = 1.90. The dissociation quotient of water in the medium of NaCI at ionic strength around 2 m and 25°C is (see Table 20): =  The pH for the formation of insoluble nickel hydroxide  69  Table 20 Dissociation quotient of water at 25°C 195 1ogQ=1ogK+  log K  -14.00  aJi  r_+bJ  1+ -qI  a  b (kg/mole)  1.022  Medium  1=O.lm  l=O.Sm  I=lm  I=2m  I—3m  I=3.5m  Lid  -0.68  -0.58  -0.54  -0.52  -0.52  /  NaCI  -0.52  -0.54  -0.35  -0.32  -0.30  I  KC1  -0.46  -0.37  -0.34  -0.30  -0.28  /  4 NaC10  I  I  -0.36  -0.33  -0.31  -0.31  Q = [HJ.[OH]  Note:  Table 21 Equilibrium quotients of nickel hydrolysis at 25°C + yH 2 xNi O =Ni(OH) +yH; 2  Species  x  y  logK  1ogQ = 1ogK +  a  b mx0.1*  m=1  m=3  NiOW  1  1  -9.86  -1.022  0.42  0.15  0.06  2 Ni(OH>  1  2  -19  -1.022  0.30  0.05  -0.04  Ni(OH)  1  3  -30  0  -0.05  -0.21  -0.26  Ni(OH)  1  4  <-44  2.044  -0.34  -0.34  -0.34  3 O 2 Ni H  2  1  -10.7  1.022  I  (0)  /  Ni ( 4 OH)  4  4  -27.74  2.044  I  -0.26  I  10.8  1.022  -0.30  -0.05  0.04  Ni(OH) ) 8 ( 2 , (log  = =  *:  Where mx is the molality of anion in all its forms. For the last row, it corresponds to the reaction: 2 Ni(OH).,,)+2 + O 2fI . Therefore, log Q,,, = log Q, fI=Ni 10 +2 log Q  §:  1 with H to produce Q is the equilibrium quotient for the reaction of a solid hydroxide M(OH) hydrolysis product, a ]. 4 Q = [M(OH)j/[H  The pH for the formation of insoluble nickel hydroxide  logQ =—14.00+  1.02  70  1.022XJL_ 1.90 ‘—O32J=1400+ 032  (169)  —14.02 For the reaction Ni(OH) () + 2H 2 = 10.8+ 310 1ogQ  1  =  2 + 2H Ni 0, the equilibrium quotient is equal to (see Table 21): 2  qy 022  —O.OOx[Cl]= 10.8+  (170)  1 022JT  In terms of dissociation of nickel hydroxide, Ni(OH) (,) = Ni + 20FF, the solubility product, 2 can be expressed as:  Q,,  log Q,, = log Q 10 +2 x log Q =  10.8  +  1.0221i  2x  i+’Ji =  —17.2  By definition,  Q, =  + 3.066  1—14.00 + 1.022JT i+’ii  x  i+-qi  0321’1 = —17.2 + 3.066Ji  )  i+’1i  0.641 (171)  0.64 x 1.90 = —16.64  ].[0H] Accordingly, the precipitation pH for the formation of nickel 2 [Ni . hydroxide can be calculated as: pH If  =  1  =  —  2  log  ] 2 [Ni 2 (H) (Q) 2  =  (172)  1 log = 5.4 (10_14.02)2 x 2.692 x 0.479 2  —  5.47 x 10 is used , the pH would be equal to: t911 1 pH=—log 2  6 5.47x10 (10_14.02)2  x 2.692 x 0.479  (173)  =6.1  These two pH values have a difference of 0.7 pH unit. Estimated from the pH titration curve, the actual pH is between these two values, yet closer to the latter. For the solution of 0.937 M NiCl 2 +2 M NaC1, [Ni]T = 0.937 M, [CuT =2 x 0.937 +2=3.874 M. Using the same Q 3 value, it can be calculated that [Ni ] = 0.294 M, [Ct] = 3.23 1 M, [NiCfl = 0.643 M and [Na] = 2 M. The real 2 ionic strength is equal to 0.5 x (0.294 x 4 + 3.23 1 +0.643 + 2) = 3.53. The dissociation quotient of water and solubility product are represented as: , = —14.00 + 4 logQ,  310 logQ  =  10.8+  1 022’Ji 1 022 x —0.3! = —14.00 +  +qi  1 0224i [Cl]T 4 +O.O 1+i1  (174) —0.3 x 3.53  —14.39 (175)  The pH for the formation of insoluble nickel hydroxide  log Q, 1,  =  =  log Q, 10 + 2 x log Q  —17.2  + 3.0 x  =  —17.2  + 3.0’_  71  0.61 + O. [Cl]T 04 (176)  0.6 x 3.53 + 0.04 x 3.874 = —17.16  And the precipitation pH for the formation of nickel hydroxide is equal to:  1O_176 1 pH=—log =5.2 (10_14.39)2 x 7352 x 0.294 2 If , 31 K  =  (177)  , the pH equals: 911 5.47 x 10 is used in calculationt  (178)  1 5.47 x 10_16 pH=—log =6.2 (10_14.39)2 x 7352 x 0.294 2  In this case, the pH estimated from the pH titration curve, viz., 5.6, is between these two pH values being closer to the former. Table 22 The p11’s for the formation of Ni(OH)S) in different solutions Solution  Temp. (‘C)  2 +2 M NaC1 0.937 M NiC1  Precipitation pH Estimated from pH titntion curve dpH/dV vs. pH  Calcd.  25  —  25  —5.9  40  —  5.5  /  60  —5.0  /  2 M NiC1 2  25  —  5.0  5.0  2 3 M NiCZ  25  —  4.4  4.4  3.92 M NiCI 2  25  —  3.7  4.0  60  —3.4  2 + 0.365 M 4 0.937 M NiC1 SO 2 Na  25  —  6.0  5.7  2 + 0.365 M NiSO 0.572 M NiCI 4  25  —  6.0  5.7  60  —5.5  25  —  2 0.937 M NiC1  0.572 M NiC1 2 + 0.365 M NiSO 4 + 0.365 M 4 SO 2 Na  §:  5.6  5.2 5.4or6.1’  6.3  . 911 This number was calculated using the solubility product from the CRC handbookt  I  / 6.0  Distribution of nickel species in aqueous so’utions as a function of pH  72  Similar calculations can be performed for other solutions. The calculated pH values together with those estimated from the pH titration curves are summarized in Table 22. The calculations at temperatures other than 25CC are not feasible, since, except for the dissociation constant of water , 61 other equilibrium quotients are not available. logK=—  4471 33 +6.0846—O.017053T T  (179)  K increases with increasing temperature. 2.3 Distribution of nickel species in aqueous solutions as a function of pH The significance of the nickel species distribution is realized in understanding what may happen in terms of the predominant nickel species in the solution at a particular pH and as the pH changes, and in interpreting the surface pH behavior. As an example, the calculation procedures are outlined in the case of 2 -Cl-SO-H solutions. With chloride and sulfate present in the solution, the Ni O following fifteen species must be taken into account over the whole range of pH (0—15), although only one or two of them may exist in a significant amount at a given pH. , NiOH, Ni(OH) 2 Ni ), Ni(OH) 2 ), Ni(OH), Ni(OH), Ni 5 OH, Ni 2 (OH), NiCl, Ni50 4 , 4 C1, H or OW, SO and HSO. The equilibrium quotients are assigned to the following reactions. 2+H Ni 0 2  =  2 + 2H Ni 0 2  NiOH + H  =  Ni (OH)) + 2H  —  [NiOH] [H] ] 2 [Ni  (180)  [Ni (OH))j [H] 2  (181)  ] 2 [Ni 2 + 3H Ni 0 2  =  Ni (OH)-i- 3H  [Ni (OH )J [H] 3  (182)  [NiJ +4H =Ni(OH)+4H 2 Ni O  —  [Ni(OH)] [HJ 4  (183)  [Ni9 2+H 21Vi 0 2  =3 OH + 2 Ni  +H  3 0 2 [N1 H ] [H]  —  —  2 + + 4H 4Ni 0 2  =  Ni4 (OH ) + + 4H  2 [NiJ [Ni ( 4 OH ) ] [H] 4  —  (184)  (185)  4 J 2 [Ni 4 + H = HSO SO  [HSO] —  2 1 WI 4 [SO  (186)  Distribution of nickel species in aqueous solutions as a function of pH 4 + 2 Ni SO=NiSQ  Ni + 2 Cr=NiCl  ] [SOt] 2 [Ni —  —  (188)  [NiC1I ] [Cu 2 [Ni  0 =H+OH’ 2 H  Q = [H’]. [OH1  Ni (OH),) = Ni 2+ + 20H  Q  pH  (187)  [NiSO]  —  2  =  (189)  [Ni ’ 2 ]. [0H1 2  (190) (191)  log(y+ [11+])  =  73  For the mass balance ofnickel and chloride concentrations, two cases must be considered separately, that is, with and without the formation of insoluble nickel hydroxide Ni(OH),). When Ni(OH) ) 5 does not form, the total sulfate concentration can be expressed as: ] = [S01• (1+ 4 4 [H + 2 1 Q ] [Ni Q ) 1 + [NiSO 4 [S0 ] 4 T = [S01 + [HSO [SO]=  (192) (193)  [SO4]T 1+Q [H] + 2 1 [Nz Q ]  Total chloride concentration is equal to: (194)  [Ni 3 Q 9 . [Cr] [CuT = [Cr] +[NiCfl = [Cr] + 2 [Cr]=  (195)  [Cl]T . [Ni 3 1+Q ] 2  Total nickel concentration is equal to: ] + [NiOH] + [Ni (011)2] + [Ni(0H)] + [Ni(OH) 2 j 4 [Ni] = [Ni (196)  +2 [Ni2 OH + 4{Ni ] 3 (OH)] + [NiSO 4 ] + [NiCfl 4 ] 21 [Ni]T=[Ni .2+2  21 +2Q  4Q —[Ni.2 4 [Hi  (I  11 Q  j  12 Q  21 2Q  (197)  •2+  + 4Q  + 24 [Ni 2 Q ]  [Ni [Cr] 3 Q ] j + 24 4 [S0  2  .2 +—[Nz J + [Jf4]  1 Q  14 Q  (198)  4 [ 2 Q ] S0 ,. +  [C1JT 3 Q  ‘  [Ni I + 2 Q j [Hi [H} [Ni 3 Q 1 ) 2 [Hi [Hi + 2 1 4 1 +Q 3 [H]  [Ni  ‘]—[Ni]=O  Distribution of nickel species in aqueous solutions as a function of pH  74  From this polynomial equation, the free nickel concentration, 24 [Ni can be solved at a given [NuT, ], [SO4]T, [CuT and pH. As Ni(OH)ZS) does not form, its concentration [Ni(OH)S)] is equal to zero. Once the free nickel concentration is known, the free chloride and sulfate concentrations and other species concentrations can be readily calculated from equations (180)-(188), (193) and (195). When Ni(OH)S) forms, the following equilibrium is assumed to be established. (199)  Ni(OH)) =Ni +2OW 2 =  [N 4 2 ]  [Ni 4 2 ] [0H1 2 =  [Ni 4 2 ] ..  =  2 [H4]  2 = QZJOQW 2 [H4] [0H1 2 = 5.47 x 10 at 25°C  (200)  (201)  2 QI[OH1  log([Nij) = logQ 2  —  2log([OH1) = logQ 2  —  , 4 2(logQ,  —  log([H4]) (202)  =  lo(]_2PH  At a given pH, the free nickel concentration can be obtained from above equation. The concentrations of other soluble species can be calculated in the same way as before. Thus, the concentration of Ni(OH),) equals: [Ni(OH)j  =  ] 24 [NuT —Ni  —  1 4 INiOH  —  Ni(OH))  —  Ni(OH)j  —  INi(OH)1 (203)  —  3 O 2 2[Ni 4 H ]  —  4[Ni ( 4 OH)4]  —  [NiSOJ  —  [NiCfl  The equilibrium quotients used to generate Figures 14-15 are listed in Tables 23-24. These quotients were derived from the data in Tables 20-21. The activity coefficients were determined  experimentally in the present work. The calculation error was controlled on the basis of the mass balance of the total nickel concentration under the condition of lA[Nu]I/[Ni]T x 100 < 10. Several important points may be summarized from the distribution curves in Figures 14-15. (1)  At a given pH, the calculation of the nickel concentration itself is very accurate as the error is controlled in the order of t[Ni]/[NiJT x 100 < 10. Accordingly, these distribution curves reflect the real situation in solution as a function of pH provided that the equilibrium quotients used in the calculations are reliable.  (2)  For 0.1-3.92 M nickel chloride solutions, it is obvious that over the pH range from 0 to 14 the predominant species are Ni 2 and NiCl in the acidic region and Ni(OH)S) in the basic region. These three species may co-exist in the transition region. The amounts of other species,  Distribution of nickel species in aqueous solutions as a function of pH  75  Table 23 Equilibrium quotients in solutions of pure nickel chloride at 25°C 2 0.937 M NiCl  2 2 M NIC1  2 3 M NiC1  3.92 M NiC1 2  log Q 11  -10.28  -10.28  -10.20  -10.11  log Q 12  -19.59  -19.82  -19.94  -20.03  log  Q 1 3 log Q 14 log Q 21 log Q  -30.45  -31.04  -31.56  -32.04  -43.45  -44.03  -44.64  -45.22  -10.11  -10.04  -10.00  -9.98  -27.04  -27.45  -27.90  -28.34  logQ  -14.02  -14.36  -14.70  -15.01  logQ,  91 -16.64or-15.26  -17.10  -17.67  -18.18  -0.17  -0.17  -0.17  8.01  33.3  96.4  ’ 3 logQ 2.69  yH  For the reaction Ni 2  §:  +  Ct  NiC1 in 2 M NaC1O . 4  =  Table 24 Equilibrium quotients in solutions of mixed nickel chloride and sulfate at 25°C 0.937 M NiC1 2+ 2 M NaCI  2+ 0.937 M NiC1 0.365 M 4 SO 2 Na  0.572 M NiC1 2+ 4 0.365 M NiSO  0.572 M NiC1 2 + 0.365 M 4 + 0.365 M 4 N1SO SO 2 Na  log Q 11  -10.29  -10.28  -10.29  -10.28  log  12 Q  -19.82  -19.64  -19.55  -19.59  log Q 13  -31.01  -30.56  -30.33  -30.45  log Q 14  -43.98  -43.57  -43.35  -43.45  log Q 21  -10.03  -10.10  -10.12  -10.11  log  Q  -27.41  -27.13  -26.97  -27.04  logQ  -14.39  -14.04  -13.99  -14.02  logQ,  -17.16  -16.66  -16.64  -16.66  log Q 1  I  0.157  0.225  0.178  log Q ’ 2  I  0.57  0.57  0.57  3 logQ  -0.17  -0.17  -0.17  -0.17  7.35  1.68  1.34  0.935  §:  For the reaction H  ¶:  For the reaction Ni 2  +  SO +  =  SO  HSO from the equation logQ 1 = 1.99— 2.036Ji’(1 =  4 in 1 M NaClO NLSO . 4  +  0.4Iij  Distribution of nickel species in aqueous solutions as a function of pH  76  110 100 90 80 70 60  (A)  —50  z 4 0 30 20 10 0 -10  o  1  2  3  4  5  6  7  8  9  10  11  12  13  14  8  9  10  11  12  13  14  pH 110 100 90 80 70  (B)  -  60.  —  50  Z  40 30 20 10 0 -10 0  1  2  3  4  5  6  7  pH Figure 14 Distribution curves of nickel species in nickel chloride solutions at 25C  Distribution of nickel species in aqueous solutions as a function of pH  77  110 100  -  90  [Ni2+]  80  [N1OH+J  70  [Ni(OH)2(aq)]  60  [Ni(OH)2(s)]  p  [N(OH)3<->]  50 (C) 40  [N(OH)4<2->]  30  [N12OH<3+>)  20  [N14(OH)4<4÷>]  10  [NiCk-]  0 -10 0  1  2  3  4  5  6  7  8  9  1011  121314  pH 110 100  (D)  90  [Ni2-*-]  80  [NiOH+]  p  70  [Ni(OH)2(aq)]  p  -60  [Ni(OH)2(s)]  50  [Ni(OH)3.<->]  40  [Ni(OH)4<2->]  30  [N 120H <3+>]  20  [N14(OH)4.c4-i->]  10  [NICI+]  p  0 -10 0  1  2  3  4  5  6  7  8  9  10  11  12  13  pH Figure 14 Distribution  curves of nickel species in nickel chloride solutions at 25°C (continued)  14  Distribution of nickel species in aqueous solutions as a function of pH  78  110 100 90  [N12+]  80  [N1OH+]  p  70  [Ni(OH)2(aq)]  p  60  [Ni(OH)2(s)J  50  [Ni(OH)3<->]  40  [Ni(OH)4<2->]  30  [N12OH<3÷>J  20  [Ni4(OH)4<4+>]  10  [NiCk-]  A  (E)  A  0 -10 0  1  2  3  4  5  7  6  8  1011121314  9  pH 110 100  -  90  [Ni2+]  80  [N1OH+]  70  [Ni(OH)2(aq)]  60  [Ni(OH)2(s)]  p  [Ni(OH)3<->]  50  (F) 40  [Ni(OH)4<2->]  30  [NI2OH<3-t->]  20  [N14(OH)4<4+>]  10  [NiCk-]  A  0 -10 0  1  2  3  4  5  7  6  8  9  10  11  12  pH Figure 14 Distribution curves of nickel species in nickel chloride solutions at 25CC (continued)  13  Distribution of nickel species in aqueous solutions as a function of pH  79  110  100 90  [Ni2+]  80  [N1OH+]  p  70  [Ni(OH)2(aq)J  p  60  [Ni(OH)2(s)J  50  [Ni(OH)3<->]  40  [Ni(OH)4<2->j  30  [NI2OH.c3+>]  20  [N14(OH)4<4-f>]  10  [NiCk-]  (0)  A  0  -10 0  1  2  4  3  5  6  7  8  10  9  1213  11  pH 110  • 0.937 M NiCI2 +2 M NaCI 100  -  (H)  90  [Ni2+]  80  [N1OH+]  p  70  [Ni(OH)2(aq)]  p  60  [Ni(OH)2(s)]  50  [Ni(OH)3.<->]  40  [Ni(OH)4<2->] XX )( XXX XX XX ) )O XX XX )( XX XX XX )( *i  30  A  [Ni2OH<3+>J  20  [Ni4(OH)4<4±>]  10  [N1CI+]  A  0 I  —10  0  1  2  3  I 4  I  I  I  I  I  5  6  7  8  9  I  I 1011  I  I  121314  pH Figure 14 Distribution curves of nickel species in nickel chloride solutions at 25CC (concluded)  Distribution of nickel species in aqueous solutions as a function of pH  80  110  0.937 M NiCI2 + 0.365 M Na2SO4  100 90  [N12-t-]  80  [N1OH+]  p  [Ni(OH)2(aq)]  p  70  [Ni(OH)2(s)]  60  [Ni(OH)3<->]  (A)  50 [Ni(OH)4<2->]  40  [NI2OHc3+>]  30  [N14(OH)4<4-i->]  20  [NiCk-]  10  [NISO4]  A  0 -10  0  1  2  3  4  5  6  7  8  9  1011  121314  pH 110  0.572 M NICI2 + 0.365 M NiSO4  100  [Ni2+]  90  80 70  [NIOH÷J  p  [Ni(OH)2(aq)]  p  [Ni(OH)2(s)]  -60 [Ni(OH)3<->]  -50  (B)  [Ni(OH)4<2->] Z  40  [N12OH<3÷>]  30  [N14(OH)4<4+>]  20  [NiCk-]  10  [NISO4]  0 -10 0  1  2  3  4  5  6  7  8  9  10  11  12  13  14  pH Figure 15 Distribution curves of nickel species in sulfate-containing nickel chloride solutions at 25°C  Distribution of nickel species in aqueous solutions as a function of pH  81  110  0.572 M NiCl2 + 0.365 M NiSO4  100 90  +  [Ni2+J  0.365 M Na2SO4  [NIOH÷]  80  p  [Ni(OH)2(aq)]  70  [Ni(OH)2(s)] 60 [Ni(OH)3<->]  (C)  50  [Ni(OH)4<2->]  40  [NI2OH<3+>j  30  [N14(OH)4.c4-i->]  20  [NICI+J  10  [NiSO4]  0 -10 0  1  2  3  4  5  6  7  8  9  10  11  12  13  14  pH Figure 15 Distribution curves of nickel species in sulfate-containing nickel chloride solutions at 25CC (concluded) Ni(OH), Ni(OH)), Ni(OH), Ni(OH), Ni OH and Ni 2 (OH) are negligible. The results 4 would be drastically different in the pH range above -P6.4 if the formation of insoluble nickel hydroxide were excluded. (3)  As a side observation, one pH titration was carried out at 25CC using a dilute nickel chloride solution (13.6 g/L 2 •6H NiC1 0 ). As seen from Figure 16, two peaks occurred upon the addition of NaOH solution. On the left of the first peak, sodium hydroxide was consumed to neutralize the free acid in the solution. Between the first and second peaks, sodium hydroxide was consumed to form insoluble nickel hydroxide. The volume between these two peaks was found to be close to the equivalent stoichiometry when the product was Ni(OH)S). On the right of the second peak, almost all of the nickel had been precipitated as insoluble Ni(OH)S) and the further addition of NaOH solution can only result in a pH rise. What is important here is that nickel hydroxide did not dissolve at all even though the pH was held at the level of the end-point for a couple of days. The pH of incipient precipitation u: 9 can be calculated simply using the solubility product at 250C[ = 5.47 x lO_16 2 =[Ni21[0H1  (204)  Distribution of nickel species in aqueous solutions as a function of pH K, 1 5.47 x 10_16 =!log pH =log 2 ] 2 2 [Ni x 13;6/237.71 2 (10’)  ..  82 (205)  =  6.99  This number compares well with Figure 16. If we assume that 99.9 % of the nickel has been precipitated, the pH would be equal to 8.49 which is also in good agreement with Figure 16. From the thermodynamic calculations, it can be known that over the pH range from 0 to 14, insoluble nickel hydroxide may form from the concentrated solutions to the dilute solutions even as low as l0 M (Figure 14-B). The formation of insoluble Ni(OH)) can be ignored only when the nickel concentration goes below  M (Figure 14-A).  10  8  -J  E 6  2: O4  >  -D 0  2  18 -o  0  12  0  (4)  4  8  12  16 20 24 28 32 1.02 M NaOH, (mL)  36  40  Figure 16 pH titration curve of dilute solution of nickel chloride (13.6 g/L •6H NiC1 0 2 , 150 mL sample, 25CC and 2 mL/min. speed)  44  One important point has been made clear through the thermodynamic calculations. The widely held electroactive species, NiOW, does not exist in a significant amount over the pH range 0— 14 under the normal nickel concentration (—1 M). NiOH becomes important only in less concentrated nickel solutions, such as low as 10 M, and at pH above 7.5.  (5)  The calculations of species concentrations have an increment of 0.2 pH unit. Therefore, when the precipitation pH is read from these distribution curves, its actual value should be plus another 0.2 pH unit in most cases.  (6)  In the case of nickel concentration polarization during electrodeposition, the precipitation pH will be a little higher than that calculated from the bulk nickel concentration, and will con tinually rise with the degree of nickel concentration polarization until reaching the limiting condition.  (7)  In solutions of pure nickel chloride, the majority of the nickel is present as free nickel ions j and the nickel chloro-complex (NiCl) when the pH is below the level where the 2 (Ni precipitation of nickel hydroxide starts to take place. The ratio of [NiClj/[Ni ’] increases 2 with the concentration of NiCl 2 or NaC1.  Distribution of nickel species in aqueous solutions as a function of pH (8)  83  In 0.937 M NiC1 2 solution, 4 Ni ( OH) may exist when the pH is between 5.2—6.6 with a maximum percentage of —1.5 % at pH 6. This will explain the later surface pH modelling where it was found that the incorporation of this species would lower the surface pH when it went above —5.  (9)  When sulfate is present in the solution, the percentage of the ion-pair NiSO 4 is quite significant and it must be taken into account in any considerations related to the surface pH. Its con centration rises with the total sulfate concentration.  (10) Whether in solutions of pure nickel chloride or mixed nickel chloride and sulfate, the pH difference between where nickel ions start to precipitate as Ni(OH)S) and where almost no soluble nickel ions are left in solution is not more than 1.5 units.  Experimental apparatus and set-up for nickel electrodeposition  84  Chapter 3 Electrodeposition of Nickel in Various Electrolytes 3.1 Experimental apparatus and set-up for nickel electrodeposition A limited number of electrowinning tests were carried out using the apparatus set-up shown in Figure 17 under conditions similar to industrial operations, in order to obtain data concerning the current efficiency of nickel deposition. The equipment used included a SOLARTRON 1286 Electrochemical Interface (i.e., potentiostat/galvanostat), a RADIOMETER COPENHAGEN ETS 822 titration system (composed of a TTT8O titrator, a PHM82 standard pH meter and an ABU8O autoburette), a COLE-PARMER peristaltic pump, a water bath and a cell. The experimental pro cedure could be computerized almost completely with little manual setup. The starting cathode substrate was copper having an area of 1.5 x 2 cm . The anode, with an area of 1.5 x 2 cm 2 , was 2 pure nickel instead of DSA in order to simplify the experimental procedure and to maintain the nickel ion concentration constant during electrodeposition. The spacing between the anode and cathode was 2 cm. The cell was simply a 200-mL beaker with 170 mL of electrolyte placed inside.  Figure 17 Schematic drawing of the apparatus for nickel electrodeposition tests Unless otherwise indicated, the circulation of electrolyte was made possible by using a peri staltic pump, and the flow rate was controlled to 10 % of the total electrolyte volume per minute. Each test was run typically for four hours. For most of the tests, the current density ranged between 200 and 1,000 A/m , temperature was at 60°C, and the pH of the electrolyte was between 1.1 and 2  Electrodeposition of nickel at 25°C  85  2.5. The pH of the electrolyte during electrolysis was held constant by adding 2.5 M HC1 solution continuously through the RADIOMETER titrator system. The tested electrolytes were NiC1 , 2  4 2 NiCl , NjSO , 4 NiC1 5 2 Na 0 , -NiSO NiC1 4 S 2 Na O , -H 2 NiC1 B 3 0 , -NiSO 2 NiC1 4 B 3 H 0 NiC1 2 NH C 4 1 and NiCl -HC1. The electrolytes were prepared using A.C.S. reagent grade chemicals and 2 deionized water. The concentration of nickel ion changed from 0.937 to 3.92 M. The study of the electrolyte composition may not seem to be very systematical; however, the results certainly reveal much useful information. The current efficiency of nickel was determined according to the weight gain of the cathode after electrodeposition. Since the current passing through the cell and the electrolysis time could be controlled precisely, and the electrodeposition was run for a long period of time, the values of the nickel current efficiency are reliable and quite accurate as regards the measurement itself. The current efficiency of hydrogen evolution can be calculated simply by subtracting the nickel current efficiency from 100. The current efficiency of nickel can also be calculated according to the volume of acid added to the electrolyte during electrodeposition. For this type of test, the electrolyte must be stirred to ensure a uniform electrolyte pH in the cell and to achieve a satisfactory agreement between the two methods of measuring the current efficiency, i.e., on the basis of weight gain and acid volume. 3.2 Electrodeposition of nickel at 25°C A limited number of electrodeposition tests were carried out at 25°C with the main purpose of confirming the later measurements of surface pH. The electrolytes were agitated mechanically rather than circulated. The agitation rate was controlled so as to be same as in the surface pH measurements. Tests were conducted under the following conditions: (1)  0.937 M NiC1 2 (55 g/L Ni ), pH 1.1, 2.0, 2.5 and 3.0 2  (2)  0.572 M NiCl 2 + 0.365 M NiSO 4 (55 g/L Ni 2 and 35 g/L SO), pH 2.5 and 3.0  (3)  0.937 M NiCl 2 +2 M NaCl at pH 2.5  (4)  0.937 M NiCl 2 + 0.485 M 3 B0 at pH 2.5 H  (5)  0.937 M NiCI 2 + 1.31 M NH. C1 at pH 2.5 4  For all of the electrolytes, the results of the electrodeposition tests were in general in good agreement with the surface pH measurements. Using 0.937 M NiCl 2 (55 g/L Ni ) at pH 2.5 as an 2 example, electrodepositions were quite successful when the current density was below 200 A/m . 2 However, at 250 A/rn , the deposit was poorer with a black and greenish surface. At 300 A/m 2 , 2 there was a very large hydrogen evolution, no metallic nickel was deposited at all, and the whole  Electrodeposition of nickel at 25°C  86  surface was covered with a layer of porous green insoluble nickel hydroxide. In 0.937 M NiC1 , 2 the average current efficiencies were —95 % at pH 2.5 and 50-150 A/m , —94 % at pH 2 and 2 100-300 Aim 2 and only —75 % at pH 1.1 and 200-500 A/rn . 2 For the solution 0.572 M NiC1 2 + 0.365 M NiSO 4 at pH 2.5, good agreement was also observed between the electrodepositions and the surface pH measurements. However, the current efficiencies of nickel were slightly lower in this solution than in 0.937 M NiCl , —94 % compared with —95 % 2 at bulk pH 2.5 and 50-150 A/m . 2 When 2 M NaC1 was added to 0.937 M NiC1 , the current efficiency of nickel was high, around 2 —99 % at pH 2.5 and 50-150 A/rn . However, when the current density exceeded 300 A/m 2 , a 2 greenish nickel hydroxide started to precipitate on the cathode surface. The reason for this phe nomenon can be understood when the activity coefficient of the hydrogen ion is taken into account. As reported in Section 2.1.1, the activity coefficient of the hydrogen ion in 0.937 M NiCl 2+2M NaC1 is almost 3 times as large as that in 0.937 M NiC1 . Thus, the concentration of free acid 2 available is only around one third of the latter at a given pH. Generally speaking, the quality of the deposits obtained at 25°C was not very satisfactory. Hydrogen gas pits were present on the cathode surface, the current efficiency of nickel was lower and the maximum feasible current density was reduced. The exception was when boric acid or ammonium chloride was added to the solutions. High nickel current efficiencies could be achieved in both of these cases. Table 25 summarizes the results. It can be seen from Table 25 that in the presence of 0.485 M (30 g/L) 3 B0 the current efficiencies are all above 97 %. High current H , efficiencies could also be achieved when 1.31 M (70 g/L) NH C1 was added to the 0.937 M NiC1 4 2 solution at pH 2.5. The slight increase in the current efficiency with current density in both solutions may result from the fact that the ratio of the nickel reduction rate to the hydrogen evolution rate increases with the cathodic overpotential. In 0.937 M NiC1 2 + 0.485 M 3 B0 the nickel deposits were quite good with a bright surface H , and no black spots at all. In one test at 2,000 Aim 2 for 2 minutes, a bright shiny nickel deposit was still obtained. Only when the current density reached 2,500 Aim 2 did extensive hydrogen evolution take place. In 0.937 M NiCl 2 + 1.31 M NH C1, however, the deposit surface appeared dark grey with many 4 crack lines across the surface. The nature of the nickel deposit surface seems to depend on the duration of the electrodeposition. For instance, an electrodeposition at 6,000 Aim 2 was carried out for 2 minutes. It was still successful with an estimated current efficiency of around 90 % without considerable obvious hydrogen evolution. However, there was H 2 evolution and over time this had a marked deleterious effect on the cathode deposit. Although the surface pH was not measured at  Electrodeposition of nickel at 60°C  87  Table 25 Current efficiencies of nickel deposition in 3 -H 2 NiC1 B 0 and 4 -NH 2 NiC1 C I at pH 2.5 and 25°C (two hours for each run)  ) 2 C.D., (A/rn  2 + 0.485 M 3 0.937 M NiCI B0 H  0.937 M NiQ 2 + 1.31 M NH C1 4  100  97.0 %  96.3 %  200  97.6  97.7  300  98.0  98.5  400  98.3  98.3  500  98.6  98.7  600  98.5  98.6  1,500  98.8  /  such a high current density, it is believed that its value was still below the precipitation pH for Ni(OH)) formation. The deposit was smooth and light grey in appearance. However, when the electrodeposition was run at 600 A/rn 2 for 2 hours, the deposit was very poor even though the current efficiency was still high. There were many cracks across the surface and the deposit peeled off from the substrate in several areas. This phenomenon seemed to be quite strange. At pH 2.5, the majority of the ammonia should be present as NH ion. Except for its buffering function from the reaction NH = H + NH , and the complexing function from NH 3 , it is not known what other effects 3 might have been operative to cause this unfavorable deposit.  3.3 Electrodeposition of nickel at 60°C In the surface pH measurements, temperature was found to have a significant effect on the electrodeposition. For example, at bulk pH 2.5, it requires a several-fold increase in current density to reach a surface pH where insoluble nickel hydroxide starts to form. The conditions for elec trodeposition which were tested at 60°C are listed below: (1)  0.971 M NiC1 2 (57 g/L Ni ) at pH 1.1, and 0.937 M NiCl 2 2 (55 g/L Ni ) at pH 1.5, 2.0 and 2 2.5  (2)  2 at pH 1.1 and 1.5 2 M NiCl  (3)  2 at pH 1.1 3 M NiC1  (4)  3.92 M NiCl 2 (230 g/L Ni ) at pH 1.1 and 2.0 2  (5)  3.555 M NiC1 2  (6)  0.971 M NiC1 2 + 2 M NaCl at pH 1.1  +  0.365 M NiSO 4 (230 g/L Ni and 35 g/L SO) at pH 1.1 and 2.0  Electrodeposition of nickel at 6OC  88  (7)  0.971 M NiC1 2 + 0.647 M 3 B0 at pH 1.1 and 0.937 M NiC1 H 2 + 0.485 M 3 B0 at pH 1.5 H and 2.5  (8)  0.937 M NiC1 2 + 1.31 M NH. C1 at pH 1.1, 1.5 and 2.5 4  (9)  0.937 M NiC1 2 + 0.365 M 4 SO (55 gIL Ni 2 Na 2 and 35 g/L SO) at pH 1.5 and 2.0  (10) 0.606 M NiC1 2  +  0.365 M NiSO 4 (57 g/L Ni 2 and 35 g/L SO) at pH 1.1 and 2.0, and  0.572 M NiC1 2 + 0.365 M NiSO 4 (55 g/L Ni 2 and 35 g/L SO) at pH 1.5 and 2.5 (11) 0.606 M NiC1 2 + 0.365 M NiSO 4  +  2 M NaC1 at pH 1.1  (12) 0.606 M NiC1 2 + 0.365 M NiSO 4 +0.647 M 3 B0 at pH 1.1 H (13) 0.572 M NiC1 2 + 0.365 M NiSO 4  +  0.365 M 4 SO at pH 1.5 and 2.0 2 Na  The results are summarized in Tables 26-29 grouped according to the level of electrolyte pH. By way of examining the data in Tables 26-29, a few important points can be summarized in terms of the current efficiency of nickel and the surface quality of the nickel deposit. Table 26 Current efficiencies of nickel deposition in various solutions at 60C and pH 1.1  f  C.D., (A/rn ) 2  200  500  750  1,000  1,500  2 (57 g/L Ni 0.971 M NiC1 ) 2  93.9  93.4  93.4  92.9  I  2 2 M NiCI  98.6  99.0  98.7  98.9  98.8  2 3 M NiCl  99.2  99.7  99.5  99.6  99.6  0.971 M NiCl 2 + 2 M NaC1  98.2  97.7  97.9  97.5  I  0.971 M NiC1 2 + 0.647 M 3 B0 H  95.1  94.4  94.1  93.8  I  0.937 M NiCI 2 + 1.31 M NH C1 4  95.4  94.4  93.9  93.7  I  0.606 M NiC1 2 + 0.365 M NiSO 4  92.1  91.5  90.8  90.8  /  0.606 M NiCI 2 + 0.365 M NiSO 4 + 2 M NaCI  97.7  97.0  96.6  96.5  /  0.606 M NiCI 2 + 0.365 M NiSO 4 + 0.647 M 3 B0 H  94.5  93.7  94.3  93.3  /  C.D., (A/rn ) 2  1,000  2,000  3,000  4,000  I  3.92 M NiC1 2 (230 g/L N1 ) 2  99.8  99.8  99.8  99.8  /  3.555 M NiC1 2 + 0.365 M NiSO 4  99.8  99.8  99.7  99.8  /  Electrodeposition of nickel at 60°C  89  Table 27 Current efficiencies of nickel deposition in various solutions at 60°C and pH 1.5 C.D., (A/rn ) 2  200  500  750  1,000  0.937 M NiCl 2 (55 g/L Ni ) 24  97.0  96.4  96.3  96.0  2 M NiC1 2  99.2  99.5  99.4  99.4  0.937 M NiC1 2 + 0.485 M 3 B0 H  96.9  97.1  96.8  96.8  0.937 M NiCI 2 + 1.31 M NH C1 4  97.0  97.2  97.3  97.4  0.937 M NiC1 2 + 0.365 M 4 SO 2 Na  95.4  95.2  94.8  94.7  0.572 M NiQ 2 + 0.365 M NiSO 4  93.6  93.3  92.8  92.2  0.572 M NiC1 2 + 0.365 M NiSO 4 + 0.365 M Na 4 S 2 O  91.3  91.0  90.1  89.5  Table 28 Current efficiencies of nickel deposition in various solutions at 60°C and pH 2 ) 2 (A/rn  200  500  750  1,000  98.2  98.8  98.7  98.2  0.937 M NiCI 2 + 0.365 M 4 SO 2 Na  98.4  98.2  98.0  98.2  0.606 M NiCI 2 + 0.365 M N1SO 4  98.7  98.8  98.2  97.9  0.572 M NiCI 2 + 0.365 M NiSO 4 ÷ 0.365 M 4 SO 2 Na  96.7  96.2  95.8  95.6  3.92 M NiCI 2 (230 WE Ni 4) 2  99.9  99.9  99.8  99.8  3.555 M N1CI 2 + 0.365 M NISO 4  99.8  99.8  99.7  99.9  C.D.,  0.937 M NiQ 2 (55  ) 2 WE Ni  Table 29 Current efficiencies of nickel deposition in various solutions at 60°C and pH 2.5 C.D., (A/rn ) 2  200  500  750  1,000  1,500  2,000  99.0  98.8  97.7  97.3  I  /  0.572 M NiCI 2 + 0.364 M NISO 4  98.8  98.6  98.7  98.2  I  /  0.937 M NiCI 2 + 0.485 M 3 B0 H  99.6  I  I  99.3  99.4  99.5  0.937 M NiC.1 2 + 1.31 M NH C1 4  99.5  99.5  99.3  99.4  /  99.4  0.937 M NiC1 2 (55  ) 2 WE Ni  Electrodeposition of nickel at 60°C  90  (1) Electrodeposition of nickel in O.937M NiC1 2 For 0.937 M NiC1 , the electrodepositions at the normal current density used in industry, viz. 2 200 A/rn 2 were all successful at pH 2.5. However, when a higher current density was considered, say up to 1,000 A/rn , the appropriate pH was 1.5. The current efficiency could still reach 96 2 97 % when the current density was changed from 200 to 1,000 A/m . The deposits looked very 2 —  good being bright with no black spots at all. pH 1.1 seemed a little too low in terms of the current efficiency, only 93 94 % being attained corresponding to 200—1,000 A/rn , although quite good 2 —  deposits could be obtained. pH 2.5 was found to be inappropriate for nickel electrodeposition in this solution except at 200 A/rn . For example, the deposit obtained at 750 A/rn 2 2 and pH 2.5 was poor. The SEM photomicrograph (Figure 18) of the cross-section of the deposit reveals that the deposit was not continuous. It appears as if the nickel deposit was adulterated with nickel hydroxide or oxide. In addition, there were cracks and shreds of black strips inside the deposit. Although the surface pH at this current density is around 4 within 100 seconds of electrodeposition (see Figure 56 in later section 4.8), it may rise with the prolonged tirne of electrodeposition due to the decrease in active electrode area frorn the adsorption of hydrogen gas bubbles on the cathode surface, especially around the corners or edges.  Figure 18 SEM photomicrograph of the cross-section of nickel deposit obtained from 0.937 M 2 at 750 A/rn NiC1 , bulk pH 2.5 and 60°C 2  Electrodeposition of nickel at 60C -0.2  -04  -  -0.3 -0.4  -0.5 0.6  91  No obvious H2 evoln.  /  opious H2  ::u no  The cathode surface  unoi  i  0  -0,7  .f -0.8 Region I & III  E  -12 -‘‘- 1.1.1.1.1.1.1.1.1.1.1.1 -200 20 40 60 80 100 120 140 160 180 200 220 240 260 Time, (mm)  Figure 19 The potential of nickel electrode as a  function of time in 0.937 M NiC1 2 at 750 A/rn , bulk 2 pH 2.5 and 6(YC  I  -1.3 —1.4  Region II  No obv,ous H2 evolutio1  -1.2  —  60  65  70  75  Copious H2 evolution 00000 I  I  I  I  80 85 90 Time, (mm)  95  100  105  110  Figure 20 Sub-section potential and nature of nickel cathode as a function of time in 0.937 M 2 at 750 A/rn NiCI , bulk pH 2.5 and 60C 2  The deposition of nickel is cyclic in nature from the viewpoint of hydrogen evolution, which observation is clearly reflected by monitoring the cathode potential (Figure 19). In the absence of the obvious hydrogen evolution, the cathode potential is very low. Hydrogen evolution has a pronounced effect on the cathode potential. There is around 0.25 volt increase in the cathode potential (equivalent to a 0.25 volt drop in the cathode overpotential) once the copious hydrogen evolution takes place. It is more interesting when we look at the corresponding colour of the cathode surface. One part of the curve from Figure 19 was selected, amplified and plotted again in Figure 20 together with the symbols indicating the observations of the cathode surface. The cathode started with a bright surface, but later on there were some black areas appearing around the edges or corners of the cathode surface. With time these areas grew and finally covered the whole cathode surface. During this period, no obvious hydrogen evolution was observed, and the cathode potential remained very low. Once the surface became completely black, in no time copious hydrogen evolution took place. Immediately, there was a sharp increase in the cathode potential. During the hydrogen evolution, the black colour gradually became faint and eventually turned bright again. At this point, the extensive hydrogen evolution stopped and the cathode potential reverted to the lowest level. Here it can be understood that hydrogen evolution is of great benefit in enhancing the mass transfer of hydrogen ions near the cathode surface. It is believed that the very large hydrogen evolution is not caused by the limiting rate of nickel mass transfer, since the electrodeposition at 1,000 Aim 2 at pH 1.5 was very successful. The only difference here is the concentration of hydrogen ions. The incubation period of hydrogen evolution seems to be quite long at pH 2 and 1,000 A/rn 2 (Figure 21). The electrodeposit was fine within the initial 90 minutes. However, at around the 98th minute, copious hydrogen evolution started to take place. As shown by the SEM photomicrograph  Electrodeposition of nickel at 60’C  92  -0.2 -0.3 -0.4 2  -0.5  ui C)  -0.6  ui  -0.7  C,)  I  -20  0  20  40  60  80  100  I  I  120  140  •  I  160  I  •  I  180 200  I  220 240 260  lime, (mm) Figure 21 The potential of nickel electrode as a function of time in 0.937 M NiCl 2 at 1,000 A/rn , bulk 2  pH2and6OC of the cross-section of nickel deposit (Figure 22), the sandwiched layers within the nickel deposit were well defined and could be easily identified. The black areas appeared to be compact and dense. However, the SEM photomicrograph of the morphology in the black area of the nickel deposit (Figure 23) reveals the isolated grains with evident gaps between them. (2) Electrodeposition of nickel in 2 M NiC1 2 and 3 M NiC1 2 When the solutions 2 M NiCl 2 and 3 M NiC1 2 were used, very high current efficiencies were achieved at pH 1.1 with an average value of 98.8 % in 2 M NiC1 2 and 99.5 % in 3 M NiC1 2 in the current density range 200-1,500 2 A/rn Besides, at pH 1.5, the average current efficiency in 2 M . 2 rose to 99.4 % in the current density range 200-1,000 A/rn NiC1 . As long as a satisfactory current 2 efficiency can be obtained, as low a pH as possible should be used. A lower pH offers many advantages, such as, improved conductivity of the electrolyte, and elimination of the possible risk of the formation of insoluble nickel hydroxide on the cathode surface. As far as the surface quality of the nickel deposit is concerned, the use of a lower pH is also beneficial in removing the hydrogen gas pits via the solution flow caused by the hydrogen evolution on the cathode surface. The surface of the deposits from these two solutions at pH 1.1 was light grey and smooth when the current density was below 1,000 A/m . The cathode deposits at 1,000 A/m 2 2 were a little rough at the bottom. There were some small nodules but no black spots on the surface at 1,500 A/m . 2  Electrodeposition of nickel at 60°C  93  31  Figure 22 SEM photomicrograph of the cross-section of nickel deposit obtained from 0.937 M 2 at 1,000 A/rn NiC1 , bulk pH 2 and 60°C 2  Figure 23 SEM photomicrograph of the morphology in the black zone of nickel deposit obtained from 0.937 M NiC1 2 at 1,000 A/rn , bulk pH 2 and 60°C 2  Electrodeposition of nickel at 60C  94  (3) Electrodeposition of nickel in 3.92 M NiC1 2 and 3.555 M NiC1 2 + 0365 M NiSO 4 In the highly concentrated solutions 3.92 M NiC1 2 and 3.555 M NiC1 2 + 0.365 M N1SO , almost 4 100 % current efficiencies of nickel were realized at pH 1.1 even when the current density ranged between 1,000 and 4,000 A/rn . A very smooth electrodeposit was achieved at a current density 2 up to 1,000 2 A/rn There were some small nodules on the cathode surface at 2,000 A/m . . At 3,000 2 and 4,000 A/rn , when the electrodeposition was carried out in 3.92 M NiC1 2 , the surface of the 2 deposits was very rough, yet the deposits were still compact and adhered well to the substrate. However, in 3.555 M NiC1 2 + 0.365 M NiSO , the deposits were poorly adherent and peeled off 4 from the substrate at 3,000 and 4,000 A/rn . It should be pointed out that the flow rate used in the 2 cell was quite slow, as the circulation rate was controlled at only 10 % of the cell volume per minute in a 200-mL cell. As the flow rate in the cell increases, the surface quality of deposit would unquestionably be improved. The operating pH in these two solutions has the potential of being further lowered in practice. (4) Electrodeposition of nickel in the presence of sulfate As regards the addition of sulfate, three compositions were tested, viz., 0.937 M NiC1 2 + 0.365 M 4 SO 0.572 M NiCl 2 Na , 2 + 0.365 M NiSO 4 and 0.572 M NiC1 2 + 0.365 M NiSO 4 + 0.365 M 4 S 2 Na . O The effect of sulfate ions is clearly reflected in the current efficiencies at pH 1.5 as shown in Table 27. The current efficiency decreased with the increase in the sulfate concentration or with the decrease in the chloride concentration. However, at pH 2, there was little difference in the current efficiency whether the solutions contained 35 g/L SO or not. When the sulfate concen tration reached 70 g/L SO, there was around 2 % drop in the current efficiency. The surface quality of the deposits was similar to that in the solutions of pure nickel chloride. One thing has been ascertained from the experiments, that is, the maximum operating current density can be raised to a higher level when sulfate is present. The appropriate pH seems to be around 2 in the solutions of mixed nickel chloride and sulfate whose composition is 55 g/L Ni 2 and 35 g/L SO with a current efficiency in the order of 98 % (Table 28). (5) Electrodeposition of nickel in the presence of2 M NaC1 When 2 M NaC1 was added to 0.971 M NiC1 2 or 0.606 M NiCl 2 + 0.365 M NiSO , substantial 4 increases in current efficiency were observed at pH 1.1 with an average of 97.8 %1 and 97.0 %2, respectively, at current densities 200-1,000 A/rn . The function of NaC1 is dual facilitating the 2 charge transfer of the nickel ion and raising the activity coefficient of the hydrogen ion. The surface quality of the nickel deposits in both solutions was good with a smooth yet slightly grey dark surface. 1 Average increase in current efficiency was 4.4 % compared with that for 0.971 M NiCI . 2 2 Average increase in current efficiency was 5.7 % compared with that for 0.606 M NiC1 2 + 0.365 M NiSO . 4  Electrodeposition of nickel in 2 M NICI 2  +  6 M HCI  95  (6) Electrodeposition of nickel in the presence of boric acid The addition of boric acid was effective too, especially in improving the surface quality of the nickel cathode. As far as the current efficiency is concerned, it increases to some degree (Tables 27 and 29). Furthermore, with the addition of boric acid, the operating current density could be raised to a much higher level, which fact is quite important when the pH must be controlled to be around 2.5 and the electrodeposition must be carried out at room temperature. As discussed previously, the addition of boric acid was very effective even at 25°C. The quality of the nickel deposit at 60°C was found again to be related to the time of electrodeposition, as the surface of the nickel deposit became rougher with time. One electrodeposition was carried out at 6,000 A/rn 2 and pH 2.5 for only 2 minutes. There was no great evolution of hydrogen and the deposit was fine and dark only around the edges. (7) Electrodeposition of nickel in the presence ofammonium chloride When ammonium chloride was present in the nickel chloride solutions, the current efficiency of nickel could also be increased. The most important feature of the addition of ammonium chloride is that it enables the use of a high current density even at pH 2.5. In one test run at 6,000 A/rn 2 for only 2 minutes, the deposit still looked fine, with the dark areas appearing only around the edges without a large hydrogen evolution during electrodeposition. Of course, the surface quality of the deposit depends on the duration of the electrodeposition. In the present tests, the deposits obtained below 750 A/rn 2 for 4 hours were found to be satisfactory. However, at 1,000 Aim , there were 2 many tiny cracks across the surface both at pH 1.1 and 1.5. One concern with the addition of NH C1 4 is the colour of the nickel deposits. The deposits always looked dark grey, not bright at all. 3.4 Electrodeposition of nickel in 2 M NiC1 2  +  6 M HC1  Some preliminary work on the electrodeposition of nickel in 2 M NiCl 2 +6 M HC1 was carried out at temperatures 25, 60 and 95°C. The incentive for this test came from the fact that the nickel activity can be raised dramatically by using an electrolyte which is highly concentrated in hydro chloric acid. Copper was used as the cathode substrate and metallic nickel as the anode. A 200-mL beaker was used for the electrolytic cell and the solution was circulated at a rate of —10 % of the cell solution volume per minute. Each electrodeposition ran for two hours. The results are sum marized in Tables 30-3 1. The current efficiencies of nickel and the anode were determined by weight gain and loss, respectively. The current efficiency of hydrogen evolution was calculated by the subtraction from 100 of the nickel current efficiency. The results listed in Tables 30-3 1 show current efficiencies which are far from optimum. Under all temperatures and current densities, there was always a copious hydrogen evolution on the cathode. At 25°C, the corrosion of copper and nickel was not  Electrodeposition of nickel in 2 M NiCI 2  +  6 M HCI  96  Table 30 Current efficiency of nickel deposition in 2 M NiCI 2 +6 M HQ at 25 and 60’C Temp., (°C)  60  25  -  C.D., (A/rn ) 2  1,000  1,000  2,000  3,000  Nickel CE,  39.0  69.9  72.7  71.9  61.0  30.1  27.3  28.1  99.2  103.7  101.3  101.7  (%)  Hydrogen CE, Anode CE,  (%)  (%)  Comments  poor deposit, good deposit, bright cracked & peeled & smooth surface, off the substrate, large H 2 evoin. large H 2 evoin.  good deposit, bright poor deposit, bright & & smooth surface, a smooth surface, yet little rough at bottom cracked in the centre, edge, large H 2 evoin. large H 2 evoin.  Table 31 Current efficiency of nickel deposition in 2 M NiCI 2 + 6 M HQ at 95°C C.D., (Aim ) 2  1,000  4,000  4,000  (Ti substrate) Nickel CE,  (%)  Hydrogen CE, Anode CE,  Comments  (%)  (%)  61.9  76.8  45.9  38.1  23.2  54.1  290.1  123.7  128.5  good deposit, bright & good deposit, bright & smooth poor deposit, peeled off smooth surface, substrate surface, rough at bottom edge, the substrate, dendrites Cu corroded severely, substrate Cu corroded severely, around edges, Ti cor anode Ni dissolved chemi- anode Ni dissolved chemically, roded severely, large H 2 cally, large H 2 evoin. large H 2 evoin. evoln.  serious, but the nickel deposit was very poor. At 60°C, the copper and nickel were corroded slightly, and although the deposits looked very good except for that at 3,000 A/rn , the current efficiencies 2 were much too low to be acceptable in practice. The worst results occurred when the temperature was raised to 95°C. At this temperature, anode nickel dissolved chernicafly very fast, and abundant hydrogen evolution resulting from the dissolution of the nickel anode could be observed. The copper substrate was corroded severely as well. The situation was even worse when titanium instead of copper was used as the cathode substrate, because the protective oxide film on the titanium surface could no longer sustain the aggressive attack by highly concentrated hydrochloric acid. The deposit obtained with the titanium substrate was also the worst. To determine the reason why the current efficiencies were so low, the activities of nickel and hydrogen ions in this solution were calculated at temperatures of 25, 60 and 95°C based on Meissner’ s  Electrodeposition of nickel in 2 M N1CI 2  +  6 M HCI  97  method and equations developed in the present work. Two sets of q values for HC1 were used. The calculated results are given in Table 32. It can be seen from the data in Table 32 that the activity of the nickel ion increases tremendously when 6 M HC1 is added. The activity of the hydrogen ion rises significantly as well. The data in Table 32 also indicate that the amount of the increase in the activity coefficients of nickel and hydrogen ions decreases with the temperature. Whether the temperature is at 25°C or at 60 and 95°C, the increase in the activity of the nickel ion is always greater than that of the hydrogen ion, as the concentration ratio of [Ni ]/[HJ is only 1/3 while the activity ratio of aN2JaH+ is, for instance, 2 3.2 at 25°C. Table 32 Calculated activity coefficients, activities and electrode potential shifts in 2 M NiCI 2+ 6 M HQ at 25,60 and 95°C =  2.33, hNI =6 and  hHc,  =  4  6.69  o qHc,(2s’c)  Temp., (°C)  11.5  25  60  95  25  60  95  a(waeT)  0.388  0.411  0.434  0.268  0.292  0.320  Y±(NiCL,,  11.10  8.18  6.03  50.8  32.5  20.8  Y±(HCI)  10.45  7.86  5.90  63.2  43.3  28.5  824  437  233  16,283  6,073  2,341  84.8  55.3  35.9  1,407  787  416  1.29  1.12  0.968  2.84  2.38  1.96  1648  874  466  32,566  12,146  4,682  509  332  215  8,442  4,772  2,496  11 aN: ,/a  3.24  2.63  2.17  3.86  2.55  1.88  ÷/a, 2 a Ni.  6.36E-3  7.93E3  1.01E2  4.57E4  5.33E4  7.52E4  -0.065  -0.069  -0.073  -0.099  -0.108  -0.114  YNi2+  H 1 ‘  1 c r  aNi 2 aJ÷  RT  4 J 2 / aH+), lnI\aNI2  (V)  Measurement of current efficiency of nickel from the acid volume  98  The cathodic reduction of nickel ions and hydrogen evolution are two competing reactions, and which one takes priority depends on its electrode potential. At 25°C, the difference in standard potentials between the nickel and hydrogen electrodes is -0.257 volt. At 60 and 95°C, the difference in the standard potentials may not be the same as at 25°C, but it will not be far off. This potential difference means that if we want to have nickel ions take precedence in the electron discharge, the negative difference in the standard potentials must be compensated by a positive shift in the term of activity quotient RT/(2F) xln(aNI2Ja,+) and/or overpotential flH —TINI. Unfortunately, the cal 2 culations ofRT/(2F) xln(aNI2.Ja+) do not give the desired result. In the last row of Table 32, this term brings not a positive but a negative shift. Although the overpotentials of nickel reduction and hydrogen evolution have not been con sidered here, the calculations seem to be compatible with the lower current efficiencies observed in the experiments. The experimental results and the calculations of the activities of nickel and hydrogen ions demonstrate that considering the nickel activity alone is insufficient, and the activity of the hydrogen ion must be taken into account as well when the operating conditions for electrolysis are selected. 3.5 Measurement of current efficiency of nickel from the acid volume The above-mentioned current efficiencies of nickel electrodeposition were all measured by the weight difference of the cathode before and after electrodeposition. These current efficiencies are accurate and reliable. However, they reflect the average current efficiency of nickel over a long period of time. They do not indicate any information about the current efficiency at different times during elecirodeposition. In this regard it should be noted that actually there are two current effi ciencies on the basis of time. The most commonly used current efficiency should strictly be called the overall current efficiency, which deals with a period of time (0 —* t). The more important current efficiency, although seldom used, is called the instantaneous current efficiency, which deals with a very short period of time (dt). The measurement of instantaneous current efficiency is not always easy. During nickel electrodeposition, there are no more cathodic reactions other than the reduction of nickel and hydrogen evolution, i.e., the total current I is equal to ‘Ni + ‘fl2• When the pH of electrolyte is controlled strictly at a constant value during electrodeposition, the current efficiency of nickel can be determined from the amount of acid added. For a galvanostatic (constant current) electrolysis, the instantaneous current efficiencies of hydrogen evolution and nickel deposition can be expressed respectively as: “2  Instantaneous CE (%)=jj-x 11  dVHC,  dt  3 1o  too  X---XC,X9650OX---j--  (206)  Measurement of current efficiency of nickel from the acid volume Instantaneous CEN (%) 1  99  100—Instantaneous 2 CEH ( %) (207) =100—  dVHC,  io  100  7 X--XCHC,X — 96500X—  dt  The symbols in these two equations are: VHC,  ---  CHCI  ---  I ‘Ni  I t  dVHc/dt  ---  ---  ---  ---  ---  volume of HC1 solution, (mL) concentration of HC1, (M) total current applied, (A) current consumed for nickel deposition, (A) current consumed for hydrogen evolution, (A) electrolysis time, (minute) acid volume change per unit time, (mL/min)  Thus, it is shown clearly in equations (206)-(207) that the instantaneous efficiency of nickel is directly proportional to the slope of the line of V, vs. time. The overall current efficiency is defined in equation (208). Overall CE(%)  where  =4- x 100 =f Instantaneous CE(%)  di’  (208)  601 t (coulomb). Therefore, the overall current efficiencies of nickel and hydrogen evolution can be calculated respectively from the following two equations. =  Overall CE (%) 11  =  , 11 dV  1  x CHC, x 96500 x !  x  o  =  VHCI x  (209) i0 XCHC,  100 lj(  Overall CEN(%)  dVHCI  x96500x&_  x-x  CHC, x96500 xJdt  (210) =  100—VHC,xlWxCffC,x96500x-  If the nickel electrodeposition is a steady-state process, the slope dV ,’dt is constant and the 11 instantaneous and overall current efficiencies will be equal to each other.  Measurement of current efficiency of nickel from the acid volume  100  0.7 0.6 0.5 -J  E  -J  E  0.4  C)  0  I  0.3 I’) C%4  U,  c’J  0.2 0.1  0.0  0  10  20  30  40  50  6070  80  90100110120  Time, (mm)  Time, (mm)  1.0  0.28  0.9  0.24  0.8 0.20  0.7 -J  -J  0.6  E  0.5  C) I  U)  c,i  0.16  0.12  0.4  U)  0.3  0.08  0.2 0.04  0.1 0.0  0102030405060708090100110120  0.00  0  102030405060708090100110120  Time, (mm)  Time, (mm)  Figure 24 The acid volume added to the electrolyte as a function of time during nickel electrodeposition from 0.937 M NiCI 2 (55 g/L Ni ) and 0.572 M NiCI 2 2 0.365 M NiSO 4 (55 g/L Ni 2 and 35 g/L SO) at 300 A/rn , different pH’s and temperatures 2 -  The apparatus shown in Figure 17 was used for the tests. A few precautions had to be exercised in order to make accurate measurements. Firstly, the electrolyte must be agitated mechanically instead of being circulated using a pump to make sure that the electrolyte within the cell has a uniform composition and constant pH. Secondly, the pH electrode must be placed in the electrolyte at the required temperature for an adequate time before tests are conducted to ensure that the pH electrode itself has become stable and its temperature has reached the electrolyte temperature. Thirdly the effect of an electric field on the pH reading is important. Depending on the current density, electrolyte conductivity and the position of the pH electrode relative to the cathode and anode in the cell, the shift in the pH reading caused by the electric field can range from ± 0.02 to ± 0.10. Such a shift in the pH reading may not be critical for other types of experiments. However, it is quite crucial when the major objective is to determine the current efficiency from the pH change of the electrolyte. In the present experiments, the pH shift from the electric field was corrected  Measurement of current efficiency of nickel from the acid volume  101  prior to conducting any tests. The experimental results are presented in Figure 24. The tests at 60°C were not very successful. Therefore, the lines for 60°C in Figure 24 were simply calculated from the weight gain of the cathode after electrodeposition. All of the data shown in Figure 24 indicate that the acid volume added to the electrolyte increases linearly with time as the electrodeposition proceeds. The poorer linearity at the lower pH especially in the electrolyte 4 -NiSO results from the limited resolution of the pH meter (±0.01). 2 NiC1 The linear relationship between the acid volume added and the electrolysis time shows that the nickel electrodeposition is quite stable in both these electrolytes. In other words, the hydrogen evolution remains relatively constant during nickel electrodeposition, and thus the instantaneous current efficiency is equal to the overall current efficiency. The possible errors resulting from the ±0.01 variation of the pH meter are analyzed as follows. By definition pH is equal to: (211)  1  pH =—log(a+)—log(yff+. CH+)  CH+=_lWm  ..  YH+  If the activity coefficient of the hydrogen ion is assumed not to change during electrodeposition, the change in the hydrogen ion concentration can be obtained as: (C+) H 2 =  10—ApH  k  (10—ApH  .  i.e., (CH+)  —  1 ACH÷=1O”•(lO”_  (CH+)  =  (CH+)l  —  1)  (212)  (213)  1)  For 200 mL of electrolyte and using 2.5 M HC1 solution to adjust the pH, the change in the acid volume is equal to: All  I  r  tCH +x200  —  200  1 pH  j — 1 ApH  (214) —  iJ  ,,  Lj..)  The corresponding change in the current efficiency for a period of two hours can be expressed as: (215)  iXV+(mL) x i0 x 2.5 x 96500 (%)= 11 ACE  72001  xlOO  The calculated errors are summarized in Tables 33-34. It can be seen from Tables 33-34 that the errors in current efficiency due to a ±0.01 pH shift are acceptable in practice. However, the errors increase at the lower pH’s and at a lower activity coefficient of hydrogen ion. As shown in Table 34, the error reaches ±5 % at pH 1.1 and 25°C in the electrolyte . 4 2 NiCl NiSO  Measurement of current efficiency of nickel from the acid volume  102  Table 33 Errors in current efficiency due to ±0.01 pH shift in 200 mL 0.937 M N1C1 2 (55 g/L N1 ) 2 at 300 A/rn 2 (0.09 A) for 2 hours (y,,+ =  pH  Temp., (‘C)  ApH  =  2.69 at 25°C, 2.35 at 40°C, 2.22 at 60°C)  -0.01  ApH = +0.01  ApH = -0.005  ApH = ÷0.005  AV S 1 MHC  ACE  IXVMHa  ACE  AVMfl  ACE  AVISMHQ  H2  (mL)  (%)  (mL)  (%)  (rnL)  (%)  (mL)  (%)  2.0  25  0.0069  0.3  -0.0068  -0.3  0.0034  0.1  -0.0034  -0.1  1.5  25  0.0219  0.8  -0.0214  -0.8  0.0109  0.4  -0.0108  -0.4  1.1  25  0.0550  2.0  -0.0538  -2.0  0.0274  1.0  -0.0270  -1.0  1.5  40  0.025 1  0.9  -0.0245  -0.9  0.0125  0.5  -0.0123  -0.5  1.5  60  0.0265  1.0  -0.0259  -1.0  0.0132  0.5  -0.0130  -0.5  1.5  60*  0.0265  0.3  -0.0259  -0.3  0.0132  0.2  -0.0130  -0.2  §:  This test was run at 1,000 Ahi . 2  Table 34 Errors in current efficiency due to ±0.01 pH shift in 200 mL 0.572 M N1C1 2 0.365 M 4 (55 g/L Ni NiSO 2 and 35 g/L SO) at 300 Mn 2 (0.09 A) for 2 hours -  (y+ = 1.09 at 25°C, 0.690 at 60°C) pH  Temp., (‘C)  ApH = -0.01  ApH = +0.01  ApH = -0.005  ApH = +0.005  AVMHØ  ACE  AVz.5MH  ACE  .AV 2 SMHCI  ACE  AV 5 1 MHCI  H2  (rnL)  (%)  (mL)  (%)  (mL)  (%)  (mL)  (%)  2.0  25  0.0171  0.6  -0.0167  -0.6  0.0085  0.3  -0.0084  -0.3  1.5  25  0.0541  2.0  -0.0528  -2.0  0.0269  1.0  -0.0266  -1.0  1.1  25  0.1358  5.1  -0.1327  -4.9  0.0675  2.5  -0.0667  -2.5  2.0  60  0.0270  1.0  -0.0264  -1.0  0.0134  0.5  -0.0133  -0.5  2.0  60*  0.0270  0.3  -0.0264  -0.3  0.0134  0.2  -0.0133  -0.2  §:  This test was run at 1,000 A/rn . 2  Chapter 4 Surface pH Measurement during Nickel Electrodeposition  103  Chapter 4 Surface pH Measurement during Nickel Electrodeposition The electrodeposition ofnickel often does not proceed at 100 % current efficiency. The balance of the current is consumed normally in hydrogen evolution. Due to this hydrogen evolution, the hydrogen ion is depleted near the cathode surface. Therefore, the pH near the cathode surface is always higher than that in the bulk electrolyte. What affects the electrode process is really the cathode surface pH rather than the pH in the bulk electrolyte. For hydrogen evolution, the reactants are hydrogen ions in acidic media and water in basic media 2H30 0 2 2H  +  +  2e  2e  =  =  2 H  +  112 +  0 2 2H  20ff  in acidic media  (216)  in basic media  (217)  The effect of potential  at the outer Helmholtz plane on the cathode surface pH may not be neglected if the ionic strength is very low. Dissolved oxygen, if present, will affect the cathode surface pH too. °2()+  4H  +  4e  =  21120  (218)  There are three basic factors which can depress the increase of the cathode surface pH. The first one is the mass transfer rate of hydrogen ions towards the cathode surface. In this regard, the lower bulk pH and the thinner thickness of the diffusion layer brought about by vigorous agitation will prevent to a large extent the cathode surface pH from rising. The second factor is the proton donating pH buffers, such as boric acid 3 B0 or the bisulfate ion, HSO. The third factor is the H , hydroxylconsuming pH buffers, such as NiOH, Ni (OH). As the cathode surface pH increases, 4 the following equilibrium reactions will shift to the right to generate more protons: HS0  —  S0+H  (219)  +H —Ni0H+H 2 Ni 0  (220)  2 + 41120 4Ni  Ni ( 4 0H) + 4H  (221)  80 2H (H + 2 ) Ni 3  (222)  —*  +3 2 Ni B0 2H  —  When the supply of hydrogen ions is unable to meet their depletion rate, the cathode surface pH will rise and eventually lead to the formation of insoluble nickel hydroxide on the cathode surface. 2 +21120 Ni  —  Ni (0H)) + 2H  (223)  The formation of insoluble nickel hydroxide must be avoided during nickel electrowinning.  Experimental apparatus and set-up for surface pH measurement  104  4.1 Experimental apparatus and set-up for surface pH measurement The measurement of the cathode surface pH was carried out using an apparatus constructed in-house. The idea of using such an apparatus originated from Romankiw’s work ’ 18 However, many improvements were made to their apparatus and experimental procedures. Solid-state electronic instruments were used in the present investigation and all measurements were almost completely computerized. A schematic drawing of the experimental arrangement is shown in Figure 25.  flat bottom p1-1 electrode  Figure 25 Schematic drawing of the apparatus for the surface pH measurement and associated equipment The whole system consisted of a potentiostat (SOLARTRON 1286 Electrochemical Interface), pH meter (RADIOMETER PHM82 standard pH meter), a pH stat (RADIOMETER ETS 822 titration system), a general combination glass pH electrode to control the pH of the bulk electrolyte, a special combination glass flat-bottom pH electrode (ORION) to measure the surface pH, a micrometer to adjust the position of the flat-bottom pH electrode, a nickel anode, a gold gauze cathode, and a computer to control the instruments and to take measurements. The whole measuring assembly was placed in the cell at an angle of around 45 degrees. The gold gauze, serving as the cathode, had an exposed diameter of —15 mm. During the experiments nickel was deposited on the front side of the gold gauze which had been preplated with a layer of nickel. The flat-bottom pH electrode was placed next to the back side of the nickel-plated gold gauze. The distance between the gold gauze cathode and the sensor of the flat-bottom pH electrode could be adjusted using a micrometer.  Experimental apparatus and set-up for surface pH measurement  105  During the measurement, they actually contacted each other. The pure metallic nickel served as the anode and had a diameter —15 mm. The nickel anode was placed directly below the gold gauze with a space of -20 mm. The cell could contain 250 mL of solution. Four types of gold gauzes were tested, viz, 100-, 200-, 500- and 1,000-mesh. However, only the results obtained with the 500-mesh gold gauze are presented in this thesis. The dimensions of the gold gauzes are listed in Table 35, both according to the manufacturer’s specifications and those estimated from the SEM photomicrographs. The calculation of the effective area will be shown. In terms of the effective area, the 500-mesh gauze was the most suitable among the four gold gauzes listed in Table 35. The diameter of the pH sensor of the flat-bottom pH electrode is around 8 mm. For 500-mesh gold gauze, as an example, this pH electrode reflects an average pH value covering over twenty thousand [1/4 x it x 80002 / (17.0  + 33.0)2  20,000] holes on the gauze surface.  Table 35 Dimensions of gold gauzes Mesh size  Estimated from SEM photos wire diameter  space between  Effective area  from manufacturer 1 wire diameter  between  Effective area (%)  space  (jim)  (jim)  (%)  (jim)  wires (jim)  100  /  /  /  19.8  234  24  200  19.5  107  45  14.7  112  34  500  17.0  33.0  89  11.4  39.4  63  1,000  12.4  12.4  118  7.4  18.3  77  wires  §: 1 mesh = 1 line per inch ¶: The gold gauzes were purchased from Buckbee-Mears Co., 245 E-6th St., 6th floor, St. Paul, MN 55101, U.S.A. The temperature was held constant at 25°C and the solution was stirred gently during pH measurement for the sake of uniform bulk pH. The bulk pH was controlled to be constant. Before the experiments, the solution was deaerated with nitrogen for 20 minutes to remove dissolved oxygen. Unless otherwise stated, the bare gold gauze was always precoated with a layer of —0.5 jim thick nickel film (equivalent to the deposition at 50 AIm 2 for 300 seconds) before any tests. This thickness was considered to be quite conservative from an examination of the nickel-coated gold gauze. After each test, the gold gauze was cleaned via anodic dissolution of the previously deposited nickel layer. The surface pH was measured as a function of time at a given current density. The deposition time was typically 150 seconds, and the surface pH values presented in this section were the readings  Electrochemical properties of gold in chloride solution  106  at the end of the experiment or averaged in the stable region. The curves of pH vs. time were recorded for each run. The surface pH’s measured with 500-mesh gold gauze are believed to be a fair representation of the true surface pH. As seen from Table 35, 500-mesh gold gauze has an 89 % effective area and a 17.0 urn wire diameter, in comparison to 45 % and 19.5 im for 200-mesh gold gauze. If the flat-bottom pH electrode was brought into direct contact with the gold gauze, the distance between the glass membrane sensor of the pH electrode and the electrochemical reaction sites was varied due to the “mushroom” shape of the wires of the gold gauze. For the 500-mesh gold gauze, the maximum distance from the “mushroom” top of the wires to the membrane sensor was around 36 % of the wire diameter (refer to Figure 30 or Figure 36), i.e., around 6.1 jim. This distance is well inside the normal diffusion layer thickness which is in the order of 100-300 p.m. To assess the reproducibility of the measurements, it was felt that the measurements were fairly good at low current densities, but small oscillations occurred at higher current densities where hydrogen evolution was significant. The reasons for this instability are due partly to the change in the hydrodynamics of the solution immediately adjacent to the gold gauze caused by hydrogen evolution, and partly, as will be discussed later on, to the surface roughness of the gold gauze from the micro point of view. The gap between the membrane sensor and the electrochemical reactions sites is inevitable. Thus trapped hydrogen bubbles could not be eliminated completely and this led to the instability in the surface pH measurements. The adsorbed hydrogen bubbles on the outside of the wires might also play a role in this instability, but the effect should be much less severe since the whole pH-measuring assembly was placed in the cell at an angle of around 45 degrees and the solution was stirred gently during the experiments. Initial experiments were carried out using a 200-mesh gold gauze as the cathode substrate. Compared with the results of 200-mesh gold gauze, the surface pH’s measured with 500-mesh gold gauze are lower due to the higher effective electrode area. Unless otherwise stated, all of the results to be presented were obtained with the 500-mesh gold gauze, 0.5 p.m nickel coated, and 150 seconds of electrolysis time. 4.2 Characterization of the gold gauze 4.2.1 Electrochemical properties of gold in chloride solution The following estimation shows that the gold ion will not precipitate before nickel if gold ions are brought into solution accidently, even though there is a strong gold chioro-complex. Based on the electrode potentials’ 911 and the solubility product of gold hydroxide at 25°C, AuCl + 3e + 3e 3 Au  —  —  Au + 4C1  Au  ° = 1.002 volts 1 E  (224)  1.498 volts  (225)  =  Investigation of new 500-mesh gold gauze  Au(OH))=Au + 3 3OW  107  the overall formation constant for the reaction Au 3 + + 4C1 —  (226)  5 = i0 K —* AuCl  3F(E —E ° 2 ) 3 x96500x(1.498— 1.002) 0 1 2.303RT — 2.303x8.314x298  will be: (227)  252  The pH for Au(OH) 3 precipitation can be expressed in equation (228). 2 4 (K 1 1 1 pH =—lod —f-- I+—log[Cl1 +—log 3 1K) 3 3 [AuC1J 4 1 1 = 7.31 +—log[Cl1 +—log 3 3 [AuClE]  1 1 = 7.71 +—log (at 2 M Cl) 3 [AuC1J  (228)  -  Equation (228) shows that the gold ion will not precipitate when the pH is below 7.71 even at a high concentration as high as 1 M. Besides, by comparing the standard potential of 1.00 volt for the AuCl/Au electrode with E 2+,NI = -0.25 volt, it is evident that gold will not dissolve preferentially 1 over nickel. Therefore, gold gauze is an inert cathode substrate and will not have any misleading effect on the surface pH measurement. 4.2.2 Investigation of new 500-mesh gold gauze Physically, the gold gauzes have a different visual appearance; one side is very shiny while the other side is dull. A series of SEM studies revealed some important information. The SEM photomicrographs taken for 500-mesh gold gauzes are shown in Figures 26-30. It can be seen from Figures 26-27 that the spaces between the wires are quite uniformly distributed over the whole gauze surface. Figures 28-29 show the SEM photomicrographs at a greater magnification for the 500-mesh gold gauze to reveal more details. Most of the important information can be ascertained from the SEM photomicrograph of the cross-section of the gold gauze (Figure 30). Figure 30 shows that both sides of gold gauze are not fully smooth, and the wire has a mushroom-like shape , contrary to the expected round shape. The top of the “mushroom” corresponds to the dull side of the gold gauze, while the “stem” side relates to the shiny side. These findings enable us to speculate that these gold gauzes were fabricated by electroforming on a certain substrate in a suitable electrolyte. In view of this unusual shape of the wires of the gold gauze, the question arises as to which side should be placed next to the flat-bottom pH electrode. It is believed that the stem side ( or the shiny side of the gauze ) is more appropriate than the top side of “mushroom” (or the dull side of the gauze). This is the way in which the experiments were carried out. However, even the stem  Investigation of new 500-mesh gold gauze  108  1  oØJJrrl  Figure 26 SEM photomicrograph of 500-mesh gold gauze (dull side) (20 kV, 500X)  [1  V I  4  Figure 27 SEM photomicrograph of 500-mesh gold gauze (shiny side) (20kv, 500X)  Investigation of new 500-mesh gold gauze  109  E  -—  Figure 28 SEM photomicrograph of 500-mesh gold gauze (dull side) (20 kV, 2,000 X)  I  I  I  kV  jm  Figure 29 SEM photomicrograph of 500-mesh gold gauze (shiny side) (20kv, 2,000 X)  Investigation of new 500-mesh gold gauze  110  Figure 30 SEM photomicrograph of 500-mesh gold gauze (cross-section) (20 kV, 4,000 X) side of “mushroom” is not perfect. As can be seen from Figure 30, the stem side is not completely smooth either. The gap between the glass membrane sensor of the flat-bottom pH electrode and the stem side of the “mushroom” will hold a small amount of electrolyte which makes possible the electrochemical reactions, including 2H + 2e = H2(g,. Depending on the viscosity, and the surface tension between the hydrogen bubbles and the gold (or nickel) and between the hydrogen bubbles and the glass membrane of the pH electrode, hydrogen bubbles can have difficulty in escaping, and can be trapped causing instability in the surface pH measurements. These trapped hydrogen bubbles were indeed observed during experiments when they grew large enough. It is believed that the technique which was developed by Romankiw 73751 for surface pH ’ 8 measurement is an excellent method when dealing with interfacial phenomena without the formation of gas bubbles. When large gas bubbles form, especially when the hydrogen bubbles tend to be adsorbed and trapped, this technique leads to surface pH measurements which are less accurate and less reproducible as reported by Romankiw and . 73751 To address this problem and to ’ 8 co-workers consider the extremely fragile mechanical strength of the gold gauze for convenient use, the apparatus itself needs to be improved to get more reliable measurements in those cases where hydrogen bubbles are a problem. The ideal apparatus would consist of gold or platinum gauze which has been embedded firnily in the glass membrane of the flat-bottom pH electrode when in fabrication. The gauze and glass membrane should be on the same level with the gauze having a  Investigation of nickel-coated 500-mesh gold gauze  111  polished surface. This kind of configuration may prove difficult to produce, but its fabrication is not impossible. It has not been found so far on the commercial market and hence was not used in the present research. Figure 31 Schematic drawing of the 500-mesh gold gauze The wire diameters 1 were estimated based on the SEM photomicrographs. The results, along with the dimensions supplied by the manufacturer, have been listed in Table 35. The effective area in Table 35 is defined to be (Real area) / (Nominal area) x 100, and is cal culated as follows. Ifit is assumed that the wires have a smooth surface and have only one side to conduct electricity, for each box in the centre of Figure 31, the nominal and real areas can be expressed as: Nominal area  =  (d + s )2  ri 1 1 1 Real area =2L7td(d+s) +1tdsj=1td(d+2s)  (229) (230)  Therefore, the effective area is as follows: Effective area (%)  =  1itd(d+2s) x 100 2 (d+s) 2  (231)  —  where: s is the space between wires, (urn); and d is the wire diameter, (jim) 4.2.3 Investigation of nickel-coated 500-mesh gold gauze A question was raised during the experiments as to whether the surface pH measured referred to the bare gold surface, the gold-nickel combination or to the nickel surface only. Practically, we  need to know the surface pH values measured on the nickel substrate such as would be encountered in industhal nickel electrowinning. One way to overcome this problem would be to use a nickel gauze instead of a gold gauze. However, the nickel gauze would introduce another problem due to its electrochemical activity in acidic media, and the nickel gauze could not be used repeatedly. The gold gauze is quite inert compared to nickel. To ascertain how thick the nickel film should be to cover completely the underlying gold surface, a series ofelectrodepositions at 50 A/rn 2 for various  1 Obviously here, diameter does not have a perfect defmition due to the mushroom-like shape of the wires.  Investigation of nickel-coated 500-mesh gold gauze  112  times was conducted. For nickel, density = 8.90 g/cm 3 and atomic weight = 58.7 gfmole. Therefore, if it is assumed that the current efficiency is close to 100 %, the nominal deposition rate of nickel at a given current density, c.d. (A/rn ), can be expressed as: 2 rate (pin/ sec)= (c.d. x l0) X 2  58.7 1 Ox 9 . 8 OO 965 x  =  3.417 x i0 x c.d.  (232)  Using the experimental set-up described in the Section 4.1, and under the conditions of 500-mesh gold gauze, NiC1 2 (55 g/L Ni ) solution, bulk pH 2, 25C, 50 A/rn 2 2 under gentle agitation, a number of galvanostatic electrodepositions were performed. nickel-coated gold gauzes were examined with EDX and SEM.  0.16 1.0  Energy, (Key)  pm  Ni  0.16  Ni  Energy, (Key)  Energy, (Key)  10.23  .-  10.23  After dissin.  Au  Au  iL  Ni 0.16  10.23  After electrodeposition, the  0.16  Energy, (KeV)  10.23  Figure 32 EDX diagrams of 500-mesh gold gauze coated with nickel layer of different thicknesses (0.05-1 jim) and after anodic dissolution (20 kV, 7,000 X) With a nickel deposit of around 0.05 j.trn, EDX could detect quite readily the existence of nickel on the surface of the gold gauze (Figure 32). However, EDX would penetrate into the surface layer of the sample up to —1 p.m. Thus, when the film thickness is less than —1 jim, EDX will produce unwanted information on the substrate even though the film may have already covered the substrate surface completely. This is evident in Figure 32 in the case of —0.5 p.m thick nickel film coating.  Investigation of nickel-coated 500-mesh gold gauze  113  As will be seen later from the SEM photomicrographs and the curve of pH vs. time, the surface of the gold gauze was quite probably completely covered with a nickel film in this case. As shown in Figure 32, EDX can verify without any doubt that --1 jim thick nickel film is sufficient to cover totally the entire surface of the 500-mesh gold gauze since no gold peaks appear. The nickel-coated gold gauze, after nickel was dissolved anodically, was also examined with EDX (Figure 32). This surface exhibited no nickel peaks. Hence one can be sure that using the potential sweep anodic dissolution method, the deposited nickel film can be completely dissolved anodically by controlling the final potential up to 0.05 volt vs. SCE. Due to the ineffectiveness of EDX when the sample thickness is less than 1 jIm, the nickel-coated gold gauzes were subjected to SEM examination. The SEM photomicrographs were taken from the surface of the gold gauze and also from the cross-section of the wires. Comparing Figures 33-35 with Figure 28, it can be seen that a —0.05 jim thick coating does not change the surface morphology of the original gold gauze very much. However, a —0.5 jim thick coating changes the surface morphology significantly. For a —1 jim thick coating (Figures 35-36), SEM photomicrographs of both the surface and cross-section of the wires demonstrate that the substrate surface has been covered completely with the nickel film. More details can be seen from Figure 36 where the nickel film was uniformly deposited along the contours of the substrate surface. If the thickness is decreased by 50 %, that is, to 0.5 jim, it is not hard to see that the surface would still be covered completely with the nickel film. Unfortunately, successful SEM work was not achieved on the sample of—0.5 jim thick nickel-coated gold gauze, let alone to the samples having a coating less than --0.5 p.m. The failure to achieve acceptable SEM data in the case of the thin film coated samples can be attributed mainly to the polishing procedure used in the preparation of the SEM samples. This procedure blurred the boundary between the nickel and the surface of the gold gauze. One feature about the coating of nickel film is that the nickel was quite uniformly deposited over the entire surface of the gold gauze even at --5 p.m thickness. This is not difficult to understand from the viewpoint of the high activation overpotential of nickel. The change in the nature of the cathode substrate will affect the magnitude of the surface pH. When passing a current through the electrodes, the cathode substrate (gold gauze) will possess excessive negative charges which will attract hydrogen ions from the electrolyte. The pH change due to this phenomenon is, however, impossible to detect using the gold gauze, since this change occurs in the electrical double layer. In the case of copper electrodeposition, the surface pH was found experimentally to be nearly equal to the bulk pH when the current density was below the copper limiting current density’ . Therefore, the pH change detected can be indicative only of 8775 the change in the hydrogen ion activity within the diffusion layer and its rise can only result from  Investigation of nickel-coated 500-mesh gold gauze  114  I  I  Figure 33 SEM photomicrograph of 500-mesh gold gauze coated with —0.05 p.m (nominal) thick nickel film (20 kV, 2,000 X) (morphology)  J  kV  EOjim  Figure 34 SEM photomicrograph of 500-mesh gold gauze coated with —0.5 jim (nominal) thick nickel film (20 kV, 2,000 X) (morphology)  Investigation of nickel-coated 500-mesh gold gauze  Figure 35 SEM photomicrograph of 500-mesh gold gauze coated with —1 urn (nominal) thick nickel film (20 kV, 2,000 X) (morphology)  Figure 36 SEM photornicrograph of 500-mesh gold gauze coated with --l jtm (nominal) thick nickel film (20 kV, 2,000 X) (cross-section)  115  Effect of nickel concentration on the surface pH in pure NiCI 2 solutions at 25C  116  the depression of hydrogen ions caused by hydrogen gas evolution. Ifthe substrate favours hydrogen evolution, a high surface pH should be expected, and vice versa. Gold is a well-known catalyst for hydrogen evolution. When starting from the bare gold substrate, the reduction of nickel ions and hydrogen evolution take place initially on the bare gold, then on the gold-nickel combination and finally on the nickel only. After complete coverage of the gold surface with a nickel film has been reached, the surface pH should not change significantly. Figure 37 Surface pH as a function of time at 50 A/rn 2 (500-mesh gold gauze, 0.937 M NiCI , 2 bulk pH 2.5, 25CC)  4.0 3.8 3.6 3.4  3.2  2.6 2.4 2.2 2.O  0  50  100  I  150  I  I  200  250  Time, (sec)  •  I  300  •  One curve of pH vs. time was chosen to reflect this trend (Figure 37). This graph indicates that the surface pH jumped to —3.6 at the beginning of electrodeposition, afterwards declined and finally reached a relatively stable level,takingaround50seconds. Forsucha short period of time, the nominal thickness of  the nickel film (assuming 100 % current efficiency) = 3.417 x  x 50 x 50  0.09 l.Lm. For such  a thin nickel film, as mentioned above, it is impossible by means of EDX or SEM techniques to confirm whether or not the surface of the gold gauze was completely covered with the nickel film. However, here the trends of the surface pH change reflect a great deal of information. Quite conservatively, it may be said that taking a factor of 6 times, i.e., 300 seconds at 50 A/m 2 for precoating the gold gauze would be sufficient, which is equivalent to —0.5 im nominal thick nickel film.  4.3 Effect of nickel concentration on the surface pH in pure NiC1 2 solutions at 25°C The concentration of nickel has a dual effect during nickel elecirocleposition. High nickel concentration enhances the rate of cathodic reduction of nickel ion by raising its activity, and depresses the hydrogen evolution at a given pH by increasing the activity coefficient of the hydrogen ion. The titration curves (Figure 38) show clearly that the amount of sodium hydroxide required to neutralize the free acid in aqueous nickel chloride solutions at pH 1 decreases dramatically with increasing NiCl 2 concentration. The pH at which insoluble nickel hydroxide starts to form decreases also with increasing NiCl 2 concentration. The surface pH’s measured in the pure nickel chloride solutions are represented in Figure39. As can be seen from Figure 39, lower surface pH’s are observed in the more concentrated nickel chloride solutions, and these lower surface pH’s mean less hydrogen evolution if the activity  Effect of nickel concentration on the surface pH in pure NICI 2 solutions at 25°C  117  7 0.937 U Ni02 6 2 U NICI2 5  3M NC12  I  2 I  0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 1.02 M NaOH, (mL)  Figure 38 pH titration curves for different NiCI 2 concentrations at 25°C (150 mL sample and 0.5 mL/min speed)  0  80  160  240  .  I  320  .  I  400  .  I  480  .  I  560  .  I  640  .  720  800  C.D., (Nm2)  Figure 39 The surface pH as a function of current density for different N1C1 2 concentrations at 25°C (500-mesh gold gauze and bulk pH 2.5)  coefficient of the hydrogen ion, ‘y +, has the same value in these solutions. It should be noted that 11 the amount of hydrogen gas formed is directly proportional to the decrease in the amount of hydrogen ions in the solutions, while the pH is equal to —log(y + [H]). Therefore, when the surface pH is 11 related to hydrogen evolution, the effect of the activity coefficient of the hydrogen ion must be taken into account. As has been shown in Section 2.1.1, the activity coefficients of hydrogen ion in 3 M NiC1 2 and 2 M NiCI 2 at 25°C are —12 and -3 times as large as that in 0.937 M NiCI 2 solution. Thus at a given pH, the concentration of hydrogen ion is considerably smaller in more concentrated nickel chloride solutions. Based on this fact, it can be understood that the depression of hydrogen evolution with the increase of nickel chloride concentration is more than just a linear relationship with the nickel concentration. In terms of nickel electrodeposition, this means that a lower bulk pH in the highly concentrated nickel chloride solutions will result in a high current efficiency such as can be reached only at a higher pH level in less concentrated solutions. Except for the adverse effect of wasting electricity from the hydrogen evolution, a lower bulk pH has many advantages, such as, increasing the con ductivity of the electrolyte, improving the surface quality of the nickel deposit by the enhancement of mass transfer from the flow of hydrogen bubbles, and reducing the likelihood of the formation of insoluble nickel hydroxide on the cathode surface. In the electrodeposition tests, it was found that the surface quality of the nickel deposit at higher pH levels was not as good as that at lower pH. The deposits looked dark with black spots on the surface at times, even though high current efficiencies were achieved. Consequently, in reality, a higher nickel current efficiency should be sought under conditions where a satisfactory nickel deposit can still be achieved.  Effect of sulfate on the surface pH in 2 4 NiCl S Na O solutions at 25°C  118  4.4 Effect of sulfate on the surface pH in 4 -Na NiC1 S 2 O solutions at 25°C Three sulfate-containing nickel chloride electrolytes were tested, that is, 0.572 M NiCI 2+ , 0.937 M NiC1 4 0.365 M NiSO 2 + 0.365 M 4 SO and 0.572 M NiC1 2 Na 2 + 0.365 M NiSO 4 + 0.365 M SO As shown in Figure 40, the presence of the sulfate ion was beneficial in terms of the surface 2 Na . 4 pH. However, the differences in the surface pH’s at different sulfate concentrations are quite marginal at bulk pH 2.5. The activity coefficients of the hydrogen ion measured previously in these solutions can be used to estimate the change in the amount of total acid 1 available for 250 mL of the solution as the pH goes from 2.5 to 5.5: 2 0.937 M NiC1  i0) x 0.250/2.69 = 2.94 x 10 mole  A[H]T =  (10’s  2 + 0.365 M NaSO 0.937 M NiC1 4  =  (10  0.572 M NiCI 2 + 0.365 M NiSO 4  =  (10 10) x 0.250/1.09 = 7.25 x 10 mole  0.572 M NiC1 2 + 0.365 M NiSO 4 + 0.365 M 4 SO 2 Na  —  —  10-) x 0.250/1.38 = 5.72 x 10 mole  —  L[HlT = (10 i0) x 0.250/0.634 = 12.5 x —  mole  7  6  J?  I NICI2-NISO4  5  5  J  I  0  4  (4  0.  3  (I)  2 0.572 NKI2 + 0.365 U NISO4  0.937 Ni2 + 0.365 U Na2SO4  2 0.572 NI2 + 0.365 NISO4 + 0.365 U Na2SO4  0.937 U NC  --  1.1  .1.1.  40  80  120  160  •  1.1,1  200 240 280 C.D., (A1m2)  320  1.1.  360  400 440  Figure 40 The surface pH as a function of current density for different sulfate concentrations at 25°C (500-mesh gold gauze and bulk pH 2.5)  0 0  1  2  3  4  5  ,zjf  N1CI2-NiSO4at6O’Cji :I  /  3  10  NiCI2-NiSO4-Na2SO4  N1CI2-Na2SO4  NiCI2 6  I,  I,  I,  6  7  8  9  10 11 12 13 14 15 16 17  1.99 M NaOH, (mL)  Figure 41 pH titration curves for different sulfate concentrations at 25 and 60°C (0.937 M NiCI , 2 0.937 M NiCI 2 + 0.365 M 4 SO 0.572 M NiCI 2 Na , 2 + 0.365 M NiSO 4 and 0.572 M NiCI 2 + 0.365 M 4 + 0.365 M 4 NiSO SO 150 mL sample and 2 Na , 0.5 mLlmin speed)  The number 2.69 is the activity coefficient of the hydrogen ion in 0.937 M NiCl , and the 2 numbers 1.38, 1.09 and 0.634 are the apparent activity coefficients of the hydrogen ion in the sulfate-containing nickel chloride solutions 0.937 M NiC1 2 + 0.365 M 4 SO 0.572 M NiC1 2 Na , 2+  1 Here total acid means the concentration of free hydrogen ion plus bisulfate ion.  Effect of sodium chloride on the surface pH in NiCI -NaCI solution at 25°C 2  119  0.365 MNiSO 4 and 0.572 MNiC1 +0.365 MNiSO 2 4 + 0.365 4 SO respectively. If the current 2 MNa , efficiency of nickel and the thickness of the diffusion layer are assumed to be of the same order of magnitude in these four solutions, the ratio of current densities to reach a surface of pH 5.5 should be around 1:1.9:2.5:4.2. Obviously, the curves in Figure 40 do not match this ratio. The reason for this, as was found in the electrodeposition tests, is that the current efficiency of nickel decreases continuously with the increase of sulfate concentration and the decrease of chloride concentration. That is to say, at a given pH and total nickel concentration, the current efficiencies have the order of CENiCIZ> > 4 + 2 CENiC1 NSO CENiC >. 4 +NiS0 12 +NSO 4 CEN+NO  Accordingly, based on the measurements of surface pH and current efficiency, it can be understood that sulfate should not be added excessively to nickel chloride solutions in nickel electrodeposition. At pH 2.5, the above calculations have indicated that the total acidity of the solution increases with increasing sulfate concentration. As the pH titration curves in Figure 41 show, the amount of total acid at pH 1 for the different electrolytes differs markedly from each other. These findings are quite consistent with the electrodeposition tests at 60°C where the dif ferences in current efficiencies of nickel in the solutions of 0.937 M NiCI 2 and 0.572 M NiC1 2+ 0.365 M N1SO 4 became larger as the pH decreased. 4.5 Effect of sodium chloride on the surface pH in 2 NiC1 NaC1 solution at 25°C Chloride ions promote the deposition of nickel, traditionally believed due to a catalysis of electron transfer via a so-called “chloride ion bridge” between Ni 2 ions and the cathode surface . 61 Piatti et a1 421 gave another account. They assumed that the nickel surface is not completely free of oxygen-containing species, and believed that it is likely that chloride ion interaction takes place through this kind of layer. It probably occurs by overlapping of the chloride ion orbitals, which are distorted due to the high local electric field strength in the electrical double layer, with part of the orbitals of nickel. However, recent theory believes that chloride ions enter into the hydration sphere of nickel ions and replace one of the associated water molecules so that nickel ions are able to move closer to the cathode surface to facilitate electron transfer” . It had been found in the electrodeposition 51 studies that the addition of 2 M NaCl increases the current efficiency of nickel deposition. This means that NaCl promotes the deposition of nickel, or in other words, inhibits the hydrogen evolution. The pH titration curve in Figure 42 shows that the free acid concentration at pH 1 is almost one-half that in pure nickel chloride solution (see Figure 38). As sodium chloride is a fairly weak complexing agent and is not a buffering agent at all, the decrease in free acid concentration at pH 1 can be ascribed to the increase in the activity coefficient of the hydrogen ion. The addition  Effect of boric acid on the surface pH in 3 -H 2 NiCI B 0 solution at 25°C  120  I  0.  3 4 1.02 M NaOH, (mL)  120  160  200  C.D., (A/m2)  Figure 42 pH titration curve for 0.937 M NiC1 2 2 M NaC1 at 25°C (150 mL sample and 0.5 mL/min speed) -  Figure 43 The surface pH as a function of current density in 0.937 M NiCl 2 2 M NaC1 at 25°C (500-mesh gold gauze and bulk pH 2.5) -  of sodium chloride would also increase the activity coefficient of nickel to a lesser extent, and thus the precipitation pH will be lower at a given nickel concentration. The measured surface pH will be the combined outcome of these two opposite effects. The results of surface pH measured for NiC1 -NaC1 (55 g/L Ni 2 , 2 M NaC1) are shown in 2 Figure 43. Again the gold gauze was precoated with —0.5 p.m of nickel film before the measure ments. The data shown here still indicate that the addition of sodium chloride is beneficial. Thus, the inhibition of hydrogen evolution by sodium chloride overrides the adverse effect of a decrease in the free acid concentration. 4.6 Effect of boric acid on the surface pH in 3 -H 2 NiCI B 0 solution at 25°C The function of boric acid in nickel electrowinning is a controversial subject. The traditional view is that boric acid serves as a pH buffer during nickel deposition. However, it has been claimed that boric acid actually serves as a homogeneous catalyst and lowers the overpotential of nickel 97 98] It has been reported that there is a complex between nickel and borate ions 6 39 deposition’ ’ [log K  -12.2  -11.1 at 55°C for reaction Ni 2 + 3 B0 = 3 2H B0 + 2H9 in mixed Ni(H 2 ) chloride-sulfate solutions based on the fact that the pH buffering capacity of the solution increases =  —  with either nickel or boric acid concentrafion . Another interesting point concerning the buffering 138 capacity of boric acid is the effect of an electric field. It was found that the true equilibrium dissociation constant of boric acid near the cathode surface is substantially larger than the corre sponding value in the bulk electrolyte. To clarify the true function of boric acid in nickelcontaining solutions, starting from the simplest case, a series ofpH titrations was conducted titrating free boric acid solution against NaOH solution. The concentration of the boric acid ranged from 5  Effect  of boric acid on the surface pH in 3 -H 2 NiCI B 0 solution at 25°C  121  to 40 g/L. As a starting point, the distribution curves of boric acid species were calculated based on the information from the literature . There are altogether four species which may exist and 37 their equilibrium reactions are as follows: 2 B H + 3 0 H 0  B0 3 2H  =  B0 3 3H  B0 3 4H  =  (233)  B(OH)-t-H  (234)  B O 2 (OH)+H  (235)  2 ( 3 B O OH)+H+2H  =  (236)  2 ( 5 4 B O OH)+2H+3H  The total concentration of boric acid can be expressed as: B0 + [B (OH)J + 2[B [H ] 0 (OH)] + 3[B 2 (OH)] + 4[B O 3 B 3 [H ] (OH)1 5 O 4 T O=3  B0 + 11 [H ] BT 3 [H I O= 3 42 4Q 2 [H]  B0 + 3 [H 4 ]  31 3Q —-  [113803] [H]  [H B 3 O] +  The equilibrium quotients  + 2Q 21  21 2Q —-  B0 3 [H 2 ] +  [H]  (  31 3Q  B0 [H 3 ] [11k]  + 4Q 42  , 1 Q 1Q ,Q 2 31  and  (238)  2 [H] (239)  1Q1  B0 + 1+) 3 3 [H 2 ] B0 [H ]  [H3803]  (237)  —  B0 T = 3 [H ]  . 37 Q 4 2 at 25°C are cited from the literature  As  BO is known, the concentration of free boric acid can be calculated at a given concentration 3 [H ] of hydrogen ion. Subsequently, the concentrations of other boric acid species can be easily calculated. The calculated results are presented in Figures 44 for 5 and 40 g/L 3 B0 at 25°C. if H the existence of B O(OH), 3 2 B ( O OH) and B (OH) is disregarded, the calculation procedure 5 O 4 becomes much simpler. The calculated result for this case is presented in Figure 45. It is believed that the information from Figure 45 is correct A number ofpH titrations were carried out to confirm this belief. A typical pH titration curve is shown in Figure 46. Over the pH range from 1 to 13.4, only two peaks occurred. On the left of the first peak, NaOH was consumed to neutralize the free acid in the solution. Between the first and second peaks NaOH was consumed to neutralize the hydrogen ions which were coming from the first-step dissociation of 3 B0 The mid-point pH in the titration H . curve (Figure 46) is almost the same as the pH at the cross-section point of two lines in Figure 45.  Effect of boric acid on the surface pH in 3 -H 2 N1CI B 0 solution at 25°C  122  110 100 90 80  70 .60  z  Q  (A’  50  040 30  20 10 0 -10 1  2  3  4  5  6  7  8  9  10  11  9  101112131415  12  13  14  15  pH 110 100 90 80 70 60  z  50  (B) 40  30 20 10  0 -10 1  2  3  4  5  6  7  8  pH Figure 44 Distribution curves of boric acid species in aqueous solutions containing 5 and 40 gIL B0 at 25°C H 3  Effect of boric acid on the surface pH in 3 -H 2 NiCI B 0 solution at 25°C  123  110 100 90 80  70 60  z  0 0  50 40 30  20 10 0 -10 1  2  3  4  5  6  8  7  10  9  11  12  13  14  15  pH Figure 45 Distribution curves of boric acid species at 25°C (considering 3 B0 and B(OH) only) H 14 12  3.5  10  -j  B -J  6  E  04  >  z U) 0  2  2.5 2.0 1.5  > 0  1.0 > 0.5 4 5 M NaOH, (mL) 3  Figure 46 pH titration curve for free boric acid at 25°C (0.485 M 3 B0 30 mL sample, and H , 0.5 mL/min speed)  ““0  z,. 5  10  15  20  25  30  40  35  [H38031, (gi)  Figure 47 Volume difference of 5 M NaOH between 2nd and 1st peaks as a function of boric acid concentration at 25°C (30 mL sample)  The difference in the volume of 5 M NaOH between the two peaks, 2 (V V ), in Figure 46 1 should be equivalent to the number of the moles of boric acid in the solution. The values of (V 2 V ) 1 were plotted as a function of the concentration of boric acid in Figure 47. The circles represent the -  -  experimental data, while the solid line represents the theoretical line under the assumption of first-step dissociation. It is surprising that the experimental data are completely on the theoretical line. This fact tells us that only 3 B0 and the monoborate anion are important, and boric acid H  Effect of boric acid on the surface pH in 3 -H 2 N1CI B 0 solution at 25°C  124  does not form polyborate anions at all, i.e., the species B o(oH); B 2 (OH) and B O 3 (oH) can 5 O 4 be ignored. If the titrations are carried out in 2 M NaC1, the pH titration curve is quite similar (Figure 48). The difference of (V 2 V,) is identical (Figure 49), despite the different volume V 1 at the first peak dpH/dV. -  14  4.0  12  3.5  10  -j  &3.0  8 -J  6  E  =  0  >  z U,  2.5 2.0  2 >  0  1.0 >  0.5 3 4 5 M NaOH, (mL)  Figure 48 pH titration curve for the boric acid in 2 M NaCI 0.485 M H,BO, at 25°C (30 mL sample, and 0.5 mL/min speed) -  0.0  0  5  10  15  20  25  30  35  40  [H3B03J, (giL)  Figure 49 Volume difference of 5 M NaOH between 2nd and 1st peaks as a function of boric acid concentration in solutions containing 2 M NaC1 (30 mL sample)  Deligianni and Romankiw 174 did some electrochemical studies with regard to the behavior of boric acid on a gold gauze substrate in 0.4 M NaCl medium’. The bulk pH was 2 and the concentration of boric acid was in the range of 0.005 —0.2 M (equivalent to 0.3 — 12.4 g/L). l’hree of their findings are worth repeating here. (1) Surface pH decreases with boric acid concentration, and its corresponding final value reaches 12 — 8 at the end of a linear potential sweep (— -1.9 volt vs. Ag/AgC1). (2) The limiting current density of the hydrogen ion reduction is independent of the con centration of the boric acid. Therefore, boric acid should not dissociate to produce hydrogen ions before the limiting current density of the hydrogen ion reduction is reached. (3) Boric acid would dissociate to produce hydrogen ions at potentials more negative than that at the limiting current density of hydrogen ion reduction. What would happen if nickel co-exists in the solution can be seen from the pH titration curve of 3 -H 2 NiC1 B 0 (Figure 50). Comparing the titration curves of pure NiC1 2 (Figure 38) and free 3 (Figure 46 or 48), the buffering capacity of 3 H,B0 -H,B0 solution increases dramatically 2 NiC1  1 Temperature was not stated in their paper, but it seems that the experiments were carried out at ambient temperature.  Effect of boric acid on the surface pH in 3 -H 2 NICI B 0 solution at 2WC  125  7 6  6  5 S 4 = 3 2 2  Oo  1  2  3  4  5  6 7 8 9 10 11 12 13 14 15 16 1.O2MNaOH,(mL)  Figure 50 pH titration curve for 0.937 M NiCI 2 0.485 M 3 B0 at 25°C (150 mL sample and H  -  0.5 mLlmin speed)  0  40  400440480520560600  C.D.,(A1m2)  Figure 51 The surface pH as a function of current density in 0.937 M NiC1 2 0.485 M 3 B0 at 25C H (500-mesh gold gauze and bulk pH 2.5) -  and the buffering range of boric acid is extended to the acidic region. This observation is supported by the formation of a weak complex between nickel and borate ions which has been reported . 381 Due to the formation of the nickel-borate complex, Ni 2 +3 B0 = Ni(H,fl0 2H 2 + 2H, more ) 3 hydrogen ions axe available in the solution. Comparing the pH tiiration curves of 0.937 M -H 2 NiC1 B 3 0 in Figure 50 and of 0.937 M NiCl 2 in Figure 38, it can be seen easily that the free acid concentration at pH 1 in 3 -H 2 NiC1 B 0 is very close to the that of pure NiC1 , indicating that 3 2 B0 H does not change the activity coefficient of hydrogen ions. But the pH at peak dpHIdV is shifted from —4.4 to —2.9 as a result of the addition of 3 B0 This also indicates that boric acid starts to H . form a complex with nickel ion and thus to produce hydrogen ions when the pH is above —2.9. The measured surface pH values are given in Figure 51. Surprisingly, the surface pH’s are much lower especially at higher current densities, and increase almost linearly with current density. As can be seen from the pH titration curve, this behavior of lower surface pH is not just the result of the buffering action of boric acid alone. It seems that boric acid also enhances the deposition of nickel, which observation appears to be in agreement with the so-called catalytic effect of boric acid. Indeed, higher current efficiencies of nickel were observed in the electrodeposition tests at bulk pH 1.1 and 60°C. Beside this catalytic effect, to account for the lower surface pH behavior, it may also be speculated that due to the very sharp pH gradient immediately away from the surface of gold gauze, as reported by Romankiw, the surface pH’s measured with the 500-mesh gold gauze axe still different to a certain degree from the real surface pH’s. Therefore, boric acid may have already played a substantial buffering role there even though it will not be apparent from the titration curve.  Effect of ammonium chloride on the surface pH in 4 -NH 2 NiCI C I solution at 25CC  126  These surface pH measurements have shown that significant benefits can be realized by adding boric acid to nickel electrolytes, especially when operating at higher current densities. As is well known, the addition of boric acid in industrial nickel electroplating industry has been practiced for several decades.  4.7 Effect of ammonium chloride on the surface pH in 4 -NH 2 NiC1 C 1 solution at 25°C The addition of ammonium sulfate or chloride is indispensable to nickel powder production via electrolysis at extremely high current densities° 100.104) Ammonium chloride is both a strong complexing agent and a pH buffer. As with boric acid, the buffer point of free ammonium chloride is in the basic region around pH 9.3, as shown in the pH titration curve of free ammonium chloride solution in Figure 52. However, the formation of strong nickel-ammonia complexes shifts this buffering range to a relatively acidic region (NH: NH 3 +H). By comparing the titration curves of 4 -NH 2 NiC1 C 1 (Figure 53) and 3 -H 2 NiC1 B 0 (Figure 46), the pH at peak dpHJdV is similarly close to —2.9, but NH C1 has a much stronger buffering action. Compared with that of pure NiCl 4 , the 2 free acid concentration at pH 1 was decreased as a result of the addition of NH C1. 4 —  0  I  0  1  2  3  .1  4 5 6 5 M NaOH, (mL)  7  8  9  Figure 52 pH titration curve for the free ammonium chloride solution at 25°C (1.31 M NH C1, 30 mL 4 sample, and 0.5 mL/min speed)  0  0  1  2  3  4  5  6 7 8 9 10 11 12 13 14 1516 102 M NaOH, (ml.)  Figure 53 pH titration curve for 0.937 M NiC1 2 C1 at 25°C (150 mL sample and 4 1.31 M NH 0.5 mLlmin speed) -  As with the addition of boric acid, when NH C1 is added, the surface pH’s are also very low 4 and increase almost linearly with the current density (Figure 54). This means that NHC1 is also quite beneficial in controlling the surface pH at a low level in nickel electrodeposition. The addition of NH C1 may not be quite feasible when the anodic reaction is chlorine evolution. 4 As chlorine is a strong oxidant, it may oxidize the ammonium ion NH in the solution to nitrogen gas. The decision whether or not to add NU C1 depends on how crucial the deleterious effects are 4 when ammonia is oxidized to nitrogen gas.  Effect of temperature on the surface pH in pure nickel chloride solution  127  7 6 5  Figure 54 The surfacepH asafunction of current density in 0.937 M NiCl 2 1.31 M NH C1 at 25°C 4 (500-mesh gold gauze and bulk pH 2.5) -  2  •1  I  0  40  .1.1  I  .1  I  I  1.1.  80 120 160 200 240 280 320 360 400 440 480 520 560 600 C.D., (A/m2)  4.8 Effect of temperature on the surface pH in pure nickel chloride solution As a starting point, three pH titrations were carried out on 0.937 M NiC1 2 solution at 25, 40 and 60°C in order to reveal the change of pH of the electrolyte with temperature. The curves shown in Figure 55 reveal two things. That is, the free acid at a given pH increases and the pH where the insoluble nickel hydroxide starts to form decreases with increasing temperature; in other words, the activity coefficient of hydrogen ion decreases with temperature and a high temperature favours the precipitation of nickel hydroxide. 7 6  6  5 5 4 0  3 3  2  2  25C (string) 40CC (no stwring) 40CC (stirring) 60CC (no stirring) —.--€-—a—.--  0 0.0  0.5  1.0  1.5  2.0 2.5 3.0 3.5 4.0 1.99 M NaOH, (mL)  4.5  5.0  5.5 6.0  Figure 55 pH titration curves for 0.937 M Ni Cl 2 at different temperatures (150 mL sample and 0.5 mL/min speed)  1  100  200  300  400 500 600 0.0., (A1m2)  700  800  900 1000  Figure 56 The surface pH as a function of current density in 0.937 M NiC1 2 at different temperatures (500-mesh gold gauze and bulk pH 2.5)  The surface pH measurements at 40 and 60°C were conducted using exactly the same apparatus and almost the same procedures as those employed at 25°C. The gold gauze was always precoated with a layer of nickel before the measurements. One exception was for tests at 60°C, where the solutions were not deaerated before measurements and not agitated during measurements in order to simulate the practical situation. Measurements at 40°C without agitation were also performed for the sake of comparison.  Effect of temperature on the surface pH in pure nickel chloride solution  128  Certain difficulties had been encountered at higher current densities as the gold gauze easily cracked when a thick layer of nickel was deposited on it. This phenomenon was quite likely to happen when the current density exceeded 1,000 A/m 2 and the elecirodeposition ran for more than 100 seconds. The amount of electricity during this period would produce a nickel deposit having a nominal thickness of around 3.6 p.m. Because of this problem, measurements were restricted to current densities up to 1,000 A/m 2 for 100 seconds of electrodeposition for each run. The measured surface pH’s in 0.937 M NiC1 2 at a bulk of pH 2.5 and temperatures of 25, 40 and 60°C are presented in Figure 56. Several things are revealed in this graph. Firstly, high tem perature does enhance significantly the rate of nickel reduction so that there is a lower surface pH. Secondly, agitation lowers effectively the surface pH by increasing the mass transferrate of hydrogen ion towards the cathode surface. Thirdly, the final surface pH’s are compatible with the pH titrations, indicating that the formation ofinsoluble nickel hydroxide on the cathode surface should be expected at those high pH levels. It should be mentioned that there is a nickel concentration polarization during nickel elecirodeposition, so that the pH at which the insoluble nickel hydroxide starts to precipitate should be somewhat higher than that estimated from the titration curves or from the solubility product based on the bulk nickel concentration. One interesting point shown here at 60°C is that without agitation the surface pH is about 0.34 unit higher than the bulk pH even at a current density as low as 100 AIm . Agitation was indeed 2 found to affect the surface pH even under no current passage when a layer of nickel was present on the surface of the gold gauze. Due to this unusual phenomenon, the potential of the nickel electrode was measured at 60°C in nickel chloride solution. Before measurements, the solution was deaerated by bubbling nitrogen gas for 10 minutes. Then, as shown in Figure 57, a current at a level of 50 A/rn 2 was passed for 20 minutes to deposit electrochemically a fresh nickel film on a mechanically polished nickel substrate (1 x 1 2 cm ) . The coulombs passed during this period were sufficient to produce a nickel deposit around 2 p.m thick, assuming a nickel current efficiency of 100 %. After the current had been passed for 20 minutes it was switched off and the potential of the nickel electrode was followed during the time when the stirrer was turned on and off for a period as indicated. The nitrogen bubbling was maintained all the time inside the cell, but the sparging tube was lifted up to the solution surface ,  when the stirrer was turned off in order not to disturb the solution near the nickel electrode. Figure 57 shows that when the agitation is stopped, the electrode potential drops about 10 mY. This potential drop can be ascribed to the chemical dissolution of metallic nickel by the hydrogen ion. In terms  Effect of temperature on the surface pH in pure nickel chloride solution -0.1  -0.1  without agitation  129  without agitation  -0.2 0  >  -0.3  Ui  -0.4 -0.5  -0.5  -0.6 0  current is on  -0.7 -0.8  I  -.  -5  0  5  current is off I  I  I  _current is 1 I  I  I  0  Time, (mm)  -0.1  current is oft  n  10 15 20 25 30 35 40 45 50 55 60 65  Figure 57 The potential ofnickel electrode vs. time in deaerated 0.937 MNiCl atbulkpH 2.5 and (iO°C 2 (50 A/rn 2 bubbling and under agitation , with N 2 except where marked)  on  5  10 15 20 25 30 35 40 45 50 55 60 65  Time, (mm)  Figure 58 The potential of nickel electrode vs. time in non-deaerated 0.937 M NiCI 2 at bulk pH 2.5 and 60°C (50 A/rn under agitation , 2 except where marked) 0.0  without agitation  -0.1  -0.2  -0.2 0  >  -0.3  Ui  Ui  -0.4  ()  C’)  -0.5 0.5  -0.6 C  0  current is on  -5  0  5  a—  -0.7 -0.8  current •is oft  I  -0.9 —1.0  101520253035404550556065  Time, (mm)  Figure 59 The potential ofnickel electrode vs. time in 0.937 M NiC1 2 at bulk pH 2.5 and 60°C (50 Aj , 2 under agitation except where marked, with prior air bubbling for 10 minutes)  currentison  ii,  -5  0  5  current is oft .5,5.5.5,5.1.1,5,  101520253035404550556065  Time, (mm)  Figure 60 The potential of nickel electrode vs. time in the non-deaerated 0.937 M NiCl 2 at bulk pH 2 and 25°C (50 Ari , without prior deaeration, under 2 agitation except where marked)  of the pH unit, this potential drop would increase the surface pH by 0.15 unit at 6(YC . This at least 1 partially explains the higher surface pH at 60°C and lower current densities where there is no agitation of the solutions. What the potential of the nickel electrode would be when the solution contains some dissolved oxygen is illustrated in Figures 58-59. The data in Figure 58 were measured before those in Figure 57, and the solution was not deaerated. On the other hand, the data in Figure 59 were measured after those in Figure 57, and the solution was bubbled with air for 10 minutes before the  1 -2.303 RT/F = -0.066 volt at 60°C, and -0.010/-0.066  0.15.  Effect of temperature on the surface pH in pure nickel chloride solution  130  measurements. Figures 58 and 59 look quite similar and indicate that agitation has an even more dramatic influence on the potential of the nickel electrode when dissolved oxygen is present. The difference in potential of the nickel electrode when the agitation is on and off is in the order of 100 mV. This is equivalent to an increase in pH of 1.5 units. Evidently, this pH increase overshoots the observed shifts in the measurements. 0.0  without agitation  without agitation  -0.1  /\  ririri  -0.1  . -0.3  w  o -0.4  C,)  -0.5  >  -0.6  -0.6 0  -0.8  -5  0  5  -0.7 -0.8  i current ison  -0.9  —10  -0.5  C  -0.7 -  -jI.-j  -0.2 0  current is off  -0.9  101520253035404550556065 Time, (mm)  Figure 61 The potential ofnickel electrode vs. time in deaerated 0.937 M NiCI 2 at bulk pH 2 and 25°C (50 Mn , with 10 minutes priorN 2 2 bubbling, under agitation except where marked)  current  4  5 1.c  •  0  I.  5  current is oil  is on I.?  1,1  •  1.1.1,  lo 15 20 25 30 35 40 45 50 55 60  Time, (mm)  Figure 62 The potential ofnickel electrode vs. time in 0.937 M NiC1 2 at bulk pH 2 and 25°C (50 A/m , 2 with 10 minutes air bubbling, under agitation except where marked)  As shown in Figures 60-62, exactly the same trends were found for the potential of the nickel electrode at 25°C. Since highly pure BDH AnalaR grade chemicals were used, there are not many options for possible electrode reactions. In all, there are five possible electrode reactions listed below together with their potential expressions at 25°C’°’° . For easy comparison with those lines 61 in Figures 57-62, all of the following potentials are expressed on the SCE scale’.  2 Ni  +  2e  =  Ni  EN.2.,,N. =  2 Ni(OH)  +  —0.498 +  2I-f  EN1(oH,N1  2 Ni0  +  (240)  4H  +  +  2e  =  RT  Ni  in +  —0.50 volt at aNI2+  1  0 2 2H  (241)  (242)  RT 2 =—0.125+j1n(aH+) =—0.125--0.0591pH —0.27 volt atpH = 2.5  (243)  2e  (244)  =  2 Ni  +  0 2 2H  1 The difference between SCE and H 2 electrode potentials is 0.241 volt at 25°C.  Effect of temperature on the surface pH in pure nickel chloride solution  EN,N.2+ =  RT 1.437 +ln— = 1.437—0.1 l8pH —  1.14 volts at pH = 2.5 and + e =  2 O.5H  (or,  0 3 H  + e =  RT l +FlnaIl+ 24 EH+,H=_O. °2(aq) +  4H  0E 21110  + 4e  =  2 05H  =  0.0591 lOga.2+ 2  131  (245)  1 (246)  + 1120)  —0.241 —O.O59lp11 =—O.39 volt atpH 2:5  =21120  (247)  (248)  RT 0.988 +lna, = 0.988  —  0.0591pH  (249)  0.84 volt at pH 2.5  Although the electrode potentials at 60°C are not exactly the same, they should be of the same order of magnitude. The changes in standard potentials for most electrode reactions are less than 1 mV per degree. For the above five electrode reactions, they have the following values : 7 ° t1 dE+,N. = +0.93 a, dE+,H2  a,  (250)  dE°N(QH,N1  mV/ C  ;  =-0.17  a,  mV/ C  (251)  dEHo  +0.90  mV/°C  ;  a,  =  +0.03  mV/°C  Using the electrode couple Ni fNi as an example, there is only +33 mV shift when the temperature 2 rises from 25 to 60°C. Looking at those values of electrode potentials at pH 2.5 and unity activity of Ni , it can be 2 seen that only the potentials of Ni /Ni and W/H 2 2 are close to the observed electrode potential in deaerated solution which has a value around -0.45 -0.41 volt (Figure 57). Thus, it is certain that —  the measured potential is either the potential of Ni fNi or H/H 2 , or more accurately their combi 2 nation, the so-called corrosion potential. By comparing the nickel electrode potentials under no agitation at 60°C where the solution contains or does not contain dissolved oxygen (Figures 57-59), it can be found that the values are quite close. This implies that the dissolved oxygen is quickly depleted locally. On the other hand, at 25°C (Figures 60-62), the differences in the electrode potentials are greater and it takes a longer time for the electrode potential to decrease when the agitation stops. In the presence of dissolved oxygen at 60°C, once the agitation is turned on, the potential increases by —100 mV, and stays within -0.2 -0.3 volt. Based on the magnitude of this value, the only electrode reaction may be Ni + 20H = Ni(0H) 2 + 2e. This sounds reasonable as dissolved oxygen may possibly oxidize metallic nickel locally to a certain degree, and in the course of nickel —  Effect of ultrasound on the surface pH  132  oxidation, the dissolved oxygen gets reduced accompanied by the consumption of hydrogen ion. This may in turn stimulate the formation of nickel hydroxide. Accordingly, it would be better to deaerate the electrolytes for nickel electrodeposition before they go to the tankhouse. For solutions of 4 -NiSO NiC1 2 NiC1 , -NaC1, 3 2 -H 2 NiCl B 0 and 4 -NH 2 NiC1 C 1 at 25°C, similar phenomena were observed as regards the electrode potential of nickel responding to agitation and dissolved oxygen. 4.9 Effect of ultrasound on the surface pH 81 in his work on the electrocleposition of NiFe alloy demonstrated that the appli Romankiw cation of an ultrasonic field could depress completely the increase of surface pH during electro deposition. In one of their tests, the surface pH dropped from —7 to the bulk pH 2.5 within 2 seconds once the ultrasound was applied. However, Romankiw did not specify how powerful was the ultrasonic device he used. 7.0  E  80A/m2  IBOA/m2 6.0  3  1 p  24  ‘E  .1  between 0-  30  2.2 -40  ON I the me  UltrasoundwasONbetweenl0oand200seconds  32  2.0  Ulasound  Gun ent was applied between 0-150 seconds  2.5 I  0  40  80  I  120 160 200 Time, (sec)  I  I  I  240  280  320  2.020  0  I  I  20  40  I  60  I  80 100 120 Time, (sec)  I  140 160  I  180 200  Figure 63 The effect of ultrasound on the surface pH in 0.937 M NiCl 2 at bulk pH 2.5, 25°C and c.d. 80 and 180 A/rn 2 A similar test was conducted during the present investigation using an 80-watt ultrasonic cleaner in 0.937 M NiCl 2 at bulk pH 2.5 and 25°C. The effect of ultrasound on the surface pH was found to be marginal. The function of ultrasound is to create a mechanical vibration near the cathode surface thereby enhancing the mass transfer rate. In the present work, the whole cell assembly was placed inside the chamber of an ultrasonic cleaner in the presence of water, and the electrolyte was not agitated mechanically. As in other tests, the gold gauze was precoated with a layer of fresh nickel film before measurements. At a current density of 80 AIm , as shown in Figure 63, the 2 ultrasound does have some effect, lowering the surface pH by around 0.2 unit. However, at 180 A/rn 2 the ultrasound is not powerful enough to depress the further increase of surface pH. In both cases, the surface pH never returned to the bulk pH level in the presence of ultrasound.  Surface pH measurements at 60°C  133  No doubt the power level of the ultrasonic cleaner affects the results. In addition, the details of Romankiw’s experiments are not given in his published work . Based on the minor depression 8 of surface pH found in the present tests and considering the possible deleterious effect of ultrasound on the DSA anodes used in nickel chloride electrowinning, the use of ultrasound in the tankhouse to lower the surface pH cannot be recommended.  4.10 Surface pH measurements at 60°C A limited number of surface pH measurements were made at 60°C using 0.937 M NiC1 2 (55 g/L ) at bulk pH 2.5,0.572 M NiC1 2 Ni 2 + 0.365 M NiSO 4 (55 g/L Ni 2 and 35 g/L SO) at bulk pH 2.5, 3.92 M NiCl 2 (230 g/L Ni ) at bulk pH 2, and 3.555 M NiC1 2 2 + 0.365 M NiSO 4 (230 g/L Ni 2 and 35 g/L SO) at bulk pH 2. The results of 0.937 M NiCl 2 have already been presented in Figure 56. Due to the aforementioned difficulties at high current densities, the maximum current density was limited to 1,000 A/rn 2 for the normal nickel concentration and 1,400 A/rn 2 for the high nickel concentration. For easy comparison, the four curves are plotted together in Figure 64. 0.937M 2 pHbuk 2.5  5.5  0  0.572 NC2 + 0.365 A NSO4  5.0  pN*2.5  4.5  3.92 IA NiC12 p*2.0 3.555 N4C12  +  0.365 U NiSO4  pithulc2.0  I  0  100 200 300 400 500 600 700 800 900 10001100120013001400  C.D., (A/m2) Figure 64 The surface pH as a function of current density at 60°C without agitation in various elec trolytes (500-mesh gold gauze)  012345678 91011121314 0.199 M NaOH, (mL)  Figure 65 pH titration curves for highly concen trated solutions at 60°C (3.92 M NiCl 2 and 3.555 M 2 + 0.365 M NiSO NiCl , 150 mL sample and 4 0.5 mL/min speed)  All of these four curves show that the surface pH’s are lower at 60°C. When the solutions contain sulfate, the surface pH’s are lowered further. From the pH titration curves at 60°C in 0.937 M 2 (Figure 55) and 0.572 M NiCl NiCl 2 + 0.365 M NiSO 4 (Figure 41), it can be seen that at a given pH more free acid is available at 60°C than at 25°C. Although the temperature will affect the diffusion coefficients of nickel and hydrogen ions and the thickness of Nernst diffusion layer, the lower surface pH’s at higher temperature can be attributed mainly to the enhanced rate of nickel  Surface pH measurements at 60C  134  discharge, as evidenced by the fact that the cathodic potential increased by —200 mV when the temperature rose from 25 to 6(YC under the conditions of 0.937 MNiC1 , bulk pH 1.5 and 300 A/m 2 . 2 The smaller activity coefficient of hydrogen ion at higher temperature can be a contributing factor. -0.1 without agtior  /\  -0.2  Figure 66 The potential of mckel electrode vs. tune in deaerated 3.92 M NiC1 2 at bulk pH 2 and 60°C (50 A/rn , with 10 minutes prior N 2 2 bubbling and under. agitation except where maiked)  -0.4 -  -0.5 -0.6 -0.7 -0.8  1-4  current is on  .1.1.  -50  •  I  current is off  1,1,1.1.  1.1,1  •  5101520253035404550556065  Time, (mm)  When the highly concentrated 3.92 M NiC1 2 solution is used, the pH behavior is quite different from that in 0.937 M NiC1 2 solution. As discussed in Section 2.1.1, a high nickel chloride concentration increases quite dramatically the activity coefficient of the hydrogen ion. This solution was found to be close to the saturation limit at room temperature and its colour was dark green and opaque. The pH titration (Figure 65) indicates that it requires very little NaOH to raise the pH of this solution from 1 to above 3. The surface pH measured in this solution increases slowly with the current density and reaches pH —2.9 at 1,400 2 A/rn and bulk pH 2. This surface pH value is considered still safe from the risk of insoluble nickel hydroxide formation. The slightly higher surface pH at the low current densities is believed to result probably from the chemical dissolution of nickel by the hydrogen ion. The measurements of the electrode potential of nickel (Figure 66) show that the electrode potential of nickel drops by —15 mV when the agitation is stopped. This potential drop can be translated to a surface pH increase of —0.23 unit at 60°C. The surface pH in the solution 3.555 M NiC1 2 + 0.365 M NiSO 4 has a similar change with the current density as in 3.92 M NiC1 2 solution, but is —0.25 pH unit lower that the latter. The surface pH at a current density of 1,400 A/m 2 and bulk pH 2 is well below the pH level where the insoluble nickel hydroxide starts to form.  Chapter 5 Modelling of Surface pH during Nickel Electrodeposition  135  Chapter 5 Modelling of Surface pH during Nickel Electrodeposition To predict theoretically with reliable accuracy the surface pH during nickel deposition would be an important objective in the surface pH measurements. At the present stage, a theoretical model has been developed for the solutions 2 -NiC 0 (NaC1)-HC1 1 -H and 4 NiC S 2 ) (NaC1)-(Na NiSO O 1 0. Due to the lack of data for diffusion coefficients and equilibrium quotients, reasonable 2 HC1-H modelling could only be carried out for the solution 0 -HC1-H 2 NiCl . The modeffing starts from the mass transport on the basis of the one-dimensional Nemst-Planck flux equation and from the chemical equilibria. The following assumptions were made: (1)  The electrodeposition is galvanostatic (viz., constant current) and has reached a steadystate,i.e., dC 0. /dtI 1  (2)  Except for nickel reduction and hydrogen evolution, no other electrode reactions occur on the cathode surface. The contribution to the total current from the cathodic reduction of dissolved oxygen is negligible. Thus  =  ‘total  +  (3)  Convection is negligible within the diffusion layer.  (4)  Precipitation of insoluble Ni(OH) 2 does not happen within the diffusion layer and/or on the cathode surface.  (5)  All chemical reactions are in equilibrium.  (6)  Temperature, diffusion coefficients, and equilibrium quotients are constant within the diffusion layer.  (7)  Activity coefficient y, is constant and the mole fraction of non-i components (incl uding H 0) is approximately equal to one within the diffusion layer. 2 —  -  1 dlna D,C N dC  dx zF  dØ dx dØ  —  1 d(ln’, + lnC) DC N (if N  zF  D  dx  RT  1 and  constant)  -‘  dØ dx  (252)  where: subscriptj refers to componentj. .1,, 1 C  -  -  -  4  -  total flux, (kmol/m • see) 2  -  concentration, 3 (kmol/m ) activity  N T  valence  R  -  -  -  diffusion coefficient, 2 /(m sec) mole fraction of non-j components absolute temperature, (°K) gas constant, (8.3 14 J/mol°K)  Modelling of surface pH for the solution 0 -HCI-H 2 N1CI activity coefficient  -  x  F  distance, (m)  -  -  -  -  136  Faraday constant, (96,500 C/equiv.) mobility, (m/sec.(volt/m))  electrical potential of solution, (volt)  5.1 Modelling of surface pH for the solution 0 -HC1-H 2 NiC1 For the solution 2 -HC1-H NiC1 0 , seven chemical species need to be considered, that is, Ni , 2 C1, NiC1, NiOH, 0H, H and Ni (OH). Their reactions are shown in graphical form in Figure 67. 4 For instance, Ni 2 (number 1) reacts with OW (number 5) to form both NiOH (number 4) and (OH) (number 7). 4 Ni (3)  (4) a)  0 2 H  NiOH  NiCl  4>1  0  CrJ_ Ni 2 (2)  OW  W  (5)  (6)  (1)  Ot a)  ‘I  A  D .0  Ni ( 4 OH) (7)  7  Figure 67 Interactions between species in the solution 2 -HC1-H NiCI 0  Figure 68 Defmition of X-coordinate for the surface pH modelling  When it comes to solving the Nernst-Planck flux equation, particular attention should be paid to the definition of the X-axis. According to the electrochemical convention, the positive current is cathodic, that is to say, the current flow is towards the cathode surface. For the convenience of mathematical calculations, the X-coordinate is defined as in Figure 68: Its origin (i.e., x =0) sits at the Nernst diffusion boundary, and its positive direction is from the bulk solution to the electrode surface. Based on this definition, the cathodic current expressions and the Nernst-Planck flux equation will have the same sign as the X-axis. For hydrogen evolution, there are four species involved in the mass transport, that is, H, OH-, NiOH and Ni (OH). Their stoichiometric relationships with the hydrogen ion are expressed in 4 the following reactions: 2 or 2 2H+2e=H 0+2e=2H 3 2H 0 +H  (253)  0 2 2H  (254)  + 2e = H 2 +20W  Ni0H+H+2e —Ni+H 0 2  (255)  Ni ( 4 0H)+4H+8e =4Ni +4H 0 2  (256)  Modelling of surface pH for the solution 0 -HCI-H 2 NiCI  137  Therefore, the total flux of the hydrogen ion according to reactions (253)-(256) is equal to: =  ) + (—J 4 (--f ) +J 5 ) (kinol/m 7 6 + (—4J • sec) 2 (257) (1 2 H NiOHh)  F  -  1 ( 2 H H)  +  F  +  1N0j  +  F  2 H 1  F  ) 2 (kA/m F  For the cathodic reduction of nickel ion, there are also four species involved in the mass transport, , NiCi, NiOH and Ni 2 i.e., Ni (OH). Their stoichiometric relationships with the nickel ion are 4 shown as follows: 2 + 2e Ni  Ni or [Ni (H 0 )612+ + 2e 2  =  NiCl + 2e  =  Ni  =  Ni  + 6H 0 2  (259)  + C1  NiOH + H + 2e  =  (258)  Ni  (260)  + 1120  (261)  Ni ( 4 OH)+4H+8e =4Ni +41120  Accordingly, the total flux of the nickel based on the reactions (258)-(261) can be presented as: =  1 +J J 3 +J 7 4 + 4J  . sec) 2 (kmol/m (262)  —  ZNi(Ni24)  2F  lNi(N1CI1)  +  2F  +  1 N i(NjOHj  2F  (ohlf 4 M(M j  +  2F  —  -  SNi  (kA/m ) 2 2F  For the chloride ion, there are two species involved in the mass transport, viz., Cl and NiCl. As the chloride ion is neither reduced nor oxidized, its net flux should be equal to zero. (263)  0 cr23  The specific flux equations for the above individual species are as follows based on the Nernst-Planck equation:  1  dC 1 ‘dx  2F RT  dC 2 dx 2  (—1)F d RT dx 22  dØ ‘dx  (264)  (265)  dC F 3 x RT  d dx  (266)  dC F 4 dx RT  dØ dx 4  (267)  Modelling of surface pH for the solution 0 -HCI-H 2 NiCI  J — D  dC 5 dx 5  (1)F d 13 c RT dx  (268)  F RT  (269)  6  dC ‘cLx  4F d4 dx 7 RT  6 dC 6  138  d 6  6&  (270)  Substituting equations (264)-(270) into the above flux equations (257), (262) and (263) forhydrogen, nickel and chloride ions, and after some appropriate rearrangements, the following three equations (271)-(273) can be obtained. 4 D dC dC 6 dC 5 D 6 41)-, dC.,  I  5 D  16D_c F d 6 D c +D D 4 57 + 6 ++[ _c 3 c D 4 7j  1 3 dC D3 D4 dC 4 dC 4137 dC-, ( 133 4 D +--+ 2C C —-+ 1 1) C + 3 1) + 4 --+---  2 D dC 3 dC 3  1  3 D  (271)  =  F 4 D  161)7 F d?p = — Ni C7Jgp; + D 1 F 1 2D  F d  =0  (272)  (273)  In view of chemical equilibria, there are altogether four reactions (274)-(277): 3 K  Ni  cr  =  (274)  NiCl  4 K  Ni + 2 OH  =  (275)  NiOH K7  4Ni+4OW  =  (276)  Ni ( 4 OH)  K  (277)  0 =H+OIf 2 H  3 K  —  aN.C,+  y — ‘y C c 3 3 —j 2 YY2 C y C 1 2 1  —  — aNi2dJcr —  —  (YN2+CN2+)  3 C c c 1 =K3=Q3  i.e.,  =Q 3 2 1 C  After differentiating equation (279) and making some rearrangements, it follows that:  (278)  (279)  Modelling of surface pH for the solution 0 NiC 2 HCI-H I  139  1 dC 2 dC dC 3 2 3 Q 1 C ---+Q =0  (280)  I  [  c 4 y  aNU,H+  4= K  = aNIZ+aOW  4 C ..  = (YNI2+CNI2+)  (YOHCOH)  ‘Y1Y5  —=Q 4 =K  i.e.,  5 C 1 y  4 ‘y  =  C 4 5 1 yjy C  (281)  —  (282)  =Q 4 5 1 C  Y4  5  Differentiating equation (282) and making some rearrangements results in:  I  1 dC dC 4 5 dC ---+ 5 4 Q 1 C -=Q 0  (283)  I = 7 K  aN. (OH )4+  7 C  tNi4OF1)’Ni ( 4 OH) =  aNI2+aOH_  4 (YNI2+CN2 4 +) (YOH_Cow)  = ()4 4 (Ys)  4 (C ) 1 (C 4 ) 5  )4  7 C 4 ) 5 4 (C ) 1 (C  =  7 K  (284)  (285) i.e.,  =  17  (C (C 7 Q 4 ) 7=1 C 4 ) 5  Differentiating equation (285) again and rearranging, we obtain: 5 7 dC dC 1 ( 7 4Q 3 ) C (C )+1 5 (C (C 7 4Q 4 ) -3 ) 5 =0 —  (287)  K = aH÷a OH = (YH÷C) (YOHCOW) = y C 6 5 ...  (288)  K =—=Q 6 C 5 1516  Differentiation of equation (288) and rearrangement gives: (289)  5 dC 6 dC 5 6 C -+C -= 0  Except for the electhcal double layer, electrical neutrality applies everywhere in the solution. (290) —  i.e.,  Ccr + CN,+ + CN.OH+  —  . 0 C 11 + CH+ + 4 CN  7 2 1 2 3 — 4 + 5 + 6 = 4C C C 0 C  (OH)  (291)  (292)  I  Modelling of surface pH for the solution 2 -HCI-H NiCI 0  140  Differentiation of equation (292) results in:  I I  (293)  1 dC dC 2 dC 5 dC 3 dC 6 4 dC 7 dC 2--—-+--+--+--+4--=0  Equations (271), (272), (273), (280), (283), (286), (289) and (293) consist of a set of 8 x 8 multilinear equations. 1 dC 2 dC 3 dC 4 dC 5 dC 6 dC 7 dC d4 --+ 1 a 1 --+ 12 -a 14 --+ a 13 + --+ a 15 --+ a 16 --+ a 17 a a 18 = b 1  (294)  2 dC 1 dC 4 dC 3 dC 6 dC 5 dC 7 dC d a--+ a+a--+ a--+ a--+ --+a--+ a  (295)  1 dC -+ 7 a 1  2 dC  1 dC  2 dC  3 dC 4 dC 6 dC 5 dC dC-, d4 a. --+ —+ 73 74 --+a a 75 --+ a --+ 76 78 77 a a 3 dC  4 dC  5 dC  6 dC  7 dC  2 b  (296) =  7 b (297)  d  8 b  The coefficients a and b. are summarized in Table 36. The boundary conditions at x =0, i.e., at the Nernst boundary layer, are as follows: (298)  c:——io” 111+  —  (299)  Q,.,  5 c ; (300) [Ni]T =  C° + C ° + c: + 4C. 3  =  C° + Q C°C° + Q 3 C°C’ + 4 4 (C°) (C°) 7 4Q 4 (301)  =  4 ( 7 4Q C) (C) 4 + C(1 + Q )+Q 0 5 C 4 ° 1 C([Ni]T + C 3  7 4Q 1 ( 4 ° 5 C )+1 (C 3 Q °  )2 +  °+Q 5 C 4 ([Ni]T 3 {+Q  —  —  °+C 5 C ) 0 6  °+1 5 C )}C 0 6 C °  —  [NuT  0  (302)  From the polynomial equation (302), C ° can be solved by iteration. Knowing the value of C, it 1 follows that values for C ,3 0 2 C ° , C° and C ° can be obtained. 7 6 C=[Ni]T+C— + 0 5 ° C C  (303)  ° 3 C  (304)  =  2 C 3 Q ° °C  Modelling of surface pH for the solution 2 -HCI-H N1CI 0  141 (305)  r°r° C: = Q4’1 ‘5  (306)  4 4 ( 7 Q C) (C)  =  =0  (307)  Table 36 The coefficients ofthe 8 x 8 multilinear equations for the surface pH modeffing ofthe aqueous solution of 2 -HCI-H NiCI 0  Eq.# (271)  1 a  11 a  a  a  a  a  a  a  0  0  0  1  D  D 4 D  1)4  1)4  (272)  a  1  0  1)3  1)4  0  0  i; (273)  a  0  1  D  41)7  a  Ic —c (  3 D  6  I) 134  +Zc ‘f 1)4  7 16D  -  7)RT  ‘\ F  ‘2  D —i,,. -;  0  0  0  0  (D 3 F C3C2J  0  j (280)  a  2 C 3 Q  1 C 3 Q  -1  0  0  0  0  0  0  (283)  a  5 C 4 Q  0  0  -1  1 C 4 Q  0  0  0  0  (286)  a  7 c c  0  0  0  0  -l  0  0  (289)  j 7 a  0  0  0  0  6 C  5 C  0  0  0  (293)  a  2  -1  1  1  -1  1  4  0  0  When the partial current densities ofnickel reduction and hydrogen evolution, bulk pH, solution temperature, total nickel concentration, activity coefficient of hydrogen ion, diffusion coefficients, the thickness of the diffusion layer (3) and equilibrium quotients are known, the calculations can be started from the Nemst boundary conditions. The thickness of the diffusion layer is first divided equally into many very small increments, Ax. The following stepwise equation is used:  From the above set of 8 x 8 multilinear equations, all of the eight unknowns (C ,C 1 ,C 3 , 4 ,C 2 ,C 5 C ,C 6 7 and ) can be solved as a function ofx within the Nernst diffusion layer. The surface pH is the pH value when x = & At this point, a judgement must be made as to whether or not insoluble nickel hydroxide has fonned. If 2 at x =6, the calculated surface pH is in good ][OHi < [Ni agreement with the experimental value. Otherwise, the surface pH should be calculated again from i = 8 and Q,,. 2 [Ni  Modelling of surface pH in 0.937 M NiCI 2 at bulk pH 2.5 and 25°C  142  Besides the surface pH, the profiles of pH vs. x, [NuT vs. x, [Ni ] vs. x, [NiCr] vs. x, [NiOtEJ 2 vs. x, [Ct] vs. x, [Ni (OH)] vs. x, p vs. x, nickel partial current densities vs. x and hydrogen partial 4 current densities vs. x can be obtained. Due to the lack of required diffusion coefficients and equilibrium quotients, in the present investigation the specific modelling of the surface pH has been limited to the following solutions: -HC1-H and 2 M 2 NiC1 0 0.937 M 2 -HC1-H at bulk pH 2.5 and 25°C. NiC1 0 5.2 Modelling  of surface pH in 0.937 M NiCl 2 at bulk pH 2.5 and 25°C  Using the procedure detailed in Section 5.1, the surface pH in 0.937 M NiC1 2 was calculated as a function of current density. This calculation embraced the following considerations. As regards diffusion coefficients, there are very few reports concerned with concentrated solutions. For the common ionic species at infmite dilution and 25°C, such as, Ni , H, Ct, data are readily available. 2 However, for the complex ions, such as NiOW, NiCl, data even at infmite dilution cannot be found. For most of the common species in concentrated solutions, their diffusion coefficients can be estimated using the Stokes-Einstein equation: p 1 D  =  k  constant  (3()9)  where D is the diffusion coefficient of species i, (m /sec) 2 jt is the absolute viscosity of the solution, (kg/m.sec) T is the absolute temperature, (°K) is Boltzmann’s constant, (1.3807 x 10 J/°K) r is the radius of the species i, (m) 1 k  However, for the diffusion coefficient of the hydrogen ion in concentrated solutions, the StokesEinstein equation is invalid due to the unique proton jump transport mechanism. Majima et a1 11081 measured the equivalent conductivity ()+) of the hydrogen ion in acidic chloride solutions. Their results demonstrated that  depended only on and decreased with the activity of water. Once is known, the diffusion coefficient of the hydrogen ion can be calculated based on the following Nernst-Einstein equation: RT  =  +  RT  •;;-i;-;E- = --I1+ =  8.314x298 2+ = 2.66 x 10 965002  (310)  The hydroxyl ion is supposed to have a similar transport mechanism as the hydrogen ion, and its diffusion coefficient may be estimated as follows:  Modelling of surface pH in 0.937 M NiCI 2 at bulk pH 2.5 and 25’C  DH+  D oir =—D°= OH  143  2.66x1O 7 x5.26x1O=1.50x1O÷ H 9.31xlO  (311)  where D° is the diffusion coefficient at infinite dilution. The following rules for the selection of diffusion coefficients were adopted: (1) If the diffusion coefficients in real solutions are available from the literature, they will be used. (2) If they are not available, except for the hydrogen and hydroxyl ions, their values at infmite dilution will be used and adjusted using the Stokes-Einstein equation. (3) If values at infinite dilution are not available, it will be assumed that they are equal to 1 x i(Y m /sec at infmite dilution with this value being adjusted using the Stokes-Einstein 2 equation. (4) The diffusion coefficients of the hydrogen and hydroxyl ions are calculated from the equivalent conductivity of the hydrogen ion which is estimated according to reference 10 using the calculated activity of water based on Meissner’s theory° . 811  .  2.2  1.25  2.0  1.20  1.8  1.15  1.6  1.10  U)  U)  8 U) >  ci)  1.4  1.05  1.2  1.00  1.0  0.95  0.8 0  0.2  0.4  0.6  0.8  1  1.2  0.90 1.4  1.6  1.8  2  N1CI2, (M)  Figure 69 The viscosity and density of aqueous NiCI -HC1 solution at 25CC (dashed lines contain no 2 HO, solid lines contain 0.1 M HO, density times a factor of , absolute viscosity 3 kg/rn times i0 kg/(m•sec), kinematic viscosity times 1O 83 /sec) 2 m Awakura et a1 831 measured systematically the viscosity and density of aqueous NiCI -HC1 2 solutions at 25CC. Their data have been plotted in graphical form in Figure 69 and fitted using some simple expressions. It was found there is a linear relationship between the density of the solution and the concentration of nickel chloride, and exponential relationships between kinematic, or absolute viscosity and the concentration of nickel chloride. The fitted equations are as follows:  Modelling of surface pH in 0.937 M NiCI 2 at bulk pH 2.5 and 25C  144  Table 37 Density and viscosity of aqueous solutions of NiC1 2 + HO at 25°C t83 HO  2 NiC1  p  v  (M)  (M)  ) 3 x io (kg,n  /sec) 2 x 10 (m  x io (kgfm•sec)  831 ExptlJ  Fitted (this work)  831 Exptl.  Fitted (this work)  83 Exptl.  Fitted (this work)  0  0.00  0.9970  1.000  0.8930  0.894  0.8903  0.896  0  0.05  1.0025  1.005  0.9196  0.908  0.9219  0.914  0  0.20  1.0205  1.022  0.9554  0.950  0.9750  0.971  0  0.50  1.0560  1.057  1.0264  1.040  1.0839  1.097  0  0.937  I  1.107  /  1.186  /  1.309  0  1.00  1.1141  1.114  1.1956  1.209  1.3320  1.343  0  1.50  1.1707  1.172  1.3784  1.405  1.6137  1.644  0  2.00  1.2260  1.229  1.6459  1.633  2.0179  2.013  0.1  0.00  0.9989  0.998  0.8976  0.895  0.8966  0.895  0.1  0.05  1.0045  1.003  0.9118  0.908  0.9159  0.913  0.1  0.20  1.0223  1.02 1  0.9576  0.950  0.9790  0.969  0.1  0.50  1.0577  1.055  1.0277  1.038  1.0870  1.093  0.1  0.937  /  1.105  I  1.182  /  1.303  0.1  1.00  1.1157  1.113  1.2015  1.204  1.3405  1.336  0.1  1.50  1.1723  1.170  1.3860  1.397  1.6248  1.633  0.1  2.00  1.2275  1.227  1.6576  1.621  2.0347  1.995  For aqueous NiC1 2 solution containing no HC1, J 2 p = 997.6+114.9 X [NiC1  ) 3 (kg/rn  with  R  =  0.9999  (312)  V=  0.8949 x 10 x exp(0.2969 x [NiC1 J) 2  / sec) 2 (m  with  R  =  0.9989  (313)  =  0.8945 x i0 x exp(0.401 1 x [NIC1 I) 2  (kg/rn see) with  R  =  0.9996  (314)  with  R  0.9999  (315)  While for NiC1 2  +  .  0.1 M HC1,  p = 999.5+114.7 x [NiClJ  ) 3 (kg/rn  Modelling of surface pH in 0.937 M NiCI 2 at bulk pH 2.5 and 25°C v=0.8944x ] 2 10xexp(0.30 ) 10x[NiCl =  (rn / 2 sec)  with  (kg/rn sec) with  0.8957 x i0 x exp(0.4048 x [NiC1,J)  145  R =0.9990  (316)  R  (317)  =  0.9997  The concentration unit of NiC1 2 is mole/L in these equations. For easy interpolation and extrapolation, part ofAwakura et al’s data is reproduced in Table 37 together with the data calculated from the above fitted equations. As shown in Table 37, the absolute viscosity of water at 25°C is 0.8903 x i0 (kg/m.sec). The average absolute viscosity of 0.937 M NiCl 2 solution is (1.309 +1.303) / 2=1.306 x 10 (kg/rn. sec). Based on these two numbers, the diffusion coefficients given in Table 38 were calculated from the Stokes-Einstein equation. Table 38 Diffusion coefficients in 0.937 M NiC1 2 at 25°C 91 ”° 108 Species  At infinite dilution /sec) 2 (m  Symbol  ExptL /sec) 2 (m  Calcd. /sec) 2 (m  2 Ni  1 D  0.705 x  cr  2 D  2.03 x i0  1.32 x i0  NiCl  3 D  1.00 x io  /  0.8903 / 1.306 x 1.00 x iO = 0.682 x i0  NiOH  1)4  1.00 x  /  0.8903/ 1.306 x 1.00 x  OW  5 D  5.26 x io  /  1.50 x iO x 262 x iO = 3.93 x 10  H  1)6  9.31 x iO  /  2.66 x i0 x262 x i0= 6.97x io  Ni ( 4 OH)  7 D  9 1.00 x i0  /  §:  0.542 x  0.8903 / 1.306 x 0.705 x io= 0.481 x iO  0.8903 / 1.306 x 2.03 x io  0.8903 / 1.306 x 1.00 x  1.38 x  =  =  0.682 x i0  0.682 x l0  In 1 M NiCI 2 aqueous solution at 25°C.  Baes and Mesmer 951 supplied several equations to correct for the effect of ionic strength on the dissociation quotient of water and the equilibrium quotients of nickel hydroxy complexes. For the dissociation quotient of water in NaC1 medium at 25°C, 1ogQ=—14.00+ Parameter b  =  (318)  1.022JT 1+ ‘41  -0.32 when 1=2 m and b = -0.30 when 1=3 m.  For the reaction Ni 2  +  0 2 H  =  NiOH  +  H, Baes and Mesrner reported that:  Modelling of surface pH in 0.937 M N1CI 2 at bulk pH 2.5 and 25C  =—9.86— 11 logQ  146 (319)  1022-/i +O.O9x[Cl]. l+-’iI _  Therefore, for the reaction Ni 2  +  0H  =  NiOH,  log Q 11 logQ = —9.86— 4 = log Q  0.09 x [Cl]T [_14.00  —  =  4.14—  +  °+  —  bI]  (320)  2.044JT bI + 0.09 x [CuT 1 +‘Ji  While for the reaction 4Ni 2  1ogQ=—27.74+  +  0 2 4H  =  Ni ( 4 OI-I)  4H, Baes and Mesmer 1 also reported that:  +  (321)  2044-41 x[Cl]T 26 —O. i+-Ji  Accordingly, for the reaction 4Ni 2  +  40Ff  =  Ni ( 4 OH)  log Q 7 = log Q —4 x log Q =  —27.74  + 2.044q1  i+4i =  28.26—  2.0444i  0.26 x [CuT —4 x 114.00+1.022-41 bI i+’Ji )  (322)  4b1 —0.26 x [Cl]  The nickel chioro-complex, NiCl, is considered and the following equilibrium quotient is applied,  2 Ni  +  Ci  NiC1  =  log Q 3  =  -0.17 (2 M NaClO 81 ) 4  Table 39 Equilibrium quotients in 0.937 M NiCl 2 at 25C Reaction 2 + C1 Ni  Equilibrium quotient =  NiCl =  2 + 0H Ni  =  NiOW  2 + 40W = Ni Ni (0H) 4 0 = W + OH 2 H Ni(OH),) = Ni 2 + 20W  [NiCfl = 4] [Cr] 2 [Ni  [NiOHI = [Nij [OH1  =  0.676  10 = 5.37 x i0  [Ni ( 4 OH)] [QJf]4 =10 = [Ni ] 24  =  1.05 x 10  Q = w4] [OH1 = 1014.02 = 9.55 x =  lNi 4 2 ] [0H1 2=  10_15262  5.47 x 10  Modelling of surface pH in 0.937 M N1CI 2 at bulk pH 2.5 and 25°C Under the conditions of [NiJ = 0.937 M and [Cl] =2 x 0.937  =  147  1.874 M, it can be calculated that  [Ni ’ 2 ] = 0.479 M, [CI] = 1.416 M and [NiC1] = 0.458 M. Hence the real ionic strength is equal to 0.5 x (0.479 x 4 + 1.4 16 + 0.458) = 1.90. Using the solubiity product of Ni(OH) (S) from the CRC 2 , the above ionic strength and expressions, the equilibrium quotients presented in t911 handbook Table 39 were calculated. The other conditions or parameters are bulk pH 2.5, temperature 298°K (i.e., 25°C), total nickel concentration [Ni]T, 0.937 (kmol/m ), current efficiency of nickel reduction 95.4 % which was 3 measured at 50-150 A/rn 2 in the present work, activity coefficient of hydrogen ion +, 11 2.69, which y was also measured in the present work. For the thickness of the Nemst diffusion layer, ö, there are certain equations for the calculation of 3 based on well-defined flows 1071: N1I4  (323)  Natural laminar flow  6=  (vD Pb ‘10.28 6= 3.23hi i., h 3 g Ap  (324)  Natural turbulent flow  (vD Pb 1.52h1 g Ap 3 h  (  (325)  hD ‘1/3 6 = O.S vd2J 4 h 6  Forced laminar flow  1 v 7181 3 D W 6= 66.7h1 I I 2Vd) V  Forced turbulent flow  —  (326)  —  where: h d v D  ---  ---  ---  ---  electrode height, (m) electrode gap, (m)  g  kinematic viscosity, 2 /(m sec)  V  Pb  ---  ---  bulk solution density, 3 (kg/rn ) standard acceleration, 9.81 (rn/sec ) 2 flow velocity of solution, (rn/see)  diffusion coefficient, 2 /(m sec) ) 3 Pb P. bulk density minus surface density, (kg/rn -  For natural convection, 6 is in the range of 100 300 Lm. When some gas is evolved simultaneously, 71 reported that 6 is inversely proportional of the square root of the volume of gas evolved Ettel’°’ per unit electrode area. -  In the present investigation, the situation is much more complicated. The electrode was not placed vertically but at an angle of around 45 degrees. Besides, mechanical agitation was applied during the measurements. Hydrogen evolution, although not vigorous in most cases, will definitely have some influence. Consequently, there is no simple way to calculate the exact thickness of the  Modelling of surface pH in 0.937 M NICI 2 at bulk pH 2.5 and 25°C  148  Nernst diffusion layer. In the present surface pH modelling, Ax was set at 0.5 pm, while the thickness of Nernst diffusion layer was adjusted to be in the probable range. The surface pH for a given current density is the pH value calculated when x =6. 7.0 6.5 6.0 5.5 5.0 0  a)  4.5  0  4.0  C/)  3.5 3.0 2.5 2.0 1.5 1.0 0  20  40  60  80  100  120  140  160  180  200  220 240 260 280 300  C.D., (A1m2)  Figure 70 Modelled surface pH in 0.937 M NiC1 2 at bulk pH 2.5 and 25°C Using a selected value of 82 pm for 6, the calculated curve of the surface pH vs. current density, as shown in Figure 70, matched the experimental data points quite well. A value of 6 larger than 82 p.m would make the surface pH rise sharply sooner, and a value of 8 smaller than 82 p.m would delay the sharp rise of the surface pH. The solid line in Figure 70 was calculated without consid eration ofthe species 4 (OH) while the dashed line was obtained with consideration of 4 Ni , (OH) Ni . It can be seen that incorporation of the species Ni (OH) in the calculation affects the results only 4 when the surface pH is above -‘5. For both cases, the calculated results are in good agreement with the experimental measurements when the surface pH is below —5. The flat part of the solid line on the right results from the formation of Ni(OH) ), whose height is dependent on the solubility product 5 of Ni(OH)>. However, the slightly increasing part of the dashed line on the right is caused by the buffering action of Ni (OH). As shown clearly in the disthbution curve (Figure 71), 4 4 (OH) Ni exists in the pH range of 5-7. In the course of the formation of this species, some extra hydroxyl ions are combined with the nickel ions so as to depress the further increase of surface pH. One comment needs to be made here. The formation of Ni (OH) may be quite slow from the viewpoint 4 of kinetics due to its complex structure.  Modelling of surface pH in 2 M NiCI 2 at bulk pH 2.5 and 25°C  149  2.0 [NIOH-i-]  p  [Ni(OH)2(aq)]  p  1.8 1.6  [Ni(OH)3..c-,.]  1.4  [Ni(OH)4.c2-.]  1.2  [NI2OH.c3+>.] [N14(OH)4<4÷>J  1.0  z  0.8 0.6 0.4 0.2 0.0 -0.2  0  1  2  3  4  5  6  7  8  9  10  11  12  13  14  pH Figure 71 Sub-section distribution curve of nickel species in 0.937 M NiCI 2 at 25°C The deviation of the surface pH above —6 can reasonably be ascribed to the instability of the surface pH on the nickel-coated gold gauze due to either the vigorous hydrogen evolution or the formation of nickel hydroxide. Nickel hydroxide alters the nature of the cathode surface or the hydrodynamics within the diffusion layer.  5.3 Modelling of surface pH in 2 M NiCl 2 at bulk pH 2.5 and 25°C The same kind of modelling of the surface pH was done for 2 M NiCl 2 at bulk pH 2.5 and 25°C. The average value of the absolute viscosity is (2.0179 + 2.0347)/2=2.026 x iO (kg/m•sec). The diffusion coefficients were calculated from the Stokes-Einstein equation and are listed in Table 40. The equilibrium quotients used in the modelling are summarized in Table 41. The other conditions are: total nickel concentration [NuT, 2 (kmole/m ), bulk pH 2.5, tem 3 perature 298°K (i.e., 25°C), current efficiency of nickel reduction 99.35 % which was measured at 100-300 A/m 2 in the present work, the activity coefficient of the hydrogen ion  1 H ’  8.01, which was  also determined in the present investigation. The step length in the calculation, Ax, was set to 0.5 I.Lm. The thickness of Nernst diffusion layer hail a different value. It was found that when 6=79 p.m. the best match between the calculated surface pH and experimental data points was achieved.  Modelling of surface pH in 2 M N1CI 2 at bulk pH 2.5 and 25°C  150  Table 40 Diffusion coefficients in 2 M NiC1 10 109] 2 at 25°C Species  Symbol  At infinite dilution /sec) 2 (m  2 Ni  1 D  0.705 x l0  0.39 1 x 10 0.8903 / 2.026 x 0.705 x 10 = 0.3 10 x i0  cr  2 D  2.03 x io  0.914 x iü  NiC1  3 D  1.00 x i0  /  0.8903 / 2.026 x 1.00 x i0  NiOH  1)4  1.00 x i0  /  0.8903 / 2.026 x 1.00 x io = 0.439 x io  OH-  5 D  5.26 x  I  1.50 x i0 4 x 196 x 10 = 2.94 x i0  H  6 D  9.31 x i0  I  2.66 x  Ni ( 4 OH)  7 D  9 1.00 x i0  I  §:  ExptL /sec) 2 (m  Calcd. /sec) 2 (m  0.8903 / 2.026 x 2.03 x  =  0.892 x 0.439 x 10  io x 196 x 10 = 5.21 x i0  0.8903 I 2.026 x 1.00 x i0 = 0.439 x i0  2 aqueous solution at 25°C. In 2 M NiC1 Table 41 Equilibrium quotients in 2 M NiC1 2 at 25°C  Reaction  2 + Cl Ni  Equilibrium quotient =  2 + OH Ni  =  2 + 40W Ni 0 2 H  =  NiC1 NiOW  =  [Nj[Cfl =  Ni(OH) ( 2 ) = Ni 2 + 20W  4]  i-°•’ = 0.676  =  (OH)4] 4 [Ni  Ni ( 4 OH)  H + OH-  =  =  =  1.02 x l0  1029 = 9.77 x 10  =  ] [0H1 24 [Ni 4  =  [114] [OH1 = i0’ = 4.37 x  =  [Ni ’ 2 ] [01112 =  _is x 10  Compared with that in 0.937 M NiC1 2 at bulk pH 2.5 and 25°C, the surface pH modelling in 2 M NiC1 2 at bulk pH 2.5, as presented in Figure 72, is not entirely satisfactory, though the general trend is consistent. The reason for this may lie in the uncertainty in the value of the diffusion coefficients employed in the calculations. The assumption of the parameter N 1 in the Nernst Planck flux equation may also result in some error, as the calculated activity of water in 2 M NiC1 2 at 25°C is around 0.854. Two points are indicated by Figure 72. Firstly, whether to incorporate  Modelling of surface pH in 2 M NICI 2 at bulk pH 2.5 and 25°C  151  the species Ni (OH)+ in the calculation or not does not make any difference. Secondly, the final 4 surface pH’s match quite well. Furthermore, the calculated and experimental data all demonstrate a lower final surface pH in the more concentrated NiCl 2 solution. 7.0 6.5 6.0 5.5 5.0 4.5  5  1.0 0  40  80  120  160  200 240 280 320 360 400  440 480 520  C.D., (A1m2) Figure 72 Modelled surface pH in 2 M NiC1 2 at bulk pH 2.5 and 25°C  560  600  Fundamentals of the rotating disc electrode technique  152  • Chapter 6 Rotating Disc Electrode Study of Nickel Electrodeposition One of the best tools for studying electrode kinetics is the rotating disc electrode. The major advantage of the rotating disc electrode is that unlike a stationary electrode, a uniform diffusion layer can be maintained over the electrode surface, and the mass transfer rate can be calculated with respectable accuracy at a given RPM. So by changing RPM, one can change in a pre-determined way the mass transfer rate towards the electrode surface. In the case of simultaneous nickel reduction and hydrogen evolution, the nickel reduction is largely activation controlled and the hydrogen evolution is primarily mass transfer controlled. Therefore, it will be important to examine the electrode kinetics of these two electrode reactions and to determine how hydrogen evolution is affected by the flow rate of the electrolyte.  6.1 Fundamentals of the rotating disc electrode technique The application of the rotating disc electrode (RDE) has become increasingly important not only in electrochemistry but also in the study of chemical kinetics. Its importance is realized in its ability to control precisely a uniform mass transport rate towards and away from the reaction site.  c) r Vr  v, Z  Vz  Comprehensive knowledge of the rotating disc electrode is contained in a monograph by Pleskov and Filinovskii” and in a special review paper by Opekar and Beran” . As shown in Figure 73, the 121 RDE is composed of a conducting disc, which is a platinum disc in the present study, embedded in the centre of an outer TEFLON cylinder. The electrode surface is polished and should be perfectly horizon tal. As the electrode rotates driven by a motor, there are three motions near the surface of the rotating disc for a viscous, incompressible liquid. Figure 73 Schematic drawing of the rotating disc elec trode  The liquid velocity vector can be divided into three components: V  radial component caused by centrifugal force azimuthal component due to the viscosity of the liquid normal component resulting from the pressure drop  Fundamentals of the rotating disc electrode technique  153  These three components of motion are a function of the rotational speed, liquid viscosity, the radial distance and the vertical distance from the disc surface. They can be expressed mathematically as : 1111 follows  V. = r  (.O•  where: 0)  F()  v = -jZj .H()  V,= r o• G()  =IS\J  (327) (328)  z  is the angular velocity of the disc  (0  = 2i RPMI6O),  V  is the kinematic viscosity, r is the radial  distance and z is the vertical distance from the disc surface. F(), G() and H() are dimensionless functions. There are two special situations worth mentioning here’ : 1 (1)  At the disc surface z=0,=0,F()=0,G()=1andH()=0. Therefore,Vr=0,V,=r0),Vz=0  (2)  3.6’I10/(2it• 2000/60) = 249 (I.un) at 2,000 rpm and 25°C  z 0, Vpp  Vr Even at z =  0, Vz  0.89’1.  F() = 0.036  —  0, G() = 0.050  —  0 and H()  0.802. The thickness of the  diffusion layer depends on the magnitude of Schmidt number (Sc = v / D)’” . 11 When 100 < Sc <250  When 250 <Sc  When Sc  ,  = 1.61))(1+O.298OSc”4O.1451Sc)  = 1.61  <  J  6—161  —>00  —  (1 +O.3539Sc ) 36  (Y  ) )  R V 1/6 Ct)-1 1 161D .  —  (329)  (330)  (331)  Equation (331) is the well-known Levich equation . For aqueous solutions, the Levich equation 31 is sufficiently accurate as Sc = v ID l0 I iO = i0. For instance, the kinematic viscosity and diffusion coefficient of the nickel ion are 1.209 x 106 1 /sec and 0.542 x 10 m 2 m /sec’° 2 respectively in 1 M NiCl 2 at 25CC. Therefore, Sc = 1.209 x 10 / 0.542 x i0 = 2231. The error resulting from using the Levich equation is only: Error (%)  1  —  (1+0.2980 x 223 l’ + 0.145 1 x 2231) X 100 = 2.4 1  (332)  The thickness of the diffusion layer in 1 M NiC1 2 at 2,000 rpm and 25CC and the limiting current density are approximately equal respectively to:  Fundamentals of the rotating disc electrode technique  =  (333  (1.209 < io 1.61(0.542 x I xlOJ l2ir•2000/60) 9 O 2 l. nFDN.2+CN.2+  154  =9.4(pn)  2 x 96500 x 0.542 x 10 x 1 x 10 = 9.4x10  (334) =  11,000 (A/rn ) 2  If no other background electrolytes are present, the NiC1 2 solution should be treated as a binary electrolyte. For a binary electrolyte, the limiting current density should be calculated according to equation  (335)[111]  =  z+FD+(l  +-_  Iz_I)ax  =11  (335)  z+zFDC  iz_iJ  eff.  where the effective thickness of the diffusion layer can be expressed as follows: 1/2  Dsaii =  \1/3f  i =1.61_)  v 2.RPM/60J i  1/2  I z. )  (336)  (337)  zD + z_  D = 0.542 x i(Y 2 /m sec and D  =  1.32 x i(Y m 2 at 25C’° . Thus from equation 91 /sec in 1 M NiC1 2  (337), D can be calculated as: =  0.542x 10x 1.32x 10(2+ 1) =0.893 x 10(rn /sec) 2 2x0.542xl0+ 1 x1.32x10  (338)  It follows from equation (336) and (335) that: =  1.61  lL1+J  0.89 3 x 1 0 (i .209 x 10 1h12 = 11.1 (Wn) 1.209 x lOJ 2000/60)  (339)  2 x 96500 x 0.542 x 10 x 1 11.1x10  (340)  X  10 =  28,300 (A/rn ) 2  Obviously, the limiting current densities calculated from equations (334) and (340) can hardly be exceeded in the normal experimental tests. All of the above-mentioned equations for the RDE are applicable only for the laminar flow. The mode of liquid flow is characterized by a dimensionless number, called the Reynolds number, Re 0r /v, where Co is the angular velocity of the disc (rad/sec), r is the radius of rotating disc (m), 2 and  is the kinematic viscosity of the liquid (m /sec). The conversion from the laminar flow to 2 turbulent flow is a gradual process. First, the disc edge is affected by turbulence and then gradually V  this effect spreads towards the centre of the disc as the rotational speed of the disc is increased. The  Fundamentals of the rotating disc electrode technique  155  critical Reynolds number for the conversion from laminar flow to turbulent flow is around 1.8 3.1 x io (h12j For the r =6 mm in the present study, the applicable maximum RPM will be in the range of: —  60 X 10 RPM <(1.8—3.1) x i0.- = (1.8—3.1) x 5 i0 2 2tr 2 x 3.1415 x (6 x 10_3)2  (341) 48,000—82,000  The above calculated rotational speed is extraordinarily fast. In practice, it is rarely exceeded in the experimental tests. If the disc vibrates vertically or radially, or if the disc surface is not perfectly smooth, turbulence may occur at a much lower Reynolds number than that calculated above. Another situation which must be avoided is when the diameter of the disc is comparable to, .5, = the thickness of hydrodynamic boundary layer which is equal to 0.249 mm at 2,000 rpm. At a sufficiently low rotational speed, the natural convection becomes significant. According to the Reynolds number, this situation will occur when Re  <  10  [h12]•  60v 60 x io RPM>10—=10 =3 2 2itr 2 x 3.1415 x (6 x 10_3)2  (342)  The ratio of the diameter of the outer insulator to the diameter of the disc must be large enough to eliminate the edge effects. Using the RDE, one of the most distinguishable features is that the thickness of the diffusion layer is known. All of general kinetic equations are applicable for the RDE. For instance, in the case of the mixed concentration and activation control (larger overpotential lii> 100 mV), the current density can be expressed as : 1141  (  (343)  1’  i =nFkCS=nFkCbj 1. \  •  i I  i i 1 1 pzFkCbiLnFkCbnFDCb/  Since the thickness of the diffusion layer,  ,  (344)  is inversely proportional to the square root of rotational  speed, the concentration polarization can be decreased at a higher rotational speed. If there is a preceding homogeneous chemical reaction, the current density is composed ofthree components : 114 k  (345)  A Ox+ne —Red  inFkCb  In the case of RDE, the thickness of diffusion layer is uniform over the disc surface and can  Experimental apparatus, procedures and conditions for the RDE tests  156  be calculated accurately in advance. Therefore, experimental conditions can be reproduced easily and the effect of the concentration polarization can be corrected readily. Experiments were undertaken to determine the dependence of the rates of nickel reduction and hydrogen evolution upon the concentrations of nickel, hydrogen and chloride ions, to determine the effect of rotational speed on hydrogen evolution, and to study the behaviour of nickel reduction and hydrogen evolution over a wide range of potential.  6.2 Experimental apparatus, procedures and conditions for the RDE tests The rotating disc electrode system used in the present study was an EG&G PARC Model 636 Electrode Rotator. Its rotational speed can be adjusted in the range of 50—10,000 rpm with an error of less than 1 %. As shown in Figure 74, the active electrode surface is platinum and its diameter is 4 mm. Therefore, the active area is equal to 7t(4/2) 2 = 12.57 = 1.257 x i0 m . With the 2 maximum current output of 2 amperes from the SOLARTRON 1286 Electrochemical Interface, the achievable current density can be as high as 2/1.257 x i0 5 = 159,000 A/m , which is well 2 beyond the maximum current density of interest for the study of nickel electrowinning. The sur rounding insulator is a TEFLON cylinder having a diameter of 12 mm.  Figure 74 Dimensions of the surface of the rotating disc electrode  Figure 75 Schematic drawing of the appara tus for the rotating disc electrode study  A schematic drawing of the experimental set-up is shown in Figure 75. The cell had a lid with five holes. These five holes positioned the working, counter and reference electrodes, a pH electrode and a gas sparger. The working electrode, although initially a platinum disc, was always precoated  Experimental apparatus, procedures and conditions for the RDE tests  157  with a nickel film (around 1 jim) before any tests. The immersion depth of the RDE into the electrolyte was about 10 mm. The counter electrode, directly below the RDE, was a pure metallic nickel disc with a diameter of 10 mm. The pH and calomel reference electrodes were placed on either side of the RDE. The electrical contact of the RDE to the potentiostat was made possible by two silver-carbon brushes in the upper part of the RDE as shown in Figure 75. The cell and electrodes were all from EG&G PARC. For all of the tests, only 100 mL of electrolyte was poured into the cell. The pH ofthe electrolyte was maintained constant through a pH electrode and a RADIOMETER titrator. One caution that has to be exercised with the rotating disc electrode is occasioned by the ohmic resistance between the working RDE and the SCE reference electrode, and between the Ag-C brush and the rotating cylinder. For most applications, it is suggested that a Luggin capillary be used and placed as close as possible to the working electrode surface in order to minimize the IR drop. For the rotating disc electrode, this method will not work well, since any objects close to the rotating disc surface will affect the hydrodynamics nearby and thus alter the mass transport equations applicable to the RDE. Furthermore, the ohmic resistance in the Ag-C brush zone will always be there. The potentiostat used in the present study was the SOLARTRON 1286 Electrochemical Interface which has two optional facilities for the ohmic drop compensation in the mode of potentiostatic operation. One is called the feedback technique and the other is called the sampling technique. When using the feedback technique, one has to know exactly the parasitic ohmic resistance between the working and reference electrodes, whose measurement can be done using an oscilloscope or the AC impedance method. There is no current interruption during this com pensation. One disadvantage with the feedback technique is that one can have less than 100 % compensation only. Once one feeds back a resistance which is equal to or greater than the parasitic ohmic resistance between the working and reference electrodes, the electronic circuits inside the SOLARTRON will become unstable. In the sampling technique, one does not need to know the parasitic ohmic resistance between the working and reference electrodes. Actually, the SOLAR TRON reads the electrode potential immediately after the current interruption (Interruption time is on the orderof27 jisec). One caution that has to be exercised is that one needs to do some preliminary test work to make sure that such a short current interruption will not affect or affect very little the electrode process which is being studied. An example is given in Figure 76 to elucidate the magnitude of the effect of the IR drop on the polarization curve,. It can be easily seen from this graph that the difference in current density at a given potential or the difference in potential at a given current density increases significantly  Experimental apparatus, procedures and conditions for the RDE tests  158  with increasing polarization. Therefore, any omission of the consideration of the ohmic drop will result in a large error especially at high current densities. Figure 76 The effect of ohmic dmp on the polarization curve (0.937 M NiCl , pH 2, 25C, 1,000 1pm, 2 5 mV/sec and bare Pt)  2000 1800 1800 1400 1200 1000 800 600 400 200 0 -1.4 1.3 -12 -1.1  -1 -0.9 -0.8 -0.7 -0.6 -0.5 0.4 -0.3 Potential vs. SCE, (volt)  Hydrogen evolution is commonly associated with nickel electrodeposition. Therefore, the contribution of hydrogen evolution to the total current must be deducted in order to obtain the authentic current for nickel reduction. The mea surement of current efficiency with a rotating disc electrode is somewhat difficult, since the deposit is so small and it is difficult to detach the electrode tip, using the technique of the weight difference of  the cathode before and after tests is not feasible. Also if too much nickel is deposited on the disc, it will make measurements less reproducible, because the hydrodynamics near the disc surface will be affected and Levich’s equation will become invalid. In-situ separation of these two currents is very difficult, although some workers have confirmed the possibility of using a very thin (16.2 I.I.m) palladium membrane as a bipolar electrode . Nickel was deposited on one side and the permeated 161 atomic hydrogen was oxidized on the other side. The collection efficiency of the hydrogen current was claimed to be around 97 %. Such a high collection efficiency can hardly be imagined when there is a copious evolution of hydrogen. Philip and Nicol 132 and Finkelstein et al 331 used a more practical technique to determine the partial current density for nickel deposition. They first deposited a layer of nickel on the platinum substrate at a constant current density and then dissolved it anodically at the same current density. The end-point for anodic dissolution was determined when the potential increased considerably to a level where other electrode reactions (probably chlorine evolution in NiCl 2 electrolyte) might take place. Therefore, the current efficiency of nickel will be equal to 100 times the ratio of the time for the anodic dissolution to the time for the cathodic deposition. Once the current efficiency is known, the partial current density for the nickel reduction can be easily determined. There are three drawbacks to this method. Firstly, when the current density is too high, the nickel deposit cannot be dissolved uniformly and completely before the potential reaches a higher level where a second reaction (e.g., chlorine evolution) may take place. Secondly, the gases, such as chlorine, generated in the anodic dissolution must be removed from the electrolyte before a second test can be carried out. Thirdly, the ohmic drop cannot be compensated readily in the mode of galvanostatic operations.  Experimental apparatus, procedures and conditions for the RDE tests  159  The technique used in the present study was based on potentiostatic operation. With the advanced computer software available, the curve of current density vs. time can be recorded and the number of coulombs can be integrated readily from such curves. The ohmic drop can be compensated using the SOLARTRON 1286 Electrochemical Interface. For the cathodic deposition, the working electrode was fixed at a constant potential and the curve of current density vs. time was recorded accordingly. Typical curves are shown in Figure 77.  120 100 d80  0  60  40 20 0  50  100  150  200  250  Time, (sec)  300  350  400  450  500  80  120  160  200  240  280  320  (sec)  Figure 77 The current density vs. time for potentiostatic operation (0.3 M N1C1 2 + 2.7 M CaC1 , 2 0.005 M HC1 <pH 0.90>, 25°C, 2,000 rpm, Ni-coated Pt) As shown in Figure 77, the current density is quite stable at a potential of -0.750 volt vs. SCE. There is always a hump at the beginning when the overpotential is higher, as in the case of -0.850 volt in Figure 77. This hump is believed to be caused by the concentration polarization. In the beginning, less concentration polarization exists; however, as the electrodeposition proceeds, the concentration polarization becomes greater and finally reaches a stable level. After a layer of nickel is deposited, it remains to be dissolved. Galvanostatic anodic dissolution is quite simple; however, as has been mentioned, it suffers from some disadvantages. Straight forward potentiostatic anodic dissolution is not good either. It has been observed that there is always a very sharp current peak at the beginning. Thus, when the measured current is integrated against the dissolution time, there will be a large error. The technique used in the present study was as follows. Potentiostatically, the anodic disso lution starts from a potential close to the equilibrium or rest potential of the working electrode. The potential of the working electrode was then increased at a pre-defined rate (mV/sec) towards the specified end potential. Once the end potential is reached, the working electrode will stay at that potential until the end of the dissolution. In this way, the initial current peak occurring in the straightforward potentiostatic anodic dissolution and the risky gas generation or substrate dissolution can be avoided. An example of the curve for anodic dissolution obtained in the present study is  Reaction orders of the rates of nickel reduction and hydrogen evolution with respect to the concentrations of electrolyte components  160  given in Figure 78. The number of coulombs for the anodic dissolution of nickel can be obtained by integrating numerically the current density against the dissolution time. The sweep rate (1--20 mV/sec) and the rotational speed (50—2,000 rpm) during the anodic dissolution were found 0.1 0 -0.1 0 >  -02  w -0.3 -0.4  C 0  -0.5 0 -  -0.6 -0.7 -0.8  0  20  40  60  80  100 120 140  160 180 200  not to affect the measurement of current effi ciency. The current efficiency of nickel can thus be calculated from the ratio of the number of coulombs obtained in the anodic dissolution to the number of coulombs obtained in the cathodic deposition. Figure 78 The current density vs. time for linear potentiostatic anodic dissolution (0.937 M NiCI 2 + 0.485 M 3 B0 pH 2, 25°C and 2,000 ipm) H ,  Time, (sec)  For all of the tests carried out with the rotating disc electrode, the electrolyte was deaerated by bubbling nitrogen gas for 20 minutes before each test. During the test, nitrogen gas was passed over the electrolyte surface. The nickel counter electrode was cemented in aLecoset7O7cold-curing resin, leaving its other side exposed. The conducting parts in the wiring were painted using MICCROSTOP stop-off lacquer (MICHIGAN CHROME and Chemical Company, 8615 Grinnell Ave., Detroit, Michigan 48213, USA). Before each test, the Pt disc was precoated with a fresh layer of nickel (—1 tm) at a low current density of 100 A/m 2 for 300 seconds in the test electrolyte. All of the tests were conducted only at 25°C due to the limitations resulting from the differential thermal expansion between the platinum disc and the TEFLON insulator. The electrolytes were prepared using the deionized water and A.C.S. reagent grade chemicals 2 •6H NiC1 0 ,2 6H 4 NiSO O , B0 4 3 H , NH C 1, NaCl, 4 SO and HC1. Short-term pre-electrolysis was carried out, even though 2 Na it was found by other workers 55 that the effect of pre-electrolysis was negligible for nickel electrodeposition.  6.3 Reaction orders of the rates of nickel reduction and hydrogen evolution with respect to the concentrations of electrolyte components When the reaction order of a certain electrode reaction with respect to the individual electrolyte component is to be determined, particular attention should be paid to its activity coefficient or the ionic strength of the electrolyte. A convenient way to deal with such a study is to use the con centration instead of the activity in an electrolyte having a constant ionic strength. For the studies in acidic nickel chloride electrolyte, calcium chloride is the preferred background electrolyte, in view of its electrochemical inertness and its certain similarity to nickel chloride. The total con-  Reaction orders of the rates of nickel reduction and hydrogen evolution with  161  respect to the concentrations of electrolyte components  centration of CaC1 2 plus NiC1 2 was maintained constant, always equal to 3 M. A lower pH electrolyte was chosen deliberately in order to produce a detectable amount of hydrogen evolution. The current density was measured at six different potentials, i.e., -0.70, -0.75, -0.80, -0.85, -0.90 and -0.95 volt vs. SCE. 50 Potential, (volt) -0.95 -0.90 -0.85 -080 -0.75 -0.70 I • —A-— vs. SCE — —— —• 40 0  c’J E  c’J 30 E  z  c’J 120  0.1  0.2  0.6  0.5  0.7  0  p  A  A  A  •  •  0  0.4 [NiCI2J, (M) 0.3  o •  10  0  0  o  o  0  A  A  .  -  p  P  —  S  •  p  p *  p  A  A  I  I  I  I  0.1  0.2  0.3  A  I  •  0.4 [N1CI2I, (M)  0.5  0.6  0.7  Figure 79 The current densities of nickel reduction and hydrogen evolution as a function of nickel concentration (NiC1 2 + CaCI 2 = 3M, pH 1.1, 25CC, 2,000 rpm and Ni-coated Pt disc) 800  180  Potential, (volt) -0.95 -0.90 -0.85 -0.80 -0.75 .0.70 —.——e— —*-- —.-— —.-- —*-— vs. SCE  700 600  0  500  p  160 0  ,,  120  p  a  140  100 400 A  A  A  A  ‘300 a  a  p  .  =  p  200  40  100  20 0  0.002  A  p  p  0  80 I 60  I  I  0.004  0.006  A  A  I  0.008 0.01 [HCI], (M)  I  A  0  • -  0.012  0.014 0.016  0  0.002  0004  0.006  0.008 0.01 [HCl], (M)  0.012  0.014 0.016  Figure 80 The current densities of nickel reduction and hydrogen evolution as a function of HQ concentration (0.3 M N1C1 2 + 2.7 M CaCl C, 2,000 rpm and Ni-coated Pt disc) 0 , 25 2 The results obtained for the effect of nickel ion concentration are presented in Figure 79. What seems clear from Figure 79 is that the rate of the nickel ion reduction is directly proportional to the concentration of nickel ion, while the rate of the hydrogen evolution is independent of the nickel ion concentration. This finding is not surprising for the reduction of nickel ion as a first order reaction is observed for most metal ion reductions. Zero reaction order with respect to the nickel  Reaction orders of the rates of nickel reduction and hydrogen evolution with respect to the concentrations of electrolyte components are  ion concentration for hydrogen evolution demonstrates that there between nickel  and hydrogen ions  162  no substantial interactions  during their simultaneous reductions. It also indicates that the  contribution of nickel hydroxy complexes to hydrogen evolution is negligible . 1 The reaction orders with respect to the electrolyte acidity respect to the nickel ion concentration. As shown in  Figure  are just  the opposite to those with  80, the current density of nickel reduction  is independent of the HCI concentration in the range of 0.002—0.015 M.  Interestingly, however,  the current density of hydrogen evolution increases linearly with increasing electrolyte  acidity.  The  independence between the current density of nickel ion reduction and the electrolyte acidity indicates that the widely held mechanism (reactions 347-349) for nickel ion reduction is not true.  +H =NiOH+H 2 Ni O  (347)  r4.s.  (348)  NiOH + e  -,  (NiOH)  (NiOH)+H+e =Ni+H 0 2  (349)  If the mechanism represented by reactions (347)-(349) were correct, the nickel ion reduction would have to be pH dependent. No clear reasons  are given in the literature as to why so many investigators  have believed that the above mechanism is correct. In Wruck’s Master’s thesis work which dealt with the reduction of nickel ion in the electrolyte 0 -NaC1-HC1-H 2 NiC1 , the above mechanism  was  still accepted even though the effect of neither pH nor the chloride concentration was investigated . 1401 800  Potential. (volt) vs. SCE  700  -0.90  -0.85  —0-—  —a—  100  -0.80  Potential, (volt) vs. SCE  90  —  -0.90 -0.85 -0.80 —e— —a-- —&—  00  600  C,  00  80 r,  0 U  0  0 0  0  0  70  0  0  0 0  —400 -  0  0000  Z  300  0  0 0 A  A  A  20  A  p  0  A  A  A  A  0.4  81  0.8  12 1.6 lclrr, (M)  2  2.4  2.8  0  I  I  0.4  0.8  A  A  10 0  Figure  1  A  A  0  0  30 A  0  0  X40  0  200 100  fl  I  I  1.2 1.6 [C]T, (M)  A  I  I  2  2.4  concentration  and hydrogen evolution as a function of chloride 2 +3 M (NaC1 + NaClO ) 4 [0.5 M Ni(C10 ) +0.005 M HC1, 25C, 2,000 rpm 4  and Ni-coated  Pt disc]  2.8  The current densities of nickel reduction  Although the reaction orders  are based  concentrations have been used, it respect to the activities of the ions in question.  on kinetic data in which molar  is more accurate to state the reaction orders with  Reaction orders of the rates of nickel reduction and hydrogen evolution with respect to the concentrations of electrolyte components  163  The effect of chloride concentration on the current densities of nickel reduction and hydrogen evolution was also studied in the present work. As shown in Figure 81, there are some fluctuations in the results; however, considering the overall trend, the current densities of nickel reduction and hydrogen evolution appear to be relatively constant over the total chloride concentration ranging from 0.2 to 2.5 M. As the ionic strength of the electrolyte was controlled by using sodium perchlorate, sodium chloride was the major source of the chloride ion. The source of nickel ion was the compound nickel perchlorate Ni(Cl0 . The zero reaction order with respect to the chloride ion concentration 2 ) 4 for the reduction of nickel ion appears to be conditional. In their studies on the effect of chloride ion concentration on nickel reduction under the conditions of 1 M Ni(C10 2 + 2 M (NaC1 + NaC1O ) 4 ), 4 54CC and 0.01-1 M [CuT, Philip and Nicol and Finkeistein et aI 331 found that the slope dlog(i M)/dlog([Cl]T) was equal to 0.87 at -0.625 volt vs. SCE, indicating almost a first order reaction. There are no reasons to suspect their results. In fact, somewhat similar results were obtained in the present investigation in the lower range of chloride ion concentration. For example, the current density of nickel reduction at 0.2 M [CuT was 476 A/m 2 at -0.90 volt vs. SCE, compared with 546 A/rn 2 at 0.4 M [CuT. However, the sloped log iN/d log[Cl]T is only 0.19, or d logi /d log[ClJ 1 is only 0.18 at -0.9 volt vs. SCE in the range of [C1IT = 0.2 —0.6 M. Furthermore, the tests at below 0.2 M [CuT were difficult, as the nickel deposit did not adhere well to the platinum substrate. 51 considered that the enhanced rate of nickel ion reduction is due to the introduction Florence’ of chloride ions into the primary solvation sphere of the nickel ions, so that the lability of the remaining primary water molecules is drastically increased. Considering the fact that the current density of nickel reduction does not increase continuously with increasing the chloride concen tration, the effect of chloride ion may come from the interfacial interaction rather than from the change in the bulk electrolyte properties. It is well known that chloride ion is a strong adsorbent. The adsorption of chloride ions on the cathode surface may decrease significantly the potential j i 1 at the outer Helmhokz plane where nickel ions accept electrons from the cathode. When the concentration polarization and Ni are taken into account, the Butler-Volmer equation can be written as follows if -i i  100/n mV:  (  i  r  (  =zFkCbI l—-- lexpi  —  1j)  ( i “i =zFkCb 1_i: ex  zF’qf F(E-V ) 1 Iexvl RT ) RT [  —________  r  (z—an)Fijf l 1 ( WZFE exp RT RT For the reduction of nickel ion, equation (350) can be transformed into: —  -  (350)  Effect of RPM on hydrogen evolution and electrode potential during nickel electrodeposition r T (2—0.5x1)F’qi l 1 iFE” NI 1 1 .2 lNI2Fk[M ]1 [1_.—_jexP[_ jexPt_RT RT Ni(L) 1  J  164  (351)  It can be seen from equation (351) that any negative shift in q1 1 will increase the current density of nickel reduction. If the cathode surface is attacked chemically due to the presence of chloride ion, the rate constant kin equation (351) will change as well. However, there is no quantitative equation to describe such a chemical change. Unlike the reduction of nickel ions, the rate of hydrogen evolution is hardly affected by the presence of chloride ions. The slight fluctuations visible in Figure 81 are likely caused by the fluctuation of the electrode potential due to sodium perchlorate. It was observed experimentally that the presence of NaC1O 4 affected the ceramic junction of the calomel electrode. The corre sponding potential change could sometimes be as high as 20 30 mV. The reason for such a phenomenon can be attributed probably to the precipitation of KC1O 4 within the ceramic junction -  as a result of its very low solubility. For the same reason, the combination glass pH electrode was affected by NaC1O . In this case, the pH reading would drop up to 0.5 unit if the pH electrode was 4 left in the electrolyte for more than 1 hour. On account of this unusual situation, a double liquid junction was used to avoid the direct contact of the calomel electrode with the NaC1O 4 solution. In addition the acidity of the electrolyte was controlled rather than the pH. Theoretically, the negative shift in  will affect the rate of hydrogen evolution. A similar  expression can be obtained from equation (350): (352) iH_Fk[H][1  ]P[RT](I?1J  It is clear from equation (352) that the effect of  is less significant for hydrogen evolution. Another  consideration has to be made which concerns the size of the hydrogen ion. Although the hydrogen ion exists in a hydrated form in aqueous solution, the degree of hydration is much less than that of the Ni 2 ion. Therefore, the hydrogen ion accepts electrons from the cathode at a distance closer to the cathode surface than does the nickel ion. Accordingly, ji 1 for the hydrogen ion is not same as that for the nickel ion. Nevertheless, the effect of Ni 1 is negligible for hydrogen evolution on the basis of the experimental results.  6.4 Effect of RPM on hydrogen evolution and electrode potential during nickel elec trodeposition One of the most interesting features of hydrogen evolution during nickel electrodeposition is the effect of agitation. The results obtained with the rotating disc electrode are shown in Figures 82 under the conditions of 25°C and -0.850 volt vs. SCE on the Ni-coated Pt disc. It is obvious here  Effect of RPM on hydrogen evolution and electrode potential during nickel e lect rode pos ition  165  that for all of the electrolytes tested, hydrogen evolution increases with the rotational speed. This prompts the belief that the rate of hydrogen evolution under all of these conditions is mainly con trolled by mass transfer while the rate of nickel reduction is more or less independent of its mass transfer rate. In other words, to improve the mass transfer, e.g., by increasing the circulation rate of the electrolyte, during nickel electrowinning is not a good way to raise the current efficiency of nickel reduction. In terms of the surface quality of the nickel deposit, however, the improved mass transfer is often desirable. Consequently, from a practical standpoint, a compromised circulation rate of the electrolyte should be employed. 100  pH 1.1 50 7MNiCI2  0.937 M NICI2 50  pH 1.1  30  20  —rio 2:  CI2  (‘ilo  w  o  05  2  3 MNiCI2°  0.5 50  100  200  500 RPM  1000  2000  2  5000  120  50  100  200  500  1000  2000  5000  RPM  20  10  Figure 82 The effect of rotational speed on the current efficiency in various electrolytes and at different PH’S (25CC, -0.850 volt vs. SCE and Nicoated Pt disc)  s  2  120  RPM  One point which needs to be mentioned here is the measurement of the nickel electrode potential. Ithas been found difficult to measure the thermodynamic equilibrium electrode potential of nickel. As shown in Figure 83, the platinum substrate was first coated with a fresh layer of nickel film at 25°C, 2,000 rpm and 200 Aim 2 (2 urn) or 400 Aim 2 (4 urn) for 300 seconds, and then the electrode potential was monitored as a function of time at different rotational speeds. The elec trolytes were deaerated before each test by bubbling nitrogen and a stream of nitrogen gas was maintained over the electrolyte surface during the test. For all of the electrolytes tested, the rotational  Effect of RPM on hydrogen evolution and electrode potential during nickel electrodeposition  RPM 2000  -0.1  0  RPM  501100120040085011600132001  0  -0.2  50 110012001400 1800 i1600i3200i  —r  -0.2  -0.3  0  -0.3  -0.4  -0.4  CI)  Ci)  -OS  -Os  >  .13  2000  -0.1  i- .i  --  166  >  .13  -0.6  C  -0.7  13.  -0.6  C  Deposi& at 200 A/m2  floposi6on a1200 A m2 1  -0.7  0  -0.8  -0.8  -0.9  0.937 0  -0.9  N1CI2 at pH 2.5  0.937MN1CI2atpH 2  100 200 300 400 500 650 700 800 900 1000 1100 1200 1300 14001500  0  100 200 300 400 500 600 750 800 900 1000 1100 1200 1300 1400 1500  Time, (see) 0  Time, (see) 0  RPM 0  2000  -0-I  RPM  50 100 200 400 80) 1600 3200  0  2000  -0.1  -02  0  50 100 200 400 800 1600 3200  0  -0.2 0  -0.3  >  -0.3  -04  C))  Cl)  -0.5 > (0 -0.6  -0.5 > (0 -0.6 Deposition  Deposition at400Afm2  at 200 Afm2  -0.9  0.937 M N1CI2 I  0  •  I  I  -0.9  0.937MN1C12 at pH 1.1  atpH 1.5  I  100 200 300 400 500 600 700 800 900 1600 1100 120013001400 l 00  I  O  Time, (see)  •  I  I  •  I  •  I  100 200 300 400 500 600 700 800 900 1000 11001200 1300 14001500  Time, (sec)  0 -0-I -02 -0.3  -O4 (I) -0.5 > (0 -0.6  -0.9  0  100 200 300 400 550 600 750 860 900 1000 1100 1200 1300 1400 1500  Time, (see)  0  100 200 300 400 500 600 700 800 900 18001100120013001400  Time, (see)  Figure 83 The effect of rotational speed on the electrode potential in electrolytes of pure nickel chloride (started with Pt substrate at 25 °C)  Polarization curves of nickel reduction and hydrogen evolution  167  speed had a remarkable effect on the nickel electrode potential. The maximum difference in the electrode potential was around 100 millivolts. Furthermore, immediately after the current was switched off, the electrode potential increased gradually even at the same rotational speed. This phenomenon indicates that the nickel electrode surface undergoes some changes, such as, chemical attack by H ion or by traces of dissolved oxygen. This finding also emphasizes the importance of the effect of agitation in studies on electrode kinetics. Comparing the electrode potentials in 0.937 M NiC1 2 at pH 2.5’- 1.1 (Figure 83), the changes in electrode potential follow almost the same contour, independent of the bulk electrolyte pH. If it is supposed that the nickel electrode behaves like a pH electrode, the electrode potentials would be -0.389, -0.359, -0.330 and -0.306 volt (vs. SCE) corresponding to pH 2.5, 2.0, 1.5 and 1.1. Also if it is assumed that under no rotation the chemical dissolution of metallic nickel, Ni + 2W = Ni 2 , reaches equilibrium, the surface pH would be around 4.23 under equilibrium with 1 M Ni 2 H 2 at 25CC. At a fast rotational speed, e.g., 3,200 rpm, it can be reasonably assumed that under no +  current passage the surface and bulk pH’s are the same or at least very close to each other. If these assumptions are true, the rotational speed should affect more significantly the electrode potential at a lower bulk pH. Also the electrode potential at a lower pH under rotation should be higher than that at a higher bulk pH. From the results in Figure 83, these two inferences are obviously wrong. Therefore, the nickel electrode is not behaving like a pure pH electrode in acidic media. The decrease in the electrode potential upon stopping the disc rotation can be understood as a result of the increase in the surface pH due to the chemical dissolution of metallic nickel Ni + 2W Ni + H 2 . However, the decrease in the electrode potential cannot be explained by the pH change 2 alone. The other interfering factor is believed to be traces of dissolved oxygen. Even though the electrolyte was deaerated by bubbling nitrogen gas before the test and a stream of nitrogen was =  maintained over the electrolyte surface during the test, a small amount of dissolved oxygen is inevitable especially when the electrode is rotated rapidly. This fact may be evident in view of the slight increase in the electrode potential with time at a given RPM. The effect of RPM on the electrode potential diminishes dramatically in 3 M NiCl . The addition of 2 M NaCl had a similar 2 effect. The more concentrated 2 NiCl electrolyte or the addition of NaCl are believed to increase the activity of the nickel ion, leading to a depressed chemical dissolution of metallic nickel. 6.5 Polarization curves of nickel reduction and hydrogen evolution Polarization curves are useful in assisting the understanding of the electrode behavior under a wide range of potential. They provide important information about the possible maximum current density applicable in practical electrowinning. Using the advanced potentiostat, the SOLARTRON 1286 Electrochemical Interface, together with the powerful computer software, the measurement of a single polarization curve can be realized in a matter of several minutes. One ofthe key parameters  Polarization curves of nickel reduction and hydrogen evolution  168  for measuring the polarization curve using the technique of linear potential sweep is the sweep rate in units of mV/sec. Generally speaking, a slower sweep rate is required for obtaining a steady-state polarization curve. However, for the rotating disc electrode, an overly slow sweep rate may produce some adverse effects if too much deposit is plated out on the disc. It was found for the cathodic reduction of nickel ions that a sweep rate at 2 mV/sec was slow enough for the measurement of polarization curves. Further lowering of the sweep rate would not create any substantial differences. A typical polarization curve for the electrolyte 0.3 M NiCl 2 + 2.7 M CaCl 2 is shown in Figure 84. The numbers in parentheses were obtained from prolonged potentiostatic tests. It is evident that the polarization curve obtained at a sweep rate of 2 mV/sec is almost in the steady-state.  51 500  E  -0.9  -0.7 -0.8 -0.6 -0.5 Potential vs. SCE, (volt)  -0.4  -0.3  Figure 84 Polarization curve at a sweep rate of 2 mV/sec (0.3 M NiC1 2 + 2.7 M CaC1 , 0.005 M HC1 2 <pH -0.9 >, 25°C, 2,000 rpm and Ni-coated Pt disc)  4500 (1) 0.937 M N1CI2  +  1.31 M NH4CI  (2) 0.937 M N1CI2  +  0.485 M H3B03  +  0.365 M Na2SO4  4000 3500  (3) 0.937 M NiCI2  3000  (4) 0.937 M NiCI2  (5) 0.937 M NICI2 +2 M NaCI  C’]  E  2500 -0.850 volt vs. SCE  b d  2000  C.D., (A1m2)  1500 1000 500  CE(Ni),  (1)  197  91.1  (2)  295  91.1  (3)  274  89.3  (4)  222  88.8  (5)  468  96.4  (%)  0 -1.3  -1.2  -1.1  -1  -0.9  -0.8  -0.7  -0.6  -0.5  -0.4  Potential vs. SCE, (volt) Figure 85 Polarization curves of combined nickel reduction and hydrogen evolution in different electrolytes (2,000 rpm, pH 2, 25°C, 2 mV/sec, Ni-coated Pt disc)  -0.3  Polarization curves of nickel reduction and hydrogen evolution  169  For the convenience of comparisons, five polarization curves in five different electrolytes are plotted together in Figure 85. The current densities in Figure 85 are the combined values of nickel reduction and hydrogen evolution. The current densities and current efficiencies of nickel measured at -0.850 volt vs. SCE are presented at the lower right inside Figure 85. By comparing these five current efficiencies of nickel, the same kind of information is revealed here as found previously in the tests of nickel electrodeposition at 60°C. That is to say, the current efficiency at the same pH is highest in the electrolyte of 2 NiC1 NaCl, and the lowest in . 4 NiC1 S 2 Na O Also, both of the additions of NH C1 and 3 4 B0 increase the current efficiency of nickel. In terms of the current H density, the highest value was achieved in the electrolyte NiC1 -NaC1 when the electrode potential 2 was not more negative than -0.96 volt vs. SCE. The addition of boric acid increased the total current density very little compared with the pure nickel chloride electrolyte if the electrode potential was above -1.02 volts vs. SCE. However, the addition of 4 SO and especially NH 2 Na C1 decreased the 4 total current density. The increase or decrease in the total current density can be attributed mainly to the increase or decrease in the activity of the nickel ion in the electrolyte resulting from the addition of the individual components. One unique feature of the polarization curves in all of the electrolytes tested except for the addition of NH C1 is that the polarization curves have a peak at a potential somewhere between 4 -0.97 and -1.15 volts vs. SCE. The reasons for the occurrence of these peaks are not quite clear. However, it is considered that the formation of insoluble nickel hydroxide or oxide on the electrode surface is very probably responsible. This speculation seems reasonable in that the height of the peak is related to the electrolyte composition explainable through the nickel ion activity, the buffering capacity of electrolyte and the rate of hydrogen evolution. For instance, in the electrolyte 0.937 M NiCl 2 + 2 M NaCl, the activity of the nickel ion was raised and the acid concentration at the same pH was reduced. These two reasons may account for the lower peak height in spite of a higher current efficiency of nickel. On the other hand, for the electrolyte 0.937 M NiC1 2 + 0.485 M B0 the peak height was raised considerably as a result of the enhanced buffering capacity of H , 3 the electrolyte in the presence of boric acid. Even though it seems somewhat controversial to say boric acid is a pH buffer during nickel electrodeposition, it is believed that boric acid does have some catalytic function at a lower surface pH in view of the higher current efficiency of nickel, and that it does behave like a pH buffer at a higher surface pH on account of the lower surface pH observed at higher current densities and the pH titration results (compare Figure 50 with Figure 38). Compared with the pure nickel chloride electrolyte, the almost double height of the peak in the presence of 0.485 M 3 B0 in Figure 85 cannot be explained only from the —2 % increase in the H current efficiency of nickel.  Polarization curves of nickel reduction and hydrogen evolution  170  When ainmonium chloride was added to the nickel chloride electrolyte, the change in the shape ofpolarization curve was substantial, indicating the disappearance of the peak. The couple NH/NH 3 has a middle-point buffer pH of around 9.25 at 25°C. Accordingly, one would not expect any substantial buffering action under acidic conditions (pH < 7). As with boric acid, this middle-point buffer pH may be shifted to the acidic region in the presence of nickel ions due to the formation of strong nickel ainmine complexes: Ni + 2 xNH  —>  (353)  2 ) 3 xH+Ni(NH  Bjerrum 6 ” 1 1 had studied the nickel ammine complexes in 2 M 3 NO and 1 M NH 4 NH C1 at 30°C. It 4 was found that the number of NH 3 bound to Ni 2 ion increased continuously with increasing NH 3 concentration, starting from Ni(NH 2 up to Ni(NH ) 3 ). The following calculations indicate the 3 pH above which nickel monoammine complex Ni(NH ’ should start to form. 2 ) 3 AG’  + NH 2 Ni 3  =  2 ) 3 Ni(NH  log  f3 = 2.82 (1  The standard formation free energies for aqueous species Nit),  M NH C1O at 25°C)’ 4 171  NH ( 3 aq)  and  Maq>  at 25°C are  -46.4 kJ/mole , -26.6 181 1061 kJ/mole’ and -79.4 kJ/mole’ , respectively. Thus, the formation free 181 energy of Ni(NH 2 is equal to: ) 3 IXG.(NH)2+  —2.303RT log f +  =  + + AG?H 2 ° + AGJ. 1 AG  =  —2.303 x 8.314 x 298 x 2.82/1000 + (—46.4) + (—26.6)  =  + AGH  (354)  =—89.1 LI If Ni(NH 2 is assumed to be the first nickel ammine complex formed, the standard free energy ) 3 change of reaction (355) should be equal to: (355) Ni + 2 NH  —>  2 ) 3 H+Ni(NH  ° 2 AG  = AG.(NH)2+  —  K  II expi  I  =  A/’° —  RT  )  =  —  4 I expi  AGJ,,+ = —89.1  —  (—46.4)  —  36.7x103 i -7 = 3.70 x 10 8.314x298)  (—79.4) = 36.7 LI  (356)  (357)  —_________  If a is designated as the initial concentration of Ni 2 ion, and b as the initial concentration of NH, and x as the concentration of Ni(NH 2 formed, it follows that: ) 3 2 K =  ] [H1 2 ) 3 [Ni(NH }• [NHJ 2 [Ni  (358) =  (a  —  x) (b  —  x)  ab  Polarization curves of nickel reduction and hydrogen evolution  171  Using the molarity to approximate the molality, the concentration x of the nickel monoainmine complex Ni(NH 2 in the solution 0.937 M NiC1 ) 3 2 1.31 M NH C1 can be calculated as follows: 4 -  x ..  fK a 2 b pH  =  ‘13.70 x i0 x 0.937 x 1.31  —logx  —log(6.74 x  lOj  =  3.2  6.74 x l0  (359) (360)  Thus, it is clear that the formation of the nickel monoammine complex Ni(NH 2 should be expected ) 3 at a pH above 3.2. By comparing the titration curves in the presence (Figure 53) and absence (Figure 38) of ammonium chloride, the existence of a buffering function is evident. The formation of the strong nickel ammine complexes actually prevented effectively the occurrence of the peak which had been observed for other electrolytes (Figure 85). The observations on the electrode surface during the linear potential sweep may be instructive. Before the occurrence of the peak, the electrode surface was observed to be bright without a sig nificant amount of gas evolution. Just at the top of the peak, the electrode surface turned gradually black starting around the edges and corners. As the black area spread over the whole electrode surface, the gas evolution became more and more massive. At the point where the current density reached a minimum and started to rise again, it was believed that water began to decompose on the cathode. If the electrodeposition was run potentiostatically at a potential between the peak and the valley, the cathode deposit was black. However, when the electrodeposition was carried out potentiostatically at a potential beyond the valley, a green deposit on the cathode was obtained. The sharp drop in the current density after the peak is probably due to the precipitation of insoluble nickel hydroxide or oxide on the cathode surface. The poorly conductive Ni(OH)<S), or NiO on the cathode surface would increase greatly the ohmic drop and possibly the activation energy for the reduction of nickel ions as well. The green deposit is obviously Ni(OH)). What is the black deposit? According to the experience of personnel at Falconbridge Ltd, the most probable com position of this black deposit is nickel oxide 1 19], which is equivalent to dehydrated nickel hydroxide. Ragauskas and Leuksminas 1461 believed that the black deposit encountered in their studies was not a basic nickel compound but highly dispersed nickel powder. The nickel powder was considered to be formed from the disproportionation of monovalent nickel ions, 2Ni = Ni 2 + Ni. The black deposit was also believed to be a mixture of nickel powder and nickel hydroxide . Such a con 201 clusion is suspect, and cannot explain why the current density drops sharply as the dispersed pul verulent nickel has a larger real active surface area and should be more active electrochemically compared to dense compact nickel.  Polarization curves of nickel reduction and hydrogen evolution  172  The black deposit was also examined by Deligianni and Romankiw 1741 using Auger spectros copy. It was found that the black deposit contained Ni, 0 and C1 if it was deposited from a nickel chloride electrolyte, or contained equal amounts of Ni and 0 if it was deposited from a nickel sulfate electrolyte. 4000 3500  3000  cJ  2500  2  2000  ci C-)  1500 1000  500  0 -1.3  -1.2  -1.1  -1  -0.9  -0.8  -0.7  -0.6  -0.5  -0.4  -0.3  Potential vs. SCE, (volt)  Figure 86 Polarization curves of combined nickel reduction and hydrogen evolution in 0.937 M NiCI 2 at different pH’s (2,000 rpm, 25°C, 2 mVlsec, Ni-coated Pt disc) The response of the shape of polarization curves to changes in the electrolyte pH is shown in Figure 86 in 0.937 M 2 NiCl at 25°C and 2,000 rpm. One characteristic of these polarization curves is that the potential at which the peak current occurs remains relatively the same, whereas the potential corresponding to the valley shifts to a more negative potential. The total current density increases with decreasing electrolyte pH because of the enhanced contribution of hydrogen evolution to the total current density. One of the most important measurements for nickel reduction and hydrogen evolution is the current efficiency of nickel over a wide range of potential so as to obtain the partial polarization curves of nickel reduction and hydrogen evolution. The measurement can be time consuming, especially at low current densities. For each measurement, the test should be started with a fresh electrode surface, and the nickel deposit on the Pt disc should not be less than 1 pm in order to have  Polarization curves of nickel reduction and hydrogen evolution  173  a manageable number of coulombs for the anodic dissolution, and not greater than 10 im so as not to interfere significantly with the hydrodynamics near the Pt disc. The results of a series of such measurements are given in Figure 87.  2  2000  100  1800  90  1600  80  1400  70  1200  60  1000  50  0 -a  0  800  40  600  30  400  20  200  10  Lii  0  I—  0  -1.3  -1.2  -1.1  -1  -0.9  -0.8  -0.7  0 -0.6  Potential vs. SCE, (volt) Figure 87 Current efficiency of nickel over the potential range covering the whole polarization curve , pH 2, 25CC, 2,000 rpm, Ni-coated Pt disc) 2 (0.937 M NiCI Each point in Figure 87 was acquired potentiostatically. Figure 87 reveals several important features of nickel electrodeposition. As is known from the equilibrium potentials, hydrogen evo lution precedes the nickel reduction. This fact is clearly demonstrated in Figure 87 when the potential is larger than -0.78 volt vs. SCE. At -0.64 volt vs. SCE, the current efficiency for nickel reduction is only 35 %, that is to say, 65 % of the current is consumed for hydrogen evolution. As the cathode potential becomes more negative, the nickel reduction becomes more dominant. At -0.78 volt vs. SCE, the current efficiency of nickel reduction reaches —90 %. Decreasing the cathode potential further does not change the current efficiency of nickel reduction very much although the total current density increases rapidly. When the current density reaches its peak, the cathode surface begins to become black starting around the edges. Subsequently, the current efficiency of nickel reduction and the total current density drop dramatically as the cathode potential becomes more negative. After reaching the minimum on the polarization curve, the current efficiency of nickel reduction continues to decline even though the total current density rises again. The increase in total current density in this section is due to the decomposition of water.  Polarization curves of nickel reduction and hydrogen evolution  174  Figure 87 provides some important information as regards the current density in commercial nickel electrowinning. Although the experimental temperature was 25°C which is not same as 60°C in the commercial operation, the current density should be chosen to be in the area where the maximum or near maximum current efficiency for nickel reduction can be achieved. Considering the situation in Figure 87, the cathode potential should be lower than -0.78 volt vs. SCE to have a reasonable nickel current efficiency. Another consideration is that the cathode potential should be away from the potential near the peak on the polarization curve. Although the nickel reduction in Figure 87 is not carried out at the limiting current density, its peak current density can be treated as equivalent to the limiting current density. The conventional practice is that the applicable current density for industrial electrowinning or electrorefmning can be chosen up to one third of its limiting current density, or of its peak current density in the case of nickel. Therefore, the data in Figure 87 suggest that the maximum applicable current density for nickel electrowinning is around 500 A/m . 2 It should be noted that the actual conditions for industrial nickel electrowinning are not exactly the same as those in Figure 87, such as, temperature, flow rate etc.; however, the principle still applies. 1800 1600 1400 1200 C\J  2  1000  ci 6  800 600 400 200 0 -1.3  -1.2  -1.1  —1  -0.9  -0.8  -0.7  -0.6  Potential vs. SCE, (volt) Figure 88 Partial polarization curves of nickel reduction and hydrogen evolution in 0.937 M NiC1 2 at pH 2, 25°C and 2,000 rpm (Ni-coated Pt disc)  From the current efficiency of nickel and the total current density in Figure 87, the partial current densities of nickel reduction and hydrogen evolution can be obtained. A series of these partial current densities at different potentials Consists of the partial polarization curves as plotted in Figure 88. The curves in Figure 88 are quite similar in shape to those of Ragauskas and Leuk  Polarization curves of nickel reduction and hydrogen evolution  175  , who used a bipolar palladium membrane electrode in 0.3 M NiC1 611 sminas 2 + 2.1 M KC1 at pH 4.5 and 25°C. One point should be noted concerning the data in Figure 88. The electrodeposition time for each data point in Figure 88 varied between 150 and 4,000 seconds depending on the magnitude  of the current density. It can be seen that the current density of nickel reduction approaches practically zero when the potential is more negative than -1.13 volts vs. SCE. In fact, the current density of nickel reduction could have already come close to zero somewhere between the peak and the minimum if the electrodeposition were run for a longer period of time. The reason is quite simple considering the way in which these measurements were made. When nickel was deposited at potentials lower than -1.03 volts vs. SCE, metallic nickel would always be deposited first because of the metallic surface and the adequate mass transfer of the nickel ion in the beginning. How long the metallic nickel deposition will last depends on the magnitude of overpotential or the degree of hydrogen evolution. On the left of the peak, this period became shorter as the potential became more negative. Consequently, if the electrodeposition is run for a longer period of time, this initial period will account for only a very small percentage of the total time. The current efficiency thus obtained will reflect more accurately the true steady-state value. The limiting current density of nickel reduction does not exist in Figure 88 since the limiting condition of hydrogen evolution comes first and subsequently the nature of the cathode surface is changed. The current density at the peak is well below the limiting current density for nickel reduction calculated from Levich’s equation. Using the parameters listed in Tables 37-3 8, the limiting current density for hydrogen evolution can be calculated as: L(H = 1 ) 2  6 v 3 0.62lnFD c o’[Hi ’’  =  6 0.621 xl x 96500 x (6.97 x l09)213 x (1.186 x lOhI  =  114 (AIm ) 2  2000x 2ic )112  lOx 10  (361)  If the diffusion coefficient at infinite dilution is used, the number will be 138 Aim . These two 2 numbers compare favorably with the current density for hydrogen evolution near the peak area in Figure 88. The section corresponding to the potential between -0.64 -0.90 volt in Figure 88 is plotted again in Figure 89 in a Tafel plot, that is, log(C.D.) vs. potential (or overpotential). The slopes -aEThlog(C.D.) determined from the linear regions in Figure 89 are summarized in Table 42. For the reduction of nickel ion, the Tafel slope in the region II(Ni) is 94 mV/decade. If it is assumed that the first electron transfer is the rate-determining step, the theoretical Tafel slope is 2.303 RT/(xF) = 0.059 1/x 0.0591/0.5 = 0.118 volt = 118 mV/decade at 25°C. It can be seen that these two numbers are reasonably close. In the non-ideal situation, the charge transfer coefficient  Polarization curves of nickel reduction and hydrogen evolution  176  1000  300  100  c..J E  6  0.3 -0.9  -0.85  -0.8  -0.75  -0.7  -0.65  -0.6  Potential vs. SCE, (volt) Figure 89 Tafel plots of the partial polarization for nickel reduction and hydrogen evolution in 0.937 M 2 at pH 2, 25°C and 2,000 rpm (Ni-coated Pt disc) NiC1 a is not exactly equal to 0.5. In the lower region I(Ni), the effect of the backward anodic dissolution of metallic nickel may exist Thus, the slope -aE/alog(C.D.) arbitrarily calculated here is not the real Tafel slope. In addition, the error of measurement at such low current densities may be large. Table 42 Tafel slopes detennined from the partial polarization curves Region on lines in Figure 89 -aE/alog(C.D.), (mV) Correlation coefficient, R 2  I (Ni)  II (Ni)  I (112)  11(112)  111(112)  55  94  88  239  112  0.9987  0.9984  0.9769  0.9934  0.9981  For hydrogen evolution, if the lower section I(H2) is ignored in view of too low current density, there are obviously two linear regions. The slope in section 111(H ) is 112 mV/decade which is 2 almost identical to the theoretical Tafel slope 118 mV/decade. In the section 11(112), the slope -aE/alog(C.D.) is 239 mV/decade. The reasons for this large slope are not well understood. The residual dissolved oxygen or noble impurities, if being reduced, would give a smaller slope. This large slope may most probably be attributed to the asymmetric electron transfer coefficient a which changes with the magnitude of the cathode overpotential.  Nickel electrowinning at high current density  177  6.6 Nickel electrowinning at high current density Regarding high current density nickel electrowinning, changes in the electrolyte composition can impose some restrictions on the electrowinning process itself or on other related processes. The addition of ammonium chloride appears to be the best candidate on the basis of the polarization curves. However, during research aimed at producing nickel granules in nickel chloride electrolytes carried out by Falconbridge Ltd., it was found that the ammonium ions were oxidized by the anodic chlorine gas. Although the addition of ammonium chloride is not compatible when the anodic reaction is chlorine evolution, it could be useful in other processes where the anodic reaction is simply the dissolution of metallic nickel, as in nickel electrorefining or in the production of nickel powder. The addition of boric acid is also worth considering in high current density nickel electro winning. As shown in Figure 85, the achievable maximum current density is extraordinarily high even at 25°C. The restriction of using boric acid is not in the electrowinning itself, but in the associated purification circuit, especially in solvent extraction. Nevertheless, boric acid has been widely used in nickel electroplating, direct nickel matte electrowinning, nickel sulfate electro winning, and nickel electrorefining. The benefits from the use of boric acid are realized mainly in a higher nickel current efficiency, stabilized bulk electrolyte pH, and especially in improved cathode surface quality. The concept of high current density nickel electrowinning has not been well defmed. What constitutes high current density electrowinning? The current density for commercial nickel electrowinning is between 200 and 240 AIm . Can it be called high current density electrowinning 2 if the operation is run at 400 A/m 2 or even at 300 A/m ? It is believed to be so, since the productivity 2 of the plant would be increased by —100 % or —50 %. Even a 50 % increase in productivity is a fonnidable increase. Since it has not been specified clearly in industry what is high current density nickel electrowinning, some investigators may have gone to the extreme, believing that the current density must be above 1,000 AIm . High current density for nickel electrowinning has been defined 2 by Ettel 11211 as 400-800 A/m . 2 If a moderate concept of high current density electrowinning is considered, say double the present commercial current density, it can be seen from Figure 85 that the addition of sodium chloride is quite beneficial. The benefits are realized in high nickel current efficiency, lower cathode overpotential and possibly lower anode overpotential, and improved conductivity of the electrolyte. The addition of sodium chloride is unlikely to have adverse effects on other processes associated with nickel electrowinning.  Hydrogen evolution on the nickel cathode in electrolytes without NiCl 2  178  The addition of sulfate may not be beneficial in terms of the total current density and the current efficiency of nickel. However, the presence of sulfate does not have a deleterious effect on these quantities. It was found from the surface pH measurements that the addition of an appropriate amount of sulfate assists in maintaining a lower surface pH.  6.7 Hydrogen evolution on the nickel cathode in electrolytes without N1CI 2 Hydrogen evolution was found to be heavily dependent on the rotational speed of the disc (i.e., mass transfer rate) in the preceding sections. In order to get a clearer understanding of hydrogen evolution, systematic measurements of the polarization curves of hydrogen evolution were made on the nickel-coated platinum rotating disc electrode at different rotational speeds and pH’s. Each test required a fresh coating of nickel film, whose thickness was controlled to around 2 pm. The coating conditions were the same for all tests, that is, 0.937 M NiC1 , pH 1.8, 25°C, 100 A/mZ, 600 2 seconds and 2,000 rpm. After the coating, the disc electrode was washed thoroughly using deionized water and immediately transferred to the test electrolyte. After each measurement, the nickel film was stripped anodically in another electrolyte similar to the coating electrolyte. Before each measurement, the electrolyte was always deaerated by bubbling nitrogen gas and a stream of nitrogen gas flow was maintained during the measurement. The potential sweep rate was controlled at 2 mV/sec. Some of the polarization curves of hydrogen evolution are summarized in Figure 90 in the range of RPM from 100 to 3,600. It can seen that all of these curves are well-defined and smooth, even though there is abundant hydrogen evolution. The potential where the current density of H reduction reaches its limiting value and where water starts to decompose shifts to a more negative value as the rotational speed increases or as the pH decreases. This phenomenon is typical for processes controlled by the mass transfer rate. The polarization curves in the electrolyte containing 50 have a similar shape. The presence of sulfate does not change the current density con 2 Na 4 siderably at lower overpotential. However, at high overpotential and in the limiting regions, the current density of hydrogen evolution increases dramatically. Since the viscosity of the electrolyte can only increase when the electrolyte becomes more concentrated, the major reason for the increased hydrogen current density is the lower activity coefficient of the hydrogen ion in the presence of sulfate. When 3 B0 is added, there are changes in three aspects. Firstly, the current density at a H potential before the limiting plateau is smaller than that in 2.5 M NaCl alone. Secondly, the potential for the decomposition of water shifts to a more positive value, meaning that the presence of 3 B0 H activates the decomposition of water. Thirdly, the limiting current density plateau is somehow not perfect, rising gradually as the potential becomes more negative. The third phenomenon is believed to be related to the dissociation of boric acid, as more protons are generated from the boric acid at  Hydrogen evolution on the nickel cathode in electrolytes without NiCI 2  179  60 50  40  10  0 -1.4  -1.3  -1.2  -1.1  -1  -0.9  -0.8  -0.6  -0.7  Potential vs. SCE, (volt)  c’J  C”  E  E  d C  C  -1.5  -1.4  -1.3  -1.2  -1.1  -1  -0.9  -0.8  -0.7  0 -1.6  -0.6  -1.5  Potential vs. SCE, (volt) 320  2.5 M NaCI + 0.365 M Na2SO4 atpH2  280  3600  160  E  100  80  40  20  -1.2  -1.1  -1  -0.9  Potential vs. SCE, (volt)  -0.7 -0.6  60 40  -1.3  -0.8  100  400  80  -1.4  -0.9  120  C)  -1.5  -1  140  C) 120  1.6  -1.1  ——  900  200  -1.2  160  RPM  1600  E  -1.3  Potential vs. SCE, (volt)  2500 240  -1.4  -0.8  -0.7 -0.6  0 -1.3  -1.2  -1.1 -1 -0.9 -0.8 Potential vs. SCE, (volt)  -0.7  -0.6  Figure 90 Polarization curves for hydrogen evolution on Ni-coated Pt electrode in 2.5 M NaC1, 2.5 M NaCI + 0.365 M 4 SO and 2.5 M NaC1 + 0.485 M 3 2 Na B0 at different RPM’s (25CC, 11 2 mV/sec, —2 p.m Ni-coated Pt disc)  Hydrogen evolution on the nickel cathode in electrolytes without N1CI 2  180  a higher pH. Comparing the results in 4.5 M NaC1 and 2.5 M NaC1, the current density at a given potential and the limiting current density are both decreased, due mainly to the fact that the activity coefficient of the hydrogen ion and the viscosity of the electrolyte are increased. However, the polarization curves in electrolytes containing NH C1 are very unusual and less reproducible. 4 According to Levich’ s equation, the limiting current density for hydrogen evolution can be described as follows:  =  L(H 1 ) 2  wb 6 v 3 0.62lnFD C 2 u  =  0.62 lFDvh!6(  =  (362)  0.62lnFD3vh!6(’ x 2lrJCb  a 112 RPM x 27t) 60  (363)  +  —-  Using 2.5 M NaC1 at pH 2 and 2,500 rpm as an example, the diffusion coefficient of the hydrogen ion is 5.73 x l0 m /sec in 2.27 M NaC1 at 25°C’ 2 , the kinematic viscosity of 2.5 M NaC1 is 221 1.18 x 10.6 m /sec at 20°C”°’, and the activity coefficient of hydrogen ion can be calculated to be 2 around 2.21 using the method discussed in Section 2.1.3 (also refer to Figure 12). Thus, it follows from equation (363) that: L(H) = t  0.621 x 96500 x (5.73 x i0’ (1.18 x l0hl6  x2 tJ 7  lOxlO  137 (AIm ) 2  (364)  This number, in spite of not being exactly the same, is close to the limiting current density found in Figure 90. For the electrolytes containing 2.5 M NaC1, 4.5 M NaCl, 2.5 M NaC1 2.5 M NaCl + 0.485 M 3 B0 and 2.5 M NaC1 + H  4 S 2 Na O 0.365 M , 1.31 M NH C1, the limiting current densities for 4 +  hydrogen evolution determined from the polarization curves are plotted in Figure 91 as a function of the square root of RPM at pH 2.5, 2, 1.5 and 1.1. Except for the ill-defined behavior of the electrolyte containing NH C1, the linear relationships between the limiting current density and 4 -sJRPM were surprisingly good in all of the electrolytes studied. The change in the slopes of these  lines reflects the combined consequence of the change in the diffusion coefficient, viscosity and the total acid concentration. If there are some pH buffers present in the electrolyte, the dissociable protons should be added to the acid concentration. Another common concern is the viscosity of electrolyte. The viscosity of the electrolyte will in most cases increase as it becomes more con centrated. In the present study, the viscosity of the electrolytes with the addition of 4 SO NaC1, 2 Na , B0 or NEI H 3 C1 should be higher than that in 2.5 M NaC1. The diffusion coefficient of the hydrogen 4 ion will accordingly decrease. As a result, the limiting current density decreases.  Hydrogen evolution on the nickel cathode in electrolytes without NICI 2  181  One important point which needs to be emphasized here is Cb in Levich’s equation. Cb is the bulk concentration rather than activity. Hence, when the comparison is made on the basis of a constant pH, the activity coefficient of the hydrogen ion must be taken into account. The activity coefficient of the hydrogen ion in sodium chloride solutions measured using a combination glass pH electrode was shown previously in Figure 12. The fact is that the activity coefficient of the hydrogen ion is larger than that in the concentrated sodium chloride solutions.  800  1800  700  1800 1400  _600  1200  500  ci  400  1000  ,800 300 600  E -J  200  400 200 30  40  0  0  10  20  30  40  50  60  70  1RPM  Figure 91 Limiting current density for hydrogen evolution as a function of the square root of RPM in electrolyte containing no nickel ions at different PH’S (25°C and —2 pm Ni-coated Pt disc) As can be seen from Figure 91, the additions of NaCl, NH C1 or 3 4 B0 all caused the limiting H current densities of hydrogen evolution to decline compared with that in 2.5 M NaC1. For 4.5 M NaC1, the reason for the decline in the limiting current density is due to the increased viscosity of electrolyte and the activity coefficient of the hydrogen ion, and the reduced diffusion coefficient of the hydrogen ion. For the electrolytes of 2.5 M NaC1 + 0.485 M 3 B0 and 2.5 M NaC1 + 1.31 M H C1, the causes for the lower limiting current density are mainly due to the increased viscosity 4 NH of the electrolyte. The buffering function of 3 B0 and NH H CI is negligible under the limiting 4 condition. When sulfate is added, the viscosity of the electrolyte will definitely increase, leading  Hydrogen evolution on the nickel cathode in electrolytes without NICI 2  182  supposedly to a reduced limiting current density. However, the actual result is just the opposite. Due the presence of bisulfate ion, the total acidity should be the sum of the free hydrogen ion plus bisuifate ion concentrations. The decreased activity coefficient of the hydrogen ion can be attributed to the increased limiting current density in the sulfate-containing electrolyte. 800  uvu  Potential, (volt) -1.00 —G—  900  Potential, (volt) -l 00 —0-—  700  800 -0.90  700  -0.85 -0.80 -0.75 -0.70  800  -0.90 -0.85 080  600  —h-— —.—  500  —.—  —i--  -0.75  400  4.5 M NaCI  ——  —.—  —*— —*--  £  400  (300  300  200  200 100  100 0 5MNaCI 0 0.005 0.01 0.015 0.02 0.025  0 0.03 0.035 0.04 0.045  0  0.005  0.010.015  0.02  0.025  0.03  0.035  0.04 0.045  0.03  0.035  0.04 0.045  [H÷], (M)  [H+1, (M) 1000  Potential, (volt) -1.00 —0-— -0.95 —s— 800 • -0.90 —h— 700 • -0.85 —.— 600 • -0.80 —a— -0.75 —*— 500 0.70 400  2.5 M NaCI  +  0.485 M H3BO3  900  300  200 100 0  0  0.005 0.01  0.015  0.02 0.025  0.03 0.035  0.04 0.045  (H+], (M)  0  0.005  0.01  0.015  0.02  0.025  [H+], (M)  Figure 92 Reaction order for the rate ofhydrogen evolution with respect to hydrogen ion concentration in the electrolytes containing no nickel ions (25°C, 2,000 rpm and -‘2 jim Ni-coated Pt disc) As shown in Figure 92, the rate of hydrogen evolution at a given potential is directly pro portional to the concentration of the hydrogen ion, same as that in nickel-containing electrolytes (Figure 80). Therefore, a lower hydrogen ion concentration, or a higher pH, should be adopted in order to reduce the hydrogen evolution. Even in the presence of concentration polarization, the reaction order obtained based on the bulk concentration as in Figure 92 is still valid. The Butler Volmer equation in the presence of concentration polarization and when -it> 100 mV can be expressed as follows: (365) II2 =  [l 8 Fk[H]  _42]exp(_’)  Hydrogen evolution on the nickel cathode in electrolytes without NiCI 2  183  Under the constant electrode potential (E) and rotational speed (RPM), the limiting current density L is proportional to [H9b. Let us assign the following symbols: 1 =  k [ 2 H]b  ,  1 k  =  (  kF exp—  (366)  ciFE”  RT  Hence, equation (365) can be rearranged as:  (  (367)  -‘  iH _ 2 kl[H1b[l_+J  ...  IH=+J  [HJb  Accordingly, the current density forhydrogen evolution is still proportional to the bulkconcentration of the hydrogen ion. The limiting current density for hydrogen evolution is presented in Figure 93 as a function of the hydrogen ion concentration under the conditions of 2,000 rpm and 25°C. It can be seen from Figure 93 that all of the lines pass through the origin, indicating that the buffering function of 3 B0 H , SO and NH 2 Na 4 C1 is negligible in sodium chloride solutions. The comparisons in Figure 93 are 4 made on the basis of the hydrogen ion concentration, thus the relative change in the limiting current density upon the addition of NaCl, 4 SO 3 2 Na , B0 and NH H C1 can be explained simply by the 4 change in the viscosity of the electrolytes and the diffusion coefficient of the hydrogen ion. 1200  36 +0.365 M Na2SO4 —0-— 32  1000 28 800  o•600  2.5 M NaCI  —o-  +0.465 M 83803  —a—  +1.31 M NH4CI  —.-—  +2MNaCI  -—  U)  —J  3)  16  12 Ci)  8 200  4 0 0  0.005  0.01  0.015  0.02 0.025 [H+],(M)  0.03  0.035  0.04 0.045  Figure 93 The limiting current density for hydrogen evolution in different electrolytes vs. the concen-  tration ofhydrogen ion (25°C, 2,000 rpm and —2 .Lm Ni-coated Pt disc)  0  0  0.01  0.02  0.03  0.04  0.05  0.06  0.07  0.08  0.09  0.1  Figure 94 The slope of (L vs. ‘JRPM) as a function of hydrogen ion activity in different electrolytes (25°C and —2 Lm Ni-coated Pt disc)  The slopes extracted from the lines in Figure 91 are plotted in Figure 94 as a function of 10 to the power of minus pH which is supposed to be equal to the activity of the hydrogen ion. The result obtained in the presence of 4 SO is interesting. On the basis of the hydrogen ion concentration 2 Na  Hydrogen evolution on the nickel cathode in electrolytes without N1CI 2  184  (Figure 93), the limiting current density for the hydrogen ion is lowest in the presence of sulfate. However, it is the highest on the basis of pH (Figure 91 or 94). The only difference on these two bases is the buffering functionof bisulfate and the activity coefficient of the hydrogen ion. The reaction order of the rate of hydrogen evolution with respect to the concentration of chloride ion is given in Figure 95. The ionic strength of electrolyte was maintained to be constant by using sodium perchlorate. The total concentration of NaC1 plus NaC1O 4 was 3 M and the concentration of acid HC1 was 0.01 M for all the tests. As expected, the hydrogen evolution does not have any interactions with chloride ions. The limiting current density and the current densities at a given potential are constant over the chloride concentration 0.2-1.2 M. The independence of hydrogen evolution with the chloride ion concentration was also observed and discussed previously in the nickel-containing electrolytes (Figure 81). Therefore, it is certain that the chloride ion will not affect hydrogen evolution in the electrolytes with or without nickel ions. 320  Potential. (vo’) IL  280  -1.0 -0.95 -0.90 -0.85 -0.80 -0.75 -0.70  —  --  -a- -a-  --  -a-  --  240 200 n  0  0  0  0  n  fl  °  °  160 C) 120  0  C)  80 A  40  p  n  — •  -  0  0.2  0.4  A  A  •  p  —  p  i.I.  .1.  0.8  12  0.6  1  1.4  INaC, (M)  -0.8  -0.75  -0.7  Potential vs. SCE, (volt)  Figure 95 The current density ofhydrogen evolution as a function of chloride concentration in 3 M (NaCl + NaC1O ) + 0.01 M HQ (25C, 2,000 rpm and 4 —2 jim Ni-coated Pt disc)  Figure 96 Tafel plot ofhydrogen evolution in 2.5 M NaC1 at pH 2, 25°C and 2,000 rpm (2 mV/sec and -2 jim Ni-coated Pt disc)  The Tafel plot for hydrogen evolution is shown in Figure 96. Figure 96 presents only a section over the potential region between -0.64 -0.9 volt vs. SCE. The Tafel slope determined from the —  straight line in Figure 96 is 172 mV/decade. In comparison with the Tafel plot in the presence of 2 (Figure 89), the current density of hydrogen evolution is of the same order of magnitude. NiCI However, the Tafel plot in the absence of NiCl 2 has only one linear region and its slope is between the two slopes obtained from Figure 89. The differences in these two situations may arise from the properties of the nickel cathode surface. In nickel-containing electrolytes, the nickel cathode surface is renewed all the time due to the continuous reduction and deposition of nickel. The coverage 8 of the cathode surface with atomic hydrogen is approximately proportional to the ratio of the current density of the hydrogen evolution to the nickel reduction Thus, the absorption of atomic  Hydrogen evolution on the nickel cathode in electrolytes without NiCI 2  185  70 60 50 ;40  d30 (.5  20 10 0 -1.4  -1.3  -1.2  -1.1  -1  -0.9  -0.8  -0.7  -0.6  Potential vs. SCE, (volt) 1600  25UNaL1+O.365UNa2SO4  —  a5UNac  —  2.5UNa+OA85MH3D3 4.5UN 2.5MNaCl+1.31MNH4c  ——  1400 1200 1000 E  ,400  ‘800 0300 0  d600 400 200  0 -1.6  -1.5  -1.4  -1.3  -1.2  -1.1  -1  -0.9  Potential vs. SCE, (volt)  -0.8  -0.7 -0.6  0 -1.6  -1.5  -1.4  -1.3  -1.2  -1.1  -1  -0.9  -0.8  -0.7 -0.6  Potential vs. SCE, (volt)  Figure 97 Polarization curves of hydrogen evolution in electrolytes without nickel ions at different pH’s (25°C, 2,000 rpm, 2 mV/sec and —2 jim Ni-coated Pt disc) hydrogen may never reach saturation. In the absence of NiC1 , —2 p.m thick nickel layer was 2 precoated electrochemically on the platinum substrate. This nickel layer might suffer physico chemically during washing with deionized water and being transferred from the precoating cell to the test cell. Furthermore, the nickel layer on the cathode is not renewed during hydrogen evolution. As a result, the adsorption and absorption of atomic hydrogen become gradually greater towards saturation as the electrolysis proceeds. All of these changes which happened to the nickel cathode surface may alter the electron transfer coefficient a and the rate constant k. If the first electron transfer is assumed to be the rate-determining step, and the effect of the coverage 0 of the cathode surface with the atomic hydrogen on the reduction of hydrogen ion is marginal, the Tafel slope should be equal to 2.303RT/czF = 0.0591/a volt at 25°C. Thus when the Tafel slope is equal to 172 mV/decade, a is equal to 0.343.  Hydrogen evolution on the nickel cathode in electrolytes without NiCI 2  186  To provide more information regarding the changes in electrode behavior resulting from the changes in electrolyte composition, a series of polarization curves is given in Figure 97 on the basis of a fixed pH and in Figure 98 according to a constant acid concentration. On these two bases, the extra protons from the dissociation of bisulfate ions must be considered for the hydrogen evolution in 2.5 M NaC1 + 0.365 M 4 SO Another important finding is the overwhelmingly large over2 Na . potential for hydrogen evolution in the presence of ammonium chloride. The degree to which the overpotential is increased here is much more pronounced than that in the presence of nickel ions (Figure 85). Therefore, it can be understood that ammonium chloride depresses the hydrogen evolution much more substantially than the nickel reduction. This result may explain why ammonium chloride has been used in the production of nickel powder at high current density.  320 280 240 200 c160 () 120 80 40  0 -1.5  -1.4  -1.3  -1.2 -1.1 -1 -0.9 Potenlial vs. SCE, (volt)  -0.8  -0.7  -0.6  1400 1200 1000  c.’J  2 800  a  600 400 200  -1.6  -1.5  -1.4  -1.3  -1.2  -1.1  -1  -0.9  Potential vs. SCE, (volt)  -0.8  -0.7 -0.6  0 -1.6  -1.5  -1.4  -1.3  -1.2  -1.1  -1  -0.9  -0.8  -0.7  Potential vs. SCE, (volt)  Figure 98 Polarization curves of the hydrogen evolution in electrolytes without nickel ions at different acid concentrations (25C, 2,000 rpm, 2 mV/sec and —2 jim Ni-coated Pt disc)  Probable mechanisms for nickel electroreduction and hydrogen evolution  187  1981 observed similar results in his experiments in that the limiting current density at a Horkans given pH was higher in sulfate electrolyte (0.33 M 4 SO than that in chloride electrolyte (0.75 M 2 Na ) NaC1). However, he did not attribute this difference to the presence of bisulfate but to the difference in the diffusion coefficient of the hydrogen ion in sulfate and chloride electrolytes. Such an explanation may not be acceptable in his studies at pH 2. According to the calculations based on the equation (90) for 0.33 M 4 SO even though the buffering point pH has shifted to 0.55, the 2 Na , concentration of bisulfate is still equal to 0.011 M. Horkans 98 also studied the effect of 3 B0 on H the polarization curves of hydrogen evolution in sulfate (0.33 M 4 SO and chloride (0.75 M 2 Na ) NaCl) electrolytes. His results, even though on the Pt electrode, reflect the same trends. He found that the addition of 0.4 M 3 B0 had little effect on the limiting current density. Actually, the H limiting current density decreased slightly. Therefore, the dissociation of boric acid is negligible when hydrogen evolution reaches the limiting conditions. The presence of 3 B0 lowered sig H nificantly the overpotential of water decomposition, which was believed to be due to the adsorption of boric acid on the electrode surface. However, the adsorption of boric acid on the Pt electrode surface was not observed during his cyclic voltammetry tests . 1  6.8 Probable mechanisms for nickel electroreduction and hydrogen evolution A completely unambiguous description of the electrode mechanism cannot be obtained from the present studies due to the fact that the intermediate species involved in the electron transfer have not been identified. Nevertheless, from a practical viewpoint, the results obtained so far do throw considerable light on the mechanisms of nickel reduction and hydrogen evolution. Saraby-Reintjes and Fleischmann assumed that nickel reduction proceeded via two consecutive one-electron charge transfer reactions with the involvement of an anion (Cl or OH) in the formation of an adsorbed complex. Taking into account the effect of the coverage 0 of the cathode with the adsorbed nickel species, they calculated theoretically the Tafel slope and reaction order (Table 43) for the following mechanism of nickel reduction: 2 Ni NiX  +  K  =  NiX  (368)  e  =  N1K  (369)  +  NiX,  +  e  =  Ni  +  K  (370)  where X can be Ci or 0H. If the effect of the coverage 0 of the cathode surface with the adsorbed nickel species and hydrogen atoms is ignored, i.e., the reduction of nickel ions can occur over the whole cathode surface, the Tafel slope and the reaction order can be derived for the different possible mechanisms as listed in Table 44. The results for the reduction of nickel ions obtained in the present study are:  Probable mechanisms for nickel electroreduction and hydrogen evolution alogi,  ] 2 a[Ni  1,  a logiN,  a[Cl-]  =  0,  alog 1 Ni  a[Fr]  =  0,  and  aE =  —  1 alogi  188  0.094 volt  Table 43 Calculated Tafel slope and reaction order for the rate of nickel reduction when the effect of the coverage of the cathode with the adsorbed nickel species is taken into account” 1 Rate-determining step  Cathode coverage by  alogi ] 2 a log[Ni  alogi a log[X1  0<0.1  1  1  0.2<0<0.8  1  1  NiX,  Ni+K—*NiX  NiK  NiX,  +  +  e  e  —*  —  NiX,  Ni  +  K  Tafel slope (mV/decade)  0>0.9  00  1  1  0<0.1  120  1  1  0.2<6< 0.8  120  0.5  1  0>0.9  120  0  1  0<0.1  40  1  1  0.2<6 < 0.8  60  0.5  0.5  0>0.9  120  0  0  As a result, the only compatible mechanism is mechanism 1(a) listed in Table 44. Such a simple  mechanism appears unusual. However, it was proposed early in 1970 by Ovari and Rotinyan 4 in their study of nickel reduction from chloride electrolytes. Nevertheless, mechanism 1(a) does not exclude the promotional effect of chloride ion on the nickel reduction. As mentioned earlier, this promotion may be the result of the adsoiption of chloride ions leading to a negative shift of potential at the outer Helmholtz plane. ji is a function of the electrolyte composition, any specific and non-specific adsorptions and the electrode potential. It should be noted that some of the nickel ion comes from the dissociation of the nickel chloro complex NiCl, as NICt’ —> Ni + Ci. This dissociation reaction may explain the fact that the current density of nickel reduction declines at a higher chloride ion concentration (see previous Figure 81). Thus, the chloride ion has two effects, one through N’i due to specific adsorption and the other through complexation with the nickel ion. In the case of nickel ion reduction, the rotating ring-disc electrode (RRDE) technique may be helpful to detect the existence of the monovalent nickel ion, even though the interference of atomic or molecular hydrogen can pose a problem. Some trial tests were carried out using a Pt-disc and Pt-ring electrode in 2.5 M NaC1 at pH 2 and 25CC. When H 2 gas was bubbled through the solution, the anodic ring current was detected when the ring potential became more positive than -0.36 volt  Probable mechanisms for nickel electroreduction and hydrogen evolution  189  Table 44 Calculated Tafel slope and reaction order for the rate of the reduction of nickel ions for different mechanisms #  Mechanism  alogi  alogi  alogi  ] 2 a[Ni  [C11  a[H]  1  0  0  1  0  0  r4.r.  1(a)  1(b)  N?  +  e  —*  Ni,  +  e  =  Ni  N?  +  e  =  Ni  Ni  r4.x.  Ni+e-*Ni 2 Ni  Ct  +  =  NiCr  Tafel slope:  —____  alogi  2.303RT F  2.303RT (l+cx)F 2.303RT  r4.z.  2(a)  2(b)  N1C1  +  e  NiCl  +  2 Ni  Ct  +  NiC1  +  e  —*  e  =  1  NiC1, Ni  1  0  Ct  +  NiCr =  1  , 4 NiCl  1  0  2.303RT (1 + (X)F  -l  2.303RT aF  r4.s.  NiC1+e —*Ni+Ct N?  +  0 2 H  =  NiOH  +  H  r4s.  3(a)  NiOH  +  e  -*  NiOH  1  0  2 NiOH+H+e= O Ni+H Ni+H O 2 =NiOW+H  3(b)  NiOH  +  e  =  NiOH,  i  o  o  2.303RT (1 + a)F  Y4j.  2 NiOH+H+e— O Ni+H  vs. SCE. This ring current became zero when N 2 gas was passed through the solution. One test was carried out using a nickel-coated Pt disc cathode and a Pt ring under the conditions of 2.5 M NaC1, 1,000 rpm, 25°C, pH 2, disc potential sweep from -0.6 to -1.35 volt vs. SCE at a sweep rate of 5 mV/see, and a ring potential 0.4 volt vs. SCE. The solution was deaerated in advance by bubbling N 2 gas. It was found that the ring current followed a similar contour to the disc current before hydrogen evolution reached the limiting condition. Using 0.937 M NiC1 2 instead of 2.5 M NaCI at pH 2 and 25°C, a similar ring current was detected. However, it was difficult to compare  Probable mechanisms for nickel electroreduction and hydrogen evolution  190  the magnitude of the ring current obtained in the absence and presence of nickel ions, since the nature of the cathode surface was not exactly the same and the ring current changed significantly with the ring potential.  Ragauskas and Leuksminas 61 carried out the RRDE studies on nickel ion reduction under the conditions of 3 M (NiCl 2 + KC1), 1,000 rpm, 25°C, pH 4.5, disc potential sweep at 1 mV/sec and ring potential 0.340 volt vs. SHE. They did find that the ring current was larger at the same disc current in the presence of nickel ions. For gas evolution, the general steps involved are the nucleation of gas bubbles, growth in size (when coalescence may occur), break-off from the cathode surface and rising in the liquid. The real electrode area and the mass transfer rate near the cathode surface may be affected during this process. However, the results obtained using the rotating disc electrode show no obvious effects by the hydrogen bubbles.  2  H+ Ni  +  e  1 NH  2 +H÷e=Ni÷H  +  ads  4  NHads= Ni  +  2 H  NiH  Figure 99 The possible routes for hydrogen evolution Hydrogen evolution in acidic solutions can be represented schematically as shown in Figure 99. The process can be divided into two steps. The first step is the reduction of the hydrogen ion to form the adsorbed hydrogen atom.: H  +  e  +  Ni  =  Ni-Ha,  (371)  where Ni represents the cathode nickel, and Ni-H is the adsorbed hydrogen atom. One should keep in mind that H in reaction (371) should have been written as H O, indicating there is always 3 a water molecule associated with the hydrogen ion. As a conventional practice, the bound water molecules are omitted in writing reactions involving hydrogen ions. The second step is either electrochemical desorption, recombination, or absorption. Electrochemical desorption is reaction 2 in Figure 99: 2 Ni-H+H+e=N i+H  Recombination desorption is reaction 3 in Figure 99:  (372)  Probable mechanisms for nickel electroreduction and hydrogen evolution Ni-H  Ni-H  +  =  2Ni  +  2 H  191 (373)  The adsorbed atomic hydrogen can also penetrate into the metal body as reaction 4 in Figure 99: Ni-H  M-H  =  (374)  Depending mainly on the operating conditions and the nature of the nickel cathode, there are four possible mechanisms for hydrogen evolution as listed in the following: (1) Slow discharge  --  fast recombination desorption mechanism r.d.s.  H+e+Ni —*Ni—H Ni-H (2) Fast discharge H  +  e  Ni-H  + --  2 Ni  +  2 H  slow recombination desorption mechanism  Ni  +  =  Ni-H  =  r4.s.  Ni  -  H + Ni  (3) Slow discharge  --  H  2Ni  —*  +H 2  fast electrochemical desorption mechanism r4.s.  H+e+Ni —*Ni—H ’ 4 Ni-H+H (4) Fast discharge H  +  e  -+  +  2 e =Ni+H  slow electrochemical desorption mechanism  Ni  =  Ni-H rds.  Ni—H+H+e —>Ni+H 2 The rate-determining step can be determined to a large extent according to the Tafel slope and the reaction order with respect to the concentration of the hydrogen ion. If the reduction of the hydrogen ion is assumed to occur over the whole cathode surface, that is to say, the effect of the coverage €) of the cathode with the adsorbed hydrogen atoms and nickel species on the reduction of hydrogen ion is negligible, the theoretical reaction order and Tafel slope can be calculated based on the various rate-determining steps (Table 45). The results obtained for hydrogen evolution in the electrolytes containing NiCl 2 are: a log 1 2 H J 2 a[Ni  =  0,  a log 1 2 H a[Cl-]  112 a log i =  0,  [H9  =  1,  and  —  aE = 0.112 volt 2 alogiH  Probable mechanisms for nickel electroreduction and hydrogen evolution  192  Table 45 Tafel slope and reaction order for the rate of hydrogen evolution with respect to the con  centration of hydrogen ion” Slowest step  H  e  +  M  M-H  +  M-H  +  =  Reaction order with respect to[If]  M-H =  2M +  ‘2  2 M-H+H+e=M+H  Tafel slope:  —_____  alogi  1  2.303 RT/((xF)  2  2.303 RT/(2F)  2  2.303RT/[(1+a)F]  The Tafel slope is taken from the linear region at a potential more negative than -0.8 volt vs. SCE. In the electrolytes containing no NiCl , the Tafel slope is somehow larger, 172 mV/decade versus 2 112 mV/decade. The chloride ion does not affect the hydrogen evolution at all whether NiCl 2 is present or not. According to these results, the rate-determining step for hydrogen evolution is most probably the first electron transfer, that is, H known exactly from the present study.  +  Ni  +  e —* Ni-H. The following step cannot be  Chapter 7 Conclusions  193  Chapter 7 Conclusions The following are the principal conclusions resulting from the study of the fundamental and applied aspects of nickel electrowinning from chloride electrolytes: (1)  The thermodynamics of nickel chloride electrolytes were examined with reference to the activity coefficients in simple and multicomponent solutions and nickel speciation in such solutions. In concentrated NiC1 2 solutions, the activity coefficient of the hydrogen ion is always greater than one and increases steadily with increasing NiC1 2 or NaC1 concentration; however, it decreases continuously with increasing sulfate concentration. In the acidic region, the predominant nickel species are Ni 2 and NiCl in concentrated pure 2 solutions and Ni NiC1 , NiCl and NiSO 2 4 in the concentrated mixed sulfate-containing NiC1 2 solutions. The concentration of the traditionally believed electroactive species NiOW is negligible. All other species such as Ni(OH)), Ni(OH), Ni(OH), Ni OH and Ni 2 (OH) 4 are negligible too over the pH range 0 to 14. The pH for the precipitation of insoluble Ni(OH)S) decreases with increasing nickel ion concentration and temperature. The effect of ionic strength on the solubility product and dissociation constant should be taken into account in such calculations.  (2)  To better understand the electrochemistry at the cathode-electrolyte interface, the cathode surface pH was measured using a flat-bottom combination glass pH electrode and a 500-mesh gold gauze cathode which had been preplated with nickel. The cathode surface pH is strongly dependent on the bulk pH, electrolyte composition, temperature, current density and agitation. Lower bulk pH, higher NiC1 2 concentration, higher temperature, application of agitation and the additions of NaC1, 3 B0 and NH H C1 result in 4 a lower surface pH. Addition of a small amount of sulfate to the electrolyte is beneficial in lowering the surface pH. However, excessive addition of sulfate is inappropriate, as the surface pH will not be further lowered and the current efficiency of nickel decreases severely. The cathode surface pH during nickel electrowinning was modelled theoretically and a general consistency with the experimental measurements was found for 0.937 M NiC1 2 and 2 M NiCI 2 at bulk pH 2.5.  Chapter 7 Conclusions (3)  194  In small-scale electrowinning experiments it was found that higher nickel concentration, and the additions of NaC1, 3 B0 and NH H C1 lead to a higher current efficiency of nickel. 4 However, the current efficiency of nickel decreases with increasing sulfate concentration. In 2 at 6(YC, the suitable bulk pH is around 1.5. At this pH, a satisfactoty nickel 0.937 M NiC1 deposit can be achieved at a current density up to 1,000 A/m 2 with a current efficiency averaging 96.4 % and without any risk of the formation of insoluble Ni(OH)) on the cathode surface.  (4)  The cathode kinetics during nickel electrowinning were studied using the rotating disc electrode. It was found that the rate of nickel deposition is first order with respect to the activity of nickel ion and zero order with respect to the activity of chloride and hydrogen ions. The rate of hydrogen evolution was observed to be first order with respect to the activity of hydrogen ion and zero order with respect to the activity of chloride and nickel ions. Polarization curves obtained in electrolytes of NiC1 , NiC1 2 -NaC1, 4 2 -Na NiC1 S 2 O and -H 2 NiC1 B 3 0 all had a characteristic peak. The height of the peak depends on the concentration of hydrogen ions. Under these limiting conditions of hydrogen evolution the cathode surface pH is raised resulting in the formation of a black deposit, most probably nickel oxide.  (5)  Concerning nickel electrowinning at moderately high current densities (up to 500 A/rn ), all 2 of the electrolytes studied appear to be suitable for producing a good quality nickel cathode with an acceptable current efficiency. When the current density is above 1,000 A/rn , the 2 addition of 3 B0 or NH. H C1, or the use of a more concentrated NiC1 4 2 electrolyte is indicated.  Chapter 8 Recommendations for Further Work  Chapter 8 Recommendations for Further Work  195  -  Due to the time limit in this thesis work, many important areas have not been explored. It is believed that these areas are worth investigating in the future from the viewpoint of the basic understanding and the practical application of nickel electrowinning. One ofthe most important areas for study is the nucleation, growth, coalescence and detachment of hydrogen gas bubbles on the nickel substrate during nickel electrowinning. One excellent technique is the optical method which has been used successfully by Bozhkov and co-workers in their studies on hydrogen evolution during zinc electrowinning. They found that the hydrogen bubbles on the cathode changed not only in size but in shape as well. Any substances which can alter the surface tension will affect the contact angle and the bubble shape. Another important area of investigation which can be both theoretical and speculative, is an AC impedance study. AC impedance during nickel electrodeposition has been studied to a certain extent, mainly by Wiart “ ‘. This technique is claimed to be very useful in identifying quali tatively the adsorbed species which often form during the electroreduction of polyvalent metal ions such as divalent nickel. The third area is not electrochemical in nature but concerns solution purification. The present study has shown the feasibility of high current density electrowinning in an electrolyte having a high nickel chloride concentration. The solution from the leaching of nickel matte by chlorine is very concentrated, containing around 230 g/L Ni . The reason for diluting this concentrated 2 solution is to facilitate its subsequent purification. An impurity of major concern is lead. Therefore, it is necessary to develop a purification process which can remove lead from concentrated nickel chloride solutions. Finally, consideration should be given to the study of the nucleation of nickel and crystal growth in the initial stages of deposition on various substrates including titanium, stainless steel, copper and nickel under conditions similar to industrial operations.  Bibliography  196  Bibliography  [1]  Desjardins, P.R., “Nickel Smelting and Refining”, CIM Bulletin, Feb. 1993, 95-99.  [2]  Conard, B.R., “Copper, Nickel and Cobalt Electrometallurgy”, A Short Course on Fundamentals and Practice of Aqueous Electrometallurgy, The Hydrometallurgy Section of The Metallurgical Society of CIM, Montréal, Québec, Canada, October 20-2 1, 1990.  [3]  Kerfoot, D.G.E. and Weir, D.R., “The Hydro- and Electrometallurgy of Nickel and Cobalt”, Extractive Metallurgy ofNickel & Cobalt”, ed. by G.P. Tyroler and C.A. Landolt, 117th TMS Annual Meeting, The Metallurgical Society/AIME, 1988, 24 1-267.  [41  Burkin, A.R., “Hydrometallurgy of Nickel Sulfides”, Extractive Metallurgy ofNickel, ed. by A.R. Burkin, John Wiley & Sons, Chichester, U.K., 1987, 98-146.  [5]  Hofirek, Z. and Kerfoot, D.G.E., “The Chemistry of the Nickel-Copper Matte Leach and Its Application to Process Control and Optimization”, Hydrometallurgy, Theory and Practice, Part A, ed. by W.C. Cooper and D.B. Dreisinger, Proc. of the Ernest Peters International Symposium, Elsevier, New York, 1992, 357-38 1.  [61  Hoare, J.P., “On the Role of Boric Acid in the Watts Bath”, Proceedings ofthe Symposium on ELECTRODEPOSITION TECHNOLOGY, THEORY AND PRACTICE, ed. by L.T. Romankiw and D.R. Turner, 1987, The Electrochemical Society, 269-284.  [7]  Gran, T.V. and Kheifets, V.L., “Electrolytic Refinement of Nickel Using a Chloride Electrolyte”, Electrometallurgy of Chloride Solutions, ed. by V.V. Stender, Consultants Bureau, New York., 1965, 91-97.  [8]  Romankiw, L.T., “pH Changes at the Cathode during Electrolysis of Ni, Fe, Cu and Their Alloys and a Simple Technique for Measuring pH changes at Electrodes”, Pro ceedings of the Symposium on ELECTRODEPOSITION TECHNOLOGY, THEORY AND PRACTICE, ed. by L.T. Romankiw and D.R. Turner, The Electrochemical Society, 1987, 301-325.  [9]  Peters, E., “Activity Coefficients in Hydrometallurgical Calculations”, Introduction to Es-pH Diagrams (Short Course Notes), ed. by E. Peters and D. Dreisinger, Dept. of Metals and Materials Engineering, The Univ. of British Columbia, Vancouver, B.C., Canada, 1990.  [10]  Stensholt, E.O., Zachariasen, H., Lund, J.H. and Thornhill, P.G., “Recent Improvements in the Falconbridge Nickel Refinery”, Extractive Metallurgy ofNickel & Cobalt, ed. by G.P. Tyroler and C.A. Landolt, 117th TMS Annual Meeting, The Metallurgical Society, 1988, 403-412.  [11]  Fujimori, M., Ono, N., Tamura, N. and Kohga, T., “Electrowinning from Aqueous Chlorides in SMM’s Nickel and Cobalt Refining Process”, Chloride Electrometallurgy [Proc. Conf.], ed. by P.D. Parker, The Metallurgical Society/AIME, 1982, 155-166.  Bibliography  197  [12)  Anonymous, “New Nickel Refining is Back on Stream at Le Havre-Sondouville”, Eng. and Mm. J., Feb. 1980, 35-36.  [13)  Stensholt, E.O., Zachariasen, H. and Lund, J.H., “The Falconbridge Chlorine Leach Process”, Nickel Metallurgy. Vo1J. Extraction and Refining ofNickel [Proc. Conf.], ed. by E. Ozberk and S.W. Marcuson, 25th Annual Conference of Metallurgists, CIM, 1986, 442-467.  [14)  Stensholt, E.O., Zachariasen, H. and Lund, J.H., “Falconbridge Chlorine Leach Process”, Trans. Inst. Mm. Metall. C, March 1986, 95, C10-C16.  [15]  Stensholt, E.O., Zachariasen, H. and Lund, J.H., “The Falconbridge Chlorine Leach Process”, Extractive Metallurgy’ 85 [Proc. Conf.], London, U.K., 1985, 377-397.  [16)  Goble, M.L. and Chapman, J.A., “Electrolytic Nickel Production in the Thompson Refinery”, Nickel Metallurgy, vol.1. Extraction and Refining of Nickel [Proc. Conf.], ed. by E. Ozberk and S.W. Marcuson, 25th Annual Conference of Metallurgists, CIM, 1986, 464-480.  [17]  Inanii, T., Ishikawa, Y., Tsuchida, N. and Hirao, M., “Effect of Slime on the Anode Potential in Nickel Matte Electrorefining”, Extractive Metallurgy ofNickel & Cobalt, ed. by G.P. Tyroler and C.A. Landolt, 117th TMS Annual Meeting, The Metallurgical Society, 1988, 429-445.  [18]  Ueda, M., “Environmental Problems and Sumitomo’s Nickel Refining Technology”, Proceedings ofthe International Conference on Mining and Metallurgy of Complex Nickel Ores, ed. by Chongyue Fu, Huanhua He and Chuanfu Zhang, International Academic Publishers, Beijing, China, Sept. 1993, 12-21.  [19]  Inami, T., Tsuchida, N., Tomoda, K. and Shigemura, H., “Studies on Anodic Dissolution of Nickel Sulfides in the Nickel Matte Electrolysis”, Nickel Metallurgy, vol.1. Extraction and Refining ofNickel [Proc. Conf.], ed. by E. Ozberk and S.W. Marcuson, 25th Annual Conference of Metallurgists, CIM, 1986, 482-511.  [20]  Yang, Y. and Meng, X., “Operating Practice and Technical Developments in Nickel Refining and Cobalt Recovery at Jinchuan Non-ferrous Metal Company”, Electro metallurgicaiPlant Practice, 20th Annual Hydrometallurgical Meeting of CIM, ed. by P.L. Claessens and G.B. Harris, Pergarnon Press, Toronto, 1990, 253-268.  [21]  Koskinen, P., Virtanen, M. and Eerola, H., “Integrated Nickel Production in Outokumpu Oy”, Extractive Metallurgy ofNickel & Cobalt, ed. by G.P. Tyroler and C.A. Landok, 117th TMS Annual Meeting, The Metallurgical Society, 1988, 355-371.  [22]  Saarinen, H. and Seilo, M., “Production of the Cathode Nickel in the Outokumpu Pro cess”, Refining Processes in Metallurgy [Proc. Conf.], 1MM Institution of Mining and Metallurgy, 1983, 267-283.  Bibliography  198  [23]  Hofirek, Z. and Hakon, P., “Production ofHigh Quality Electrowon Nickel at Rustenburg Base Metals Refiners (Pty.) Ltd”, Electrometallurgical Plant Practice, 20th Annual Hydrometallurgical Meeting of CIM, ed. by P.L. Claessens and G.B. Harris, Pergamon Press, Toronto, 1990,233-251.  [24]  Ettel, V.A. and Tilak, B.V., “Electrolytic Refining and Winning of Metals”, Compre hensive Treatise of Electrochemistry, vol.2 (Electrochemical Processing), ed. by I. O’M. Bockris, B.E. Conway, E. Yeager and R.E. White, Plenum Press, New York, 1981, 327-3 80.  [25]  Hougen, L.R., Parkinson, R., Saetre, 3. and Van Weert, G., “Operating Experiences with a Pilot Plant for the Electrowinning of Nickel from All-Chloride Electrolyte”, CIM Bulletin, June 1977, 136-143.  [26]  Gong, Q., Liao, H., Yang, T., Hu, J. and Fang, Z., “Electrodeposition of Nickel from All-Chloride Electrolyte”, Trans. Nonferrous Met. Soc. China, 1992,2(3), 46-52.  [27]  Alekseeva, R.K. and Zenkevich, V.V., “Permissible Impurity Content of Nickel Elec trolytes When Operating at High Current Densities”, Tsvetn. Met., 1977, (9), 23-25.  [28]  Gogia, S .K. and Das, S.C., “The Effect of Mg , Mn 2 , Zn 2 2 and Al 3 on the Nickel Deposit during Electrowinning from Sulfate Bath”, Metall. Trans. B, 1988, 19B(6), 823-830.  [29]  Ettel, V.A., “Fundamentals, Practice and Control in Electrodeposition An Overview”, Application of Polarization Measurements in the Control ofMetal Deposition, ed. by I.H. Warren, Elsevier Science Publishers, 1984, 1-14.  [30]  Ostanina, T.N., Murashova, I.B. and Pomosov, A.V., “Production of Electrolytic Nickel Powders under Conditions of a Linearly Increasing Current”, Soy. Powder Metall. Met. Ceram., 1988, 27(7), 523-526.  [31]  Teschke, 0. and Galembeck, F., “Large Nickel Particle Powder Formed by Electrodeposition on a Metal Surface Partially Covered by Polytetrafluoroethylene Layer”, J. Electrochem. Soc., 1984, 131(8), 185 1-1853.  [32]  Philip, H.I. and Nicol, M.J., “The Kinetics and Mechanism of the Deposition of Nickel from Chloride Solutions”, Chloride Hydrometallurgy [Proc. Conf.], Benelux Metallurgie, Brussels, 1977, 250-269.  [33]  Finkeistein, N.P., Philip, H.I., Nicol, M.J. and Balaes, A.M.E., “The Mechanism of the Electrodeposition and Dissolution of Nickel from Chloride Media”, NIM Report No .1 796, National Institute for Metallurgy, 1 Yale Rd., Milner Park, Johannesburg, South Africa, April 21, 1976, pp. . 31  [34]  Pittie, W.H. and Overbeek, G., “The Electrowinning of Base Metals from Chloride Solutions with Insoluble Anodes”, Chloride Hydrometallurgy [Proc. Conf.], Benelux Metallurgie, Brussels, 1977, 283-293.  -  Bibliography  199  [35]  Sabot, J.L., Musanda, K., Bauer, D. and Rosset, R., “Electrorefining in Calcium Chloride (5M) at 98°C: Applications to Purification of Nickel and Lead’, Chloride Hydrometal lurgy, [Proc. Conf.], Benelux Metallurgie, Brussels, 1977, 270-282.  [36]  Teschke, 0. and Menez Soares, D., “Electrodeposition of Nickel by an Asymmetric Periodically Reversed Step Current”, J. Electrochem. Soc., 1983, 130(2), 306-310.  [37]  Baes, Jr., C. F. and Mesmer, R. E., The Hydrolysis of Cations, John Wiley & Sons, New York, 1976, 104-111.  [38]  Tilak, B.V., Gendron, A.S. and Mosoiu, M.A., “Borate Buffer Equilibrium in Nickel Electrorefining Electrolytes”, J. Appi. Electrochem., 1977, 7(6), 495-500.  [39]  Hoare, J.P., “Boric Acid as Catalyst in Nickel Plating Solutions”, J. Electrochem. Soc., 1987, 134(12), 3102-3103.  [40]  Wruck, W.J., “Electrodeposition of Nickel from Chloride Solutions”, M.S. Thesis, Uni versity of Wisconsin, Madison, WI, 1978.  [41]  Epelboin, I. and Wiart, R., “Mechanism of the Electrocrystallization of Nickel and Cobalt in Acidic Solution”, J. Electrochem. Soc., 1971, 118(10), 1577-1582.  [42]  Piatti, R.C.V., Arvia, A.J. and Poclesta, J.J., “The Electrochemical Kinetics Behavior of Nickel in Acid Solutions Containing Chloride (Ci) and Perchlorate (C1O) Ions”, Electrochim. Acta, 1969, 14, 541-560.  [43]  Chevalet, J. and Zutic, V., “The Effect of Chloride Ions on the Electrochemical Behavior of Nickel at Mercury Electrodes”, J. Electroanal. Chem. Interfacial Electrochem., 1973, 44, 411-423.  [441  Saraby-Reintjes, A. and Fleischmann, M., “Kinetics of Electrodeposition of Nickel from Watts Bath”, Electrochim. Acta, 1984, 29(4), 557-566.  [45]  Ovari, E. and Rotinyan, A.I., “Electrode Processes of a Nickel Electrode in Nickel Chloride Solutions”, Soy. Electrochem., 1970, 6(4), 5 16-520.  [46]  Ragauskas, R. and Leuksminas, V., “Possible Reason for the Peak Appearing in Polar ization Curves for Nickel Ion Discharge in Chloride Solutions”, Soy. Electrochem., 1988, 24(6), 675-682.  [47]  Vilche, J.R. and Arvia, A.J., “Kinetics and Mechanism of the Nickel Electrode I. Acid Solutions Containing a High Concentration of Chloride and Nickel Ions”, Corros. Sci., 1975, 15, 419-43 1.  [48]  Chassaing, E., Joussellin, M. and Wiart, R., “The Kinetics of Nickel Electrodeposition: Inhibition by Adsorbed Hydrogen and Anions”, .1. Electroanal. Chem. Interfacial Electrochem., 1983, 157(1), 75-88.  -  Bibliography  200  [49]  Epelboin, I., Joussellin, M. and Wiart, R., Impedance Measurements for Nickel Depo sition in Sulfate and Chloride Electrolytes”, J. Electroanal. Chem. Inteifacial Electro chem., 1981, 119, 61-71.  [50]  Epelboin, I., Joussellin, M. and Wiart, R., “Impedance of Nickel Deposition in Sulfate and Chloride Electrolytes”, J. Electroanal. Chem. Interfacial Electrochem., 1979, 101, 281-284.  [51]  Joussellin, M. and Wiart, R., “Anion Dependence of Ni Electrodeposition in Acidic Electrolytes”, Proc. ofthe Symp. on Electrocry stalllzation, ed. by R. Weil and R.G. Barradas, The Electrochemical Society, 1981, vol.81-6, 111-123.  [52]  Hurlen, T., Kinetics and Thermodynamics of Ni/Ni(ll) Reactions in Concentrated Solutions of Nickel and Calcium Chlorides”, Electrochim. Acta, 1975, 20, 499-505.  [53]  Vagramyan, A.T., Zhamagortsyan, M.A., Uvarov, L.A. and Yavich, A.A., “Effect of Temperature on the Electrochemical Behavior of Nickel in Chloride Solutions”, Soy. Electrochem., 1970, 6(6), 733-738.  [54]  Kozlov, V.M., “Importance of Hydrogen Evolution in the Development of Structural Imperfections during Nickel Electrocrystallization”, Soy. Electrochem., 1982, 18(10), 1203-1207.  [55]  Yeager, 3., Cels, J.P., Yeager, E. and Hovorka, F., “The Electrochemistry of Nickel: I. Codeposition of Nickel and Hydrogen from Simple Aqueous Solutions”, J. Electrochem. Soc., 1959, 106(4), 328-336.  [56]  Tamm, L.V., Tamm, Yu.K. and Past, V.E., “Effect of Acid Concentration on Hydrogen Overvoltage at Nickel”, Soy. Electrochem., 1974, 10(1), 75-77.  [57]  Chen, B., Shu, Y. and Xie, S., “Study on the Inhibition of Gas Holes in Nickel Cathode by H ”, J. Central-South Inst. Mi Metall., 1989,20(3), 3 17-323. 0 2  [58]  Fossi, P., Gandon, L., Bozec, C. and Demarthe, J.M., “Refining of High-Nickel Concen trates”, CIM Bulletin, July 1977, 188-197.  [59)  Diard, J.P. and Le Gorrec, B., “Identifications of the Parameters of the Simultaneous Cathodic Reduction of H and Ni 2 Ions in a Chloride Medium of pH 2.1. Potentiostatic Identification”, Surf. Technol., 1981, 13, 127-144.  [60]  Dorsch, R.K., “Simultaneous Electrodeposition of Nickel and Hydrogen on Rotating Disk Electrode”, J. Electroanal. Chem. Interfacial Electrochem., 1969, 21, 495-508.  [61]  Ragauskas, R. and Leuksminas, V., “Hydrogen Evolution during Nickel Discharge in Chloride Solutions”, Soy. Electrochem., 1987, 23(3), 295-301.  [62]  Polukarov, Yu.M., Grinina, V.V. and Antonyan, S.B., “Nickel Electrodeposition under the Combined Effect of AC and DC”, Soy. Electrochem., 1980, 16(3), 361-363.  Bibliography  201  [63]  Semenova, A.I., Orlova, E.A. and Andrushchenko, V.N., “Effects of Current Reversal on Pit Formation during the Electrodeposition of Nickel”, Soy. Non-Ferrous Met. Res., 1981, 9(6), 510-512.  [64]  Nakahara, S. and Felder, E.C., “Defect Structure in Nickel Electrodeposits”, J. Electro chem. Soc., 1982, 129, 45-49.  [65]  Subramanyan, P.K., “Electrochemical Aspects of Hydrogen in Metals”, Comprehensive Treatise of Electrochemistry, Vol.4, Electrochemical Materials Science, ed. by J. O’M. Bockris, B.E. Conway, E. Yeager and R.E. White, Plenum Press