Thermodynamics of gold electrowinning
The following electrochemical reactions are of relevance during the gold electrowinning process (E values quoted for metal ion concentrations of 10-4 mol.dm-3, NaCN concentration of 0.2% and NaOH concentrations of 2%):
[1] Au(CN)-2 + e- ↔ Au + 2 CN- ; E = - 0.672 V SHE
[2] O2 + 2 H2O + 4 e- ↔ 4 OH- ; E = 0.419 V SHE
[3]2H2O + 2 e- ↔ H2 + 2OH- ; E= -0.809 VSHE
[4]Cu(CN)2-3 + e- ↔ Cu + 3CN- ; E = -0.760 VSHE
The gold, which is present in the solution in the form of aurodicyanide (Au(CN)2-), is reduced to metallic gold, according to Reaction [1] at potentials more negative than the reversible potential.
Reaction [2], representing oxygen reduction in alkaline solutions, is another cathodic reaction competing with golddeposition. This is mainly because the electrolyte is likely to be saturated with oxygen due to the oxygen evolution occurring at the anode.
Reaction [3] represents the evolution of hydrogen, which occurs at a significant rate at potentials more negative than -0.96 VSHE when working at pH values above 10.1 The evolution of hydrogen should be under kinetic control at a significant range of potentials more negative than -0.96 VSHE and therefore consumes a high proportion of the cathodic current.
The major anodic reaction is the oxidation of water to oxygen, also indicated by Reaction [2].
The reduction reaction of cuprous cyanide (Cu(CN)3 2-) to metallic copper is indicated in Reaction [4]. The more negative potential for the reduction of cuprous cyanide to metallic copper than that for the reduction of aurodicyanide to metallic gold, signifies that gold should plate preferentially to copper at these conditions. However, copper may codeposit with gold if high overpotentials are applied, or whenthe copper concentration is high relative to that of gold.
Morphology and adhesion of the plated gold
From an industrial point of view, it is essential to obtain loosely adherent precipitates that can be easily removed with high pressure water. If this cannot be achieved, some gold will get locked up inside the electrowinning cells, which can result in electrowinning being uneconomical. The electrolyte temperature, current density, and conductivity of the electrolyte have typically proved to be the main variables that influence the adhesion of gold precipitates. Co-deposition of base metals, such as copper, may also have an effect on the morphology and adhesion of the precipitate.
Experimental
The electrodeposition of gold in industrial electrowinning cells is typically operated under mass transfer control. To realistically simulate this and to obtain repeatable results, it is necessary to control the rate of mass transfer to the electrode. This can be conveniently done using a rotating disc electrode for which mass transfer correlations are well known. The rate of mass transport to the electrode surface is a function of the angular velocity of the rotating disc as indicated by the Levich equation below:
[5] iL = 0,62 n Fv-1/6 D2/3 Cw 1/2
where, iL = limiting diffusion current density (A/cm2)
n = number of electrons transferred
F = Faraday’s constant (96 485 C/mol electrons)
v = kinematic viscosity (m2/s)
D = diffusion coefficient (m2/s)
C = concentration (mol/ dm3)
ω = angular velocity (rad/s).
From Equation [5] it is evident that a direct relation exists between the limiting current and the square root of the angular velocity. The angular velocity was varied and the subsequent limiting current measured, in order to determine if the reaction is mass transfer controlled. On basis of this, an angular velocity of 1200 rpm. was selected for all test work.
For the rotating disc tests, a 500 ml Perspex beaker, in which the rotating disc was placed together with two carbon counter electrodes symmetrically on either side of the rotating disc, was used as cell. The working electrode, counter electrodes and reference electrode were connected to an ACM Gill AC potentiostat. The cathode was an 8 mm diameter, 304 stainless steel rod with length of 7.5 mm, mounted in a non-conductive plastic cylinder, leaving an exposed area of 0.5 cm2. The cathode was polished, to a mirror finish, with 0.25 micron diamond paste. Electrical connection to the electrode was achieved through a threaded rod screwed into the back of the cylinder. Contact between the sides of the electrode and the electrolyte was prevented by fitting it into a Teflon sleevewith a diameter of 15 mm. As reference electrode, a silver/silver chloride electrode with a Luggin tube was used, the end of the Luggin tube placed just beneath the rotating disc. The Luggin tube salt bridge was filled with a saturated potassium chloride solution. Impedance tests were done
before each test to determine the resistance of the electrolyte between the cathode and the tip of the Luggin probe, and the potentials were adjusted using Ohm’s law.
Synthetic solutions were prepared by AngloGold’s West Wits assay laboratories to simulate plant conditions. The electrolyte was prepared by adding 2 per cent caustic soda together with 0.1 per cent CN, which simulates conditions normally encountered in practice. Chemical compositions
typically ranged from low to high copper and gold concentrations in order to determine at which concentrations the influence of copper became problematic. The solution temperature was kept constant at 25°C for all test work by immersing the Perspex beaker in a water bath. The influence
of gold and copper at various concentrations was studied by conducting potentiodynamic tests using an ACM Gill AC potentiostat. Potentiodynamic scans were done from 0 to -2.5 V at a scan rate of 3 mV/s. The applied potential and the resulting current were stored on a personal computer using data acquisition software.
The adhesion of the precipitates was determined by a pull-off test using Scotch tape. The degree ofa dhesion was rated on a scale of 1 to 5, with 1 being very loosely adherent and removable with slow flowing water, 2 fully removable with Scotch tape, 3 partially removable with Scotch tape (more than 50 per cent removed), 4 partially removable with Scotch tape (less than 50 per cent removed), and 5 highly adherent and only removable with 600 grid sanding paper.
The adhesion of the precipitates was determined by a pull-off test using Scotch tape. The degree of adhesion was rated on a scale of 1 to 5, with 1 being very loosely adherent and removable with slow flowing water, 2 fully removable with Scotch tape, 3 partially removable with Scotch tape (more than 50 per cent removed), 4 partially removable with Scotch tape (less than 50 per cent removed), and 5 highly adherent and only removable with 600 grid sanding paper.
