You are on page 1of 12

Hydrometallurgy 80 (2005) 1 – 12

www.elsevier.com/locate/hydromet

Short review

Kinetics and reaction mechanism of gold cyanidation:


Surface reaction model via Au(I)–OH–CN complexes
G. Senanayake *
A.J. Parker Cooperative Research Centre for Hydrometallurgy, Department of Mineral Science and Extractive Metallurgy,
Murdoch University, Perth, WA 6150, Australia
Received 25 February 2005; received in revised form 20 July 2005; accepted 3 August 2005
Available online 15 September 2005

Abstract

The current status of the mechanism of gold cyanidation based on diffusion and surface adsorption–reaction models are
reviewed. Published rate data based on chemical oxidation from flat gold surfaces in pure aerated cyanide solutions are
analysed to show a reaction order of 2.7 with respect to cyanide at low concentrations. At higher cyanide concentrations, the
reaction rate reaches a limiting value of R Au(lim) = 7.3  10 6 mol m 2 s 1, independent of the cyanide concentration and
stirring rate. This chemically controlled dissolution of gold in pure cyanide solutions is considered to be different from the
widely reported cyanide or oxygen diffusion controlled dissolution of gold, depending on their relative concentrations. The
proposed reaction mechanism to rationalise this behaviour involves the formation of a heterogeneous redox transition state
(Au.H2O)2.(CN)2–3.(O2) which produces the intermediate Au(I)(OH)(CN) on the gold surface. Oxygen is reduced to
hydrogen peroxide which may degrade in three ways: (i) oxidize gold to produce the same gold(I) intermediates on surface,
(ii) oxidize cyanide to cyanate (iii) disproportionate to water and oxygen. The surface adsorbed Au(I) intermediate reacts with
cyanide to produce more stable Au(CN)2 in solution. The proposed surface chemical model rationalises the reaction order of
c 3 at low cyanide concentrations and calculates an intrinsic rate constant of k Au = 8.6  10 6 mol m 2 s 1 for gold
cyanidation by oxygen. This value is in reasonable agreement with the value of k = 6.9  10 6 mol m 2 s 1 based on the
model proposed by Wadsworth et al. [Wadsworth, M.E., Zhu, X., Thompson, J.S., Pereira, C.J., 2000. Gold dissolution and
activation in cyanide solution: kinetics and mechanism. Hydrometallurgy, 57, 1–11.], which considered the mass transfer away
from the active crystalline gold surface site followed by fast charge transfer, combined with two-electron reduction of oxygen
on the gold surface.
D 2005 Elsevier B.V. All rights reserved.

Keywords: Gold cyanidation; Reaction mechanism; Surface transition state; Kinetics; Rate constants; Gold(I) speciation

* Tel.: +61 8 93602833; fax: +61 8 93606343.


E-mail address: G.Senanayake@murdoch.edu.au.

0304-386X/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.hydromet.2005.08.002
2 G. Senanayake / Hydrometallurgy 80 (2005) 1–12

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Current status of reaction mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1. Limitations of diffusion model due to effect of impurities . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2. Formation of surface films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3. Surface adsorption–reaction models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.4. Need for further studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3. Gold(I) speciation: justification and importance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
4. Analysis of rate data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
4.1. Levich equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
4.2. Rate data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
5. Reaction mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
6. Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

1. Introduction The theoretical and practical aspects of gold cya-


nidation have been frequently reviewed (Cornejo and
Gold cyanidation has been reported to involve the Spottiswood, 1984; Nicol et al., 1987; Li et al., 1992;
chemical reactions shown in Eq. (1) (Bodländer, 1896) Fleming, 1992), while the fundamental aspects of gold
and Eq. (3) (Elsner, 1846), where Eq. (3) can be cyanidation reaction and their relevance to leaching of
treated as the sum of the two partial reactions shown gold from ores have been advanced by Cathro (1963),
in Eqs. (1) and (2). Hydrogen peroxide produced at Cathro and Koch (1964), MacArthur (1972), Nicol
the interface by reduction of oxygen can react with (1980), Kirk et al. (1980), Dorin and Woods (1991);
gold (Eq. [(2)), or with cyanide ion (Eq. (4)), or Osseo-Asare et al. (1984), Lorenzen and van Deven-
disproportionate to H2O + 0.5O2; while cyanate pro- ter (1992); Zheng et al. (1995), Guan and Han
duced in solution further degrades to other products. (1994), Crundwell and Godorr (1997), Wadsworth
et al. (2000), Jeffrey and Ritchie (2000a,b, 2001)
2Auþ4CNþO2 þ2H2 O¼2AuðCNÞ 
2 þ2OH þH2 O2
and Xue and Osseo-Asare (2001). Significant pro-
gress has been made in recent years on mixed
ð1Þ
electrochemical-transport (diffusion) models that
are capable of explaining the effect of oxygen
2Au þ 4CN þ H2 O2 ¼ 2AuðCNÞ
2 þ 2OH

ð2Þ pressure, cyanide concentration and agitation (Li
et al., 1992; Wadsworth et al., 2000).
4Au þ 8CN þ O2 þ 2H2 O ¼ 4AuðCNÞ
2 þ 4OH

Despite long-term interest and industrial applica-
ð3Þ tion, the reaction mechanism of gold cyanidation by
oxygen is still being debated and/or investigated—
CN þ H2 O2 ¼ CNO þ H2 O ð4Þ especially in the following three areas:

Kudryk and Kellogg (1954) highlighted the impor- (i) stoichiometry and chemical or diffusion con-
tance of understanding the rate controlling factors of trolled nature of the reaction,
gold cyanidation which would allow correct choice of (ii) nature of the passivation layer on gold surface,
conditions such as agitation, temperature, and the (iii) effect of host minerals and impurities in the
reagent concentrations. They showed the electroche- solid state or solution.
mical nature of the gold cyanidation reaction and that
the rate is determined by the rate of diffusion of Although a number of different leaching models
cyanide or dissolved oxygen to the gold surface, have been presented, it is difficult to categorise them
depending on their relative concentrations. due to the fact that each model considers a number
G. Senanayake / Hydrometallurgy 80 (2005) 1–12 3

of different factors such as diffusion, adsorption, gold cyanidation reported by previous researchers to
charge transfer, surface films etc. Nevertheless, it the different methods of stirring and the presence of
is important to revise these models in order to impurities in cyanide or other electrolytes—as well as
understand the current status of the reaction mechan- impurities in the gold discs used in kinetic studies. For
ism of pure gold in aerated cyanide solutions in the example, alloyed silver and copper (Choi et al., 1991;
absence of impurities in solid and aqueous phases. Sun et al., 1996; Xue and Osseo-Asare, 2001; Breuer
This is useful to develop a surface chemical model et al., 2005), and even the minor amounts of lead(II)
that would rationalise the role of metal ions in contaminated with analytical grade sodium cyanide
aqueous phase, alloyed metals, and the effect of and sodium perchlorate (Jeffrey and Ritchie, 2001)
host minerals. can affect the rate of gold cyanidation. Both silver(I)
This work describes a surface chemical model on and lead(II) cause a positive effect on gold dissolu-
the basis of surface adsorbed species such as Au(OH)0 tion; but excess dissolved lead(II) (Lorenzen and van
and Au(OH)(CN) to rationalise the results reported Deventer, 1992) and silver(I) (Wadsworth and Zhu,
by Jeffrey and Ritchie (2001) and to compare with the 2003) retards the leaching kinetics.
adsorption-charge transfer model proposed by Wads-
worth et al. (2000). 2.2. Formation of surface films

It has been widely reported that the anodic oxida-


2. Current status of reaction mechanism tion of gold in cyanide media is initiated by the
adsorption of cyanide onto the gold surface, while
2.1. Limitations of diffusion model due to effect of passivation and the peaks in polarization curves
impurities have been attributed to the formation of adsorbed
gold(I) and gold(III) species such as Au(I)(CN)0,
Based on electrochemical studies, Kirk and Au(I)(OH)0, Au(I)(OH)(CNx ), Au(III)(OH)(CN)3 and
Foulkes (1978) described gold dissolution as a reac- Au(III)(OH)30 (Cathro and Koch, 1964; Kirk et al.,
tion controlled by aqueous boundary layer diffusion 1980; Nicol et al., 1987; Wadsworth et al., 2000; Xue
(mass transfer) of cyanide and oxygen to the gold and Osseo-Asare, 2001). It is widely accepted that the
surface. Zheng et al. (1995) showed that the rates of cyanidation of gold is a slow process due to passiva-
gold cyanidation based on a rotating disc electrode tion of the gold surface by a film of AuCN (Cathro
quartz crystal microbalance (REQCM) were lower and Koch, 1964; MacArthur, 1972; Nicol, 1980;
than the predicted values based on the Levich equa- Zheng et al., 1995; Jeffrey and Ritchie, 2001). The
tion. They related this to the effect of a boundary film anodic passivation of gold can also be a result of the
on the gold surface and the impurities in gold and/or adsorbed hydroxide and the formation of AuOH on
solution. Li et al. (1992) showed how the mixed the surface (Kirk et al., 1980). Supporting the views of
potential theory can be successfully used to model Nicol et al. (1987), Jeffrey and Ritchie (2001) pro-
the cyanidation kinetics reported by Kudryk and Kel- posed the formation of a chain-like film of AuCN to
logg (1954) by combining charge transfer with cya- account for the low rates of cyanidation of gold in the
nide ion diffusion to the surface. absence of impurities. While the dissolution of AuCN
Jeffrey and Ritchie (2000b) found that the mea- only occurs at the chain ends, the presence of impu-
sured value for the rate of gold cyanidation by oxygen rities such as lead ions in the cyanide solutions accel-
in an air saturated solution of 20 mM cyanide on the erates this process.
basis of oxygen transfer and stoichiometry in Eq. (3) Nicol et al. (1987) noted the difficulties in obser-
was smaller than the predicted value. Moreover, Jef- ving passivation of gold in aerated/oxygenated plant
frey and Ritchie (2001) showed that pure gold has a liquors, despite electrochemical evidence for an
very low rate of dissolution in aerated ultra-pure adsorbed layer of AuCN, because the liquors contain
cyanide solutions, while the reaction is chemically trace heavy metal ions that disrupt the formation of
controlled rather than diffusion controlled. Thus, AuCN. Guan and Han (1994) failed to identify the
they attributed some of the conflicting results on species that caused passivation, because it was an
4 G. Senanayake / Hydrometallurgy 80 (2005) 1–12

unstable intermediate. Yet, Crundwell and Godorr Wadsworth et al. (2000) reviewed the early literature
(1997) reported the dominance of a passivating in support of various adsorption–reaction paths of
layer of AuCN on the gold surface at latter stages gold dissolution. They proposed the involvement of
of batch leaching experiments. They presented a two (or more) gold atoms as active sites shown by
kinetic model using batch leaching data based on Au2(s) in Eqs. (13)–(15), and bridging cyanide ions in
half order reaction rates with respect to cyanide and the rate-determining step, leading to higher reaction
oxygen. orders up to 3 with respect to cyanide. Based on this
model, incorporating a two-electron reduction of oxy-
2.3. Surface adsorption–reaction models gen on gold surface, they derived the rate equation
given by Eq. (17).
According to the widely accepted combined diffu-
sion–adsorption–oxidation model, the cyanidation of
Au2ðsÞ þ 2CN ¼ Au2 ðCN Þ2ðadsÞ ð13Þ
gold follows the five steps described by Eqs. (5)–(8),
which involve (i) the diffusion of cyanide from bulk
solution to interface denoted by b and i, (ii) surface Au2 ðCN Þ2ðadsÞ þ CN ¼ Au2 ðCN Þ3ðadsÞ ð14Þ
adsorption equilibrium, (iii) anodic oxidation, (iv)
stabilisation/desorption of surface products and (v)
Au2 ðCN Þ3ðadsÞ ¼ AuCN 
ðadsÞ þ AuðCN Þ2ðadsÞ
diffusion of products into the bulk solution, leading
to the overall reaction given in Eq. (10).  ðrate determining stepÞ ð15Þ
CN 
ðbÞ Y CNðiÞ ð5Þ
AuCN 
ðadsÞ þ CNðadsÞ
AuðsÞ þ CN ¼ AuðCNÞ ð6Þ ¼ AuðCNÞ 
ðiÞ ðadsÞ 2 þ e ðfast charge transfer stepÞ
ð16Þ
AuðCNÞ 0
ðadsÞ ¼ AuðCNÞðadsÞ þ e

ð7Þ
RAu ¼ k a Utot ½CN 3 = f1 þ K½CN 3 g ð17Þ

AuðCNÞ0ðadsÞ þ CN
ðiÞ Y AuðCNÞ2 ðiÞ ð8Þ
where k a = rate constant for the anodic reaction (mol
m 2 s 1), U tot = total number of adsorption sites shown
AuðCNÞ 
2 ðiÞ Y AuðCNÞ2 ðbÞ ð9Þ
by Au2(s) (bare), Au2(CN)2(s) and Au2(CN)3(s) in

Eqs. (13)–(14), and K is the product of equilibrium
AuðsÞ þ 2CN
ðbÞ ¼ AuðCNÞ2 ðbÞ þ e

ð10Þ constants of Eqs. (13) and (14).
Xue and Osseo-Asare (2001) used the mass trans-
A series of equations, similar to those in Eqs. (5)–(9), fer law and the Butler–Volmer equation to show that
can be written for the cathodic reduction of oxygen the reaction order with respect to cyanide concentra-
(Hiskey and Sanchez, 1990). Juttner (1984) reported tion would depend on the rate-determining step. For
that the two-electron reduction of O2 to H2O2 predo- example, a slow rate controlling discharge step shown
minates on Au substrate. Guan and Han (1994) con- by the forward reaction of Eq. (18) would give a
sidered the oxygen reduction as a two-stage process reaction order of 1, as observed by Xue and Osseo-
described by Eqs. (11)–(12). More recent studies by Asare (2001) and Guan and Han (1994). In contrast,
Wadsworth et al. (2000) also considered a two-elec- an equilibration of the discharge step in Eq. (18),
tron reduction process and showed that dissolved followed by the forward reaction of the adsorbed
cyanide depresses the rate of oxygen reduction on a species with cyanide shown in Eq. (8), will give rise
gold surface. to an overall reaction order of 2 with respect to
O2 þ H2 O þ 2e ¼ HO  cyanide concentration.
2 þ OH ð11Þ
0
HO 
2 þ H2 O þ 2e ¼ 3OH

ð12Þ AuðsÞ þ CN
ðbÞ ¼ AuðCNÞðadsÞ þ e

ð18Þ
G. Senanayake / Hydrometallurgy 80 (2005) 1–12 5

2.4. Need for further studies Table 1


Stability constants of gold(I) and silver(I) complexes
It is important to extend these models, which have Complex Ionic strength log b a
been largely originated from electrochemical studies, Ag(CN)
2 1(NaClO4) 20.1
to develop a surface chemical reaction model for the Ag(OH)(CN) 1(NaClO4) 12.8
Ag(OH) 0 3.6 (4.2 at 18 8C, 0.2 KNO3)
dissolution of gold in oxygenated cyanide solutions. 2
Ag(OH)0 0 (or dil) 2.3 (3.9)
Wadsworth et al. (2000) showed how the calculated Au(CN)
2 0.025(KCN) 36.6, 38.3
rates based on anodic and cathodic currents of gold Au(OH)(CN) 23.3b
oxidation and oxygen reduction combined with the Au(OH)
2 22
adsorption-charge transfer model described in Eqs. Au(OH)0 10.2, 20.6
Au(CH3CN)+2 3.1
(13)–(17) were in good agreement with the measured
Au(CH3CN)(OH)0 10.7
data using rotating discs in aerated alkaline cyanide a
At 25 8C, Hogfeldt (1982); Sillen and Martell (1964); Kissner
solutions. A model that can be used to compare and
et al. (1997); Stefánsson and Seward, 2003; Nicol et al., 1987.
contrast the results reported by previous researchers, b
See text.
and to rationalise the effect of solid state and solution
impurities, would also need to consider the following
issues, described in the present study. example, the stability constant b{Ag(CN)(OH)} =
1012.8 corresponds to E8{Ag(CN)(OH)/Ag} = 0.044
(i) intermediate gold(I) species involved in the sur- V, based on E8{Ag + /Ag} = 0.799 V at 25 8C and
face reaction with oxygen and cyanide Eq. (20).
(ii) a reaction mechanism involving the simulta-
neous reaction of gold and oxygen on gold E8fMðCNÞðOHÞ =Mg¼E8fMþ=Mg

(iii) a rate constant for the intrinsic surface reaction.  0:059 log bfMðCNÞðOHÞ g
ð20Þ

3. Gold(I) speciation: justification and importance For gold, this corresponds to E8{Au(CN)(OH)/
Au} = 0.314 V and b{Au(CN)(OH)} = 1023.3 based
The reported evidence for intermediate silver(I) and on Eqs. (19) and (20) and E8{Au + /Au} = 1.69 V at
gold(I) complexes such as Ag(OH)0, Ag(OH)2, 25 8C. The value of 1023.3 for b{Au(CN)(OH)} is of
Ag(CN)(OH), Au(OH)0, Au(CH3CN)(OH)0, Au(CH3 the same order as one of the values (1020.6) reported
CN)2+ and Au(OH)2 (Table 1) shows the possible for- for b{Au(OH)0}. However, both values are fifteen or-
mation of intermediates such as Au(OH)0 and ders of magnitude smaller than 1038 for b{Au(CN)2}
Au(CN)(OH) in addition to Au(CN)0 in surface (Table 1). Of the two values reported for b{Au(OH)0}
reactions during cyanide leaching of gold. Previous in the literature (1010.2, 1020.6), the lower value shows
studies (Finkelstein and Hancock, 1974; Senanayake a better fit to the linear relationship with stability
et al., 2003) highlighted problems associated with constants of other gold(I) complexes (Senanayake,
measuring the stability constants of gold(I) com- 2004). Thus, Au(OH)(CN) would seem to be a
plexes and showed the importance of the following more plausible intermediate compared to Au(OH)0.
relationship between the standard reduction poten- In the case of thiosulphate leaching of gold, the
tials of Au(I)/Au(0) and Ag(I)/Ag(0) redox couples: intermediate complex Au(NH3)(S2O3) has a stability
constant of 1020 compared to 1024 for the stable com-
E o fAuðIÞ=Auð0Þg ¼ 1:79E8fAgðIÞ=Agð0Þg þ 0:236 plex Au(S2O3)23. Yet, the rate of anodic oxidation of
ð19Þ gold in ammoniacal thiosulphate can be modelled on
the basis of the formation of Au(NH3)(S2O3) as an
This relationship can be used to predict the stability intermediate (Senanayake, 2005). Kissner et al. (1997)
constants of unstable gold(I) complexes by using the noted that the high stability of Au(CN)2 is due to the
published (Hogfeldt, 1982; Sillen and Martell, 1964) moderate basicity, minor hardness and k-acceptor cap-
values of stability constants of Ag(I) (Table 1). For ability of cyanide ligand, while hydroxide, which is a
6 G. Senanayake / Hydrometallurgy 80 (2005) 1–12

hard and strongly basic ligand, is a good donor for Au(I). Au(OH)(CN) may be formed as an ionic intermediate
They also noted that OH and NH3 are basic and strong on the surface, it would not be an insoluble film, but
j-bond donors which should coordinate in the same would readily desorb/dissolve as the more stable com-
way with Au(I). These views are supported by the plex Au(CN)2. These results support the formation of
stability constants reported in Table 1. Thus, it is rea- an insoluble film such as Au(OH), in addition to AuCN,
sonable to analyse the rate of gold oxidation on the in dilute cyanide solution and highlights the need to
basis of formation of reaction intermediates such as consider surface reactions on the basis of such species.
Au(OH)0 and Au(CN)(OH) adsorbed onto the gold
surface. However, they are eventually converted to the
more stable Au(CN)2 with a higher stability constant 4. Analysis of rate data
of 1038.
Support for such adsorbed species also come from 4.1. Levich equation
electrochemical/nanobalance studies. Jeffrey and
Ritchie (2001) measured the anodic polarization The Levich (1962) equation (Eqs. (21)) has been
curve for pure gold in a pure 20 mM cyanide solution widely used for the interpretation of rate data based
maintained at 25 8C. The measured current density of on the electrochemical or chemical dissolution of
0.03 A m 2 in the potential region for leaching ( 0.1 metal from a rotating disk, under diffusion controlled
V) was independent of stirring rate. Assuming that the conditions:
measured current was due to the oxidation of gold to
2=3
AuCN, they calculated a mass increase of 650 ng in a J X ¼ 0:62DX x1=2 t 1=6 ½X  ð21Þ
40 min scan. This was close to the measured mass
increase of 605 ng of the rotating gold disc. However, where, J X = flux of reactant (mol m 2 s 1); x = rotation
Fig. 1 shows these two values as well as the predicted rate of the disc = rpm. 2k / 60 (s 1); t = kinematic visco-
mass increase on the basis of formation of other gold(I) sity (0.89.10 6 m2 s 1 for water); D X = diffusion coef-
species shown in Table 1 and other intermediates such ficient of X(m2 s 1); [X] = concentration of X(mol m 3
as Au2O (Bard, 1973). It is clear that the measured or mM).
increase in mass is much closer to the predicted values Thus, the pseudo first order dependence of the rate
on the basis of solids such as Au(OH)0 + Au(CN)0 or of gold dissolution with respect to cyanide concentra-
Au(OH)0 than that based on Au(CN)0 alone. Although tion and oxygen pressure could be the result of a

800
-
Au(OH)(CN)

Measured increase in mass


0
Au(CNO)
Predicted increase in mass / ng

0
-

Au(CN) +Au(OH)
+Au(OH)(CN)
0
Au(OH)

0
Au(CN)

700
0

0
Au(OH)

Au2O

600

500
Predicted Gold(I) species formed on anode
Fig. 1. Comparison between measured and predicted increase in mass of a gold anode due to passivation by different gold(I) species (see text).
G. Senanayake / Hydrometallurgy 80 (2005) 1–12 7

chemically or diffusion controlled surface reaction tion of cyanide was maintained constant at 20 mM.
(Sun et al., 1996; Guan and Han, 1994; Choi et al., The rate of gold dissolution was determined by mea-
1991). For example, at steady state, the rate of diffu- suring the loss of mass with time. Wadsworth et al.
sion is equal to the rate of surface reaction (Leven- (2000) used a rotating gold disc in cyanide solutions
spiel, 1972). Thus, the Levich equation given in the maintained at pH = 10.5 and 300 rpm. The progress of
form of Eqs. (22) can be used to compare the rate data reaction was measured by determining dissolved
obtained under diffusion controlled conditions, where Au(I) in solution using ICP analysis. The results
J Au(I) represents the flux in each case with m being the were in excellent agreement with the results reported
relevant stoichiometric factor that should satisfy the by Jeffrey and Ritchie (2001). However, since the pH
overall mass balance. For example, the relevant values used by the two groups were different, the concentra-
in the case of Eqs. (1) are m = 2 for CN and 0.5 for tion of free cyanide [CN]free was calculated in the
O2 whereas for Eqs. (3), the relevant values are m = 2 present study using [CN]total and pK a (HCN). A plot
for CN and 0.25 for O2. of [CN]free vs. [CN]total gave a slope of 0.9 con-
firming that only a small fraction of cyanide was in
logfRAuðIÞ g ¼ logfJ AuðIÞ g
the form of HCN.
¼ logf0:62mt 1=6 ðDX Þ2=3 x1=2g þ log½X  Fig. 2 shows a log–log plot of R Au vs. [CN]free.
ð22Þ The slope at low cyanide concentration is 2.7. There-
fore, it is clear that rates at low cyanide concentrations
do not follow Eq. (22). This indicates that the rate is
4.2. Rate data controlled by a surface chemical reaction that involves
two to three cyanide ions. This behaviour of pure gold
Jeffrey and Ritchie (2001) used a rotating electrode in pure dilute cyanide solutions is different from
quartz crystal microbalance to determine the progress results reported previously which indicate that gold
of cyanidation of freshly plated gold at 25 8C, in cyanidation was generally first order with respect to
aerated solutions of different cyanide concentrations cyanide concentration, oxygen partial pressure and
(pH c 10) at a disc rotation speed of 300 rpm (x 1/2 = square root of disc rotating speed with a low activa-
5.61 s 1/2). They have also reported results obtained tion energy E a = 22–29 kJ mol 1 (Sun et al., 1996;
using 4 different rotation rates, while the concentra- Guan and Han, 1994; Choi et al., 1991). By contrast,

-4.5
Jeffrey and Ritchie, 2001
Wadsworth et al., 2000
{log RAu /mol m-2 s-1}

-5.5

-6.5 y = 2.74x - 7.37


R2 = 1.00

-7.5
0 0.4 0.8 1.2 1.6
- -3
log {[CN ]free / mol m }

Fig. 2. Log–log plots of gold oxidation rate vs. [CN]free at 25 8C. Data from Jeffrey and Ritchie, 2001 (pH = 10) and Wadsworth et al., 2000
(pH 10.5).
8 G. Senanayake / Hydrometallurgy 80 (2005) 1–12

the higher activation energy of E a = 47–55 kJ mol 1 is important to consider a reaction mechanism that can
(Thurgood et al., 1981) for oxidation of a pure gold describe this behaviour of gold.
anode in cyanide solutions, or E a = 47 F 2 kJ mol 1
for pure gold cyanidation by oxygen (Jeffrey and
Ritchie, 2001), as in the present example, supports a 5. Reaction mechanism
chemically controlled reaction.
The value of R Au reaches a limiting value R Au(lim) It is possible to consider the chemical dissolution of
at higher concentrations of [CN] (Fig. 2). According gold in oxygenated cyanide solutions as a reaction that
to the mixed potential theory, the limiting rate at involves the simultaneous reduction of oxygen and
higher cyanide concentrations is a result of the reac- oxidation of gold as shown in Fig. 3. The three cases
tion being controlled by oxygen diffusion to the inter- shown in Fig. 3 consider the involvement of 2 gold
face, so that the anodic dissolution rate of gold is atoms per oxygen molecule and (a) reaction without
matched by the cathodic reduction rate of oxygen. the involvement of cyanide producing Au(OH)0, (b)
Thus, the oxygen diffusion, which is enhanced at reaction with only one cyanide ion producing
higher rotation rates, should result in an increase in Au(OH)0 + Au(OH)(CN), and (c) reaction with two
R Au with increasing x 1/2. This is not observed, as cyanide ions producing Au(OH)(CN). Eqs. (23)–(26)
noted by Jeffrey and Ritchie (2001), and the limiting consider only the case described in Fig 3b because it
rate remains independent of rotation rate. Therefore, it is representative of the other two extreme cases.

I II III IV V

OH2
Au Au(OH)0
OH2 O OH
 (a) cyanide not involved
OH2 O OH
Au Au(OH)0
OH2

CN-
Au Au(OH)(CN)-
OH2 O OH
 (b) cyanide involved in one site
OH2 O OH
0
Au Au(OH)
OH2

CN-
Au Au(OH)(CN)-
OH2 O OH
 (c) cyanide involved in both sites
OH2 O OH
Au Au(OH)(CN)-
CN-

Fig. 3. Formation of Au(I) intermediates in surface reaction with O2 and (a) 0; (b) 1; (c) 2 CN ions (I) gold surface, (II) adsorbed water and/or
cyanide, (III) oxygen, (IV) adsorbed Au(I) species after reaction, (V) hydrogen peroxide.
G. Senanayake / Hydrometallurgy 80 (2005) 1–12 9

Surface equilibration–redox reaction h ¼ K ads ½O2 ½CN n =1 þ K ads ½O2 CN n ð30Þ

2pAuðsÞ þ 2H2 O þ CN þ O2 RAu ¼ k Au  ð31Þ



¼ 2pAuðH2 OÞ:ðCN Þ:ðO2 Þads ð23Þ
RAu ¼ k Au K ads ½O2 ½CN n =f1 þ K ads ½O2 ½CN n g
2pAuðH2 OÞ:ðCN Þ:ðO2 Þads ðfrom Eqs:ð30Þ; ð31ÞÞ: ð32Þ

¼ pAuðOHÞads þ pAuðOHÞðCN Þads þ H2 O2
At lower values of [O2][CN]n
ð24Þ
f1 þ K ads ½O2 ½CN n g c 1 ð33Þ
Desorption/stabilisation
Then,
pAuðOHÞads þ CN ¼ pAuðOHÞCNÞ
ads=aq þ OH


ð25Þ RAu ¼ k Au K ads ½O2 ½CN n ðfrom Eq: ð32ÞÞ: ð34Þ

General rate equation


pAuðOHÞðCN Þads=aq þ CN ¼ p þ AuðCNÞ
2

þ OH : ð26Þ ½CN n ¼ k Au K ads ½O2 ðRAu Þ 1  K ads ½O2 


ðfrom Eq: ð32ÞÞ: ð35Þ
The equilibration shown in Eq. (23) can be con-
sidered as the formation of a surface adsorbed transi-
tion state followed by the redox reaction which
produces the adsorbed gold(I) species and hydrogen In the case of solutions of low cyanide concentra-
peroxide. The two unstable species Au(OH)0 and tions, Eq. (32) simplifies to Eq. (34). This explains the
Au(OH)(CN) formed on the surface react with cya- value of n between 2 and 3 obtained as the reaction
nide to produce more stable Au(CN)2 in solution as order with respect to cyanide at low concentrations
shown in Eqs. (25)–(26), while hydrogen peroxide (Fig. 2). Moreover, Eq. (32) can be rearranged to Eq.
can react in the manner described before. The reaction (35) that can be used to plot [CN] n vs. (R Au) 1 to
model in Fig. 3 shows a maximum of 2CN ions obtain values for K ads and k Au. Fig. 4a shows the
involved in the surface reaction mechanism. It is relevant plots for n = 1 and 2 while Fig. 4b and c
important to rationalise the reaction order of 2–3 represent n = 3 and n = 4 respectively. In all cases,
demonstrated by the slope of linear relationship in the dissolved oxygen concentration can be considered
Fig. 2 at low cyanide concentrations. Thus, a general as c0.25 mM, because the experiments were carried
form of the surface reaction is shown in Eqs. (27)– out in air saturated solutions. The best linear relation-
(28), which involves nCN ions. The equilibrium ship is given by n = 3 (Fig. 4b) with slope of 3  10 8
constant for Eqs. (23), surface coverage (h), and the and intercept in the range  3.3  10 3 to  3.7 
rate expression are given by Eqs. (29)–(31). 10 3 for the two data sets reported by Jeffrey and
Ritchie (2001) and Wadsworth et al. (2000). The value
General surface reaction of k Au = slope/(-intercept) = 8.6  10 6 mol m 2 s 1
based on the average intercept is close to the limiting
2Au þ 2H2 O þ nCN þ O2 rate of R Au(lim) = 7.3  10 6 mol m 2 s 1 calculated
from the plateau in Fig. 2 described in Section 4.
¼ ðAu:H2 OÞ2 :ðCN Þn :ðO2 Þ ð27Þ
Thus, the limiting rate of gold cyanidation in pure
impurity-free solutions at high cyanide concentra-
ðAu:H2 OÞ2 :ðCN Þn :ðO2 Þ Y Products ð28Þ tions is approximately equal to the intrinsic rate
constant. The slope showing a reaction order of 2.7
indicates the predominant reactions represented by
K ads ¼ h=fð1  hÞ½O2 ½CN n g ðfor Eq: ð27ÞÞ ð29Þ Eqs. (36)–(38), which involve 2 or 3 cyanide ions in
10 G. Senanayake / Hydrometallurgy 80 (2005) 1–12

(a)
0.6
Solid lines: Jeffrey and Ritchie, 2001
Dashed lines: Wadsworth et al., 2000

n=1
0.4
{[CN-]free}-n

0.2

n=2

0
0.E+00 2.E+06 4.E+06 6.E+06
(RAu)-1

(b)
0.2
Solid line: Jeffrey and Ritchie, 2001
Dashed line: Wadsworth et al., 2000 y = 3E-08x - 0.0033
R2 = 0.9999
{[CN-]free}-n

n=3
y = 3E-08x - 0.0037
R2 = 0.9998

0
0.E+00 2.E+06 4.E+06 6.E+06
(RAu)-1

(c)
0.05
Solid line: Jeffrey and Ritchie, 2001
Dashed line: Wadsworth et al., 2000
0.04
{[CN-]free}-n

0.03
n=4
0.02

0.01

0
0.E+00 2.E+06 4.E+06 6.E+06
(RAu)-1

Fig. 4. Plot of {[CN]free} n vs. (R Au) 1 using data from Fig. 2 to examine the validity of Eq. (35). (a) n = 1 or 2, (b) n = 3, (c) n = 4.
G. Senanayake / Hydrometallurgy 80 (2005) 1–12 11

the transition state. The intermediate Au(OH)(CN) ! At higher cyanide concentrations, the reaction rate
formed at the interface will rapidly react with cya- approaches a limiting value of R Au(lim) = 7.2  10 6
nide to produce the most stable Au(CN)2. mol m 2 s 1 which is close to the intrinsic rate
constant k Au = 8.6  10 6 mol m 2 s 1.
ðAu:H2 OÞ2 :ðCN Þ2 :ðO2 Þ ¼ 2AuðOHÞðCNÞ þ H2 O2
ð36Þ References

ðAu:H2 OÞ2 :ðCN Þ3 :ðO2 Þ Bard, A.J., 1973. Encyclopedia of Electrochemistry of the Elements,
vol. IV. Marcel Dekker, New York.
¼ AuðCNÞ
2

þ AuðOHÞðCNÞ þ H2 O2 þ OH Bodländer, G., 1896. Die chemie des cyanidverfahrens. Z. Angew.
ð37Þ Chem. 9, 583 – 587.
Breuer, P.L., Dai, X., Jeffrey, M.I., 2005. Leaching of gold and
copper minerals in cyanide deficient copper solutions. Hydro-
ðAu:H2 OÞ2 :ðCN Þ3 :ðO2 Þ ¼ 2AuðOHÞðCNÞ metallurgy 78, 156 – 165.
 Cathro, K.J., 1963. The effect of oxygen in the cyanide process for
þ CNO þ H2 O ð38Þ
gold recovery. Proc. Aus. I.M.M., vol. 207, pp. 181 – 205.
Cathro, K.J., Koch, D.F.A., 1964. The dissolution of gold in cyanide
Wadsworth et al. (2000) reported the two values for solutions. Proc. Aus. I.M.M., pp. 111 – 126.
the terms k aU tot = 6.9  10 6 and K = 5.3  10 3 in Choi, Y.U., Lee, E.C., Han, K.N., 1991. The dissolution behaviour
Eq. (17), based on an electrochemical model which of metals from Ag/Cu and Ag/Au alloys in acidic and cyanide
involved the anodic oxidation of gold and cathodic solutions. Metall. Trans. 22B, 755 – 764.
Cornejo, L.M., Spottiswood, D.J., 1984. Fundamental aspects of the
reaction of oxygen via reactions in Eqs. (11)–(16). gold cyanidation process: a review. Miner. Energy Resour. 27,
The slight differences in rate constants determined in 1 – 18.
this work (8.6  10 6 mol m 2 s 1) and from their Crundwell, F.K., Godorr, S.A., 1997. A mathematical model of the
model (6.9  10 6 mol m 2 s 1) may be partly attrib- leaching of gold in cyanide solutions. Hydrometallurgy 44,
uted to the use of free cyanide concentration in the 147 – 162.
Dorin, R., Woods, R., 1991. Determination of leaching rates of
present analysis. Nevertheless, the good agreement in precious metals by electrochemical techniques. J. Appl. Electro-
rate constants supports the validity of both models and chem. 21, 419 – 424.
highlights the possibility of extending the surface Elsner, L., 1846. Über das verhalten verschiedener metalle in
chemical model to rationalise the cyanidation kinetics einer wässrigen lösung von zyankalium. J. Prakt. Chem. 37,
of silver and gold alloys, which will be presented in 441 – 446.
Finkelstein, N.P., Hancock, R.D., 1974. A new approach to the
future communications. chemistry of gold. Gold Bull. 7, 72 – 77.
Fleming, C., 1992. Hydrometallurgy of precious metals recovery.
Hydrometallurgy 30, 127 – 162.
6. Summary and conclusions Guan, Y.C., Han, K.N., 1994. An electrochemical study on the
dissolution of gold and copper from gold copper alloys. Metall.
Mater. Trans. 25B, 817 – 827.
! The rate of chemical dissolution of gold in pure Hiskey, J.B., Sanchez, V.M., 1990. Mechanistic and kinetic aspects
cyanide solutions is controlled by the surface che- of silver dissolution in cyanide solutions. J. Appl. Electrochem.
mical reaction between gold, cyanide and oxygen 20, 479 – 487.
via a transition state (Au.H2O)2.(CN)2–3.(O2). Hogfeldt, E., 1982. Stability Constants of Metal–Ion Complexes,
2nd Supplement, IUPAC Chemical Data Series No. 2, Part A,
! The simultaneous reaction produces the surface
Inorganic Ligands. Pergamon, Oxford.
intermediate Au(OH)(CN) while oxygen is re- Jeffrey, M.I., Ritchie, I.M., 2000a. The leaching of gold in cyanide
duced to hydrogen peroxide. solutions in the presence of impurities: I. The effect of lead.
! The intermediate Au(OH)(CN) reacts with cyanide J. Electrochem. Soc. 147, 3257 – 3262.
ions to produce more stable Au(CN)2 in solution. Jeffrey, M.I., Ritchie, I.M., 2000b. The leaching of gold in cyanide
Hydrogen peroxide may react with gold or cyanide, solutions in the presence of impurities: II. The effect of silver.
J. Electrochem. Soc. 147, 3272 – 3276.
or disproportionate to oxygen and water. Jeffrey, M.I., Ritchie, I.M., 2001. The leaching and electrochemistry
! The reaction order for cyanidation at low cyanide of gold in high purity cyanide solutions. J. Electrochem. Soc.
concentrations is close to 3. 148, D29 – D36.
12 G. Senanayake / Hydrometallurgy 80 (2005) 1–12

Juttner, K., 1984. Oxygen reduction electrocatalysis by under- cesses. Green Processing ’2004. Aus. I.M.M., Melbourne,
potential deposited metal atoms at different single crystal pp. 113 – 122.
faces of gold and silver. Electrochim. Acta 29, 1597 – 1604. Senanayake, G., 2005. Catalytic role of ammonia in the anodic
Kirk, D.W., Foulkes, F.R., 1978. A study of anodic dissolution of oxidation of gold in copper-free thiosulfate solutions. Hydro-
gold in aqueous alkaline cyanide solutions. J. Electrochem. Soc. metallurgy 77, 287 – 293.
125, 1436 – 1443. Senanayake, G., Perera, W.N., Nicol, M.J., 2003. Thermodynamic
Kirk, D.W., Foulkes, F.R., Graydon, W.F., 1980. Gold passiva- studies of the gold(III)/(I)/(0) redox system in ammonia-thiosul-
tion in aqueous alkaline cyanide. J. Electrochem. Soc. 127, phate solutions at 25 8C. In: Young, C.A., Alfantazi, A.M.,
1962 – 1969. Anderson, C.G., Dreisinger, D.B., Harris, B., James, A.
Kissner, R., Welti, G., Geier, G., 1997. The hydrolysis of gold(I) in (Eds.), Hydrometallurgy 2003, Leaching and Solution Purifica-
aqueous acetonitrile solutions. J. Chem. Soc. Dalton Trans., tion, vol. 1. TMS, Warrendale, pp. 155 – 168.
1773 – 1777. Sillen, L.G., Martell, E., 1964. Stability constants of metal–ion
Kudryk, V., Kellogg, H.H., 1954. Mechanism and rate controlling complexes. Special Publication, vol. 17 and 26. Chemical
factors in the dissolution of gold in cyanide solutions. J. Met. 6, Society, London.
541 – 548. Stefánsson, A., Seward, T.M., 2003. The hydrolysis of gold(I) in
Levenspiel, O., 1972. Chemical Reactions Engineering. Wiley, aqueous solutions to 600 8C and 1500 bar. Geochim. Cosmo-
New York. chim. Acta 67, 1677 – 1688 (and references therein).
Levich, V.G., 1962. Physico-Chemical Hydrodynamics. Prentice Sun, X., Guan, Y.C., Han, K.N., 1996. Electrochemical behaviour
Hall, Englewood Cliffs, NJ. of the dissolution of gold–silver alloys in cyanide solutions.
Li, J., Zhong, T., Wadsworth, M.E., 1992. Application of Metall. Mater. Trans. 3B, 355 – 361.
mixed potential theory in hydrometallurgy. Hydrometallurgy Thurgood, C.P., Kirk, D.W., Foulkes, F.R., Graydon, W.F., 1981.
29, 47 – 60. Activation energies of anodic gold reactions in aqueous alkaline
Lorenzen, L., van Deventer, J.S.J., 1992. Electrochemical interac- cyanide. J. Electrochem. Soc. 128, 1680 – 1685.
tions between gold and its associated minerals during cyanida- Wadsworth, M.E., Zhu, X., 2003. Kinetics of enhanced gold dis-
tion. Hydrometallurgy 30, 177 – 194. solution: activation by dissolved silver. Int. J. Miner. Process.
MacArthur, D.M., 1972. A study of gold reduction and oxidation in 72, 301 – 310.
aqueous media. J. Electrochem. Soc. 119, 672 – 677. Wadsworth, M.E., Zhu, X., Thompson, J.S., Pereira, C.J., 2000.
Nicol, M.J., 1980. The anodic behaviour of gold. Gold Bull. 13, Gold dissolution and activation in cyanide solution: kinetics and
105 – 111. mechanism. Hydrometallurgy 57, 1 – 11.
Nicol, M.J., Fleming, C.A., Paul, R.L., 1987. The chemistry of the Xue, T., Osseo-Asare, K., 2001. Anodic behaviour of gold, silver,
extraction of gold. In: Stanley, G.G. (Ed.), The Extractive and gold–silver alloys in aqueous cyanide solutions. In:
Metallurgy of Gold, vol. 2. S. African Inst. Min. Metall., Young, C.A., Twidwell, L.G., Anderson, C.G. (Eds.), Cyanide:
Johannesburg, pp. 831 – 905. Social, Industrial and Economic Aspects. TMS, Warrendale,
Osseo-Asare, K., Xue, T., Ciminelli, V.S.T., 1984. Solution chem- pp. 563 – 576.
istry of cyanide leaching systems. In: Kudryk, V., Corrigan, Zheng, J., Ritchie, I.M., La Brooy, S.R., Singh, P., 1995. Study of
D.A., Liang, W. (Eds.), Precious Metals: Mining, Extraction gold leaching in oxygenated solutions containing cyanide–cop-
and Processing. Metall. Soc. AIME, Warrendale, pp. 173 – 197. per–ammonia—using a rotating quartz crystal microbalance.
Senanayake, G., 2004. Fundamentals and applications of metal– Hydrometallurgy 39, 277 – 292.
ligand complexes of gold(I/III) in non-cyanide gold pro-

You might also like