Solvent extraction of silver by LIX® 79: experimental equilibrium study
Ana María Sastre,1 Anil Kumar2 and Francisco José Alguacil3
1
Departament d´Enginyeria Quimica, ETSEIB, Universidad Politécnica de Catalunya,
Diagonal 647, 08028 Barcelona, Spain
2
Prefe Plant, Bhabha Atomic Research Center, Prefe Lab. Tatapur, P.O.Ghivali 401502
M.S.India
3
Centro Nacional de Investigaciones Metalúrgicas (CSIC), Avda. Gregorio del Amo 8,
Ciudad Universitaria, 28040 Madrid, Spain. E-mail: fjalgua@cenim.csic.es
Abstract: The extraction of silver(I) from argentocyanide aqueous cyanide solutions
using LIX 79 has been studied. Different variables that could affect the extraction
system were evaluated: equilibration time, aqueous pH, metal and extractant
concentrations and organic phase diluent. The extraction of the Ag(CN)2- complex
with respect to other metal-cyano complexes has also been studied on both synthetic
and real leach solutions. Silver experimental data have been analysed numerically to
determine the stoichiometryof the extracted species and its equilibrium constant. It
was found that silver(I) was extracted into the organic phase by the formation of the
species RHAg(CN)2 (LIX 79= R).
Keywords: LIX 79; extraction; silver; cyanide
1 INTRODUCTION
The use of hydrometallurgy has helped in the recovery of metallic valuables from various
resources; this is the case for copper, in which the circuit leaching-solvent extractionelectrowinning process is the most economical way to produce copper.1,2
The hydrometallurgy of silver is associated to the hydrometallurgy of gold which
traditionally used two leaching media: cyanide or chloride, the former of major
importance. Once silver is dissolved, the purification and/or concentration of the pregnant
solution is accomplished by one (or a combination) of these procedures: carbon adsorption
and zinc cementation.
Silver solvent extraction has not known a major degree of application because the
available extractants (eg amines and quaternary ammonium salts) do not completely meet
the two main objectives: (i) silver extraction from alkaline cyanide leach solutions and (ii)
ease of silver stripping.
However, in recent years, a number of extraction systems have been developed to meet the
above requirements. These extraction systems include: organophosphorous derivatives,
2
3-10
guanidine derivatives and mixtures of amines-organophosphorous derivatives.
In the present work, the extraction of silver from cyanide aqueous media by LIX 79 has
been studied. Extraction of metal has been carried out under different experimental
conditions, giving the stoichiometry and the value of the extraction constant of the species
formed into the organic phase.
2 MATERIALS AND METHODS
2.1 Materials
The extractant, LIX 79, was used as supplied by the manufacturer (Cognis); the active
substance of the extractant is based on the guanidine functional group, the basic strucutre
of which is shown below
where Rn represents alkyl groups. All other chemicals were of AR grade except the metalcyano complexes of silver (Ag(CN)2-) and copper (Cu(CN)43-) which were prepared from
the corresponding cyanide salts (AgCN or CuCN) by adding the respective CNequimolecular concentration to form each complex.
A number of experiments were performed using two real leach solutions obtained from the
treatment of silver bearing materials: a concentrate from Brazil and electronic scraps
(ECB). The conventional cyanide leach of such materials gave solutions of varying
compositions, as shown in Table 1.
2.2 Procedure
All experiments were carried out at 20C. Extraction tests were performed by shaking the
appropriate organic and aqueous solutions at an O/A phase ratio of 1 for 30 min (unless
otherwise stated). After settling, the metal content in the equilibrated aqueous phase was
analysed by AAS. Metal in the organic phase was calculated by mass balance. The
accuracy of the estimation of the metal concentration in the loaded organic phase by mass
balance was checked by experiments in which complete stripping (using 0.4 mol dm-3
NaOH) of the loaded organic solution was carried out and by analysing the stripped
solution. An average accuracy of 97%, between calculated and analytical results, was
regularly obtained.
3
3 RESULTS AND DISCUSSION
3.1 Influence of equilibration time
To study the influence of equilibration time on silver(I) extraction using LIX 79,
experiments were performed using different initial extractant concentrations. The organic
solutions were composed of LIX 79, 0.25 or 0.38 mol dm-3 in n-heptane, whereas the
aqueous phase contained 0.09 mmol dm-3 silver.
Results obtained shown that equilibrium is nearly reached (0.99 fractional approach of
equilibrium) within 5 min of contact. Beyond this little further improvement is achieved.
3.2 Influence of initial silver concentration
The influence of initial silver concentration was studied with organic phases of LIX 79,
0.25 mol dm-3 in n-heptane, and aqueous phases which contained different initial metal
concentrations ranging from 0.09 to 1.4 mmol dm-3.
Figure 1 shows the variation of silver distribution against the initial metal concentration. It
was found that the variation in the initial silver concentration does not significantly
influence the metal extraction, thus, it is also deduced that, as will be further
demonstrated, there is no formation of polynuclear silver complexes in the organic phase.
Experimental data obtained at other pH values, though not presented in this work, showed
the same behaviour.
3.3 Influence of organic phase diluent
Cumene and n-heptane were used to determine their effect on the extraction of silver using
LIX 79. Aqueous solutions contained 0.25 mmol dm-3 silver and organic phases contained
0.25 mol dm-3 LIX 79 in each diluent. Results obtained show that the change in the
organic diluent has little effect on the extraction of silver, since the pH50 values obtained
for cumene and n-heptane are 10.1 and 10.3, respectively.
3.4 Influence of the extractant concentration
Figure 2 shows the variation in silver extraction against pH for experiments carried out
with aqueous phases of 0.1 mmol dm-3 silver and organic phases which contained different
LIX 79 concentrations in n-heptane. As expected, as the extractant concentration is
increased, the corresponding silver extraction line is shifted to more alkaline pH values,
whereas the slopes obtained from this plot are near (minus)one, which are in accordance
with the stoichiometry proposed for the extraction reaction (see Section 3.6). The
behaviour of LIX 79 has been compared with other extractants, as shown in Table 2.
4
79-Ag(CN)2-
3.5 Selectivity of the LIX
extraction system
The selectivity of the present extraction system against the extraction of different metalcyano complexes was studied with organic phases of 0.25 mol dm-3 LIX 79 in n-heptane
and aqueous solutions with a 0.25 mmol dm-3 metal concentration.
Results are shown in Fig.3, which represents the variation in log D against equilibrium
pH, the aurocyanide complex is extracted at more alkaline pH values that the
argentocyanide ion does, but Ag(CN)2- is extracted preferably to other metal-cyano
complexes. From results shown in this figure it can be seen that an extraction sequence
can be tentatively established for these metal-cyano complexes: M(CN)2- > M(CN)4n- >
M(CN)6n-; thus the extraction of these complexes depends on the metal coordination
number. In general, those complexes with lower coordination numbers were extracted
preferentially over those with higher numbers and lower charge complexes are extracted
preferentially over higher charge complexes.4,6
Furthermore, extraction experiments were performed on the real leach solutions whose
compositions are shown in Table 1; the corresponding results in terms of the separation
factors (ßAg/metal) against pHeq are shown in Table 3. It can be seen that, independent of the
aqueous solution treated, the apparent order follows the series ßAg/Au < ßAg/Cu < ßAg/Fe.
3.6 Silver extraction mechanism
A preliminary computer simulation, using the program MEDUSA,11 was performed to
define the metal speciation in the aqueous pH value. The program solves the mass balance
equations of the different components of the chemical system in terms of equilibria and
stoichiometric formation constants. The results of this simulation, under the experimental
conditions used in the present work, shows that, at this alkaline pH values, the Ag(CN)2species is predominant in the aqueous solution. This result agreed with literature data.12
The expected general equilibrium from which the Ag(CN)2- complex is extracted by LIX
79 can be represented by the reaction:
R org + Haq + Ag(CN )2aq  RHAg(CN )2org
+
-
(1)
where R denotes the active substance of the extractant. Assuming ideal behaviour for the
reaction in eqn(1) in organic and aqueous phases, the stoichiometric equilibrium constant
can be written as:
Kext =
[RHAg(CN )2 ]org
[R ]org [H+ ]aq [Ag(CN )-2 ]aq
(2)
Polynuclear silver(I) complexes in the organic phase have been excluded due to the fact
5
that the experimental values of log D versus [Ag]Initial coincide for the different total silver
concentrations (Fig.1).
The experimental data were treated numerically with the program LETAGROP-DISTR13
in order to obtain the composition and extraction constants of the extracted species (eqns
(1) and (2) as well as to look for new models of species which could improve the goodness
of the fits. The error square sum U over all Np defined as:
U =  (log Dcal - log Dexp )2
(3)
was used in the minimization process. Np is the number of experimental points, Dexp is the
experimental distribution coefficient and Dcal is the value calculated by the program for the
model tested after solving the mass balance equations of the components of the system.
Therefore, the best model is the one which gives the lowest value of U. The calculations
were performed by taking a set of species and extraction constants as the starting input and
considering the influence of the minimized function when partially varying or adding new
species to the model. The results of the numerical calculation are given in Table 4; for this
system, the program fits the existence of one species in the silver-loaded organic phase,
the corresponding stoichiometry is represented by RHAg(CN)2.
Table 5 shows the values of log Kext for the silver-LIX 79 extracted species when using
different diluents for the organic phase.
4 CONCLUSIONS
The use of LIX 79 extractant allowed silver extraction from alkaline cyanide media. Silver
extraction seems to be independent on the initial metal concentration and the organic
diluent and more dependent on the extractant concentration and aqueous pH values. The
argentocyanide complex is extracted preferentially over other metal-cyano complexes, but
not over the Au(CN)2- complex, at alkaline pH values.
Silver(I) is extracted into the organic phase by formation of the RHAg(CN)2 species. No
polynuclear silver complexes are extracted in the experimental conditions used.
ACKNOWLEDGEMENTS
The authors wish to thank Cognis Corp for its gift of a sample of LIX 79 extractant. This
work has been supported by MCYT (PPQ2002-04267-C03-03) and CIRIT (2001-SGR
00249) (Generalitat de Catalunya). The support of CSIC (Spain) is also acknowledged.
REFERENCES
1 Szymanowski J, Hydroxyoximes in Copper Hydrometallurgy, CRC Press, Boca Raton
6
(1993).
2 Kordosky G, Proceedings International Solvent Extraction Conference (ISEC´02),
Cape Town. pp 853-862 (2002).
3 Mooiman MB and Miller JD, Selectivity considerations in the amine extraction of gold
from alkaline cyanide media, Min. and Metal. Proc. August:153-157 (1984).
4 Mooiman MB and Miller JD, The chemistry of gold solvent extraction from cyanide
solution using modified amines, Hydrometallurgy 16:245-261 (1986).
5 Miller JD, Wan RY, Mooiman MB and Sibrell PL, Selective solvation extraction of
gold from alkaline cyanide solution by alkyl phosphorous esters, Sep. Sci. Technol.
22:487-502 (1987).
6 Alguacil FJ, Hernández A and Luis A, Study of the KAu(CN)2-amine Amberlite LA2
extraction equilibrium system, Hydrometallurgy 24:157-166 (1990).
7 Kordosky GA, Sierakoski JM, Virnig MJ and Mattison PL, Gold solvent extraction
from typical cyanide leach solution, Hydrometallurgy 30:291-305 (1992).
8 Caravaca C, Alguacil FJ and Sastre A, The use of primary amines in gold(I) extraction
from cyanide solutions, Hydrometallurgy 40:263-275 (1996).
9 Alguacil FJ and Caravaca C, Synergistic extraction of gold(I) cyanide with the primary
amine Primene JMT and the phosphine oxide Cyanex 921, Hydrometallurgy 42:197-208
(1996).
10 Virnig MJ and Wolfe GA, Proceedings International Solvent Extraction Conference
(ISEC´96), Melbourne. pp 311-316 (1996).
11 Puigdomenech I, Medusa, Royal Institute of Technology, Stockholm (2002).
12 Marsden J and House I, The Chemistry of Gold Extraction, Ellis Horwood, London
(1992).
13 Liem DH, High-speed computers as a supplement to graphical methods.12.Application
of LETAGROP to data for liquid-liquid distribution equilibria, Acta Chem. Scand.
25:1521-1534 (1971).
7
Table 1. Composition of the real leach solution used in extraction experiments
Brazilian
concentrate
Electronic
circuit board
All values in g dm-3
Au
Ag
Fe
Cu
0.03
0.02
0.004
0.003
0.02
0.0003
0.005
0.08
8
Table 2. Comparison of extractants on the argentocyanide complex extraction
Extractant (type)
Concentration
Diluent
[Ag] (mmol dm-3)
pH50
Reference
Tridecylamine (PA)
Primene 81R (PA)
Primene JMT (PA)
0.8M
0.8M
0.5M
Xylene
Xylene
Xylene
0.25
0.25
0.25
9.7
8.0
7.7
8
8
8
PrimeneJMT + Cyanex
0.25M+0.25M
Xylene
0.25
9.6
9
Adogen 283 (SA)
Amberlite LA2 (SA)
DBBP (PE)
0.05M
0.05M
Undiluted
Xylene
Xylene
-
5
1.5
5
6.3
4.9
50% extraction at
0.1M NaOH and
0.5M NaCN
3
6
5
Guanidine derivative
(GD)
0.01M
0.07
10.9
7
LIX 79
0.25M
18% kerosene
72% mixed xylene
10% tridecanol
n-heptane
0.1
10.3
This work
921 (PA+PO)
PA:primary amine; SA:secondary amine; PO:phosphine oxide; PE:phosphonic ester; GD:N,N´-bis(2-ethylhexyl) guanidine
9
Table 3. Values ofAg/metal obtained in the extraction by LIX 79
Aqueous
solution
pHeq
Ag/Au
Ag/Cu
Ag/Fe
Synthetic
9.47
9.86
10.34
9.70
0.8
0.6
0.6
0.4
4.3
8.0
8.6
4.9
6.6
11.5
20.0
6.4
10.20
0.2
2.5
5.6
Brazilian
concentrate
Electronic
circuit board
Organic phase: LIX 79 0.25 mol dm-3 in n-heptane
10
Table 4. Results of the numerical treatment of experimental data in the extraction of
silver(I) by LIX 79
Species
log Kext
(log Kext)
U

RHAg(CN)2
R2HAg(CN)2
R3HAg(CN)2
10.84±0.12
11.58±0.14
12.82±0.21
0.04
0.05
0.07
0.55
0.78
1.6
0.175
0.210
0.306
11
Table 5. Values of log Kext for the extraction of the argentocyanide complex by LIX 79
in two diluents
Diluent
log Kext
(log Kext)
U

n-heptane
cumene
10.84±0.12
10.69±0.05
0.04
0.02
0.55
0.001
0.175
0.04
12
Figure 1. Silver extraction by LIX 79 in n-heptane at various initial metal
concentrations. pHeq 10.05±0.05.
Figure 2. The influence of the concentration of LIX 79 on silver extraction.
Figure 3. The extraction of metal-cyano complexes by LIX 79.