The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on

Base Metals
Jűrgen Gnoinski

SKORPION ZINC: OPTIMISATION AND INNOVATION

J. Gnoinski
jgnoinski@skorpionzinc.com.na

Namzinc (Pty) Ltd, Private Bag 2003, Rosh Pinah

Namibia

ABSTRACT

The Skorpion process comprises three processing steps, atmospheric leaching in sulphuric
acid, zinc solvent extraction and electrowinning to produce special high grade zinc. The
major mode of impurity transfer was attributed to crud runs in the solvent extraction circuit.
Crud may be described as a heterogeneous, partly inorganic, partly organic mass of
material which accumulates predominantly at the interface between organic and aqueous
solutions. The paper describes the route causes of the crud formation in the solvent
extraction circuit and optimisation strategies which are being followed to prevent impurity
excursions in electrowinning.

Keywords: zinc oxide, solvent extraction, crud formation, colloidal silica

Page 329
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

1.0 Introduction

The Skorpion process was developed to treat a complex non-sulphide oxide ore body using
atmospheric leaching followed by solvent extraction and electrowinning of zinc (1,2,3) .
Sauconite is the largest contributor to the zinc grade with lesser amounts of zinc
hydroxides. The gangue minerals are mainly quartz, K-feldspar and muscovite mica. The
plant operation achieves an annualised zinc production of 150 000 tonnes special high
grade zinc. The choice of SX technology was dictated by the need to upgrade the zinc tenor
for electrowinning, to reject dissolved halides and to minimise soluble losses of zinc in the
filtration step.

The refinery was commissioned during the first half of 2003 and ramped up to full
production in 2005. The full production run was interrupted during 2005 and August 2006
due to hydrogen fires in electrowinning caused by the transfer of significant concentrations
of impurities from the solvent extraction circuit, namely nickel and copper, into the
electrolyte solution. The major mode of impurity transfer was attributed to crud runs in the
solvent extraction circuit. Crud may be described as a heterogeneous, partly inorganic,
partly organic mass of material which accumulates predominantly at the interface between
organic and aqueous solutions.

The paper describes the route causes of the crud formation in the solvent extraction circuit
and optimisation strategies which are being followed to prevent impurity excursions in
electrowinning.

2.0 Process chemistry

2.1 Leaching

Leaching of zinc oxide ores by sulfuric acid are generally described in terms of the
following reaction equations:

ZnO.SiO2 + H2SO4 + H2O = ZnSO4 + H4SiO4 (1)

Zn2SiO4 + 2H2SO4 = 2ZnSO4 + Si(OH)4 (2)
Zn4Si2O7(OH)2. H2O + 4H2SO4 = 4 ZnSO4+ 2 Si(OH)4 +2 H2O (3)

Leaching is carried it out in a series of agitated tanks at about 50oC. The final pH is
maintained between 1.8 - 2.0 to maximise stability of colloidal silica. After leaching the pH
is raised to 4.0 to ensure to maximise agglomeration and the formation of precipitate.

Page 330
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

Dissolved siliceous species, namely, Si(OH)4 (monosilicic acid) remain in solution under
conditions, where its solubility is not exceeded. Depending on pH, temperature and
dissolved salts concentrations siliceous species will tend to polymerise to form colloidal
silica. These siliceous polymers, sized from 1 nm to 1000 nm, have a massive surface area,
but are still small enough to be unaffected by gravity. Colloidal-sized, polysilicic molecules
can cause serious problems in a solvent extraction (SX) system by stabilising emulsions.
The formation of colloidal silica is irreversible and follows a dehydration process;

nSi(OH)4 = (SiO(OH)2)n.H2O + (n-1)H2O ( 4)

2.2 Solvent extraction

During solvent extraction the organic extractant used is 40% v:v di-2-ethyl hexyl
phosphoric acid (D2EHPA, symbolized as RH) diluted in the solvent ESCAID 100. The
extraction of zinc is by cation exchange:

Zn2+(aq) + 2RH (org) = 2H+(aq) + ZnR2 (org) (5)

The organic phase is stripped of zinc using spent electrolyte:

ZnR2(org) + 2H+(aq) = Zn2+(aq) + 2RH(org) (6)

The extraction versus pH curves provide an indication of the selectivity of the organic
extractant under different conditions. A typical E-pH curve measured is shown in Figure1
(4). From reaction (5) it is seen that the acidity of the aqueous phase increases as the
extraction proceeds, with two protons released for each metal ion extracted. For typical
Skorpion operating conditions of 30 g/l Zn in the PLS and 10 g/l Zn in the raffinate, this
corresponds to approximately 30 g/l H2SO4 being generated across the extraction circuit.
The general order of selectivity with increasing pH in this system is (4):

Fe(III) < Zn < Ca < Al(III) < Mn < Cd ~ Cu < Mg < Co < Ni.

Page 331
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

Figure 1: Effect of pH on the selectivity of extraction of base-metal cations in the Skorpion

Zinc system. Conditions: 40 vol.% D2EHPA in Escaid 100, 40 oC 30g/l Zn in feed (4).

2.3 Electrowinning

The last step in the Skorpion process s the electrodeposition of zinc onto aluminium
cathodes form the loaded electrolyte:

2 Zn2+ + 2 H2O = Zn + 4H+ + O2 (7)

The produced spent electrolyte is sent back to the stripping stage of SX.

3.0 Plant description

A simplified Skorpion flow sheet of the process is given in Figure 2 (5). A detailed
discussion of the circuit is given in (5). Following an atmospheric sulphuric acid leach;
iron, aluminium and silica are precipitated from the pregnant leach solution in the
neutralisation section. This neutralised and clarified pregnant liquor solution (PLS) is sent
to the solvent extraction circuit (SX), (Figure 3). Extraction is performed over three stages
operating at 400C with an O:A ratio between of 1 to 1.5. Zinc is transferred to an organic
phase that contains 40% v/v D2EHPA diluted in kerosene. The acidic raffinate is sent back
to leaching. The first two stages of the washing circuit use demineralized water to wash out
physically entrained impurity species; spent electrolyte is employed as a scrub liquor in the
third wash stage to ‘‘crowd off’’ co-extracted impurity species from the loaded. The loaded
organic phase is stripped from its zinc content by contact with spent electrolyte. Following
electrowinning , the end product is SHG zinc cathodes (99.995% Zn).

Page 332
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

Figure 2: Simplified process flowsheet of the Skorpion Zinc refinery.

Page 333
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

Figure 3: Schematic representation Skorpion SX circuit.

Iron and aluminium are also extracted preferentially to zinc, but cannot be easily stripped
from the organic phase. For this reason, a small portion of the stripped organic is bled for
regeneration. In the regeneration step, the organic is brought into contact with a strong
hydrochloric solution (6 mol/l HCl) to remove all the iron. The resulting iron-chloride
solution is neutralised and is returned to the leaching section.

The operating and engineering availability of the refinery was significantly increased by
decoupling the electrowinning and solvent extraction sections as well as the SX and leach
sections by installation of additional capacity. An additional thickener was installed to
ensure flexibility and low solids transfer into SX.

4.0 Defining the problem

After more than 15 months of stable operation, a significant excursion of metal impurities
namely, Cu and Ni into the electrowinning circuit forced an unscheduled shut of the
refinery in August 2006. After restart of the refinery, intermittent impurity transfer across
SX continued impacting on the performance of the electrowinning circuit and the output of
the Refinery (Figure 4). The relative quantities of Ni, Cu, Co, and Cd transferred were
consistent with the chemical composition of the PLS. A salient feature of Figure 1 is that
the intermittent impurity excursions appeared as well defined spikes against an otherwise
steady background levels rather than an overall increase in the impurity background
suggesting a physical rather than a chemical transfer mechanism. Transfer of PLS across
SX may either occur by aqueous entrainment and/or crud or another mechanism involving
some degradation of the extractant or diluent. The contributions of these mechanisms are
discussed below.

Page 334
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski
1600 9000

1400 8000
Ni

[Cu] mg/l in loaded electrolyte

[Ni] mg/l in loaded electrolyte

7000
1200 Cu
6000
1000
5000
800
4000
600
3000
400
2000

200 1000

0 0
1-Jul

16-Jul

31-Jul

15-Aug

30-Aug

14-Sep

29-Sep

14-Oct

29-Oct

13-Nov
Figure 4: Evolution of impurities concentrations in loaded electrolyte (1 July -18
November 2006).

4.1 Effect of aqueous entrainment

Routine aqueous entrainment measurements averaged between 100 and 200 mg/l for the
organic and aqueous extraction and washing stages. Mass balance calculations showed that
these levels were too low to have caused the observed impurity levels in the loaded
electrolyte.

4.2 Organic health

Standard shake-out tests on the organic phase confirmed that chemical extraction of metal
impurities by the organic was within design limits, thus excluding chemical transfer as a
primary transfer mechanism. Also, co-extraction was not likely to occur in the form of
sporadic spikes. Sensitive Fourier Transform Infrared spectroscopy FTIR) and Nuclear
magnetic Resonance (NMR) spectroscopic analysis of the organic phase could not detect
any significant degradation products.

The viscosity of the organic phase was however moderately elevated and minor changes in
phase disengagement times were apparent. These alterations were suggestive of a gradual
change in the physical properties of the organic phase, but were not expected to impact
significantly on the hydraulic behavior. Detailed chemical analysis showed a build-up of

Page 335
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

elements such as scandium (100 mg/l) and yttrium (100 mg/l) on the extractant. It is well
known that trivalent species such as Fe3+ can polymerize D2EHPA and change the viscosity
of the organic phase. In the Skorpion Zinc flowsheet a bleed of the organic is treated with 6
M HCl to maintain the organic-phase iron loading at 300 mg/l. Subsequent to the detection
of Sc and Y, the runtime on the HCl plant was changed to a ensure removal of these species
form the organic phase.

4.3 Crud transfer

Crud in hydrometallurgical SX can be defined as stable, non-coalescing deposits occurring

within the settler of an SX operation and originate from semi-stable aqueous-organic
emulsions formed in SX settlers (7,8,9).

The formation of emulsions in the extraction stages was observed prior and during the
metal impurity excursions and was accompanied by unusually high levels of mobile
floating crud and accumulation of large quantities of bottom crud in the extraction settlers.
Elemental analysis of the floating crud indicated the presence of significant amount of Si
(9.85%), Zn (4.1%), Al (1.79%), Ni (0.03%) and Cu (0.06%). The moisture content was
37%. The crud contained therefore a significant amount of entrained PLS. The most likely
cause of impurity transfer was thus the sporadic transfer of floating crud (and with it
entrained PLS) from the extraction stages across the washing circuit into the stripping
circuit With intermittent crud runs as the main impurity transfer mechanism, the observed
spikes in nickel and copper in the loaded electrolyte can be explained.

4.4 Causes of crud formation

The precursor to crud formation is the development of an emulsion (10). The emulsion
progresses to a crud through repeated mixing, and continuous inclusion of solids and
colloids. Once formed, its dispersion and separation is difficult to achieve under operating
conditions. It is generally accepted that the contributing factors to crud formation are the
composition and components of the pregnant liquor stream (PLS), the components or
degradation products of the organic phase, and the operating equipment or operating
conditions within the circuit (10). Available information on other systems indicate that the
components of pregnant liquor stream with significant influence on crud formation are fine
solids, silica, manganese, and foreign organic matter (10). Considering the chemical
composition of the crud namely, Si ,and Al it was mooted that crud formation was linked to
solids carryover and the presence of colloidal silica. Solids may enter the solvent extraction
(SX) circuit with the aqueous feed or precipitate during the process, while colloidal-sized,
polysilicic molecules can cause serious problems in a SX system by stabilising emulsions..

Page 336
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

The solids content entering SX is monitored by routine measurement of suspended solids

(TSS). The frequency distributions of the TSS values during the impurity and after the
excursions are shown in Figure 5. The average TSS value during the excursion was 52 mg/l
compared to 39 mg/l after the impurity excursions. After the impurity excursions a
significant fraction (34%) of the TSS measurements was smaller than 23 mg/l. This must be
compared to the targeted value of <10 mg/l to minimize crud formation in SX.

0.02
0.02
0.02
0.01
After impurity excursion
0.01
Density

0.01
Before impurity excursion
0.01
0.01
0.00
0.00
0.00
0 50 100 150 200 250
TSS mg/l

Figure 5: Frequency distributions of total suspended solids (TSS) entering SX.

An important observation during the impurity excursion was a noticeable change in the
physical appearance of the crud from coarse (bottom crud) to voluminous, highly
gelatinous type crud (floating crud) during the impurity excursion. In literature this type of
crud has often been associated with fine clay and silica (10). The predominance of Si
(27,859 mg/l) among the contaminants in the floating crud left very little doubt that it was
the main contributing factor in the crud formation. Analyses of the crud suggested that it
was entering the circuit in at least two different forms: soluble species (colloidal silica) and
suspended solids.

Crud transfer was intrinsically linked with nickel excursions in the electrolyte. Bearing in
mind that D2EHPA is not selective towards extraction of nickel under the operating
conditions (Figure 1) and no evidence of organic degradation could be determined, nickel
was used in the analysis below as a tracer in crud transfer.

Page 337
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

Comparisons of the silicon concentration during and after the excursion are shown in
Figure 6. There was no significant change in the mean of the concentration (60 and 63 mg/l
prior and after the excursion) confirming that the form of silicon (colloidal or suspended)
had a more significant impact on crud formation than the actual concentration. This was
also confirmed in laboratory testwork (10): Dissolved silica had no notable effect either on
the phase disengagement (PD) time or interfacial characteristics. However colloidal silica
(130 mg/L) led to a slightly longer PD time and persistent cloudiness of the aqueous phase
including the interfacial region. It is reasonable to suggest that this could eventually lead to
emulsification, a known precursor of crud formation. These clearly demonstrated that while
dissolved silica is not a major contributing factor to crud formation, colloidal silica is
certainly a major contributing factor.

As indicated (section 2.1) above low pH (<2.0) in the leach section is expected to reduce
the selectivity of the leaching process and increase the proportion of amorphous silicates
and impurities in the PLS. Considering this, a decision was made to control the pH at 2.0 in
order to maximise the stability of colloidal silica.

Page 338
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski
0.045

0.040

0.035

0.030

Density
0.025

0.020

0.015

0.010

0.005

0.000
0 50 100 150
Si (mg/l)
a)
0.045

0.040

0.035

0.030
Density

0.025

0.020

0.015

0.010

0.005

0.000
0 50 100 150
Si (mg/l)
b)
Figure 6: Frequency distributions of Si in PLS a) during impurity excursions b) after
impurity excursions.

The frequency distributions of the pH values in the last leach reactor recorded for the
periods during and after the impurity excursions are shown in Figure 7. The period after the
excursions refers to the time when no impurity excursions were observed and the pH was
controlled to 2.0. The average pH in neutralisation over the same periods remained the
same at 4.2, although the variance decreased after optimisation of the pH control.

Page 339
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

5.00
4.50
4.00 pH after im purity
excursions
3.50
3.00
Density
pH during im purity excursions
2.50
2.00
1.50
1.00
0.50
0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00
pH in leach

6.0
pH after impurity excursions
5.0

4.0
Density

3.0
pH during
impurity
2.0 excursions

1.0

0.0
3.50 4.00 4.50 5.00
pH in neutralisation

Figure 7: Frequency distributions of pH a) in last leach tank b) in neutralisation.

The significant impact of the leach pH on Ni concentrations in the electrolyte is shown in

Figure 8. At the low leach pH, Ni in the electrolyte averaged 120 ppb compared to 80 ppb
after the pH control was optimised at pH 2.0. Bearing in mind that the targeted value in
electrowinning is 80 ppb Ni unacceptable high levels of Ni were detected during the
impurity excursions. Subsequent pH optimisation resulted in far better Ni control and no
impurity excursions were observed since.

Page 340
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

0.030

0.025

0.020
Density [Ni] after excursion

0.015

0.010
[Ni] during impurity
0.005 excursions

0.000
0 100 200 300 400
Ni ppb

Figure 8: Frequency distributions of Ni in electrolyte.

5.0 Conclusions

The major mechanism of impurity transfer across the SX circuit has been attributed to
intermittent episodes of aqueous entrainment that were associated with the build up and
release of crud in the extraction stages and transfer to the electrowinning circuit. The
predominance of Si among the contaminants in the floating crud left very little doubt that it
was the main contributing factor in crud formation. Low pH (<2.0) in the leach section
reduced the selectivity of the leaching process and increased the proportion of colloidal
silica and impurities in the PLS. After adjusting and controlling the leach pH at 2.0 crud
transfer episodes have reduced significantly. However, the current TSS values are still
above the target of 10 mg/l. Skorpion Zinc is currently evaluating a number of solid-liquid
separation technologies such as pinned bed clarification to achieve this.

6.0 Acknowledgements

The author would like to express thanks to staff at Skorpion Zinc, Técnicas Reunidas and
Anglo Research and Curtin University of Technology for their assistance and helpful
discussions.

Page 341
The Southern African Institute of Mining and Metallurgy The Fourth Southern African Conference on
Base Metals
Jűrgen Gnoinski

6.0 References

1. Martín, D., Díaz, G., García, M. A. and Sánchez, F., “Extending zinc production
possibilities through solvent extraction,” Proceedings of the International Solvent
Extraction Conference, ISEC 2002, Vol. 2, Sole, K. C., Cole, P. M., Preston, J. S.
and Robinson, D. J., Eds., Johannesburg: South African Institute of Mining and
Metallurgy, 2002, 1045-1051.

2. García, M. A., Mejias, A., Martín, D. and Díaz, G., “Upcoming zinc mine projects:
The key for success is ZINCEX solvent extraction”, Lead-Zinc 2000, Dutrizac, J.
E., Gonzalez, J. A., Henke, D. M., James, S. E. and Siegmund, A. H.-J., Eds.,
Warrendale, PA: The Minerals, Metals & Materials Society, 2000, 751-761.

3. Díaz, G., Martín, D. and Lombera, C., “Modified Zincex Process: the clean, safe
and profitable solution to the zinc secondaries treatment,” Res., Conserv. Recyc.,
Vol 10,1994, 43-57.

4. Sole, K. C., “Preliminary investigation of the behaviour of impurity elements and

the stripping of iron(III) in the Skorpion solvent-extraction process“, Anglo
American Research Laboratories Blue Cover Report, Project No. R24692/3, Report
No. 8.

5. Gnoinski J., Bachmann T., Holtzhausen S., “Skorpion Zinc: Defining the cutting
edge of zinc refining,” Proceedings of the International Symposium on Lead and
Zinc Processing, Koto, Japan, 2005, 1315-1325.

6. Cole P.M., Sole K., “Zinc solvent extraction in the process industries“, Mineral
Processing & Extractive Metall. Rev., 2003, 24: 91-137,

7. Ritcey, G. M., “Crud in Solvent Extraction Processing - A review of causes and

treatment“, Hydrometallurgy, 1980, vol. 5, pp. 97 - 107.

8. Burniston, T., Greenshields, J. N. & Tetlow, P. E. 1992, “Crud control in copper SX

plants“, Engineering and Mining Journal, vol. 193, no. 1 (January), pp. 32 - 35.

9. Hughes, C. A., Evans, H. A. & Warren, L. J. 1996, “Investigation of the role of

silica in the formation of cruds in a hydroxyomine / sulfuric acid solvent extraction
syste“', in ISEC '96, eds. Shallcross, D. C., Paimin, R. & Prvcic, L. M., Elsevier,
Amsterdam, Melbourne, Australia, pp. 1277 – 1282.

10. Ibana, D.C., Helm D.J., Collard J.M., Steffens M.J.. (2006), “Assessment of crud
formation in the Skoprion Solvent Extraction Circuit“ , Anglo American Research
Blue Cover Report, Project No. 49600, Report No. 2.

Page 342