Metallurgical Testwork to Define the Design Criteria for the

Electrowinning System for KCGM Fimiston Stage III Gold Room

Including Initial Commissioning and Operational Experience
1 2
M Costello , C Waller and P Mason 3

TABLE!
ABSTRACT
Summary ofdesign criteria for Williams Mine System.
The Phase I electrowinning system at KCGM's Fimiston Operation
utilised the well proven technology of electrowinning precious metal
values onto steel wool followed by calcining and direct smelting to Parameter. Units Value
bullion. The above process, while reliable and successful, was to some Cell tvoe Custom equipment
extent labour intensive. The Willillms Mine in Ontario, Canada had
published conditions under which precious metal values are electrowon Cell volume m3 3.5
onto woven stainless steel and directly removed using high pressure Number of anodes/cell 20
sprays. Number of cathodes/cell 18
The current paper describes the testwork which was carried out to Number of cells 4
check the applicability of the Williams criteria to KCGM eluates arising Cell flow rate m3/h1cell 3.2
from the elution of highly loaded carbon from the calcine circuit and
lowly loaded carbon from the float tails circuit. Cell voltage V 2.5-3.5
Information is also given on the commissioning and operational Cell current amos 650
2
experience which occurred. Current density amos/m 1.35-1.94
Wt of stainless steel/cathode iT 2800
INTRODUCTION Number of lavers 24
Cathode dimension 'mm 875 x 875
Kalgoorlie Consolidated Gold Mines (KCGM) have recently
Calculated linear velocitv m\min 0.07
upgraded their operations. As a part of that undertaking, they
Feed l!rade AUiT\m3 350-400
have consolidated the elution and electrowinning of gold from all
Barren grade AUiT\m3 3-5
their operations within the Kalgoorlie region into one facility. In
order to streamline the operation and minimise handling it was
decided to deposit the gold as a sludge onto stainless steel
cathodes, from which the gold sludge could then be hosed and
collected in a filter for subsequent smelting.
Eluate samples and stainless steel mesh
The KCGM system design was based on published information Samples of actual eluate produced at the Fimiston plant were
on the design and operating experience at the Williams Mine, in used in this testwork series. A Fimiston eluate assaying 30 g/m3
Canada, which at the time was the only known, low silver was used for study on low grade solutions while a Gidji eluate
operation achieving this aim. Details were published by the assaying 820 g1m3 was used for the study of high grade solutions.
Society of Mining, Metallurgy & Exploration Inc (Bames and Detailed analyses are given in Tables 3 and 4.
Rapone, 1990). A summary of the design criteria for the Williams Double woven double layered stainless steel mesh was
Mine system is given in Table 1. obtained from Wire Mesh Industries, Sydney for use in the
testwork. Wire diameter was 1541lm.
TEST PROGRAM In the low grade Fimiston eluate the solution pH was 12.3 and
it contained 0.22 'per cent w/v NaCN and 1.43 per centw/v
The design of the new KCGM system was to be carried out by NaOH. The NaOH level was calculated by titrating the total
Minproc Engineers who specified a matrix of tests given in the alkalinity to pH 7.0 and subtracting the contribution of the
following Table 2. The objectives were (0 examine the pass cyanide and expressing the difference as NaOH.
efficiency (gold plated from solution in a single pass through the In the high grade Gidji eluate the solution pH was 12.0 and it
cathode) and the adherence (or otherwise) of the gold deposit contained 0.34 per cent w/v NaCN and 3.3 per cent w/v NaOH.
onto the steel wool over a range of electrolyte flow rates, current Again the NaOH was calculated from the total alkalinity.
densities and gold concentrations. The ranges selected
encompased widely different operating points (Telfer and Pilot electrowinning cell and rectifier
Williams plants).
A perspex pilot electrowinning cell was used for testwork and a
50 mm x 50 mm cathode cross sectional area was chosen to give
suitable eluate flow rates in the linear velocity range that was to
be investigated, ie 0.07-0.35 m1min. Anode to cathode
1. AMMTEC Limited, 6 MacAdam Place, Balcatta WA 6021. separation was the same as that used in conventional cell design,
ie75 mm.
2. KCGM Pty Ltd, Private Mail Bag 27, Kalgocirlie WA 6430. ~ower supply to the cell was supplied by a Hinco Engineering
Model GPR 3060D. The constant current facility of this unit was
3. Minproc Engineers Ltd, 22 Stirling Highway, Nedlands WA used to maintain stated target current densities. Electrolyte
6009. recirculation was via a Promintec diaphragm pump using
adjustable stroke and frequency for flow control. The electrolyte
and cell were temperature controlled at 5D-60°C.

The AuslMM Annual Conference Perth. 24 - 28 March 1996 155 .

M COS1ELLO, C WALLER AND P MASON

TABLE 2
Testwork matrix to investigate sludge formation and current efficiency using four variables of(a) Solution feed grade - Au g/m3;
2
(b) Current density - Alm ; (c) Layers ofknitted SS wool; (d) Electrolyte superficial velocity - mlmin.

Test (a) J!lm3 (b)A1m2 (c) (dhnlmin Condition

1 520 3 32 0.Q7 Williams criteria for Gidji strips
2 70 3 32 0.Q7
3 10 3 32 0.Q7
4 25 3 32 0.14 Higher flow for Fimiston
5 10 3 32 0.14
6 3 3 32 0.14
7 25 4 48 0.14 Effect of increased plating area
8 10 4 48 0.14
9 3 4 48 0.14
10 520 6 32 0.14 Desired criteria
11 190 6 32 0.14
12 70 6 32 0.14
13 25 6 32 0.14
14 10 6 32 0.14 Desired criteria
15 3 6 32 0.12
16 520 6 32 0.35 Usual design flow, eg Telfer
17 25 6 32 0.35
18 3 6 32 0.35
19 520 18 16 0.36 Plant test point (Fimiston)
20 25 18 16 0.35
21 3 18 16 0.35

TABLE 3 Removal of the gold was initially accomplished using a Hardie

Analysis ofFimiston low grade eluate. Pope heavy duty hand held garden spray operating at 70 KPa and
passing 13 Llmin. This system was adequate for removing
poorly attached gold plated from low grade Fimiston eluates, but
Element AI! Au AI Ca Co Cu Fe could not remove gold which was well attached and plated from
Cone fdm3 13.7 30.0 2.2 1.4 0.18 11 38 high grade Gidji eluates.
Element K Mg Na Ni Zn Si - The nozzle intended for use at plant scale was a spraying
Cone fdm3 9 8.8 10000 2 1 15 - system nozzle SS CO BSPTH1I4 USS VEE-JET 50-20. A high
pressure pump was used to supply 12 Llmin to this spray at 700
KPa but this was also unable to remove Gidji gold.
TABLE 4 The final spray removal trials were conducted using very high
Gidji high grade eluate. pressure sprays operating in the range 14-28 000 KPa and
supplied by a commercial KARCHER unit, and under these
conditions competent well attached gold was successfully
Element Ag Au AI Ca Co Cu Fe removed. This method was therefore selected for the actual plant
Cone fdm3 84 820 3.6 5.8 0.24 305 9 operation.
Element K Mg Na Ni Zn - -
Concg/m3 7 0.2 18000 35 1.2 - - Inability to plate from high grade solutions at low
current densities
In accordance with the agreed testwork matrix an attempt was
Experimental method
made to plate from a high grade Gidji eluate at low current
Plating experiments were carried out in a single pass, pass out, density. The target conditions were as follows - gold
mode using a single cathode and two anodes. Cell pass efficiency concentration 760 g/m3 linear velocity 0.14 m/min, current
was checked hourly by taking simultaneous cell in and cell out density 6 amps/m2 and 32 layers of steel wool. It immediately
samples. became obvious that Gidji eluates showed a quite different
Current efficiencies were calculated from the cell assay data response to Williams eluate and that the cell potential applicable
and the flow rate to determine the weight of gold plated per run. to that current density was below the deposition potential for gold
The theoretical current required to plate this gold at 100 per cent and no plating occurred at all. The opportunity was taken to vary
efficiency was calculated and compared with the actual current the current density and the resulting data is given in the following
passing to give the actual current efficiency. Table 5. •
The degree of attachment of gold was determined by drying the The data in Table 5 clearly show that at low current densities
cathode at 120°C before use and then the process was repeated (in the Williams region) gold will not plate from Gidji eluates,
after plating and after spray cleaning to calculate the percentage and that both current efficiency and pass efficiency are direct
of gold still attached. Micro photographs were also taken to functions of applied current density.
illustrate the degree of attachment of the gold.

156 Perth, 24 • 28 March 1996 The AuslMM Annual Conference

 METALLURGICAL TESTWORK FOR ELECTROWINNING SYSTEM AT KCGM FIMISTON STAGE III GOLD ROOM

TABLE 5
Electrowinning characteristics ofhigh grade Fimiston eluate at variable current densities.

Time Cell current Current density Cell voltage Cell flow Cell temp Gold in Au GoldoutAu Fraction
(amns) (A1m2) (V) (ml/min) ("C) (!!1m 3 ) (!!1m3 ) nlated
0800 0.3 6 1.3 344 53 761 761 Nil
0830 0.3 6 1.3 348 57 761 761 Nil
0900 0.3 6 1.3 350 57 761 761 Nil
1000 0.9 18 1.8 356 59 761 612 0.196
1030 0.9 18 1.8 340 57 761 640 0.159
1100 0.9 18 1.8 360 57 761 587 0.228
1130 0.9 18 1.8 350 56 761 598 0.214
1200 1.2 24 2.0 360 58 761 537 0.294
1230 1.2 24 2.0 352 57 761 520 0.317
1300 0.6 12 1.5 352 57 761 713 0.063
1330 1.2 24 2.0 352 57 761 518 0.319
1400 1.2 24 2.0 356 57 761 509 0.331

Gold plated % 16.95

Cathode in g 54.42
Cathode out g 69.12
Attachment Au g 14.7
% 87

TABLE5a TABLE 6
Summary ofdata from Table 5. Current voltage datafor 650 g/m3 Au Gidji eluate.

Current de~ity Current efficiency Fraction plated

(amns/m-) (%) Steel Linear Cell temp Cell volts Cell amps Current
6 Nil Nil
layers velocity ("C) (V) (i) densit;~
(m/min) (amns/m
12 23 0.063 16 0.35 55 1.01 0.09 2.6
18 48 0.200 16 0.35 55 1.23 0.16 4.5
24 57 0.315 16 0.35 55 1.33 0.19 5.4
16 0.35 55 1.38 0.23 6.5
16 0.35 55 1.55 0.39 11.1
HIGH GRADE GIDJI ELUATES 14.6
1.5
16 0.35 55 1.67 0.51
16 0.35 55 1.88 0.63 17.9
1.4
HYDROGEN
DISCHARGE AREA
16 0.35 55 2.05 0.75 21.3
1.3
16 0.35 55 2.21 0.92 26.1
1.2
16 0.35 55 2.33 1.14 32.3
1.1
16 0.35 55 2.46 1.40 39.7
.
CIl

~ 0.9
OXYGEN DISCHARGE PLUS
GOLD PLATING AREA
!<
Id
0.8
a:
a:
0.7
":J
u
0.6
Examination of the current voltage curve allows the following
Id
u
conclusions to be drawn. In the potential range 1.0-1.4 V
0.5 corresponding to the current density interval 2.6-6.5 amps/m2 no
OXYGEN DISCHARGE
0.4 AREA gold plating occurs and the current is carried by the reversible
0.3 oxygen discharge reaction:
0.2 401£ ~ 2H20 + 02 + 4e- EO = +O.45V
0.1
At anode to cathode voltages in the range 1.4-2.QV
corresponding to current densities in the range 6.5-21.3 amps/m2
0
1.01 1.23 1.33 1.38 1.55 1.67 1.88 2.05 2.21 2.33 2.46
CELL VOLTAGE VOLTS gold plating and the oxygen to hydroxide reactions occur
simultaneously at the cathode.
FIG 1 - Current voltage curve. At anode to cathode voltages above 2.QV the hydrogen
discharge reaction:

. Production ofa current voltage curve for Gidji eluates 2H20 + 2e- = H2 + 201£ EO = -0.78V

As a result of the realisation that it would not be possible to plate
high grade Gidji eluates at low current densities it was decided to
investigate the shape of the current voltage curve for this system.
The resulting data is presented in the following Table 6 and
Figure 1.
starts and hydrogen can be seen gassing off the cathodes. The
data would seem to support earlier Australian practise that current
densities in the range 5-15 amps/m2 are optimum for gold
electrowinning.

The AuslMM Annual Conference Perth, 24 - 28 March 1996 157

 M COSTELLO, C WALLER AND P MASON

Reason for the inability to plate high grade Gidji TABLE 7

eluates Relationship between current efficiency and gold grade.
Overall summary.
The minimum recommended caustic soda concentration by
AARL is 5 glL. The figure quoted as used at the Williams Mine Run No Cell in Calculated current
is 10 gIL while the low grade Fimiston eluate contained 14 glL (Au l!frn3) efficiency (%)
and the high grade Gidji eluate 33 gIL. This would cause the cell 15· 5 1.7
resistance to be much lower for Gidji eluate and the oxygen '21 8.6 2.7
reaction would proceed earlier and carry the current to the 8 15.3 5.5
exclusion of gold.
13 26.7 9.6
4 30.1 12.0
Interrelationship between current density, current 11 177 25.1
efficiency, solution grade and pass efficiency 19B 493 58
In gold electrowinning the ability to plate the gold clearly
depends on having sufficient current arpving at the cathode
surface to plate the gold at the rate that it is arriving at the
cathode surface. At the opposite extreme, when plating is
completed and gold grades have fallen to very low levels, then,
for the same current passing, the current efficiency approaches 70 ~-,-------------------,
zero. Paul, Filmer and Nicol (1983) demonstrated that because of
the reversible oxygen reaction occurring simultaneously with
gold plating, current efficiencies for gold cannot be greater than 60
approximately 50 per cent. In cases where currents are
deliberately limited by choosing to operate at low current
50
densities then it can be shown by simple calculation that the rate
of gold plating will be limited by this choice and that low pass
efficiencies must therefore result. 40

In order to demonstrate the relationship between current

efficiency and solution grade the data for those plating runs
30
which were not limited by choosing too low a current density
were calculated an the data is presented in the accompanying
Table 7 and Figure 2. 20

Examination of this data led to the conclusion previously

published Costello (1986), that there is an approximately linear 10
relationship between current efficiency and cell head grade and
that for these electrolytes the relationship was:
oLL_ _- L_ _- l . .L..-_ _-L-_ _.....L---l
Current Efficiency = 0.12 x Cell Head Grade 200 400

The second conclusion was that for high grade solutions a CElL HEAD GRADE I>JJ 91m3

maximum possible current efficiency was in the range 58-66 per

cent at 493-714 glm3 Au, a result which is close to that predicted
by Paul, Filmer and Nicol (1983).
Summary of testwork tesults
The testwork matrix described in Table 2 was completed and an
overall summary of all results is given in the following Table 8.
Major conclusions arising from the testwork
The testwork program led to the following major conclusions.
• Low grade Fimiston eluates can be successfully plated at low
current densities and the resulting bullion can be hosed off
using low pressure (about 70 kPa) water jets.
• The first use of previously unused cathodes causes a strong
gold to steel bond to occur and this gold cannot be removed
using low pressure water jets.
• High grade Gidji eluates cannot be successfully plated at low
current densities «6 amp/m2) which is in contrast to the
procedures followed at the Williams Mine. The reason for
this appears to be that Gidji eluates contain over three times
the caustic level as the Williams eluates. This causes the
caustic to oxygen reversible reaction which is
thermodynamically more favourable to carry all of the current
at Iow current densities hence the deposition potential for
• There appeared to be no advantage in running at very low
gold is not reached and no plating occurs.
FIG 2 - Current efficiency vs cell head.
• The plate quality resulting from electrowinning from high
grade Gidji eluates (+500 g/m3 Au) is so high that plate
removal by low pressure sprays is not possible.
• Very high pressure sprays (about 30 000 kPa) generated by
special units of the Karcher type can remove all types of gold.
• The ability to remove gold even with very high pressure
sprays is limited by the number of layers of double woven
stainless steel used and 16 layers was recommended as the
starting point for the Fimiston design.
• The design fraction per cathode pass for a 16 layer cathode
was recommended to be 0.15 when not limited by restrictions
on current density.
• While running at low current densities is adequate for plating
from low grade solutions there is clear evidence that this will
severely restrict system performance at high grades. The new
elution system was to operate on the split elution principle so
that even higher grades were expected. For this reason the
recommended desigu maximum current density was 20
2
amps/m •
linear velocities and 0.25 mlmin was recommended as the
design value.

158 Perth, 24 - 28 March 1996 The AuslMM Annual Conference

 METALLURGICAL TESTWORK FOR ELECTROWINNING SYSTEM AT KCGM FIMISTON STAGE III GOLD ROOM

TABLES
Summary oftestwork results.

Run Run No Auin Auout Fraction Current Current Linear Cell amps Cell volts Layers Attachment
(g1m3) (g1m3)
order plated dens~~
(AIm
efficiency
(%)
velocity
(m/min)
(A) (V) 31688 (%)

I 4(1) 30.1 24.7 0.195 3 12 0.14 0.15 1.5-1.6 32 85.5

2 13 26.7 16.5 0.380 6 9.6 0.14 0.30 1.9 32 Nil
3 17 22.0 17.8 0.195 6 4.9 0.35 0.30 1.8-2.1 32 Nil
4 5 19.1 16.1 0.157 3 5.7 0.14 0.15 1.4-1.6 32 Nil
5 14 18.8 15.8 0.158 6 2.8 0.14 0.30 2.1 32 Nil
6 3 17.1 11.3 0.339 3 5.5 0.07 0.15 1.7-1.8 32 5.9
7 8(1) 15.3 9.5 0.379 4 5.5 0.14 0.30 1.9 48 100
8 9 5.9 3.9 0.326 4 1.8 0.14 0.30 1.6-1.7 48 12.3
9 6 5.5 3.6 0.350 3 3.7 0.14 0.15 1.7 32 -
10 15 5.0 3.3 0.352 6 1.7 0.14 0.30 1.9 32 80.5
11 20(1) 28.7 23.0 0.206 18 8.7 0.35 0.45 2.0 16 82.0
12 18 8.7 8.6 Nil 6 Nil 0.35 0.30 1.5 32 Nil
13 21 8.6 6.9 0.194 18 2.7 0.35 0.45 2.1 16 3.2
14 lOA 761 - - - - 0.14 - - - -
15 lOB 651 552 0.152 8-12 24-47 0.14 0.4-0.6 1.6-1.8 32 14
16 19 714 variable 0.160 36 66(2) 0.25 0.9 2.4 16 78
17 19A 588 567 0.036 12 24 0.25 0.45 1.9 16 100
18 19B 493 417 0.154 18 58 0.25 0.68 2.2 16 87
19 II 177 144 0.188 18 25 0.25 0.68 2.2-2.4 16 54.3

(1) New cathodes used for these run resulting in a much higher attachment than recycled cathodes.
(2) Operating at very high current densities with attendant co-evaluation of hydrogen gas caused plated gold to strip off the cathode and leave the cell.

PLANT DESIGN As difficulty was experienced in depositing gold at the low end
from the high grade Gidji solutions, it becomes necessaIj' to
The objective in the new Fimiston gold room was to deposit gold design for operation at high current density, up to 18 AIm. It
as a sludge onto knitted stainless steel wire cathodes. This sludge was thus necessary to increase the rectifier capacity to 5000 amp
was to be washed off with water sprays, filtered, dried and per unit and the design criteria as shown in Table 9 was
smelted. Eluate from a range of carbons was to be treated, with established.
anticipated pregnant solution grades ranging from 25 g1m3 TABLE 9
(Fimiston carbon) to 550 g/m3 (Gidgi carbon).
KCGM electrowinning design criteria.
Prior to conducting testwork at Ammtec, Minproc based the
design on information from the Williams mine in Canada.
Williams pregnant solution grades fall in the range 200-600 glt, Carbon source Gidgi Fimiston
Gidji carbon eluate was expected to run around 550 glt. Williams Carbon batch size. t 8 II
achieve a good single pass efficiency through their cells by Sinl!le bed volume, m
3
16 22
runnin~ a low solution flow (0.07 rn/min) and low current density
No of bed volumes 9 8
(3 AIm). Their computed current efficiency is in the range
Pregnant volume, m3 144 176
25-40 per cent.
Carbon loadinl!:, fdt 9800 400
For flexibility at Fimiston, Minproc provided for a flow range
Pregnant grade, g/m3 544 25
from 0.07 to 0.35 rn/min, the latter being more typical. They
initially specified 3000 am~s (variable) per cell of 24 cathodes, Cathode area, m 2 1.0 1.0
which would permit 6 AIm if 32 layer cathodes were used, or Electrolyte velocity, 0.25 0.25
12 Alm2 if only 16 layers were used. A current efficiency of 21 rnImin
per cent would be sufficient at a flow of 0.14 rn/min (twice that of No of cells _ 2 2
Williams). Cathodes per cell 24 24
Testwork was initiated at Ammtec to establish a suitable Single pass time, h 4.8 5.9
operating point within the design range. The aim was to establish Current, amp/cell 5000 5000
a relationship between pass efficiency, sludge production, Tar~et barren grade, 5.0 1.0
solution grade, current density, and solution velocity. fdm'
It became evident that the low current density used by Williams Current efficiencv, % 22 1.0
was insufficient to achieve gold deposition from Gidji solution. Wire diameter, mm 0.152 0.152
3
At low voltage a competing cell reaction, probably reduction of Steel density, kfdm 8020 8020
oxygen as discussed earlier, carries the current. As the voltage is Cathode lavers 16 16
increased a change of slope (Figure 1) indicates where gold Weightllaver/m2, g 200 200
deposition starts. At this low end, there would appear to be an Cathode weil!ht, kl!: 3.2 3.2
operating window in which a loosely adhering gold will deposit, 2
Cathode wire area, m 10.5 10.5
the point at which Williams operate. At the higher voltage end,
hydrogen evolution will interfere with the gold deposition, thus Current density, 19.8 19.8
amp/m2
another operating region is available.

The AuslMM Annual Conference Perth, 24 - 28 March 1996 159

 M COSTBLLO, C WALLER AND P MASON

COMMISSIONING AND OPERATION apparently little effect on pass efficiency, which held at around 70
per cent until solution gold grades dropped below 10 glm3•
To-date commissioning and operation of the Fimiston stainless The first batch of high grade Gidji carbon stripped in the new
steel electrowinning circuits has been carried out in three phases elution circuits produced an eluate grading 1050 glm3 gold. At
dictated by project circumstances and operational requirements. these high gold tenors electrowinning at 2000-3000 amps and six
volts (three volts per cell) produced a pass efficiency of around
Phase 1 : Pre-commissioning 85 per c«nt in the early sti!!ges.
The new gold room complex at KCGM's Fimiston plant was Under all conditions run the cells produced a soft metal deposit
installed as part of a $115 million project which substantially that was easily removed from the stainless steel cathodes using a
increased plant throughput and gold production. Karcher pressure washer with rotary nozzle attachment.
Pregnant eluate for commissioning the two stainless steel A site visit by Mike Costello on the 9 August 1994 to review
electrowinning circuits was to be provided by two new ll-tonne available data and conduct a survey of electrowinning conditions
high temperature (l30°C) split AARL elution circuits. The confirmed that the circuits were performing somewhat below that
stainless steel electrowinning circuits and elution circuits were to specified in the process design criteria. In particular it appeared
be commissioned in parallel, replacing Fimiston's existing that, at the lower range of rectifier currents then in use (2000
conventional steel wool electrowinning circuit and low amps, 7.4 amp/m2), that the pass efficiencies were being current
temperature (l05°C) AARL elution circuit. density controlled.
When construction delays threatened to push back By the second half of August a single operating regime had
commissioning of the new elution circuits to a point where the been identified which, although not offering the final desired
commissioning team would be faced with additional pressure performance of a 'single pass and out' pass efficiency, did enable
from increasing gold production, it was decided to attempt to the circuits to be put onto a routine operating basis with
pre-commission the electrowinning circuits on eluate from the minimum operator intervention required for the wide range of
existing Fimiston elution circuit. eluate grades processed.
As the new gold room was remote from the existing eluate The operating parameters established were:
storage tanks, eluate transfer was undertaken using a bulk Cell current: 4000 amps
chemical tanker. The large volume of eluate to be transferred, 2
over 150 cubic metres, and the requirement for temporary load-in Current density: 14.7 amps/m
and load-out pumps and pipelines made the transfer a time Rectifier voltage: Adjusted to achieve required
consuming and cumbersome process. In event only a single amperage. Usually 6.4-7.8 volts.
batch of low grade eluate from a Fimiston CIL carbon strip was
transferred. Cell voltage: 3.2-3.9 volts
That single batch, however, enabled the commissioning team to Eluate flow rate: 23m3/hr/cell
prove the operation of the electrowinning circuits under
Electrowinning time: ten hours on circulation then
something approaching operating conditions. Although frequent
single pass to tail.
stoppages for equipment adjustment and, particularly, control
system modification and reprogramming rendered the When compared with the process design criteria, the operating
metallurgical performance results meaningless, commissioning parameters were just above the mid-point on the current settings
and operating staff were able to approach phase 2 in a much more but with a significantly higher flow rate than design
confident manner. Leaks were fixed, instruments calibrated, (23 m3/hr/cell vs 15). No attempt had been made to optimise
control philosophy proven and the circuits were ready to go! caustic and cyanide levels or to evaluate the effects or levels of
silver and copper in the eluates.
Phase 2 : Achieve acceptable operating conditions Af this point the electrowinning circuits were handed over to
the operating personnel with further optimisation of the process
Practical completion of the new elution circuits was reached at conditions to be carried out at a later date.
about the time that the Fimiston plant expansion achieved its first Under these conditions the following performance was
step increase in gold recovery at the Fimiston site. achieved over the first three weeks of September.
Site modifications to the carbon handling systems required that
. the old elution circuit and gold room be decommissioned as soon TABLE 10
as possible. Thus it was essential that the new gold room and Performance achieved over the first three weeks ofSeptember 1994.
stainless steel electrowinning circuits achieve acceptable
operating conditions in as short a time as possible.
Eluate No of Avehead NaOH NaCN Recovery
Electrowinning of eluate from the new elution circuits tvne e/wins 3
f!lm Au (%) (%Au)
commenced on 21 July 1994. The eluate was, once again, from Fimiston 11 117 1.7 0.57 97.2
Fimiston circuit carbon and graded 28 g/m3 gold. A current of
Mt 9 226 1.5 0.57 98.7
2100 amps at a rectifier setting of 5.5 volts (2.75 volts per cell) Charlotte
was used and a cell flow rate of 15 m3/hr. The electrowinning
Gidii 10 860 1.5 0.54 99.3
took 14 hours but was hindered by a number of circulation pump
trips triggered elsewhere.
Factory settings on the rectifiers were restricting the units to an
output of 2100 amps holding the voltage down to only 5.5 volts. Phase 3 : Optimisation
These were adjusted to allow the rectifiers to run up to currents in Optimisation of the stainless steel electrowinning circuits will
excess of 5000 amps with output voltages then limited to 7.5 to 8 require a consideration of the following variables:
volts (3.75 to 4 volts per cell).
• current density,
The second batch of eluate was from higher grade carbon from
• eluate flow rate,
the new Mt Charlotte circuit and graded 133 glm3 gold.
Electrowinning was carried out at various amperages with • cyanide concentration,
• caustic concentration,

160 Perth. 24 - 28 March 1996 The AuslMM Annual Conference

 METALLURGICAL TESTWORK FOR ELECTROWINNING SYSTEM AT KCGM FIMISTON STAGE III GOLD ROOM

TABLE 11 CONCLUSIONS
Comparison ofcurrent Fimiston operating conditions with
Williams System. A comparison of current Fimiston operating conditions with the
Williams system from which it was developed is shown in
Parameter Units Williams Fimiston Table 11.
Cell volume m3 3.5 3.85 Clearly Fimiston's operating conditions are significantly
Number of 19 25 different to those established at the Williams Mine. This
anodes/cell demonstrates the robustness of the process with a broad range of
Number of 18 24 operating conditions available to suit individual operators needs.
cathodes/cells However, a greater knowledge and understanding of the effect of
Number of cells 4 2 per circuit process variables is required. It is hoped that this will be gained
Cell flow rate m3/h1cell 3.2 15-24 over the coming months. One final comment on the use of
Cell voltage V 2.5-3.5 3.2-3.9 stainless steel cathodes for electrowinning eluates with a high
Cell current amos 650 4000 gold to silver ratio. There was some concern following the
laboratory scale testwork that the plating process may work so
Current density amps/m2 1.35-1.94 14.7
well that the gold plate would be difficult, if not impossible, to
Wt of stainless g 2800 3500 remove from the cathode material. Indeed several of the test
steel/cathode
cathodes plated using the laboratory test cell would not have
Number of 24 16 looked out of place in a jewellers window! Fortunately this
lavers
phenomena has not been repeated on a plant scale with gold
Cathode mm 875 x 875 lOOOx 1000 deposits readily washed from the cathode 'and recovered for
dimension
filtration and smelting.

• degree of gassing, REFERENCES

• silver and copper levels and control, and
• additional layers of cathode material.
Under the direction of Fimiston metallurgical staff ongoing
electrowinning operations are being closely monitored and key
operating parameters varied.
Although the electrowinning circuits have not yet achieved
their process design criteria performance the robustness of the
initial design has meant that their performance is more than
adequate for ongoing operations. This, inevitably, results in
metallurgical effort being diverted to other areas of the plant with
a higher priority.
Barnes, D and Rapone, T R, 1990. Electrowinning and refining at the
Williams mine using knitted stainless steel cathodes, Minerals and
Metallurgical Processing, August(5):128-134.
Costello, M C, 1986. Experience with the use of electrowinning gold
from pregnant eluates in the gold mining industry in Australia, in
Proceedings Symposium on the Electrochemistry of Gold Processing
(RACI, Murdoch University, Perth).
Paul, R L, Filmer, A 0 and Nicol, J V, 1983. The recovery of gold from
concentrated aurocyanide solution, in Proceedings 3rd International
Symposium on Hydrometallurgy, pp 640-704.

The AuslMM Annual Conference Perth, 24 - 28 March 1996 161

162 Perth, 24 - 28 March 1996 The AuslMM Annual Conference