You are on page 1of 7

Minerals Engineering 54 (2013) 14–20

Contents lists available at SciVerse ScienceDirect

Minerals Engineering
journal homepage: www.elsevier.com/locate/mineng

Recovery of nickel and cobalt from laterite leach tailings through resin-in-pulp
scavenging and selective ammoniacal elution
Patrick Littlejohn ⇑, James Vaughan 1
School of Chemical Engineering, Level 3, Chemical Engineering Building (74), College Road, St. Lucia, Queensland 4072, Australia

a r t i c l e i n f o a b s t r a c t

Article history: In the hydrometallurgical processing of nickel laterite ore, a significant amount of dissolved nickel and
Available online 18 March 2013 cobalt can be lost to neutralized tailings slurry during counter current decantation. Resin-in-pulp (RIP)
scavenging using ion exchange resin functionalized with iminodiacetic acid (IDA) has been proposed
Keywords: as an effective way to recover lost metal value. However, the tendency of this resin to co-load impurities
Ion exchange makes integration of metal value back into the plant challenging. In this paper, the application of bis-pic-
Resin-in-pulp olylamine functionalized resin to RIP scavenging is explored. Both types of resin load the valuable metals
Nickel laterite
effectively but bis-picolylamine functionalized resin has a much lower affinity for unwanted impurities.
Nickel and cobalt
Picolyl
After loading, selective elution of nickel and cobalt from the new resin was carried out using an ammonia
Ammonia based eluent. The resultant eluate was high in nickel and cobalt with ppm levels of impurities, making it
appropriate for downstream integration in existing nickel/cobalt refineries.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction et al., 1991). However, the authors are unaware of any commercial
application of resin-in-pulp in the base metal industry.
Every nickel laterite processing flow sheet is different, but in The goal of resin-in-pulp scavenging is to selectively recover va-
general, a leaching stage is employed to dissolve valuable nickel lue metal from slurry by contacting the slurry with relatively large
and cobalt followed by separation of the solution from unwanted bead chelating ion exchange resin. After value metal has loaded
solids using counter current decantation. Counter current decanta- onto the resin, the resin can be physically separated from the slurry
tion (CCD) of this material is challenging with the process residence by sieving. Once the resin is isolated, it is placed into columns
time being long requiring large CCD tanks, occupying a large plant where the metal value is eluted from the resin through contact
footprint and requiring high capital investment (Adams 2007; Tay- with an appropriate solution. By reducing losses of metal value
lor and Jansen, 2000). Depending on the settling characteristics of to tailings, overall metal recovery is improved and CCD throughput
the precipitate, 5% or more of the leached nickel and cobalt can can potentially be increased.
be lost to the slurry underflow through solution entrainment, Previous researchers have evaluated potential ion exchange res-
co-precipitation, and sorption processes on the high surface area ins for this process on the basis of resin capacity, resin selectivity and
solids. For a site producing 40,000 tonnes per annum nickel and loading/elution kinetics (McKevitt and Dreisinger, 2011; Mendes
2500 tonnes per annum cobalt this represents yearly losses of and Martins, 2004, 2005). Resin capacity and selectivity work in tan-
approximately $40 million USD, given current LME spot prices (as dem to dictate the amount of resin required to load a given amount
of August 2012). The use of ion exchange resin and resin-in-pulp of nickel and cobalt. In a scavenging process, only small amounts of
loading has been suggested as a method of scavenging this other- nickel and cobalt are present against a complex background of un-
wise lost nickel and cobalt value from CCD tailings (Adams, 2007; wanted impurities. As a result, the useful resin capacity is deter-
Duyvesteyn et al., 2002; Mendes, 2010; Wassink et al., 2006; mined not only by the number of sorption sites on the resin, but
Wyethe et al., 2002; Zainol, 2005; Zontov, 2001). Ion exchange resin also by the tendency of those sorption sites to load unwanted impu-
has been used to recover value metal in the uranium and gold rities instead of the desired nickel and cobalt (Littlejohn and Vaugh-
industry through resin-in-pulp/leach and similar carbon-in-leach/ an, 2012). Resin particle size distribution is important for both
pulp processes for decades (Fleming and Nicol, 1980; Prasad separation of resin from pulp as well as for kinetics – larger average
bead size with narrow particle size distribution makes resin sieving
easier but slows down resin kinetics by requiring longer intraparti-
⇑ Corresponding author. Tel.: +61 (07) 3365 8398; fax: +61 (07) 3365 4199.
cle mass transfer. A final requirement is for the resin to have a high
E-mail addresses: p.littlejohn@uq.edu.au (P. Littlejohn), james.vaughan@uq.
backbone strength to reduce resin breakage and attrition in the
edu.au (J. Vaughan).
1
Tel.: +61 (07) 3365 3850; fax: +61 (07) 3365 4199. aggressive resin-in-pulp environment (Halle and Rossoni, 2005).

0892-6875/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.mineng.2013.02.002
P. Littlejohn, J. Vaughan / Minerals Engineering 54 (2013) 14–20 15

Several researchers have investigated potential resins for this The goal of this research is to investigate the possibility of using
process and have concluded that a large bead resin functionalized BPA resin for resin-in-pulp scavenging, followed by ammoniacal
with iminodiacetic acid (IDA) (Adams, 2007; McKevitt and Drei- elution to recover nickel and cobalt in an eluate compatible with
singer, 2011; Zainol and Nicol, 2009) is the most appropriate for existing downstream nickel refining processes. Two primary ques-
scavenging nickel and cobalt from laterite tailings. Examples of tions drove this research:
commercially products using this functional group are Lanxess
MonoPlus TP 207 XL, Purolite S930, or Dow Amberlite IRC748. Each  Can bis-picolylamine functionalized resin effectively scavenge
of these resins has a high particle size and narrow particle size dis- nickel and cobalt from laterite tailings?
tribution. The backbone structure is macroporous which enhances  Can nickel and cobalt be selectively recovered through ammoni-
the ion exchange kinetics compared with gel type resin. Addition- acal elution?
ally, all have the high resin capacity characteristic of IDA function-
alized resin. IDA functionalized resin is available at relatively low 2. Experimental methods
cost, on the order of $10,000 USD per cubic meter.
IDA functionalized resin seems the natural choice for nickel lat- 2.1. Resin characteristics
erite resin-in-pulp. However, several characteristics of the resin
make integration of an IDA resin-in-pulp unit into existing nickel The resin used in this work is Lewatit MonoPlus TP 220 (hereaf-
laterite flow sheets difficult. First, while IDA resin is relatively ter referred to as TP 220) made by Lanxess ION in Germany. TP 220
selective for nickel and cobalt over impurities, in scavenging re- uses a bis-picolylamine functional group and a backbone of poly-
sin-in-pulp the concentration of impurities is orders of magnitude styrene crosslinked with divinylbenzene. The mean particle diam-
higher than that of base metal value. As a result, IDA resin from eter of the spherical shaped resin is 0.6 mm with 90% of the
scavenging circuits is typically only 30–50% loaded with nickel particles in the range of 0.55–0.65 mm. The capacity of the resin
and cobalt. Many functional groups are filled with undesirable depends on solution composition, but previous research has shown
impurities such as iron, aluminium, manganese, calcium and mag- the operational capacity of TP 220 to be 0.9–1.1 mol of equivalent
nesium (Littlejohn and Vaughan, 2012; Littlejohn et al., 2012; Was- charge per liter of tapped wet settled resin (Littlejohn and Vaugh-
sink et al., 2006; Zainol, 2005; Zainol and Nicol, 2009). Secondly, an, 2012) (note: all future resin volumes refer to tapped wet settled
IDA resin is most readily eluted using strong sulphuric acid. The resin). Resin was received protonated with sulphuric acid and was
quantitative nature of this elution recovers not only the nickel subsequently rinsed with de-ionized water prior to use. The selec-
and cobalt, but also the significant amounts of impurities co- tivity order of the resin is as follows (Lanxess, 2011; Littlejohn and
loaded onto the resin. This solution can be difficult to integrate into Vaughan, 2012):
the flow sheets of existing nickel refineries. Residual acid adds sig-
Cu  Ni > FeðIIIÞ > Co > Mn  K > Ca > Na > Mg > Al
nificant neutralization cost to the process of refining contained
nickel into mixed hydroxide precipitate, and high impurity content
makes the solution inappropriate for downstream integration close 2.2. Resin handling and synthetic resin loading
to final metal production. More sophisticated elution configura-
tions involving multiple columns or pachuca columns can help To explore resin elution chemistry, samples of TP 220 were
overcome these challenges, but add complexity and cost of the cir- loaded with nickel and cobalt from solutions containing these met-
cuit. Due to the incompatibility of acidic eluate with most down- als prepared in the laboratory. This loading was carried out by con-
stream nickel refining methods, scavenged value must be tacting a known volume of tapped wet settled resin in the
integrated back into existing flow sheets through upstream recycle protonated form with 10 times the amount of solution containing
(Littlejohn et al., 2012). 5 g/L Ni or Co as dissolved sulphate salts. The resin and solution
Downstream integration of scavenged metal value would im- were gently agitated at room temperature and the pH was adjusted
prove the commercial viability of nickel–cobalt resin-in-pulp. As through addition of 2 M NaOH until it remained stable at pH 4, typ-
several nickel refineries use ammonia based processes, one ically requiring 4–6 h to reach equilibrium. Once loaded, the resin
possibility is to take advantage of nickel and cobalt’s affinity for was placed in a column and washed with 20–30 bed volumes of
ammonia relative to impurity metals to elute resin into ammoni- de-ionized water to remove any solution entrained in resin pores.
acal solutions. Some success in this area has been achieved by This process yielded resin loaded with 30 g Ni or 19 g of Co per liter
eluting scavenge loaded IDA resin with solutions of ammonia- of metal form resin.
ammonium sulphate-magnesium sulphate (Littlejohn and Vaugh- Determination of loaded resin composition was carried out by
an, 2010; Littlejohn et al., 2012). While technically feasible, the quantitatively eluting 2–5 mL of resin. This small volume of resin
presence of magnesium in a ‘‘pure’’ solution is unwelcome at was wet filtered and contacted with 100 mL of 200 g/L H2SO4 in
best. in a baffled Erlenmeyer flask. The flask was set in a shaking incuba-
One alternative to IDA functionalized resin for scavenging resin- tor at 25 °C rotating at 150 rpm for 24 h. The efficacy of different
in-pulp is bis-picolylamine (BPA) resin. BPA resin was first devel- eluate compositions was tested in a similar fashion. To determine
oped to remove small amounts of copper and nickel from cobalt the kinetics of the elution reaction, samples of the solution were
electrolytes (Grinstead, 1984; Rosato et al., 1984). At first glance taken at regular intervals.
BPA resin seems unsuitable for nickel/cobalt resin-in-pulp as it
has a lower capacity than IDA resin (Littlejohn and Vaughan, 2.3. Tailings material – industrial and synthetic
2012). However, while the total capacity of BPA resin is roughly
half that of IDA, BPA is orders of magnitude more selective for base Resin-in-pulp loading experiments were carried out in batch
metals such as nickel, cobalt, copper and zinc over important later- reactors using both synthetically generated pulp and tailings from
ite impurities. As a result, the useful capacity of BPA resin under re- an industrial nickel refinery employing a high pressure acid leach-
sin-in-pulp scavenging conditions is similar to that of IDA resin. ing process. Synthetic pulp was prepared from a starting solution
The use of BPA resin to recover nickel from nickel heap leach liquor of 20 g/L Fe(III), 4 g/L Al, 2 g/L Mn, and 20 g/L Mg with 0.5–1 g/L
has been successfully trialled (Marston and Rodgers, 2009). Addi- Ni and 50–500 mg/L Co in a sulphate matrix. The initial pH of this
tionally, BPA resin has been proven to be amenable to ammonia solution was adjusted to 1.5 using sulphuric acid. This solution was
based elution (Grinstead, 1984). heated to 80 °C on a hotplate in a baffled batch reactor with vigor-
16 P. Littlejohn, J. Vaughan / Minerals Engineering 54 (2013) 14–20

ous stirring via overhead impeller. The pH was increased to precip- After every stage of both elution and loading, the resin was
itate iron through addition of CaCO3 as slurry (20% w/w). Iron pre- rinsed in the column with 20 bed volumes of de-ionized water to
cipitation was carried out over 5 h with a terminal pH of 3.9–4 and remove entrained solution. Particular care was necessary during
an extent of reaction of greater than 99.5%. The composition of the the second pass of metal bearing decant solution resin that had
resultant solution phase in the slurry (20% w/w solids) is shown in been eluted with ammonia. One bed volume of 1 g/L H2SO4 was
Table 1. Resin-in-pulp contact with this slurry was carried out passed through the column to insure that the metals in the decant
immediately after precipitation was complete and the slurry had solution would not immediately precipitate on contact with the
cooled to 60 °C. high pH resin.
The industrial plant samples used in this study came from three
points in the process. The first was slurry from after the acid leach
3. Results and discussion
but before final neutralization. The second was slurry from after fi-
nal neutralization before discharge to tailings pond. The third was
3.1. Eluent composition – synthetic batch flasks and resin columns
decant solution from the overflow of the tailings pond to the evap-
oration pond. Fig. 1 shows the source of each sample in the process
To determine the optimal composition of the eluent, solutions
schematic.
containing various amounts of ammonia and ammonium sul-
The composition of these materials is summarized in Table 2.
phate/carbonate were equilibrated with the resin. The results of
Slurries 1 and 2 are 38% and 41% w/w solids respectively, while
these batch experiments are presented in Table 3. It can be seen
the decant solution contained only 0.1% w/w solids.
that nickel is readily eluted by even dilute ammonia solutions, with
or without ammonium sulphate or carbonate. The extent of nickel
2.4. Resin-in-pulp scavenging tests elution is proportional to ammonia concentration. The presence of
ammonium sulphate in solution did not significantly affect nickel
The same experimental method was used to test resin-in-pulp elution but the presence of ammonium carbonated appears to sup-
scavenging on both the industrial and synthetic material. Once press the elution slightly.
the slurry was heated to the desired temperature, wet filtered TP To investigate the kinetics of ammonia elution, samples of resin
220 resin was added to the slurry. Resin was contacted with slurry were contacted with ammonia solution using the batch flask resin-
for 3 h under moderate mixing to ensure homogenous distribution solution contact methodology described in Section 2.2, with fre-
of the resin while avoiding resin breakage. No significant resin quent samples taken for the first 2 h and a final sample after
breakage was observed, but particle size distribution was not 24 h to show equilibrium behaviour. The kinetic curves generated
explicitly measured. The pH of the reactor was monitored inline are shown in Fig. 2. It is clear that the ammonia elution reaction is
and was controlled through addition of 20% w/w CaCO3 slurry. considerably faster than acidic elution of the same resin, though at
The slurry reached the pH set point of 4.0 within 20–30 min and equilibrium all three solutions recover greater than 95% of the
was stable for the duration of the 3 h resin loading experiment. nickel on the resin. Increasing the concentration of ammonia im-
On completion, final slurry and solution was sampled while resin proved reaction kinetics, but not drastically so.
was separated from slurry via sieves and washed thoroughly with Next, experiments were carried out to determine the elution
water. The resin was sub-sampled and assayed using the acid elu- behaviour of nickel and cobalt in the continuous flow column reac-
tion procedure described in Section 2.2, while the rest was set tor configuration. Two columns were run, one with pure nickel
aside for column elution or further reloading. loaded resin and the other cobalt. Samples of resin were eluted
with 6.8% w/w NH4OH solution at 25 °C. As the previous kinetic
2.5. Column procedures tests indicated a relatively rapid rate of resin elution, a relatively
high elution flow rate of 10 BV/h was chosen. The results of these
Column contacting was used for ammonia elution of loaded re- column tests are shown in Fig. 3. This figure shows that a high
sin. The column used was 15 mm in diameter. Depending on the recovery of nickel and cobalt from TP 220 resin can be achieved
bed volume size and elution flow rate, the linear velocity of flow in a small number of bed volumes at a relatively high elution flow
ranged from 0.8 to 1.6 m/h. In any case, eluent was pumped into rate. Previous researchers have tended to elute IDA resin in sulphu-
the top of the column using a peristaltic pump, passed over the re- ric acid at 2 BV/h (Nicol and Zainol, 2003; Wassink et al., 2006; Zai-
sin, and eluate was collected and sampled. As this was ammonia nol, 2005; Zainol and Nicol, 2009) whereas in this work 5 BV/h was
based elution, care was taken to keep the eluent reservoir, column, found to be adequate. Grinstead identified that nickel is amenable
and eluate reservoir air tight to prevent excess volatilization of to elution from bis-picolylamine resin with ammonia, but found
ammonia. Additionally, the column was water jacketed to allow that cobalt was not (Grinstead, 1984). These results demonstrate
temperature control when necessary. the contrary – that both metals can be recovered effectively with
Column contacting was also used to scavenge metal from the ammoniacal elution.
decant solution described in Table 2. This decant solution con-
tained roughly 0.1% solids on receiving, which were removed from 3.2. Resin-in-pulp loading
solution via vacuum filtration using quantitative cellulose filter pa-
per. As a result, this material was suitable for use in columns. In Resin-in-pulp scavenging of metal from both the synthetic and
this case, the tailings decant solution was passed over 50 mL of re- industrial samples described in Tables 1 and 2 was evaluated. The
sin at a rate of 5 bed volumes per hour at 25 °C. For this volume of results of these resin loading experiments are shown in Tables 4
resin and flow rate, the linear velocity of flow was 1.4 m/h. Solution and 5. In every case, recovery of nickel was greater than 95% and
samples of permeate were collected periodically. recovery of cobalt greater than 80%. While not shown in these ta-
Table 1 bles, resin-in-pulp contact at 25 °C yielded similarly high recovery.
Composition of solution in synthetic pulp, sulphate matrix. However, the higher temperature changed the slurry rheology and
facilitated resin-slurry mixing and subsequent sieving.
mg/L Ni Co Fe Al Mg Mn Ca
In each test, resin-slurry contact time was 3 h. While the kinet-
Synth L 480 63 68 28 13,000 1459 612
ics of resin loading were not rigorously studied, solution samples
Synth H 1212 151 120 33 16,461 2044 725
were taken at the 0.5 and 1.5 h mark. In each test the solution
P. Littlejohn, J. Vaughan / Minerals Engineering 54 (2013) 14–20 17

Fig. 1. Schematic flow sheet of material sampling points.

Table 2
Composition of solution in industrial pulp and decant solution.

Samples Solution Concentration (mg/L) Solids (mg/kg) pH


Ni Co Fe Al Cr Mg Mn Ca Cu Zn Ni Co
Slurry 1 489 46 3043 944 102 21409 1505 747 17 1.3 641 1.3 1.9
Slurry 2 478 45 2136 67 4 19069 1349 731 0.2 2.6 695 0.3 4.5
Decant 549 60 4928 799 44 31797 2251 991 2 7.9 – – 2.4

Table 3
Equilibrium recovery of nickel from synthetic loaded resin with various solutions, 5 mL resin, 100 mL solution, 25 °C.

Test NH4OH (% w/w) (NH4)2SO4 (g/L) (NH4)2CO3 (g/L) [Ni] (mg/L) % Ni recovery
1 2.80 – – 890 75
2 5.60 – – 1027 84
3 8.43 – – 1132 92
4 11.24 – – 1187 99
5 5.60 289 – 1089 91
6 5.60 198 – 1091 89
7 5.60 99 – 1097 90
8 5.60 – 218 1073 85
9 5.60 – 152 1083 86
10 5.60 – 76 1128 91

Fig. 2. Kinetics of eluting synthetic nickel loaded TP 220 in batch flasks, ammonia Fig. 3. Column recovery of synthetic loaded pure nickel and pure cobalt resin.
vs. sulphuric acid, 25 °C.
As a result, ferric is exchanged for nickel as the resin undergoes
successive loading stages. This was demonstrated with tests 4a/
concentration of nickel at 30 min was 20–40 ppm, indicating that 4b. The resin in test 4a was loaded with 10.3 g/L Ni and 2.94 g/L
the majority of nickel was recovered in the first 30 min of contact. Fe. A portion of this resin was assayed and the remainder was con-
Cobalt followed a similar profile. tacted with another batch of tailings slurry (4b). After the second
Evidence of the high selectivity of TP 220 for nickel and cobalt contact the loaded resin contained 15.2 g/L Ni and 1.3 g/L Fe. These
can be seen in Table 5, which shows that the useful capacity of results indicate that ferric iron can be displaced by nickel in resin-
the resin (defined as mass of nickel and cobalt per liter resin over in-pulp contact. Marston and Rodgers observed similar behaviour
mass of all loaded metals) ranges from 75% to 90%. The most signif- during recovery of nickel from laterite heap leach liquor (Marston
icant impurity loaded by TP 220 was ferric iron. However, this and Rodgers, 2009). Despite nickel being ahead of iron in the BPA
loading is not irreversible – it can be seen from the resin selectivity selectivity order a certain amount of iron will co-load with nickel
order presented in Section 2.1 that nickel is higher than ferric iron. on the resin.
18 P. Littlejohn, J. Vaughan / Minerals Engineering 54 (2013) 14–20

Table 4
Resin-in-pulp scavenging of Ni and Co from synthetic and industrial materials, pH 4, 60 °C, 3 h contact time.

# Starting slurry Vol % resin (v/v) Initial Sol’n (ppm) Final Sol’n (ppm) Overall recovery (%)
Ni Co Ni Co Ni Co
1 Slurry 1 2.1 489 46 8 6 98.4 87.0
2 Slurry 1 2.9 489 46 18 9 96.3 80.4
3 Slurry 2 2.0 478 45 9 9 98.1 80.0
4a Slurry 2 3.0 478 45 13 10 97.1 77.8
4b Slurry 2 2.8 478 45 12 8 97.5 82.2
5 Synthetic 220L 4.4 480 63 7 1.8 98.5 97.1
6 Synthetic 220H 5.3 1212 151 14 6.6 98.8 95.6

Table 5
Composition of loaded resin from resin-in-pulp tests in Table 4.

Resin # Ni (g/L) Co (g/L) Fe (g/L) Al (g/L) Cr (g/L) Mg (g/L) Mn (g/L) Ca (g/L) Cu (g/L) Zn (g/L)
1 12.35 0.70 2.78 0.10 0.06 0.21 0.02 0.13 0.01 0.29
2 12.49 0.54 3.60 0.18 0.15 0.09 0.01 0.04 0.01 0.26
3 12.17 0.91 1.88 0.02 0.01 0.07 0.01 0.06 0.00 0.65
4a 10.25 0.71 2.94 0.10 0.04 0.46 0.07 0.03 0.00 0.59
4b 15.24 0.58 1.27 0.05 0.01 0.14 0.03 0.12 0.00 0.64
5 2.91 0.51 0.08 0.08 – 0.07 0.07 0.05 – –
6 6.67 0.97 0.10 0.12 – 0.51 0.45 0.09 – –

After being loaded through resin-in-pulp contact, resin was


eluted from the resin with 6.8% w/w aqueous ammonia. Fig. 4
shows the elution response of the combined resin from tests 1
and 2. Only Ni, Co, Cu and Zn are shown in the figure, as the elution
of other metals was minimal. Peak concentrations of Al, Fe, Cr, Mg,
Mn, and Ca in the eluate were 1 ppm or less. After 4 bed volumes of
eluent, recovery of nickel and cobalt to the solution phase was 94%
and 83% respectively. This level of selectivity and nickel and cobalt
recovery was achieved in all columns of resin-in-pulp loaded resin.

3.3. Column loading from tailings decant solution

Recovery of nickel and cobalt value from tailings pond decant


solution was investigated. The industrial sample used was taken Fig. 5. Column loading of TP 220 with tailings decant solution, first contact.
from tailings decant en route to crystallization ponds, and as such
contained virtually no solids. As a result, nickel and cobalt could be
recovered onto resin from this solution using simpler column ion
exchange rather than resin-in-pulp. The column contacting took
the following form: TP 220 resin in the proton form was first con-
tacted with decant solution, then resin was eluted using aqueous
ammonia, followed by a second contact with decant solution.
Figs. 5 and 6 show breakthrough curves for Ni, Co, and Fe for each
contact. Breakthrough curves for Al, Ca, Cr, Mg, Mn and Na are ne-
glected, as these elements showed no affinity for the resin in these
tests. The small amount of Zn and Cu in solution followed the same
curve as that of Ni.

Fig. 6. Column loading of TP 220 with tailings decant solution, second contact.

The results indicate that nickel and cobalt load preferentially


onto the resin despite the solution being at pH 2.4. Subsequent
ammonia elution of the column loaded resin showed the same high
selectivity and recovery as with the resin loaded from pulp. It can
be seen that the resin is less selective for cobalt than nickel, indi-
cated by the earlier breakthrough curve of cobalt. Towards the
end of each column experiment, cobalt concentration in the per-
meate was greater than that of the head solution, indicating that
Fig. 4. Ammonia elution response of resin-in-pulp loaded TP 220. cobalt loads early but is displaced by nickel.
P. Littlejohn, J. Vaughan / Minerals Engineering 54 (2013) 14–20 19

Table 6
Composition of resin after column loading from decant solution and elution with 6.8% w/w aqueous ammonia.

Resin (g/L) Al Ca Co Cr Cu Fe Mg Mn Na Ni Zn
1st Load 0.11 0.23 0.41 0.01 0.03 4.26 0.29 0.01 0.41 13.78 1.44
1st Elution 0.05 0.20 0.09 0.01 0.00 4.20 0.18 0.01 0.20 1.36 0.08
2nd Load 0.02 0.30 0.40 0.03 0.01 5.80 0.09 0.01 0.04 12.41 1.60
2nd Elution 0.01 0.19 0.16 0.02 0.00 5.70 0.06 0.01 0.03 1.53 0.16

Of note is the behaviour of iron. The total concentration of iron The technical feasibility of scavenging nickel and cobalt from
in the head solution was 5.8 g/L; however, most of this was present laterite leach tailings was demonstrated on both synthetic material
as ferrous iron, indicated by the relatively low potential recorded and samples from an industrial nickel refinery. BPA functionalized
between a platinum electrode and a Ag/AgCl reference electrode TP 220 resin was found to be effective at recovering nickel and co-
(373 mV Pt vs. Ag/AgCl). Ferrous iron does not bond strongly with balt both in batch resin-in-pulp tests on industrial neutralized tails
BPA resin, but ferric is second only to nickel and copper in the slurry as well as in column contact with clarified tailings decant
selectivity order. As a result, it can be seen from iron levels in solution. The resin’s high selectivity for nickel and cobalt over
the permeate in Figs. 5 and 6 that iron loads onto the resin. How- impurities lead to useful resin capacity of 75–90%, considerably
ever, by the end of the second contact with resin the amount of higher than that of iminodiacetic acid functionalized resin. Ferric
iron in the permeate is greater than that of the head, indicating iron was the most significant impurity to be loaded, but does not
that iron is being displaced off the resin by nickel. This result sug- overwhelm the resin. Instead, iron on the resin is gradually dis-
gests that the ratio of nickel to iron on the final loaded resin is placed by nickel due to its rank in the resin selectivity order. Re-
dependent on the number of bed volumes passed and the flow rate sin-in-pulp loading kinetics were rapid, with over 95% of the
of solution, with higher bed volumes passed leading to higher nick- metal ultimately loaded being taken up by the resin in the first
el to iron ratio on the loaded resin. 30 min of contact. The scavenging of nickel and cobalt from tailings
After each contact with decant solution, resin was eluted with decant offers a potential method to recover value metals that
6.8% w/w aqueous ammonia in the same fashion as the resin-in- would otherwise be lost to waste while obviating the need for
pulp loaded resin. The results of elution were similar to that of more complicated resin-in-pulp contact.
the resin-in-pulp loaded resin. Nickel and cobalt recovery in 2 Selective elution of nickel and cobalt from BPA functionalized
bed volumes of eluate was 90% and 75% respectively, with ppm resin was effective. 90–95% of nickel and 75–85% of cobalt was
levels of impurities. A summary of resin compositions at each stage eluted using 2–4 bed volumes of 6.8% w/w NH4OH solution. Elution
of the sequential column loading is presented in Table 6. kinetics were rapid and the elution response curve was sharp, with
the majority of metal being recovered in the first 0.5–1 bed volume
3.4. Resin fouling of eluate. Impurities such as iron, aluminium, chromium, manga-
nese, magnesium and sodium demonstrate low affinity for ammo-
As the ammoniacal elution method described above necessarily nia solutions, and are limited to 5 ppm or less in the eluate. Due to
involves contact of a high pH solution with resin loaded with met- their affinity for both BPA resin and ammonia solutions, zinc and
als that are not soluble in ammonia, potential exists for resin foul- copper follow nickel and cobalt throughout the process. Further re-
ing via hydroxide precipitation. Aluminium and chromium are search investigating the possibility of resin fouling via hydroxide
unlikely to cause issues as they have a very low affinity for the re- precipitation during ammoniacal elution is warranted.
sin during loading and thus only minimal amounts are present dur- Resin-in-pulp scavenging of nickel and cobalt from laterite tail-
ing elution. Iron may be more problematic as BPA resin displays ings has not been commercialized to date, in part due to the diffi-
affinity for ferric iron and thus significant amounts of iron may culty of integrating acidic eluate from IDA resin into existing nickel
be present on the resin during ammoniacal elution. This issue refining flow sheets. The use of BPA resin and ammonia elution of
could be managed in several ways. If resin-in-pulp contact with nickel and cobalt offers new possibilities for the integration of re-
tailings slurry is used, dissolved ferric iron in the initial slurry sin-in-pulp into existing refinery flow sheets. As the eluate con-
can be minimized through neutralization of the load slurry. Addi- tains high concentrations of nickel and cobalt and trace amounts
tionally, ferric iron loading on the resin can be minimized through of impurities, downstream integration of value metal recovered
contacting loaded resin with fresh slurry, as shown in tests 4a/4b from tailings may be possible.
in Section 3.2. In practice this would require counter-current RIP
operation. A second method of managing ferric iron would be to re-
Acknowledgements
duce it to ferrous iron using a suitable reductant, such as sodium
sulphite. As BPA resin displays very low affinity for ferrous iron
The authors would like to acknowledge Lanxess ION for provid-
compared to ferric, once reduced iron can be rinsed off the resin
ing resin samples. The help of Chen Han during experimentation is
(Marston and Rodgers, 2009). Finally, a periodic wash with sulphu-
appreciated. The support of UniQuest is also recognized.
ric acid could be employed to dissolve hydroxide precipitates
should they form.
References

4. Conclusions Adams, M.A., 2007. Towards a virtual metallurgical plant – CCD vs RIP case study.
In: Proceedings of ALTA Ni/Co/Cu 2007, Perth, Australia.
Duyvesteyn, W., Neudorf, D., et al., 2002. Resin-in-pulp method for recovery of
This program of experimental work aimed to answer the fol-
nickel and cobalt. US Patent No. 6350420. Washington, DC: U.S.
lowing questions: Fleming, C.A., Nicol, M., 1980. A comparative study of kinetic models for the
extraction of uranium by strong-base anion-exchange resins. J. S. Afr. Inst. Min.
 Can BPA resin scavenge nickel and cobalt from laterite tailings? Metall. 80 (2), 89–99.
Grinstead, R., 1984. Selective absorption of copper, nickel, cobalt and other
 Can nickel and cobalt be selectively recovered through ammoni- transition metal ions from sulphuric acid solutions with the chelating ion
acal elution? exchange resin XFS 4195. Hydrometallurgy 12, 387–400.
20 P. Littlejohn, J. Vaughan / Minerals Engineering 54 (2013) 14–20

Halle, O., Rossoni, D., 2005. A new (monodisperse) chelating resin designed for Mendes, F., 2010. Resin-in-leach process to recover nickel and/or cobalt in ore
mining application: Lewatit Monoplus TP 207. ALTA Ni/Co/Cu 2005, Perth, leaching pulps. US Patent No. 2010/0196190. Washington, DC: U.S.
Australia. Nicol, M., Zainol, Z., 2003. The development of a resin-in-pulp process for the
Lanxess, 2011. Lewatit TP 220 Technical Sheet, viewed September 11th, 2012 http:// recovery of nickel and cobalt from laterite leach slurries. Int. J. Min. Procss. 72,
www.lewatit.com/common/ 407–415.
download.php?file=e8a10ef8026615d3c32a955b1612e16f. Prasad, M.S., Mensah-Biney, R., et al., 1991. Modern trends in gold processing –
Littlejohn, P., Vaughan, J., 2010. Selective ammoniacal elution of nickel and cobalt overview. Min. Eng. 4 (12), 1257–1277.
from iminodiacetate cation exchange resin. ALTA Ni/Co/Cu 2010, Perth, Rosato, L., Harris, G.B., et al., 1984. Separation of nickel from cobalt in sulphate
Australia. medium by ion exchange (XFS 4195 resin). Hydrometallurgy 13, 33–44.
Littlejohn, P., Vaughan, J., 2012. Selectivity of commercial and novel mixed Taylor, Jansen, 2000. Future Trends in PAL Plant Design for Ni/Co Laterites. ALTA Ni/
functionality cation exchange resins in mildly acidic sulfate and mixed Co/Cu 2000. Perth, Australia.
sulfate–chloride solution. Hydrometallurgy 121–124, 90–99. Wassink, B., Neufeld, M., et al., 2006. Towards a resin-in-pulp process for recovery
Littlejohn, P., Vaughan, J., et al., 2012. Elution strategies for recovery of nickel and of nickel and cobalt from laterite leach tailings: an iminodiacetic acid ion
cobalt from laterite tails through scavenging resin-in-pulp. ALTA Ni/Co/Cu exchange resin as a prospective resin. International Water Conference 2006.
2012, Perth Australia. Wyethe, J.P., Kotze, M.H., et al., 2002. Cobalt, Nickel and Copper Recovery with
Marston, C., Rodgers, M., 2009. Nickel recovery from lateritic ore in the presence of Resin-in-Pulp. ALTA Ni/Co/Cu 2002, Perth, Australia.
ferric iron using bispicolylamine chelating resin. Conference of Metallurgists. J. Zainol, Z., 2005. The development of a resin-in-pulp process for the recovery of
Budac. Sudbury, Canada, CIM. nickel and cobalt from laterite leach slurries. Perth, Murdoch University. Ph.D.
McKevitt, B., Dreisinger, D., 2011. A comparison of large bead ion exchange resins Zainol, Z., Nicol, M., 2009. Comparative study of chelating ion exchange resins for
for the recovery of base metals in a resin-in-pulp (RIP) circuit. In: Proceedings of the recovery of nickel and cobalt from laterite leach tailings. Hydrometallurgy
the South African Institute of Mining and Metallurgy Base Metals Conference. 96 (4), 283–287.
Mendes, F., Martins, A., 2004. Selective sorption of nickel and cobalt from sulphate Zontov, N., 2001. Potential Benefits of Resin-in-Pulp for PAL Plants. ALTA Ni/Co/Cu
solutions using chelating resins. Int. J. Min. Proc. 75, 359–371. 2001, Perth, Australia.
Mendes, F., Martins, A., 2005. Recovery of nickel and cobalt from acid leach pulp by
ion exchange using chelating resin. Min. Eng. 18, 945–954.

You might also like