Int. J. Appl. Ceram. Technol.

, 13 [2] 274–279 (2016)

DOI:10.1111/ijac.12515

The Separation of Indium and Copper from Spent Cu-In Targets

Wei-sheng Chen*
Department of Resources Engineering, College of Engineering, National Cheng Kung University, No.1, University Rd.,
Tainan, Taiwan 70101, Republic of China

Shu-mei Huang
Sustainable Environment Research Laboratory, National Cheng Kung University, No.1, University Rd., Tainan, Taiwan
70101, Republic of China

Nowadays, the CIGS solar cells produced process, only 30% of the original target material is used, while the rest 70% need
to be recycled in order to achieve the resources sustainable. Common methods to separate and recover indium include precipita-
tion, solvent extraction, ion-exchange, and electrolytic processes. This research is to investigate the recovery and purification of
indium from spent sputtering targets so that the recovered and purified indium can be reused in the thin film and semiconductor
industries. The processes selected include ion-exchange resin, solvent extraction, and zone refining. Eventually, indium recovery
using ion exchange and solvent extraction has high recoveries. Further purification was conducted by electrowinning and zone
refining process. The indium product of zone refining can be used as raw material of indium sputtering target.

Introduction Methods to separate and recover indium from the

spent targets include precipitation,3 solvent extraction,4
Indium is a rare metal, with its content in the crust ion-exchange, and electrolytic5 processes.
about 10
4%. Indium is widely distributed in sphalerite The common cation-exchange resins include cal-
at a concentration of 0.1% or less. The indium reserve ixarene-based resin with its ion-exchange capacity of
of the world amounts to only 16,000 tons.1 In recent 109 mg/g,6 and chelating resin of IDA,7 Amberlite XAD-
years, indium has been extensively used in both solar 4-HMPN,8 D4189 with their ion-exchange capacity in
and electronic industries. The thin film solar cell is con- the range of 1–14 mg/g. However, none of them is highly
sidered as having a higher potential in the solar energy selective for copper and indium. Although Levextrel resins
technology, given its advantages of lightweight, thinness, are more selective, but the preparation is quite compli-
low cost, flexibility, and diversified design. The thin film cated. Lee10 used D2EHPA to separate indium from gal-
types include copper–indium–gallium–selenium (CIGS) lium in sulfuric acid solution. It was showed that the
solar cell and cadmium telluride solar cell. The CIGS is distribution coefficient of indium and gallium increased
a direct band gap material, which has higher absorptivity with pH and D2EHPA concentration. The indium and
than amorphous silicon. Copper–indium alloy sputtering gallium recoveries were 99.9% and 26%, respectively, in
targets are applied to CIGS solar cell process as a light a single-stage extraction, and they could be separated
energy conversion absorption layer. The development completely by a second-stage extraction. This research is
and application of indium is likely to produce a signifi- to investigate the recovery and purification of indium
cant amount of process waste as the demand increases in from spent sputtering targets so that the recovered and
the future. Therefore, the recovery of indium metal has purified indium can be reused in the thin film and semi-
attracted more attention.2 conductor industries. The processes selected include ion-
Nowadays, the CIGS solar cells are produced mostly exchange resin, solvent extraction, and zone refining.
through vacuum evaporation or evaporation deposition. This research is to investigate the recovery and
In the process, only 30% of the original target material purification of indium from spent sputtering targets so
is used, while the rest 70% need to be recycled in order that the recovered and purified indium can be reused in
to achieve the resources sustainable. the thin film and semiconductor industries. The pro-
cesses selected include ion-exchange resin, solvent extrac-
*[email protected]
© 2016 The American Ceramic Society
tion, and zone refining.
 17447402, 2016, 2, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/ijac.12515 by Zhejiang University, Wiley Online Library on [04/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ceramics.org/ACT The Separation of Indium and Copper 275

Experimental experiments were conducted and analyzed in triplicate

samples.
Reagents used in the study are PA418 anion- The resulting solution from the solvent extraction
exchange resin, InCl3 (98%,Alfa Aesar), CuCl2.2H2O study contained about 700 ppm indium. The solution
(97%,Dau-jiou Drugs), H2C2O4.2H2O (Sigma-Aldrich), was electrolyzed at 0.8 V to produce cathode indium of
D2EHPA (di-2-ethyl-hexylphos-phoric acid) (95%), and 99% in purity. Both the anode and cathode were made
HCl (37%,Merck). of stainless steel. The cathode indium serves as the feed
Instruments used include Analyst 400 Atomic for the subsequent zone refining test. In the zone refining
Absorption Spectrophotometer (AAS) and Phenom SEM. tests, the sample of indium with 99% in purity was
Feed solution of indium and copper was prepared placed in a ceramic boat and heated until it melted. The
by dissolving indium chloride and copper chloride in loaded boat was then transferred to a quartz tube within
hydrochloric acid solution. Each of the indium and cop- a zone refining furnace, in which the pressure was set to
per concentration in the feed solution was approximately 10
1 to 10
2 torr. Argon gas was introduced to protect
300 ppm. Analytical grade reagents and distilled water the indium sample from oxidation. Induction heating
were used throughout the experiments. The tests of ion was applied to one end of the sample, and the speed of
exchange were carried out with 1.00 g of resin placed in the molten zone adjusted to 2–5 cm/h.
an Erlenmeyer flask with 100 mL of the feed solution.
Mixing of the resin and solution was provided for 24 h
at room temperature by a magnetic stirring bar. After Results and Discussions
the reaction, concentrations of metal ions in the solution
were analyzed by AAS. The ion-exchange rate was (R) Indium ions form negatively charged indium chlo-
calculated by the equation below (Eq. 1): ride complexes in hydrochloric acid solution and then
exchange with the chloride ions provided by PA418.11,12
ðCo
CeÞ After the addition of 5% of oxalic acid, oxalate will form
R¼  100% ð1Þ
Co complexes with indium ions are negatively charged and
The metal exchanged by the resin was eluted with then use the PA418 chlorine ion exchange. In such an
sodium hydroxide solution, sodium chloride solution, or environment, the copper ions do not react with PA418,
hydrochloric acid solution. The efficiency of elution is which can be effectively separated from the copper and
calculated by the equation below (Eq. 2): indium. Prediction of reaction is as follows:

Cd  Vd InðC2 O4 Þð3
2nÞ $ ð2n
3ÞCl
ðn = 2Þ ð3Þ

E¼  100% ð2Þ n
ðCo
CeÞ
The exchange isotherm was established to determine
where Co, Ce, Cd, Vd represent for the initial concen- the nature of the interactions between indium complex
tration, concentration of the equilibrium, concentration and the surface of the resin. The experiments were car-
of elution agent, and the volume of the elution agent, ried out in flasks containing 100 mL of solution with
respectively. initial indium concentrations ranging from 30 to
Elution experiment focuses on the results of differ- 150 ppm. Dried resin (1.00 g) was added in each flask
ent types and different concentrations of desorbent and then shaken at room temperature for 24 h. After
indium. separation of the resin by filtration, the filtrate was ana-
The solution containing copper and indium was lyzed using AAS. From the initial and equilibrium con-
extracted and then stripped. The extraction experimen- centration of indium in aqueous solution, the amount of
tal parameters include initial solution pH value, O/A indium uptaken by the resin, q (mg /g dry resin), was
ratio (organic phase/aqueous phase volume ratio), oscil- calculated according to the relation (Eq. 4). The Lang-
lating frequency, and D2EHPA concentration. Aqueous muir isothermal model (Eq. 5) was adopted to describe
and organic phases were mixed in a separating funnel the experimental results, where a and K are the capacity
oscillator oscillating at room temperature for 30 min. of the single layer adsorption capacity (mg/g) and con-
After separation of the phases, concentrations of metal stant, respectively.
ions in the solution were analyzed by AAS. The extrac-
tion was then calculated to determine the extraction ðCo
CeÞ
q¼ V ð4Þ
effect. To evaluate the reproducibility of the results, all m
 17447402, 2016, 2, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/ijac.12515 by Zhejiang University, Wiley Online Library on [04/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
276 International Journal of Applied Ceramic Technology—Chen and Huang Vol. 13, No. 2, 2016

Ce 1 Ce
¼ þ ð5Þ
q aK a
As shown in Fig. 1, the plot of Ce v.s. Ce/q is of
high liner. The capacity of PA418, a, calculated from
Eq. (5) is 55 mg/g, which is very high in comparison
with other adsorbents (Table I).
In the kinetic tests, various amounts of the dried
resin and 2.000 L of solution containing various initial
Fig. 2. The results of kinetic ion exchange experiments.
indium concentrations, as shown in Fig. 2, were applied.
The resin was preconditioned with ultrapure water at
room temperature and 150 rpm for 24 h. Solution sam-
ples (10 mL) were taken at various time periods (0, 5,
10, 20, 40, 60, 120, 240, 480, 720, 1440, 2160, 2880,
4320 min), and the results are shown in Fig. 2. The
experimental results show that after 10 h of reaction, the
concentration of the indium in the aqueous does not
vary significantly, which implies that the reaction reach
equilibrium after 10 h. Also, the time needed to reach
equilibrium is irrelevant to the initial indium concentra- Fig. 3. Loading curve of indium on resin.
tion and resin amount. In the result, it can be considered
the reaction of the ion-exchange resin with aqueous con-
the experimental results are shown in Fig. 3: The Fig. 3
taining indium will reach equilibrium after 10 h.
shows that after 10 h of experiment runs through the
The 60 mL, weight 42.65 g PA418 filled column,
point, completely through about 25 h later, adsorption
the column inner diameter of 1.5 cm, length 50 cm.
capacity is 56.55 mg/g, this result is consistent with the
Solution flows into the column at a flow rate of 20 mL/
Langmuir model results.
min (SV = 20); the indium concentration is 143 ppm;
Elution experiments are conducted with different
types and concentration of elution agents. About 1.00 g
of resin was placed in Erlenmeyer flasks with 50 mL of
elution agent. They were mixed for 24 h at room tem-
perature, and the results are shown in Table II. The
results show that different elution agents result different
elution rates. The optimum elution agent and concentra-
tion were found to be 1M hydrochloric acid with more
than 99% of elution efficiency.
Tests of solvent extraction were conducted with vari-
ous initial pH of solution under the condition of fixed
Fig. 1. Langmuir isothermal curve. O/A = 1/1, 0.05M D2EHPA, oscillating frequency of

Table I. Comparison of Indium Ion Exchange by Different Resin

Adsorbent Solution pH Saturated ion exchange
P507 levextrel resin11 HCl 1.0–1.5 2.33 mg/g
P204 levextrel resin10 H2SO4 2.5 16.7 mg/g
IDA chelating resins7 The solution containing Br
2.5 1.23 m mol/g
The solution containing ClO4
1.64 m mol/g
Chelating resins of D4189 HCl – 14.35 mg/g
Humic acid17 Simulated wastewater 4–5.5 3 mg/g
Calixarene-based cation exchange resins6 HNO3 3 109 mg/g
Amberlite XAD-4-HMPN8 – 4.5–8 1.27 mg/g
 17447402, 2016, 2, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/ijac.12515 by Zhejiang University, Wiley Online Library on [04/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ceramics.org/ACT The Separation of Indium and Copper 277

Table II. Results of the Elution Experiments Fig. 5. The results show that the extraction efficiency
changes slightly when the oscillating frequency was 200
Indium
to 300 per min. When the oscillating frequency up to
concentration
300 per min, the extraction percentage increased with
in the solution
the oscillating frequency. There were better mixing and
Elution Elution after elution Elution
extraction between the two phases, so that the extraction
solution concentration (ppm) rate (%)
percentage of indium increased. When the oscillating fre-
NaOH 8% 1.56 0.51 quency increased to 500 per min, the extraction percent-
12% 2.34 0.77 age of indium was > 99%. The extraction is completed
NaCl 0.5M 166.63 34.61 when oscillating frequency is over 500 per min.
1M 357.4 72.23 Under the conditions of fixed solution pH = 2.5,
HCl 0.5M 310.45 64.48 O/A = 1/1, 0.05M D2EHPA and oscillating frequency
1M 352.44 >99 of 500 per min, the oscillating time was adjusted to 5,
15, 30, and 60 min, respectively. The extraction rate of
different oscillating time is shown in Fig. 6. The results
500 per min and oscillating time of 30 min. Figure 4 show that the indium extraction was > 99% at 5 min.
shows the effect of aqueous phase pH on extraction. The Under the conditions of solution pH 2.5, O/A = 1/
optimal pH was observed. The results show that the 1, oscillating frequency of 500 per min and oscillating
extraction of indium increases with an increase in aque- time of 5 min, the concentration of D2EHPA were set
ous solution pH. The D2EHPA is an acidic extractant. to 0.01, 0.05, 0.1, 0.2, 0.5, or 2M. The effect of
In the solvent extraction of metal ions, the extraction D2EHPA concentration on the extraction efficiency was
reaction can be expressed as follows (13): also observed. As shown in Fig. 6, the extraction of
indium increased with an increase of D2EHPA concen-
3 þ
aq þ ðHAÞ2;O ¼ MA 3;O þ 3Haq
M3þ ð6Þ tration up to 0.1M. The extraction was more than 99%
2 when the D2EHPA concentration was 0.1M. The extrac-
HA: Extractant D2EHPA; M3+: Trivalent metal ion; tion percentage of copper did not increase significantly
aq: aqueous phase; O: organic phase. when the concentration was <0.2M. It increases to 40%
Therefore, higher acidity of aqueous phase is unfa- of extraction of copper when the D2EHPA concentra-
vorable to the indium extraction by D2EHPA. Figure 4 tion is higher than 1M.
shows that the copper extracted to organic phase is low From the extraction experiments, it was observed
at all pH. Because the D2EHPA exchanges hydrogen ion that D2EHPA could extract indium selectively from the
with trivalent indium ion selectively, the copper ions solution containing both copper and indium. The
mostly exist in monovalent and bivalent forms. Hence, indium in organic phase was then stripped to aqueous
the copper and indium can be separated by the extrac- phase by different stripping reagents. Six stripping solu-
tion process. tions were prepared. The results of stripping studies are
Under the conditions of pH value = 2.5, O/A = 1/ shown in Table III. The results indicate that 2M
1, 0.05M D2EHPA, oscillating time of 30 min, various hydrochloric acid solution can be used to efficiently strip
oscillating frequencies were set to 200, 300, 400, 500, the indium from organic phase, so as to facilitate subse-
and 550 per min. The effect on extraction is shown in quent refining processes.

Fig. 4. Effect of solution pH on extraction efficiency. (0.05M Fig. 5. Effect of Oscillating frequency on extraction efficiency.
D2EHPA;O/A:1/1; Oscillating frequency: 500 per min; Oscillat- (Aqueous phase: pH = 2.5; 0.05 M D2EHPA; O/A = 1/1; Oscil-
ing time: 30 min). lation time: 30 min).
 17447402, 2016, 2, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/ijac.12515 by Zhejiang University, Wiley Online Library on [04/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
278 International Journal of Applied Ceramic Technology—Chen and Huang Vol. 13, No. 2, 2016

Fig. 6. Effect of D2EHPA concentration on extraction efficiency.

(Aqueous phase: pH = 2.5; O/A = 1/1; Oscillating frequency:
500 per min; Oscillating time: 30 min). (a)The surface of the
zone refining product.(1:1000). (b) The bottom of the zone
refining product.(1:1000).

Table III. Stripping Efficiency of Various Stripping (1)The surface of the zone refining product.(1:1000)
Reagents
Stripping reagent Stripping rate (%)
0.1M HCl 0
0.5M HCl 57.7
2M HCl >99.9
0.1M HNO3 12
0.5M HNO3 97
1M HNO3 99.7

The subsequent electrolysis of the strip liquor

resulted in an indium product of 99% in purity. This
indium product serves as the feed material for the zone
refining tests. The zone refining results show that the
amount of oxygen at the top is much higher than that at (2)The bottom of the zone refining product.(1:1000)
the bottom (Fig. 7). It can be considered there may be
Fig. 7. Morphology of the samples taken at various locations.
oxidation at the top from the color of the indium oxide.
Besides the material itself, oxygen also comes from the
Table IV. AAS Analysis of Impurities in Various
equipment. Furthermore, there are some impurities
Locations of the Indium Product
found floating on the surface as indicated by arrows
shown in Fig. 7(1). Impurities (ppm) A B C D
By high-frequency induction heating that a small
section of solid melted into liquid, at the solid/liquid Al 0.56 – – –
boundary, the impurity atoms will diffuse to the liquid Si 1.50 – – –
phase. Thus, by passing a crystal boule through a thin Ca 1.85 – – –
section of furnace very slowly, such that only a small Fe – – 2.71 0.16
region of the boule is molten at any time, the impurities Ni – – 0.14 –
will be segregated at the end of the crystal. Because of Sn – 0.78 3.48 1.38
the lack of impurities in the leftover regions which solid- Pb – 0.98 0.13 –
ify, the boule can grow as a perfect single crystal if a seed
crystal is placed at the base to initiate a chosen direction Conclusions
of crystal growth. Table IV shows that after melting
through zone refining, the purity of the beginning was This research is to separate and recover indium and
significantly higher than the end. The indium product of copper from spent Cu-In targets. The experiments of
purity in 99.995% can be obtained by zone refining. separation and enrichment of indium were conducted
 17447402, 2016, 2, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/ijac.12515 by Zhejiang University, Wiley Online Library on [04/11/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ceramics.org/ACT The Separation of Indium and Copper 279

using ion-exchange resin and solvent extraction processes techniques in resource recycling of CIGS thin solar cell
in a hydrochloride acid system. The ion-exchange experi- module wastes” (MOST 103-2218-E-006-025).
mental results show that the characteristics of indium
ion exchange fit the Langmuir isotherm model. The ion-
exchange capacity and equilibrium time are 55 mg/g and References
10 h, respectively. Hydrochloric acid solution of 1M was
found to be an effective desorbing agent. Indium is suc- 1. Available at: http://minerals.usgs.gov/minerals/pubs/commodity/indium/ (ac-
cessed January 24, 2015).
cessfully recovered from acid medium by D2EHPA in 2. B. Gupta, N. Mudhar and I. Singh, Sep. Purif. Technol., 57 294–303
kerosene with 99.4% recovery. The loaded organic phase (2007).
3. H. Minamisawa, K. Murashima, M. Minamisawa, N. Arai and T. Okutani,
can be stripped by 2M hydrochloride acid solution effec- Anal. Sci., 19 401–404 (2003).
tively. Further purification was conducted by electrowin- 4. M. Tuzen and M. Soylak, J. Hazard. Mater., 129 79–185 (2006).
ning and zone refining process. The indium product of 5. L. Medvecky and J. Briancin, Chem. Pap., 58 93–100 (2004).
6. B. B. Adhikari, M. Gurung, H. Kawakita and K. Ohto, Chem. Eng. Sci., 78
zone refining can be used as raw material of indium 144–154 (2012).
sputtering target. 7. A. Yuchi, T. Sato, Y. Morimoto, H. Mizuno and H. Wada, Anal. Chem., 69
2941–2944 (1997).
8. S. Kamal, Z. Ehsan, S. Mojtaba and S. N. Masoud, Acta Chim. Slov., 57
222–229 (2010).
9. J. S. Liu, et al., Nonferrous Metals, 3 32–35 (2006).
Acknowledgments 10. M. S. Lee, J. G. Aha and E. C. Lee, Hydrometallurgy, 63 269–276 (2002).
11. J. A. Schufle and H. M. Eiland, J. Am. Chem. Soc., 76 960–962 (1954).
12. C. B. Onishi, X. F. Zhao, H. Doi and S. Yamazaki, Indium adsorbent and
The authors would like to thank the Ministry of indium separation methods, China Patent CN 101014407A, 2007.
Science and Technology “The plan of research of key 13. C. H. Xiong and C. P. Yao, Hazard. Mater., 170 1125–1132 (2009).