Nickel and cobalt recycling from lithium-ion batteries by electrochemical processes

Waste Management 25 (2005) 215–220 www.elsevier.com/locate/wasman Nickel and cobalt recycling from lithium-ion batteries by electrochemical processes C. Lupi *, M. Pasquali, A. DellÕEra Dipartimento ICMMPM, Università di Roma ‘‘La Sapienza’’, Via Eudossiana 18, 00184 Roma, Italy Accepted 17 December 2004 Abstract The presence of LiCoO2 and LiCoxNi(1
x)O2 in the cathodic material of Li-ion and Li-polymer batteries has stimulated the recovery of Co and Ni by hydrometallurgical processes. In particular, the two metals were separated by SX method and then recovered by electrochemical (galvanostatic and potentiostatic) processes. The metallic Ni has been electrowon at 250 A/m2, pH 3–3.2 and 50 °C, with 87% current efficiency and 2.96 kWh/kg specific energy consumption. Potentiostatic electrolysis produces a very poor Ni powder in about 1 h with current efficiency changing from 70% to 45% depending on Ni concentration in the electrolyte. Current efficiency of 96% and specific energy consumption of 2.8 kWh/kg were obtained for Co at 250 A/m2, pH 4–4.2 and 50 °C, by using a solution containing manganese and (NH4)2SO4. The Co powder, produced in potentiostatic conditions (
0.9 V vs. SCE, pH 4, room temperature) appears particularly suitable for Co recycling as cobaltite in new batteries. Ó 2005 Elsevier Ltd. All rights reserved. 1. Introduction Lithium and lithium-ion batteries are used in several fields: audio–video, communications, computers, uninterruptible power supplies, etc. The recycling of these batteries, containing LiCoO2 and LiCoxNi(1
x)O2 in the cathodic material (Broussely, 2004; Broussely et al., 1999; Cho et al., 2000), is today very important, both from an environmental and an economic point of view (Wiaux, 2001). Indeed, for the battery industries it could be very interesting to recover some battery materials to recycle them in the production of new ones. Particularly Toxco and Sony/Sumitomo have developed recycling processes for lithium batteries. The Toxco process (US patent, 1994) uses a preliminary cryogenic treatment of batteries by liquid nitrogen * Corresponding author. Tel.: +39 6 4458 5636; fax: +39 6 4458 5641. E-mail address: carlalupi@uniroma1.it (C. Lupi). 0956-053X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.wasman.2004.12.012 (
195 °C), to considerably reduce the material reactivity. After this treatment, the batteries are crushed and then treated with an alkaline solution. The produced cake is treated to obtain lithium salts. This process is industrially exploited for batteries of all sizes. The Sony/Sumitomo process has been presented in 1996. It includes two main steps: battery incineration carried out at Sony plants at about 1000 °C to open them, and the cobalt extraction performed at Sumitomo plants. The incineration opens the batteries and eliminates the flammables such as electrolyte and separators. Subsequently the batteries are broken and sieved. The residue of the last operation contains Fe, Cu and Al pieces that could be magnetically separated. The powder leaving the sieve is principally made up of carbon and LiCoO2 and/or LiCoxNi(1
x)O2. The Co is recovered from this powder and reintroduced in the production cycle of batteries. The Co and Ni recovery process studied in this paper is perfectly adaptable with both cryogenic (Toxco 216 C. Lupi et al. / Waste Management 25 (2005) 215–220 process) and incineration (Sony/Sumitomo process) treatment of batteries. Nevertheless, the cryogenic treatment of batteries seems to be more appropriate because in this case it is easier to separate the different battery components such as cathode, anode, separators, can and electronics. In this paper the electrolytes are made up of Ni and Co solutions coming from the separation of the two metal ions operated by solvent extraction. These electrolytes are utilized to recover the two metals by electrochemical techniques. 2. Experimental The cobalt and nickel containing solutions have been prepared by dissolving the cathode material, recovered from Li-ion exhausted batteries, with H2SO4 and H2O2 pure reagents in an appropriate ratio (Pasquali and Lupi, 2001). The leach liquor was treated to separate Ni from Co by solvent extraction methods. The selective extraction of Co has been performed by saponified 0.5 M CYANEX 272 in kerosene. By using an O/A ratio equal to 3 with three stages the aqueous raffinate contains practically only Ni. The raffinate produced has been used as Ni electrolyte after treatment with active carbon. Co electrolytic solution has been obtained by stripping with an H2SO4 aqueous solution, Co from the organic phase (Lupi et al., 1999; Lupi and Pasquali, 2002). Fig. 1. Photo of the two-compartment cell with an anionic membrane. The volume of each compartment is 200 ml. The holes allow sampling and the introduction of a pH meter and a reference electrode. The cathode was an aluminum net or foil and the anode was a stainless steel net for Co and titanium net for Ni. The quantitative analyses of Co and Ni were made by A.A.S. UNICAM 969. The pH solutions were monitored by an Amel pH-meter mod.334-B. The Co and Ni deposits were analyzed by a Philips PW 1390 diffractometer and a SEM Hitachi S2500 equipped with EDS quantitative analysis KEVEX apparatus. 3. Results and discussion 2.1. Galvanostatic apparatus 3.1. Cobalt recovery The electrowinning tests were carried out at constant current density by using an Amel mod.553 galvanostat, an Amel mod.721 coulombmeter and a Linseis mod.6510 recorder. A 1000 cm3 metacrylate cell was designed and equipped with an aluminum or an AISI 316L stainless steel cathode and a Pb–8 Sb anode. Cathodic and anodic compartments were separated by a polypropylene membrane and the electrodes were 3 cm apart. The catholyte pH was controlled by potassium hydroxide or sulfuric acid additions. Boric acid was added to the catholyte as buffer. The temperature was kept constant with a JULABO thermostat. 2.2. Potentiostatic apparatus The potentiostatic electrolysis tests were carried out by potentiostat EG&GPAR 270 at a constant potential of
0.85 to
0.90 V vs. SCE for Co and of
1.20 to
1.50 V vs. SCE for Ni. The potentiostat is connected with a computer for data acquisition and analysis. In the electrolytic cell the anodic and cathodic compartments, 200 ml each, were separated by a BDH anionic exchange membrane (Fig. 1). 3.1.1. Galvanostatic tests Preliminary tests were performed to individualize both cathode material and current density to be utilized. In particular AISI 316L stainless steel and aluminum were considered as a cathode. The former gives deposits having high internal stress that worsened the deposit quality. The 250 A/m2 current density has been chosen that represents a good compromise between kinetics and deposit consistency. The electrolyte composition has been selected considering that the highest yields are reached (both with and without Mn) at the highest concentrations, as shown in Fig. 2. The Mn presence in the electrolyte has been tested because it is frequently present as an impurity in rechargeable batteries. As can be seen in Fig. 2, this presence is favorable for Co electrowinning because the current efficiency increases independently of the Co concentration. Furthermore the deposit morphology is positively influenced as shown in Fig. 3. The electrowon cobalt is globular in both cases, but with manganese containing electrolyte the grains are finer, thus giving a more homogenous deposit. C. Lupi et al. / Waste Management 25 (2005) 215–220 the advantage. Additions of ammonium sulphate in the presence of Mn cause a consistent decrease of the current efficiency at room temperature, but also a gain in cell voltage and specific energy consumption in comparison with the additive-free electrolyte. By raising the temperature up to 50 °C, there is no decrease in the yield, while an increase in cell voltage of about 100 mV is obtained, this resulting in a good value of about 2.8 kWh/kg specific energy consumption. C.E. (%) 100 99 25 ˚C 98 25 ˚C, 1g/l Mn 217 97 96 95 94 93 0 10 20 30 40 Co (g/l) Fig. 2. Results of electrowinning tests carried out at 250 A/m2 with solution containing different Co concentration and with or without Mn. Fig. 3. SEM micrographs of electrowon Co obtained from (a) Mn free and (b) Mn containing electrolyte. After selecting a Co concentration of 33.3 g/l several tests were carried out to determine the best operative conditions. In particular, temperature and presence of additives (Mn and (NH4)2SO4) were investigated (Table 1). Ammonium sulphate was added as a buffer. A temperature increase slightly decreases the current efficiency, as expected, while the effect on cell voltage is greater, as it decreases of about 300 mV in the range 25–50 °C. As already mentioned, the Mn presence in the electrolyte has a positive effect on current efficiency, but the cell voltage also increases, this counterbalancing Table 1 Results of cobalt electrowinning using additives Mn (g/l) 1 1 1 – 1 (NH4)2SO4 (g/l) T (°C) gI (%) DV (V) Specific energy consumption (kWh/kg) – – 15 – 15 – 15 25 25 25 40 40 50 50 96.56 97.22 94.10 94.58 97.04 95.69 95.81 3.476 3.610 3.319 3.243 3.081 3.055 2.964 3.276 3.384 3.209 3.120 2.889 2.905 2.815 3.1.2. Potentiostatic tests Aluminum, net or sheet, represents the best cathode material if compared with lead, stainless steel or nickel for Co. Even if lead is a good material as a cathode, considering the high hydrogen over-potential, the recovery of metallic Co from its surface is difficult. The use of net or aluminum foil depends on the type of Co recycling: using foil, electrolysis has longer duration and allows recovery of Co as metal, while using Al net the electrolysis duration decreases but Co cannot be completely recovered as metal. The pH conditions for cobalt deposition are not critical. Indeed, it is possible to work in a larger pH interval (4–6), without any problems. It must be considered that pH 6 is a limit for Co electrolysis because of the Co(OH)2 precipitation near the cathode when the electrolysis starts. The high activation energy for cobalt reduction initially promotes the hydrogen discharge (2H3O+ + 2e
! H2 + 2H2O; E0 = 0.00 V) that locally produces a pH increase, thus reaching the condition of cobalt-hydroxide precipitation. Preliminary electrolysis tests have been performed at
0.9V vs. SCE and pH 4–6 by using both aluminum net and foil cathode in order to evaluate the best operative conditions. Quite high and constant current values for both pH utilized are observed, but they are more quickly reached at pH 4. The electrolysis time is shorter when aluminum net is used due to higher surface area. With Al foil, the time needed to have the same Co depletion in the electrolyte doubles, even in the same conditions (Table 2). The electrolysis carried out at pH 6 showed that it is easy to reach 100% current efficiency by using both Al net and foil, while working at pH 4, 100% current efficiency is only obtained by using aluminum sheet; current efficiency decreases to 80% if Al net is used (Table 2). Metallic cobalt powder treated at 300 °C in air is transformed into Co3O4. A mix of this Co-oxide with Li2CO3 or LiOH compounds produces a fresh LiCoO2 ready to be used again in new lithium-ion batteries. Even if the greater aluminum net electrode area reduces the electrolysis times, cobalt recovery is more complicated; in fact it is possible to recover either metallic powder by net rubbing, or (CH3COO)2Co and CoSO4 cobalt 218 C. Lupi et al. / Waste Management 25 (2005) 215–220 Table 2 Results of Co potentiostatic electrolysis (
0.9 V SCE) on Al net and foil at different pH T (min) DT 0 30 60 90 120 150 180 210 240 258 0 30 30 30 30 30 30 30 30 18 Al net, pH 4 Al foil, pH 4 Al net, pH 6 Al foil, pH 6 C (ppm) g% C (ppm) g% C (ppm) g% C (ppm) g% 1270.00 1015.00 858.50 339.50 79.85 – 63.00 70.00 83.00 78.00 1272.00 1270.00 1227.50 1215.00 1068.00 918.00 495.00 234.30 125.90 – 1223.00 1208.00 1193.00 980.50 291.00 79.40 – 1152.00 – salts by crystallization of the solutions obtained by washing the net with the respective acids. The deposit morphology of Co is shown in Fig. 4, for two different pH values: 4 (a) and 6 (b) where a poorly compact deposit with branched lamellae is visible. 3.2. Ni recovery 3.2.1. Galvanostatic tests Titanium, aluminum and AISI 316L cathodes have been tested in preliminary electrowinning experiments in order to choose the best cathode material. The deposits obtained on Ti cathode at any temperature in the 25– 60 °C range were characterized by internal stresses and broke off from the titanium surface quite rapidly. Aluminum presented analogous results. Better results were obtained using AISI 316L as cathode, so it has been se- 0.16 43.00 28.24 59.00 63.00 96.00 100.00 94.00 1.00 63.00 95.00 100.00 98.00 933.00 >100 910.00 >100 544.00 100 239.00 171.00 98 95 lected for the galvanostatic tests reported in this work. The more significant results are reported in Table 3. From Table 3 it can observed that temperature has an effect on deposition time. It must be noted that the test duration was set at 2 h, and shorter test durations were caused by early deposit detachment from cathodic support or breaking caused by relevant internal stresses. The addition of boric acid doubles the time for deposit detachment, maintaining almost the same results in terms of current efficiency and cell voltage. Increasing current density is not useful in the absence of boric acid. The deposit morphology does not show remarkable changes in the presence of boric acid, as highlighted in Fig. 5. The deposit looks compact and homogenous with rounded particles in both cases. In Fig. 5(a) some marks of bubbles due to gas evolution are evident. Fig. 4. SEM micrographs of cobalt deposit obtained in potentiostatic conditions. Table 3 Results of nickel electrowinning with AISI 316L cathode by using a solution containing 49.5 g/l Ni Time (hours) H3BO3 (g/l) Catholyte pH C.D. (A/m2) T (°C) gI (%) DV (V) SEC (kWh/kg) 1.40 0.84 1.22 1.70 2.00 – – 20 20 20 3.2 3.2 3.2 3.2 3.2 125 250 250 250 250 40 40 25 40 50 84.11 89.43 89.83 88.22 86.91 2.630 2.936 3.124 2.986 2.820 2.86 3.00 3.18 3.09 2.96 219 C. Lupi et al. / Waste Management 25 (2005) 215–220 Table 4 Hydrogen discharge potential on the different materials Cathode materials Potential of hydrogen discharge Ni Pt Al
0.40
0.57
0.90 The current efficiency of the process was calculated by the real/theoretical Ni deposited quantity ratio g¼ ½NiðNH3 Þ6 2þ þ 2e
! Ni þ 6NH3 ðaqÞ E0 ¼
0:480 V 2H3 Oþ þ 2e
! H2 þ 2H2 O E0 ¼ 0:000 V while the anodic reaction is: 6H2 O ! O2 þ 4H3 Oþ þ 4e
E0 ¼ 1:229 V The hydrogen evolution could be reduced by using an appropriated cathodic material having a high hydrogen reduction over-potential. Preliminary cyclic voltammetry tests have been carried out on different cathode materials (platinum, nickel and aluminum) and using the same Ni-free blank solution. Table 4 shows that aluminum has the highest hydrogen over-potential. After choosing the aluminum as cathode the best results in terms of deposit quality and electrolysis time are reached by working at
1.5 V SCE. Indeed, about 80 min are needed to leave less then 100 ppm Ni in the electrolyte and to produce a thick gray Ni powder on the Al net cathode. 1800 80 1600 70 1400 60 1200 50 1000 40 800 30 600 400 20 200 10 0 C.E.(%) 3.2.2. Potentiostatic tests The reactions occurring at the cathode are: where Qc are the coulombs calculated from the grams of Ni deposited and Qs are the coulombs experimentally passed in the cell. Every 10 min the D (ppm) of Ni in the electrolyte has been measured and then Qc calculated, while the apparatus supplied Qs. Fig. 6 shows the current yield and nickel depletion during the electrolysis. It can be seen that after 80 min the nickel concentration is 84 ppm and the current efficiency is about 45%; this value is acceptable considering the very low content of the Ni. The X-ray analysis performed on Ni powder electrowon potentiostatically demonstrated that it was pure metal. Ni in solution (ppm) Fig. 5. SEM micrographs of Ni deposit obtained without (a) and with (b) boric acid addition. Qc ; Qs 0 0 10 20 30 40 50 60 70 80 90 Time (min.) Fig. 6. Results of the test carried out at
1.5 V SCE for Ni deposition. Fig. 7. SEM micrographs of Ni deposit obtained with potentiostatic method. 220 C. Lupi et al. / Waste Management 25 (2005) 215–220 The deposit morphology is shown in Fig. 7. The SEM micrographs show a characteristic dendritic structure that represent the starting point for globular nickel growing. The Ni deposit is not compact, but it is a powder that can be easily recovered from the aluminum electrode.  The electrolysis at constant potential of a partially depleted solution of Ni produces a very pure powder leaving less than 100 ppm of nickel in the solution. Acknowledgments 4. Conclusions A hydrometallurgical process to recycle Li-ion and Li-polymer batteries containing both LiCoO2 and LiCoxNi(1
x)O2 cathode material has been investigated. The operations involved in the process are: cathodic paste leaching, cobalt–nickel separation by solvent extraction with modified Cyanex 272 in kerosene, Co and Ni metal recovery by galvanostatic electrowinning and Co and Ni recovery by potentiostatic electrolysis carried out on partially depleted electrolyte. The experimental results lead to the following conclusions:  Cobalt electrowinning performed at 250 A/m2 current density produces a good Co deposit with a CE and a SEC of about 96% and 2.8 kWh/kg, respectively.  The electrolysis at constant potential appears particularly suitable for cobalt recycling as cobaltite in new batteries, with a yield of 100%.  The electrochemical filter gives a solution containing less than 1 ppm Co.  Nickel electrowinning performed at 250 A/m2 current density produces an attractive Ni deposit with a CE and a SEC of about 87% and 2.96 kWh/kg, respectively. The authors thank the Italian National Research Council (CNR) for providing financial support to this work. References Broussely, M., 2004. Li-ion batteries for HEV/EV and other industrial applications. In: Nazri, G., Pistoia, G. (Eds.), Science and Technology of Lithium Batteries. Kluwer. Broussely, M., Biensan, P., Simon, B., 1999. Lithium insertion into host materials: the key to success for Li-ion batteries. Electrochim. Acta 45, 3. Cho, J., Jung, H., Park, Y., Kim, G., Lim, H., 2000. Electrochemical properties and thermal stability of LiaNi1
xCoxO2 cathode materials. J. Electrochem. Soc. 147, 15. Lupi, C., Cavallini, M., Pilone, D., 1999. Studio di processi idrometallurgici per il riciclaggio degli accumulatori esausti Ni–Cd. ANPA Research Project. Lupi, C., Pasquali, M., 2002. Electrolytic nickel recovery from lithiumion batteries. Miner. Eng. 16, 537–542. Pasquali, M., Lupi, C., 2001. Electrolytic cobalt recovery from lithiumion batteries. In: Rao, S.R., Amaratunga, L.M., Kondos, P.D., Richards, G.G., Kuyucak, N., Kozinski, J.A. (Eds.), Waste Processing and Recycling in Mineral and Metallurgical Industries IV. CIM, Toronto, Canada, pp. 41–54. US Patent 5.345.033, 1994. Wiaux, J.P., 2001. In: Pistoia, G., Wiaux, J.P., Wolsky, S.P. (Eds.), Used Battery Collection and Recycling. Elsevier.