Recovery of Lithium from Simulated Secondary Resources (LiCO 3 ) through Solvent Extraction

: Lithium extraction is currently too ine ﬃ cient to be economical or marketable. The objective of this work was to ﬁnd the best extractant and the most inexpensive approach to recover lithium chemically from lithium ion batteries containing other desired metals using the solvent extraction technique. The extraction e ﬃ ciency of various extracting types was investigated. The highest extraction e ﬃ ciency of lithium ion from aqueous solution was obtained with bis(2-ethylhexyl) phosphate (DEHPA), with 75% recovery. Studying the e ﬀ ects of selected extractants in this experiment, it was found that the acidic extractant group provided better extraction e ﬃ ciency than solvating extractants. Further investigation of inﬂuential variables was carried out, including extraction time, pH of aqueous solution, and initial concentration. The results indicate that 6 h of extraction brings the system to equilibrium, and pH 1.5 is the best for extraction e ﬃ ciency.


Introduction
Lithium is in the alkaline metals group and can be found in brines (about 59%); another source of high-grade lithium is hard rock (about 29%) [1]. Its unique electrochemical properties, high redox potential, and higher heat specification give it strong potential for battery and power resource applications. Research on the distribution of lithium production in various countries suggests that those that produce and export the most lithium ore materials are Australia and Chile. It has also been reported that China and Chile have noteworthy resources of lithium ore compared to other countries [1]. The global market of lithium usage in 2016 was reported to be about 35% consumed in batteries, followed by ceramics and industrial glass, at about 32% [1]. Li ion batteries (LIBs) have become very popular in several types of electronic devices, including mobile phones, computers, batteries, and hybrid vehicles. The forecast for their use in mobile devices is rapidly expanding, and an extended range of electronic devices made with lithium ion batteries has become greatly needed. The lithium world market outlook published by Taiyou Research in 2014 on the forecasted Li ion demand showed that the demand significantly increased to 30 billion USD from 2008 to 2020 [2].
As a result of expanded demand for various applications with LIBs, lithium resource management and recycling recovery also need to be considered. The increased lithium ion battery demand was studied by Navigant Consulting with regard to vehicle usage, and the demand is expected to be about 221 billion USD by 2024 [3]. There are also expected expansions in mobile devices and portable batteries, aside from vehicles. The demand scenario was examined in several studies, showing that the supply of LIBs will not be able to meet the demand by 2023 [1]. Although the published reports indicate that lithium resources in the world will not be enough to support the need in the near future, lithium resource management is still unclear due to the limited resources and lack of waste management. In early 2000, several reports mentioned that LIB waste disposal was proceeding without treatment Sustainability 2020, 12, 7179 3 of 13 regarding the result of using solvating extractant in experiments to recover trivalent lanthanides using tributylphosphate (TBP). The results showed that better extraction efficiency could be achieved with extractants with higher atomic numbers [20]. For our experiment, we selected n-butanol to study extraction efficiency compared with other groups.

•
Chelating Extractant: The chelating structure is created when an organic extractant forms chelating properties in the organic phase in order to extract metals from the aqueous phase, such as the process of pretreating organic liquid with a chelating agent to form the chelation [18,20,21]. • Synergistic Extractant: The expectation was that a synergistic extractant would show better performance [18,19,23,24]. Synergistic systems can be found in mixtures of acid extractants and neutral extractants, such as in a metal ion extraction study using the synergistic system of trioctylphosphine oxide (TOPO) and trialkylphosphine oxide (TRPO) extractants [25], and the extraction of trivalent rare earth in a mixture of sec-octylphenoxy acetic acid (CA12) and Cyanex 301 in n-heptane [26].
In this work, we prepared lithium and cobalt in aqueous solution from lithium carbonate standard solution, and cobalt from cobalt (II) carbonate basic. The mixed solution of Li and Co was prepared in a ratio of 1:8, referring to the leaching solution from previous research by Jha et al. [8], showing the composition analysis of spent lithium ion batteries; the obtained leaching solution was 23.67% cobalt, 2.87% lithium, 22.13% copper, 4.30% aluminum, and 0.26% nickel.
The Li-Co ion solution was extracted using extraction solutions composed of n-butanol, trioctylamine (TOA), trioctyl phosphine oxide (TOPO) mixed with kerosene, dichloro-methane, 4-methyl-2-pentanol (MIBC), and the organo-phosphoric acid group, bis (2-ethylhexly) phosphate (DEHPA). The optimum conditions for lithium ion extraction were studied to determine the highest separation efficiency. The effects of extraction time, pH, extractant concentration, and co-ions in a mixed solution of Li and Co were investigated. Moreover, a separation mechanism of liquid-liquid extraction was also proposed, and the concentration of lithium ion was analyzed by inductively coupled plasma mass spectrometry (ICP-MS), Fourier-transform infrared spectroscopy (FTIR), and ion chromatography (IC).

Extraction Efficiency of Various Extracts
A stock solution of 10 mg/L lithium carbonate was prepared by using 99% of Li 2 CO 3 (Inorganic Ventures, Christiansburg, VA, USA) with deionized water in a volumetric flask. Cobalt (II) carbonate 98% (Carlo Erba reagents S.A.S, Val-de-Reuil, France) was prepared at 10 mg/L by dilution with 37% of HCl (Merck KGaA, Darmstadt, Germany) and deionized water. Subsequently, both Li and Co solutions were prepared to mix the Li and Co solution at 10 mg/L to represent the co-ion found in lithium ion batteries. Six types of extractants were used in this experiment: 98% of n-butanol (Merck KGaA, Darmstadt, Germany) 92.5% of trioctylamine (TOA) (Merck KGaA, Darmstadt, Germany), 98% of 4-methyl-2-pentanol (MIBC) (Merck KGaA, Darmstadt, Germany), 97% of bis(2-ethylhexyl) phosphate (DEHPA) (Sigma-Aldrich Pte Ltd., Singapore) and 98% of trioctylphosphine oxide (TOPO) from (Merck KGaA, Darmstadt, Germany) with 99% of kerosene (Sigma-Aldrich Pte Ltd., Singapore). The extraction experiment was done by sanity testing with heating at 32 • C and using a ratio of aqueous to extractant of 1:1 at equal extraction time of 2 h. The extraction efficiency was determined by Equation (1): where C i is the initial concentration and C f is the final concentration.

Studying Extraction Equilibrium
After the best extraction efficiency in the extractant group was defined, the following experiment to investigate the effect of extraction time was carried out. This experiment used pure DEHPA solution, with sample collection time points ranging from 30 min to 12 h at room temperature without heating system.

Effect of pH
The effect of pH was studied in the range of 1.5-3 using HCl acid and NaOH for pH adjustment in the aqueous phase prior to the extraction experiment. DEHPA 0.06 M with kerosene was used as the extractant with a phase volume ratio of A:E (aqueous to extractant) at 1:1. The extraction experiment was performed at room temperature in test tubes covered by tube caps. The samples were shaken by the lab equipment at 150 rpm for 60 min to make sure the two phases were mixed simultaneously. Then, the samples rested for 40 min at room temperature for phase separation and were left to reach extraction equilibrium. Sample collection was done only in the aqueous phase prior to ICP-MS testing.

Studying Extraction Selectivity
The effect of the co-ion was studied using a lithium and cobalt mix solution at a ratio of 20:80 wt.% The experiment was designed to understand the extraction efficiency when there are other metals in the solution. We performed a selectivity experiment using two concentrations, 10 mg/L of lithium/cobalt and 2 mg/L of lithium/cobalt solution. The Li-Co mix is used to represent the co-ion in leached solution of spent batteries. The volume ratio of aqueous and organic phase was 1:1 by volume. DEHPA 100% was used in this experiment. The sample solutions were extracted and samples were collected after 30 min to 12 h at room temperature. ICP-MS was used to identify the concentration of lithium and cobalt at each time interval.

Effect of Initial Concentration of Lithium
The effect of the different initial concentrations of lithium was investigated at the same time as the selectivity experiment by using 10 and 2 mg/L of lithium solution. The concentration of the stock solution was validated by ICP-MS prior to the experiment. The two concentrations were extracted by pure DEHPA solution in an equal A:E (aqueous to extractant) ratio. This experiment was aimed at studying extraction efficiency with lower lithium concentration. Sample collection was conducted time intervals from 30 min to 24 h to ensure that extraction time was efficient for both experiments.

Result of Extraction Efficiency
Solvent extraction can be classified in several ways-for example, by type of extractant and its functional group or chemical properties, or by the removal mechanism. In our experiment, we also aimed to study the efficiency of different extractants so that we could find the one with the highest efficiency to use for further experiments. In our research, after the extraction experiment for 2 h, it was found that recovery of nearly 75% of lithium was achieved by leaching in DEHPA solution as the solvent structure, as shown in Figure 1. In our research, after the extraction experiment for 2 h, it was found that recovery of nearly 75% of lithium was achieved by leaching in DEHPA solution as the solvent structure, as shown in Figure  1. The result of n-butanol extraction, as represented by the alcohol group from the solvating extractant, showed the lowest effectiveness compared to the others because its solubility in the aqueous phase was higher. The alcohol properties in the solvate reaction are not suitable to remove lithium from the aqueous phase based on our experiment. TOPO is also classified as a solvation extractant and achieved higher performance compared to n-butanol due to the higher atomic number of the extractant, and includes the effect of oxygen solvating in the aqueous phase, which can remove lithium directly compared to the c-chain in butanol. DEHPA is represented as an acidic extractant in the organo-phosphoric acid group at pH 2.63. The others (MIBC and TOA) are normally used as synergistic extractants. The experimental result of the synergistic extractant group shows that its lithium recovery is better than that of the solvating group due to its stronger acid solution and lower pH and electron donating ability. On the other hand, leaching in DEHPA, which also has shorter reaction times, achieved higher performance, as shown in Figure 2. The mechanism could be explained by the proton donating effect of the selective ion through the organic phase, which is non-soluble in the aqueous phase. The ion exchange happens at the hydroxyl group of DEHPA, which replaces hydrogen with lithium ion, as shown in Figure 3. The separation mechanism is combined between solubility and acidic properties of the extractant. The relative solubility of ions in the organic phase of each extractant leads to the separation of ions from the aqueous phase to the organic phase, while the higher acidity conditions with high proton dissociation also lead to an increase in the formation of the complex compound between metal ion and extractant. We also used FTIR spectroscopy to characterize an organic functional group of DEHPA before and after extraction, as shown in Figure 4a. From the FTIR spectrum, we observed three major absorption peaks of the DEHPA compound before extraction. The peaks at 2885-2865 cm −1 show the symmetric stretching vibration of the CH3 group, and the asymmetric stretching vibration of CH3 occurs at 2975-2950 cm −1 . The broad medium-intensity peaks at 2700-2560 cm −1 show the P-OH bond. The vibrational band of P=O is revealed at 1226 cm −1 . The peaks after extraction The result of n-butanol extraction, as represented by the alcohol group from the solvating extractant, showed the lowest effectiveness compared to the others because its solubility in the aqueous phase was higher. The alcohol properties in the solvate reaction are not suitable to remove lithium from the aqueous phase based on our experiment. TOPO is also classified as a solvation extractant and achieved higher performance compared to n-butanol due to the higher atomic number of the extractant, and includes the effect of oxygen solvating in the aqueous phase, which can remove lithium directly compared to the c-chain in butanol. DEHPA is represented as an acidic extractant in the organo-phosphoric acid group at pH 2.63. The others (MIBC and TOA) are normally used as synergistic extractants. The experimental result of the synergistic extractant group shows that its lithium recovery is better than that of the solvating group due to its stronger acid solution and lower pH and electron donating ability. On the other hand, leaching in DEHPA, which also has shorter reaction times, achieved higher performance, as shown in Figure 2. In our research, after the extraction experiment for 2 h, it was found that recovery of nearly 75% of lithium was achieved by leaching in DEHPA solution as the solvent structure, as shown in Figure  1. The result of n-butanol extraction, as represented by the alcohol group from the solvating extractant, showed the lowest effectiveness compared to the others because its solubility in the aqueous phase was higher. The alcohol properties in the solvate reaction are not suitable to remove lithium from the aqueous phase based on our experiment. TOPO is also classified as a solvation extractant and achieved higher performance compared to n-butanol due to the higher atomic number of the extractant, and includes the effect of oxygen solvating in the aqueous phase, which can remove lithium directly compared to the c-chain in butanol. DEHPA is represented as an acidic extractant in the organo-phosphoric acid group at pH 2.63. The others (MIBC and TOA) are normally used as synergistic extractants. The experimental result of the synergistic extractant group shows that its lithium recovery is better than that of the solvating group due to its stronger acid solution and lower pH and electron donating ability. On the other hand, leaching in DEHPA, which also has shorter reaction times, achieved higher performance, as shown in Figure 2. The mechanism could be explained by the proton donating effect of the selective ion through the organic phase, which is non-soluble in the aqueous phase. The ion exchange happens at the hydroxyl group of DEHPA, which replaces hydrogen with lithium ion, as shown in Figure 3. The separation mechanism is combined between solubility and acidic properties of the extractant. The relative solubility of ions in the organic phase of each extractant leads to the separation of ions from the aqueous phase to the organic phase, while the higher acidity conditions with high proton dissociation also lead to an increase in the formation of the complex compound between metal ion and extractant. We also used FTIR spectroscopy to characterize an organic functional group of DEHPA before and after extraction, as shown in Figure 4a. From the FTIR spectrum, we observed three major absorption peaks of the DEHPA compound before extraction. The peaks at 2885-2865 cm −1 show the symmetric stretching vibration of the CH3 group, and the asymmetric stretching vibration of CH3 occurs at 2975-2950 cm −1 . The broad medium-intensity peaks at 2700-2560 cm −1 show the P-OH bond. The vibrational band of P=O is revealed at 1226 cm −1 . The peaks after extraction The mechanism could be explained by the proton donating effect of the selective ion through the organic phase, which is non-soluble in the aqueous phase. The ion exchange happens at the hydroxyl group of DEHPA, which replaces hydrogen with lithium ion, as shown in Figure 3. The separation mechanism is combined between solubility and acidic properties of the extractant. The relative solubility of ions in the organic phase of each extractant leads to the separation of ions from the aqueous phase to the organic phase, while the higher acidity conditions with high proton dissociation also lead to an increase in the formation of the complex compound between metal ion and extractant. We also used FTIR spectroscopy to characterize an organic functional group of DEHPA before and after extraction, as shown in Figure 4a. From the FTIR spectrum, we observed three major absorption peaks of the DEHPA compound before extraction. The peaks at 2885-2865 cm −1 show the symmetric stretching vibration of the CH 3 group, and the asymmetric stretching vibration of CH 3 occurs at 2975-2950 cm −1 . The broad medium-intensity peaks at 2700-2560 cm −1 show the P-OH bond. The vibrational band of P=O is revealed at 1226 cm −1 . The peaks after extraction indicated the decline of intensity of the P-OH bond, as shown in Figure 4b. The shift of the P=O band between 1226 and 1211 cm −1 also reveals the bonding formation between lithium ion and P=O. In addition, the inorganic phosphates band was also indicated at 1015 cm −1 after extraction; the band shifted from 1030 cm −1 , the aliphatic phosphates band before extraction.
Sustainability 2020, 12, x FOR PEER REVIEW 6 of 13 indicated the decline of intensity of the P-OH bond, as shown in Figure 4b. The shift of the P=O band between 1226 and 1211 cm −1 also reveals the bonding formation between lithium ion and P=O. In addition, the inorganic phosphates band was also indicated at 1015 cm −1 after extraction; the band shifted from 1030 cm −1 , the aliphatic phosphates band before extraction.   indicated the decline of intensity of the P-OH bond, as shown in Figure 4b. The shift of the P=O band between 1226 and 1211 cm −1 also reveals the bonding formation between lithium ion and P=O. In addition, the inorganic phosphates band was also indicated at 1015 cm −1 after extraction; the band shifted from 1030 cm −1 , the aliphatic phosphates band before extraction.

Result of Extraction Equilibrium
After we defined the extractant with the highest efficiency, we then prolonged the extraction time to find the equilibrium. The experiment of extraction equilibrium shows that the metal concentration in aqueous solution decreased as a function of time and then became saturated. The result shows that

Result of Extraction Equilibrium
After we defined the extractant with the highest efficiency, we then prolonged the extraction time to find the equilibrium. The experiment of extraction equilibrium shows that the metal concentration in aqueous solution decreased as a function of time and then became saturated. The result shows that extraction of 10 mg/L lithium in lithium carbonate solution reached maximum efficiency after 4-6 h and saturated after that, as shown in Figure 5.

Result of pH Effect
The effect of pH on lithium recovery by DEHPA extraction was determined by using a solution of 10 mg/L lithium carbonate and 10 mg/L cobalt and mixed solution of Co-Li, referring to the same mass ratio in batteries. The result of lithium extraction indicated that lithium and cobalt in solution could be recovered better in lower pH conditions using the DEHPA extractant, as shown in Figure 6. The efficiency of DEHPA in lithium recovery shows significantly greater improvement over cobalt recovery. The recovery of lithium was above 60%, while cobalt recovery was lower than 40%. The lowest pH that we studied, pH 1.5, obtained the best recovery performance. The result shows that with increased pH, the recovery of lithium and cobalt is decreased. This happened in acidic conditions during extraction. It is concluded that the better synergistic bonding factor between selective ions in lower pH introduces a high level of free protons in the organic phase compared to the aqueous phase. Several researchers are studying the effect of pH on metal recovery. Nan et al. (2005) [15] published the results of their study of metal recovery from spent lithium ion batteries. They studied the effect of each parameter on extraction efficiency, including extraction temperature, phase ratio, extraction concentration, extraction time, and pH. The purpose of the experiment was to recover

Result of pH Effect
The effect of pH on lithium recovery by DEHPA extraction was determined by using a solution of 10 mg/L lithium carbonate and 10 mg/L cobalt and mixed solution of Co-Li, referring to the same mass ratio in batteries. The result of lithium extraction indicated that lithium and cobalt in solution could be recovered better in lower pH conditions using the DEHPA extractant, as shown in Figure 6. The efficiency of DEHPA in lithium recovery shows significantly greater improvement over cobalt recovery. The recovery of lithium was above 60%, while cobalt recovery was lower than 40%. The lowest pH that we studied, pH 1.5, obtained the best recovery performance. The result shows that with increased pH, the recovery of lithium and cobalt is decreased. This happened in acidic conditions during extraction. It is concluded that the better synergistic bonding factor between selective ions in lower pH introduces a high level of free protons in the organic phase compared to the aqueous phase.

Result of Extraction Equilibrium
After we defined the extractant with the highest efficiency, we then prolonged the extraction time to find the equilibrium. The experiment of extraction equilibrium shows that the metal concentration in aqueous solution decreased as a function of time and then became saturated. The result shows that extraction of 10 mg/L lithium in lithium carbonate solution reached maximum efficiency after 4-6 h and saturated after that, as shown in Figure 5.

Result of pH Effect
The effect of pH on lithium recovery by DEHPA extraction was determined by using a solution of 10 mg/L lithium carbonate and 10 mg/L cobalt and mixed solution of Co-Li, referring to the same mass ratio in batteries. The result of lithium extraction indicated that lithium and cobalt in solution could be recovered better in lower pH conditions using the DEHPA extractant, as shown in Figure 6. The efficiency of DEHPA in lithium recovery shows significantly greater improvement over cobalt recovery. The recovery of lithium was above 60%, while cobalt recovery was lower than 40%. The lowest pH that we studied, pH 1.5, obtained the best recovery performance. The result shows that with increased pH, the recovery of lithium and cobalt is decreased. This happened in acidic conditions during extraction. It is concluded that the better synergistic bonding factor between selective ions in lower pH introduces a high level of free protons in the organic phase compared to the aqueous phase. Several researchers are studying the effect of pH on metal recovery. Nan et al. (2005) [15] published the results of their study of metal recovery from spent lithium ion batteries. They studied the effect of each parameter on extraction efficiency, including extraction temperature, phase ratio, extraction concentration, extraction time, and pH. The purpose of the experiment was to recover Several researchers are studying the effect of pH on metal recovery. Nan et al. (2005) [15] published the results of their study of metal recovery from spent lithium ion batteries. They studied the effect of each parameter on extraction efficiency, including extraction temperature, phase ratio, extraction concentration, extraction time, and pH. The purpose of the experiment was to recover valuable metals from spent batteries, targeted at Cu, Co, and Li metals. They used sulfuric acid to dissolve metal from scrap batteries, then made a deposition of cobalt by oxalate, and finally studied the extraction of Cu using Acorga M5640 and Co using Cyanex 272 as the selected extractants. They also studied the effect of pH on the extraction ratio of Cu and Co. The result indicated that the highest efficiency of copper extraction achieved was 97% at pH 1 with a 1:1 phase ratio. On the other hand, the highest efficiency of cobalt extract by Cyanex 272 was obtained at pH 5.5, resulting in recovery of more than 96%.
Jian et al. (2012) [28] also studied the recovery of lithium and cobalt salt from waste batteries using an acid leaching process with sulfuric acid and peroxide followed by solvent extraction using P507 as the organo-phosphoric extractant. In an experiment with acid leaching, the effect of pH was also investigated. The result of the effect of pH on acid leaching showed that the separate factors of lithium and cobalt achieved better performance in higher pH conditions. Their experiment studied pH in the range of 1-6, and the result seemed stable after pH 4, so the best condition they chose was pH 3.5 in order to avoid the loss of the selected ion during the leaching process. The result also suggests that leaching efficiency is also influenced by other effects, such as acid concentration. Therefore, both pH and acid concentration need to be considered and optimized.
Virolainen et al. (2017) [29] published the results of leachate separation of lithium ion batteries by solvent extraction using Cyanex 272 and PC-88A as extractants, targeting the recovery of lithium, nickel, and cobalt from the leachate of spent batteries. They studied three separate steps, starting with a loading step using Cyanex 272 to remove cobalt and nickel, followed by a lithium scrubbing process using NiSO 4 to remove lithium impurity, and for the last step, they used H 2 SO 4 for selective stripping of nickel, leaving the cobalt in the leachate. The loading step by Cyanex 272 reflected the specific removal of Ni and Co over lithium. They also studied the effects of phase ratio and pH isotherm of Cyanex 272 and PC-88A with and without solution modification using TOA and TBP before the selectivity experiment. The results suggest that the selectivity of Co, Ni, and Li using these two extractants was not significantly different in the pH range of 5-7, which was determined from the pH isotherm experiment. The extraction efficiency indicated that the selected extractant with 5% modification by TOA achieved 100% extraction of cobalt and about 99% of nickel, while lithium ion showed better selectivity to be extracted faster compared to cobalt and nickel, so they decided to remove lithium in the first step of the experiment.
We also reviewed a report by Harvianto et al. (2016) [17]. They studied lithium extraction from seawater using TTA-TOPO in kerosene as the extractant, followed by lithium ion stripping after the extraction experiment. The effects of pH and various acid types were investigated. The stripping result showed that lithium ion was successfully removed in lower pH of 2-4. Efficiency above 90% was achieved at a pH of about 3. Then they observed that the efficiency dropped significantly to below 20% in pH above 3.5.
From the literature, we learned that both phase ratio and pH are critical parameters for solvent extraction. Previous studies on separation of spent batteries mostly focused on removing cobalt using various types of extraction. Several types of extractants are well known for their selectivity in the extraction of cobalt. One reason might be that the weight composition of cobalt in batteries is much higher compared to lithium and nickel (Co 14.2-14.5, Ni 0.49-0.51, and Li 2.7-2.9 g/L) [29]. However, there are more influential factors that affect recovery efficiency. In our experiment, the phase ratio was fixed at 1:1, which we selected from our screening experiment. The pH effect was introduced because the pH value of selected extractants from the extractant experiments suggested that we could achieve better efficiency in lower pH. The pH of the solution was adjusted in the aqueous phase before the extraction experiment. Based on the extractant concentration, we studied a modification by 0.06 M DEHPA with kerosene to observe the extraction efficiency. Therefore, the result of our experiment would apply with speculative conditions only. We also reviewed a report by Nascimento et al. (2015) [30]. They suggested that the concentration of extractant is impactful for extraction efficiency, as a high volume of extractant is increases possibilities of metal formation with the extractant. Other factors need to be considered and focused on to improve the extraction efficiency, which could promote the effects of pH specifically.

Result of Initial Concentration Effect
We investigated the effects of initial concentrations of 2 and 10 mg/L of lithium in aqueous solution without pH pre-treatment before experimenting. The concentration of 2 mg/L represented lithium wt% in LIBs at a low concentration of about 1-2 mg/L. After the highest efficiency was defined as 10 mg/L lithium, the same experimental conditions were used for extraction as for low concentration, including the sampling plan, extraction volume, sample collection method, and extraction time. The results are shown in Figure 7. The maximum extraction efficiency of 10 mg/L lithium was achieved at 73%. The extraction reaction drastically increased starting from 30 min to 4 h. The extraction experiment reached equilibrium after 6 h. On the other hand, the extraction of 2 mg/L lithium concentration obtained the highest efficiency at 35% and reached equilibrium after 4 h of extraction. We also observed that 2 mg/L lithium seemed to maintain the relative starting efficiency after 2 h of extraction.

Result of Initial Concentration Effect
We investigated the effects of initial concentrations of 2 and 10 mg/L of lithium in aqueous solution without pH pre-treatment before experimenting. The concentration of 2 mg/L represented lithium wt% in LIBs at a low concentration of about 1-2 mg/L. After the highest efficiency was defined as 10 mg/L lithium, the same experimental conditions were used for extraction as for low concentration, including the sampling plan, extraction volume, sample collection method, and extraction time. The results are shown in Figure 7. The maximum extraction efficiency of 10 mg/L lithium was achieved at 73%. The extraction reaction drastically increased starting from 30 min to 4 h. The extraction experiment reached equilibrium after 6 h. On the other hand, the extraction of 2 mg/L lithium concentration obtained the highest efficiency at 35% and reached equilibrium after 4 h of extraction. We also observed that 2 mg/L lithium seemed to maintain the relative starting efficiency after 2 h of extraction. We also performed ion chromatography analysis of the solution after sample preparation prior to our experiments. Figure 8 shows chromatograms of 2 and 10 mg/L of Li 2 CO 3 solution before extraction. The standard cation solution of was mixed and prepared as the standard curve prior to analysis. The R-square of the standard curve was required at a minimum of 99.95% confidence. The ion content in standard solution included Li, Na, K, Mg, and Ca. The IC result shows that the intensity peak of 2 mg/L is at about 150, while that of 10 mg/L is above 450 in the range of retention time between 3.5 and 4.5 min. We further confirmed this with an IC test of the sample in the aqueous phase after the extraction experiment and observed that the related peak intensity as a result of the lithium amount decreased and was removed to the organic phase.  We also performed ion chromatography analysis of the solution after sample preparation prior to our experiments. Figure 8 shows chromatograms of 2 and 10 mg/L of Li 2 CO 3 solution before extraction. The standard cation solution of was mixed and prepared as the standard curve prior to analysis. The R-square of the standard curve was required at a minimum of 99.95% confidence. The ion content in standard solution included Li, Na, K, Mg, and Ca. The IC result shows that the intensity peak of 2 mg/L is at about 150, while that of 10 mg/L is above 450 in the range of retention time between 3.5 and 4.5 min. We further confirmed this with an IC test of the sample in the aqueous phase after the extraction experiment and observed that the related peak intensity as a result of the lithium amount decreased and was removed to the organic phase.

Result of Initial Concentration Effect
We investigated the effects of initial concentrations of 2 and 10 mg/L of lithium in aqueous solution without pH pre-treatment before experimenting. The concentration of 2 mg/L represented lithium wt% in LIBs at a low concentration of about 1-2 mg/L. After the highest efficiency was defined as 10 mg/L lithium, the same experimental conditions were used for extraction as for low concentration, including the sampling plan, extraction volume, sample collection method, and extraction time. The results are shown in Figure 7. The maximum extraction efficiency of 10 mg/L lithium was achieved at 73%. The extraction reaction drastically increased starting from 30 min to 4 h. The extraction experiment reached equilibrium after 6 h. On the other hand, the extraction of 2 mg/L lithium concentration obtained the highest efficiency at 35% and reached equilibrium after 4 h of extraction. We also observed that 2 mg/L lithium seemed to maintain the relative starting efficiency after 2 h of extraction. We also performed ion chromatography analysis of the solution after sample preparation prior to our experiments. Figure 8 shows chromatograms of 2 and 10 mg/L of Li 2 CO 3 solution before extraction. The standard cation solution of was mixed and prepared as the standard curve prior to analysis. The R-square of the standard curve was required at a minimum of 99.95% confidence. The ion content in standard solution included Li, Na, K, Mg, and Ca. The IC result shows that the intensity peak of 2 mg/L is at about 150, while that of 10 mg/L is above 450 in the range of retention time between 3.5 and 4.5 min. We further confirmed this with an IC test of the sample in the aqueous phase after the extraction experiment and observed that the related peak intensity as a result of the lithium amount decreased and was removed to the organic phase. Another interesting result from the IC test is the chromatogram comparison. When we plotted the overlay chromatogram between the lithium solution and lithium mixed with cobalt, as shown in Figure 9, we observed that the retention time of lithium ion shifted from 4 to 3.5 min, and the intensity of pure lithium had a sharp peak compared to the peak from the mixed solution. The lithium ion peak from the mixed solution shows a decline by about three times and broader intensity compared to lithium without cobalt in the solution. The reason for the intensity pattern is the co-ion effect from other metals. We speculate that there might be some reaction happening, since the cobalt solution was pre-mixed with HCl. When the solution becomes more acidic, it might affect the free ions. The tendency of lithium is to be removed faster. The co-ion effect is another reason that could explain this phenomenon, although the cobalt ion required more time to separate from the IC due to its higher atomic number compared to lithium. However, when there were other ions in the solution besides cobalt, since the solution included HCl as well, it might directly affect the lithium ion. However, the IC result suggests that we can remove lithium from the aqueous phase using DEHPA, as indicated by the intensity change and the effect of retention time. Another interesting result from the IC test is the chromatogram comparison. When we plotted the overlay chromatogram between the lithium solution and lithium mixed with cobalt, as shown in Figure 9, we observed that the retention time of lithium ion shifted from 4 to 3.5 min, and the intensity of pure lithium had a sharp peak compared to the peak from the mixed solution. The lithium ion peak from the mixed solution shows a decline by about three times and broader intensity compared to lithium without cobalt in the solution. The reason for the intensity pattern is the co-ion effect from other metals. We speculate that there might be some reaction happening, since the cobalt solution was pre-mixed with HCl. When the solution becomes more acidic, it might affect the free ions. The tendency of lithium is to be removed faster. The co-ion effect is another reason that could explain this phenomenon, although the cobalt ion required more time to separate from the IC due to its higher atomic number compared to lithium. However, when there were other ions in the solution besides cobalt, since the solution included HCl as well, it might directly affect the lithium ion. However, the IC result suggests that we can remove lithium from the aqueous phase using DEHPA, as indicated by the intensity change and the effect of retention time.

Effect of Co-Ion on Extraction Efficiency
The selectivity of the DEHPA extractant on lithium ion recovery was studied by using two types of lithium in solution. The first was lithium carbonate solution without other heavy metals to be compared with the result of a lithium and cobalt solution, representing the other metals in batteries. The result shows that the selectivity of DEHPA for lithium was almost constant in each solution ( Figure 10). After 6 h, 72% and 73% extraction efficiency was achieved without cobalt and with cobalt in the solution, respectively. The extraction efficiency of lithium recovery was not significantly different when we added mixed ions to the solutions in our experiment. The result indicates that the selected concentrations of extractant in our research were not affected by the selectivity of lithium extraction when there was cobalt in the solution.

Effect of Co-Ion on Extraction Efficiency
The selectivity of the DEHPA extractant on lithium ion recovery was studied by using two types of lithium in solution. The first was lithium carbonate solution without other heavy metals to be compared with the result of a lithium and cobalt solution, representing the other metals in batteries. The result shows that the selectivity of DEHPA for lithium was almost constant in each solution ( Figure 10). After 6 h, 72% and 73% extraction efficiency was achieved without cobalt and with cobalt in the solution, respectively. The extraction efficiency of lithium recovery was not significantly different when we added mixed ions to the solutions in our experiment. The result indicates that the selected concentrations of extractant in our research were not affected by the selectivity of lithium extraction when there was cobalt in the solution. Another interesting result from the IC test is the chromatogram comparison. When we plotted the overlay chromatogram between the lithium solution and lithium mixed with cobalt, as shown in Figure 9, we observed that the retention time of lithium ion shifted from 4 to 3.5 min, and the intensity of pure lithium had a sharp peak compared to the peak from the mixed solution. The lithium ion peak from the mixed solution shows a decline by about three times and broader intensity compared to lithium without cobalt in the solution. The reason for the intensity pattern is the co-ion effect from other metals. We speculate that there might be some reaction happening, since the cobalt solution was pre-mixed with HCl. When the solution becomes more acidic, it might affect the free ions. The tendency of lithium is to be removed faster. The co-ion effect is another reason that could explain this phenomenon, although the cobalt ion required more time to separate from the IC due to its higher atomic number compared to lithium. However, when there were other ions in the solution besides cobalt, since the solution included HCl as well, it might directly affect the lithium ion. However, the IC result suggests that we can remove lithium from the aqueous phase using DEHPA, as indicated by the intensity change and the effect of retention time.

Effect of Co-Ion on Extraction Efficiency
The selectivity of the DEHPA extractant on lithium ion recovery was studied by using two types of lithium in solution. The first was lithium carbonate solution without other heavy metals to be compared with the result of a lithium and cobalt solution, representing the other metals in batteries. The result shows that the selectivity of DEHPA for lithium was almost constant in each solution ( Figure 10). After 6 h, 72% and 73% extraction efficiency was achieved without cobalt and with cobalt in the solution, respectively. The extraction efficiency of lithium recovery was not significantly different when we added mixed ions to the solutions in our experiment. The result indicates that the selected concentrations of extractant in our research were not affected by the selectivity of lithium extraction when there was cobalt in the solution.

Conclusions
The extraction efficiency results of various types of extractants show that the highest efficiency was achieved by bis (2-ethylhexyl) phosphate (DEHPA) with the pH of the solution at 2.63. The results of the extraction equilibrium indicated that extraction efficiency was almost constant after 4-6 h of extraction time. The effects on various types of extractants suggest that the highest efficiency of lithium recovery is obtained from acidic extractants, followed by synergized extractants. Solvating extraction had the lowest efficiency in our experiment. We also observed that pH played an important role in the extraction mechanism as a result of the acidic solution leading to more free ions in the solution, promoting better extraction efficiency. However, the effect of the initial concentration of lithium in solution of 2 to 10 mg/L reflected an interesting result, suggesting that when the concentration is changed, the extraction conditions need to be considered and optimized. There are enhancement techniques for removing metal from extractant after an extraction experiment, such as acid stripping, washing, and scrubbing, following the solvent extraction step. The additional steps help to improve the efficiency of recovery because the characteristics of each technique will allow more conditions to be optimized. Since the application that we are looking at uses lower concentrations of lithium, about 1-2 mg/L, the extraction conditions and other recovery techniques in lower concentrations will be our area of interest for further investigation.
Generally, there will be other metal ions present in waste lithium ion batteries. We also found from our experiment that the extraction efficiency of lithium removal is maintained even when other ions are present in the solution. The effect of mixed ions was not clear with regard to how cobalt affects extraction efficiency, but selectivity is another factor that needs to be considered in further experiments.
Moreover, it was found that solvent extraction using organo-phosphorous acid can be applied for lithium recovery from spent lithium ion batteries, as we used DEHPA in our study. The results from our experiment suggest that lithium recovery can be done in proper extraction conditions. In addition, there is room to improve the efficiency and enhance the selectivity when the lithium concentration is low. The combination of synergistic and chelating extractants is one possible approach to improve our experiment. We are also interested in combinations with other techniques, such as electrodialysis, which will be the subject of one of our enhancement studies in the future.