Hydrometallurgy 151 (2015) 78–83

Contents lists available at ScienceDirect

Hydrometallurgy

j ourna l homepage: www.e lsev ie r .com/ locate /hydromet

Lithium recovery system using electrostatic field assistance

Taegong Ryu a, Dong-Hee Lee b, Jae Chun Ryu b, Junho Shin c, Kang-Sup Chung a,⁎, Young Ho Kim b,⁎a Mineral Resources Research Division, Korea Institute of Geoscience and Mineral Resources, Daejeon 305-350, Koreab Department of Fine Chemicals & Applied Chemistry, Chungnam National University, Daejeon 305-764, Koreac Department of Biological Engineering, Inha University, Incheon 401-751, Republic of Korea

⁎ Corresponding authors. Tel.: +82 42 868 3645.E-mail addresses: ksc@kigam.re.kr (K.-S. Chung), yh_k

http://dx.doi.org/10.1016/j.hydromet.2014.11.0050304-386X/© 2014 Published by Elsevier B.V.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 24 April 2014Received in revised form 7 November 2014Accepted 12 November 2014Available online 18 November 2014

Keywords:Lithium recoveryLithium adsorbentElectrostatic field assistanceMCDI

In this study, we suggested a novel recovery system for lithium using electrostatic field assistance (EFA) toimprove the conventional adsorption method, which requires a long adsorption time. To verify our suggestedconcept, a lithium selective electrode was prepared using a lithium selective adsorbent and hydrophilic PVAbinder, and a test set was fabricated by modifying the conventional membrane capacitive deionization (MCDI)system. The adsorption performance by EFA was compared to that of the conventional adsorption process. Theresults showed that the adsorption performance by EFA increased with increased cell potential up to 1.0 V andinitial lithium concentrations. The adsorption performance for lithium by EFA was superior to that byphysisorption in the range of initial lithium concentration tested. The adsorption time to reach an equilibriumstate by EFA was reduced compared to that of physisorption, which showed the acceleration effect of EFA. Thesuggested system also showed good reproducibility and durability during the repeated adsorption/desorptioncycles. This system could be an alternative recovery system for lithium from aqueous solution containing lithiumions.

© 2014 Published by Elsevier B.V.

1. Introduction

Lithium is a crucial resource used in the raw materials of secondarybatteries, light aircraft alloys, catalysts and fuel for nuclear fusion reac-tions (Dang and Steinberg, 1978; Epstein et al., 1981; Hartley et al.,1978). It is expected that the exhaustion of lithium resources due tothe rapid increase in its demand will act as an obstacle for human civi-lization in the near future. There have been many efforts to develop re-covery technology for lithium fromvarious sources, such as ore, salt lakeand seawater (Chitrakar et al., 2001). Of these, the recovery technologyof lithium by the adsorption method from seawater has received muchattention because seawater is a vast source containing approximately2.5 × 1014 kg of lithium, although the average lithium concentration islow (0.17 mg/L) (Abe et al., 1985; Chitrakar et al., 2001; Miyai et al.,1988).

Inorganic adsorbents with high selectivity for lithium ions in aque-ous solution have been investigated, of which spinel-type lithiumman-ganese oxides, such as LiMn2O4, Li1.33Mn1.67O4, and Li1.6Mn1.6O4, areconsidered to be promising lithium-selective adsorbents (Ammundsenet al., 1995; Feng et al., 1992; Ooi et al., 1990; Sagara et al., 1989;Wang et al., 2006). However, there are drawbacks to overcome for thepractical use of the lithium recovery technology using these adsorbentsbecause the recovery of the powder type adsorbent after the applicationis difficult and the adsorption process demands a long time, from a few

im@cnu.ac.kr (Y.H. Kim).

days to a few weeks, when the concentration of lithium is low. In addi-tion, the use of toxic agents such as hydrochloric acid is inevitable in thedesorption process. This may lead to environmental problems and a de-crease in the recovery efficiency because the addition of a large amountof alkali-solution to adjust the pH in the following separation step of thelithium ions is required after the desorption process (Shi et al., 2011).

In our previouswork (Ryu et al., 2013a),we suggested an alternativedesorption process for lithium ions using EFA without acidic solutionafter the adsorption reaction of the lithium ions in which a modifiedMCDI technology with the adsorbent membrane was applied. UsingEFA, the required time in the desorption process decreased, and the pro-cedure was simplified. However, a drawback in this system with a con-ventional adsorption process by ion-exchange reaction remains due tothe slow adsorption reaction. To accelerate this adsorption reaction,we also applied EFA in the adsorption process. In the proposed adsorp-tion method, the adsorption reaction of lithium ions can be acceleratedby the enrichment of lithium ions onto a lithium adsorbent layer due toEFA, as depicted in Fig. 1. Other ions that are not adsorbed onto the ad-sorbent layer are removed by flushing them with distilled water underno EFA. Finally, the concentrated lithium solution is obtained by desorp-tion with EFA, as mentioned in our previous work (Ryu et al., 2013a).

The objective of this study is to verify the feasibility of the proposedconcept. A spinel type of LiMn2O4 was chosen as the lithium-selectiveadsorbent. A test set was fabricated by modifying that of conventionalMCDI system, in which, the modified test set consisted of the carbonelectrode, adsorbent electrode, anion exchangemembrane and a dielec-tric spacer. The effect of various experimental parameters such as the

Fig. 1. The conceptual diagram of the proposed lithium recovery process using EFA: (a) current collector, (b) lithium-selective adsorbent layer, (c) anion exchange membrane, and(d) activated carbon layer.

79T. Ryu et al. / Hydrometallurgy 151 (2015) 78–83

initial concentrations of the lithium ions and the applied cell potentialswere examined. The experimental results were compared to thatof the conventional adsorption process without EFA under identicalconditions.

2. Experimental methods

2.1. Preparation of the electrodes

A spinel-type lithium manganese oxide (LMO, LiMn2O4) was pre-pared by a conventional solid-state reaction using Li2CO3 and MnCO3

as reactants (Ryu et al., 2013b). The resulting LMOpowderwas then im-mersed and stirred in a 0.5mol/L HCl solution to extract the lithium ionsby ion exchange reaction, which yielded the lithium-selective adsor-bent. Polyvinyl alcohol (PVA, Mw 31,000–50,000, 98–99% hydrolyzed,Sigma-Aldrich) was chosen as a binder for the adsorbent powderbecause it is known as a typical hydrophilic polymer for facilitatingthe migration of ions onto the surface of electrodes in aqueous solutionand because the cross-linked PVA polymer is not dissolved by water(Park and Choi, 2010; Yeom and Lee, 1996). PVA solution was preparedby dissolving PVA powder in deionized water. The adsorbent powder(particle size b2 μm) was mixed with the PVA solution to give a totalbinder content of 15 wt.%. Glutaraldehyde (GA, 25 wt.% solution inwater, Samchun Chemical), a cross-linking agent, was added into themixture of adsorbent powder and PVA solution and was vigorouslystirred at room temperature for 1 h to obtain a homogeneous slurry.The resulting slurry was then coated onto a graphite sheet (150 μmthickness, Dongbang Carbon Co.) of the adsorbent electrode with athickness of 250 μm using a doctor blade equipped with an automaticspeed controller and was then dried in a vacuum oven at 50 °C to re-move the water. HCl solution was sprayed over the dried electrode toaccelerate the esterification between PVA and GA. It was then uniformlypressed using a roll press. The resulting adsorbent layer was approxi-mately 200 μm thick, and the loaded amount of adsorbent on the graph-ite sheet was approximately 1.0 g per 100 cm2 (10 cm × 10 cm in size).To prepare the carbon electrode having a high capacitance as a cathodeof this system (Kim and Choi, 2010), the activated carbon powder(ACP, P-60, specific surface area = 1260 m2/g, Daedong AC Corp.)

and poly(vinylidenefluoride) (PVdF, Sigma-Aldrich) dissolved indi-methylacetamide (DMAc, Sigma-Aldrich) were mixed to give atotal binder content of 15 wt.% and coated onto the graphite sheetusing a doctor blade. The coated electrode was dried in a vacuumoven at 50 °C overnight to remove the organic solvent. After heatingthe dried electrode at 90 °C, it was uniformly pressed using a rollpress. The resulting carbon layer was approximately 200 μm thick, andthe loaded amount of activated carbon on the graphite sheet wasapproximately 0.9 g per 100 cm2 (10 cm × 10 cm in size).

The crystalline forms of the obtained adsorbents were determinedusing an X-ray diffractometer (XRD, D/MAX 2200, Rigaku), and themorphology was examined using analytical scanning electron micros-copy (SEM, S-4700, Hitachi).

2.2. Fabrication of the test cell

The test cell was fabricated bymodifying the conventional MCDI cellthat consists of carbon electrodes, an anion exchange membrane and adielectric spacer (Biesheuvel and Van der Wal, 2010; Kim and Choi,2010; Li et al., 2008). Fig. 2 shows a schematic diagram of the test celland the experimental device. The test cell consisted of four parts,i.e., the carbon electrode, adsorbent electrode, anion exchange mem-brane (Neosepta AMX, Tokuyama Soda Corp.) and a dielectric spacer(porous urethane foam, 200 μm thick). Each part was assembled be-tween two Plexiglas plates as follows: lower plate/carbon electrode/AEM/dielectric spacer/adsorbent electrode/upper plate. The resultingdistance between the two electrodes was approximately 150 μm dueto the flexibility of the dielectric spacer. The upper plate was designedto give two inlet holes positioned diagonally opposite at each cornerand one outlet hole at the center of the plate with an inner diameterof 2 mm so that solution can flow in contact with the inner sides ofthe electrodes. The solutionwas supplied to the test cell using a peristalticpump (Masterflex L/S, Cole-Parmer Co.). A cell potential was appliedusing a potentiostat device (Zive MP2, WonATech Corp.). The conductiv-ity of the effluent solution was automatically measured at a time intervalof 2 s using a conductivity meter (Vernier Software & Technology) with adata acquisition system. The ion concentration of the samples collected in

Fig. 2. Schematic diagram of the test cell and experimental device: (a) influent reservoir, (b) peristaltic pump, (c) test cell, (d) conductivity meter, (e) effluent reservoir, (f) computer, and(g) potentiostat.

80 T. Ryu et al. / Hydrometallurgy 151 (2015) 78–83

the solution was analyzed using an inductively coupled plasma atomicemission spectrophotometer (ICP-AES, Optima 7300D, Perkinelmer).

3. Results and discussion

In the preliminary experiments, the adsorption performance wastested to fix the adsorption time. In applying the cell potential, the con-ductivity of effluent solution in the general CDI cell rapidly decreaseswith time due to capacitive adsorption onto electrode surface and thesubsequent depletion of ions. Then the conductivitywith time increasesslowly as the sorption capacity became saturated. The further reductionin the conductivity occurs with the strength of applied cell potential(Lee et al., 2010).

In this system, as soon as LiOH solution was supplied to the cell, theion exchange reaction between Li+ and H+ in the adsorbent layer oc-curred slowly without applying the cell potential. In order to clarifythe acceleration effect of electrostatic field on the adsorption of Li+,the cell potential was simultaneously applied with supplying the LiOHsolution. This leads to gradual increase of conductivity as will be seenin Fig. 3 (period 1). The test was conducted at an applied cell potentialof 1.0 V, and aflow rate of 20mL/min using LiOH solution (initial lithiumconcentration of 50mg/L) inwhich the conductivity change in the efflu-ent solution was monitored. From the performance test, the conductiv-ity curve of the effluent solution reached a plateau state within 40 min,in which the concentration of the effluent solution reached an initiallithium concentration within 40 min indicating that the saturation of

0 1000 2000 3000 4000 5000 6000 7000 80000

200

400

600

800

1000

1200

1400

1600

1800

2000

Con

duct

ivity

(µS

/cm

)

Con

cent

ratio

n (p

pm)

(c) desorption(b) washing with distilled water

(a) adsorptionPeriod 3Period 2Period 1

Physisorption 1.0 V

Time (sec)

0

10

20

30

40

50

Fig. 3. The conductivity change and ICP determination of the effluent solution duringthe adsorption and desorption processes (initial lithium concentration: 50 mg/L, cellpotential: 1.0 V, and flow rate: 20 mL/min).

ions onto the electrodes occurred, as seen in Fig. 3. In the followingexperiments, the adsorption test with comparative experiments wasconducted at a fixed adsorption time of 40 min.

Fig. 3 shows the conductivity change of the effluent solution duringthe adsorption and desorption processes with and without EFA in theadsorption step. In addition, ICP determination of the effluent solutionby EFA was conducted to confirm the Li+ concentration change, andthe result was plotted in Fig. 3. The test was conducted in the followingthree steps. First, the LiOH solution was fed into the test cell for 40 minwith and without a cell potential of 1.0 V in which the adsorbent elec-trode was negatively charged and the carbon electrode was positivelycharged while the adsorption process of lithium ions proceeds (Period1). Second, the distilled water was supplied to the test cell to rinseaway the remaining lithium ions that were not adsorbed until the con-ductivity in the effluent solution reached that of distilled water (Period2). Finally, a cell potential was reversely applied to the test cell to posi-tively charge the adsorbent electrode and negatively charge the carbonelectrode under the flow of distilled water (Period 3). The adsorbedlithium ions were released and extracted into the distilled water inthis step.

The conductivity curve obtained with EFA (in Period 1) reached theplateau state slower than that by physisorption. This result can be ex-plained that Li+ ions passing through the test cell were captured ontothe adsorbent electrode due to capacitive adsorption by the appliedcell potential, which resulted in the ion depletion of the effluent solu-tion. Consequently, the conductivity increase in the effluent solutionup to the initial value of LiOH solution was delayed (approximately40 min) compared to that of physisorption (approximately 20 min). Inthe second step (Period 2), the lithium ions, which could be capturedin the pores or the interface between adsorbent particles and/or PVAbinder on the electrode, not adsorbed to the lattice position of thelithium-selective adsorbent need to be completely removed by flushingwith distilled water. As shown in Fig. 3 (Period 2), the conductivity ofeffluent solution decreased to that of distilled water within 40 min byflushing with distilled water. This result indicated that the evolutionpeak in Period 3 was due to the release of lithium ions adsorbed in thelattice of the adsorbent.

The comparison test with the conventional MCDI system, in whichthe adsorbent electrode was replaced to that manufactured with themixture of PVA and activated carbon powder without the adsorbenton the graphite sheet, was also conducted under identical adsorption/washing/desorption process. From the result, the detection level ofextracted lithium amount without the adsorbent (less than 5%) by ICPanalysis was negligible. This result confirmed that ions not beingadsorbed can be removed by washing with distilled water. The moni-tored conductivity curve reaching the plateau state was also checkedby ICP analysis, which indicated the initial lithium concentration and

0.5 1.0 1.5 2.00.0

0.5

1.0

2.0

2.5

3.0

3.5

4.0

5 ppm50 ppm500 pm

Ext

ract

ed a

mou

nts

(mg/

g)

Applied potential (V)

Physisorption

Fig. 5. Extracted lithium amount obtained by various initial Li+ concentrations with andwithout an applied cell potential in the adsorption process.

81T. Ryu et al. / Hydrometallurgy 151 (2015) 78–83

no lithium ion after the adsorption andwashing times (40min), respec-tively. Fig. 4 shows the conductivity change of the effluent solution andthe amounts of extracted lithium ions obtained by the desorption stepwith the applied potential from 0.5 V to 2.0 V in the adsorption stepand 3.5 V in the desorption step. The extracted amount of lithium(mg/g) in the desorption process was calculated by the mass of lithiumextracted (mg) over themass of applied adsorbent (g) in Period3. The ex-tracted amounts of lithium ions increased at the cell potential of up to1.0 V and slightly decreased at cell potentials of 1.5 V and 2.0 V inthe adsorption step at the initial lithium concentration of 50 mg/L.The maximum amount obtained with 1.0 V in the adsorption stepwas 1.36mg/g under identical desorption conditions. The extracted lith-ium amount at the relatively low cell potential of 0.5 V was larger thanthat by physisorption. This result indicates that EFA enhanced the ad-sorption reaction on the lithium-selective adsorbent, whichmight be at-tributed to the enrichment of lithium ions on the adsorbent electrode byelectrical force.

The initial concentration of ions in an aqueous solution is a crucialfactor in the adsorption reaction. As the initial lithium concentration in-creases, the amount of adsorbed lithium tends to increase. To evaluatethe effect of the initial lithium concentration on the adsorbed amountin our system, a series of experiments was conducted with varyinginitial lithium concentrations and cell potentials in the adsorption stepunder identical desorption conditions, as displayed in Fig. 5. Theincrease in the adsorbed lithium amount was observed from bothphysisorption and adsorption by EFAwith increasing initial Li+ concen-tration. This result could be caused by increase of Li+ diffusion rate intoadsorption site and pH of LiOH solution as the initial Li+ concentrationincreased. The pH of the initial lithium concentrations of 5 mg/L,50 mg/L, and 500 mg/L were 10.8, 11.6, and 12.6, respectively, whichis known to be a favorable pH range for ion-exchange reaction betweenLi+ and H+ (Wang et al., 2009). The amount of lithium adsorbed by EFAwas higher than that of physisorption in the range of initial Li+ concen-tration tested. The highest lithium amount adsorbedwas obtained at anapplied potential of 1.0 V in case of initial lithium concentrations of50 mg/L and 500 mg/L, and at 2.0 V in case of 5 mg/L. As shown inFig. 5, the lithium amount adsorbed tends to increase with appliedpotential up to 1.0 V. This result was attributed to the increase of Li+

concentration onto adsorbent electrode surface by capacitive adsorp-tion with increasing cell potential because the diffusion rate of Li+ iongenerally increases as the ion concentration increases. This phenome-non can be observed in the conventional MCDI process presented byKim et al., in which the removal amount of salt increased with theincrease of applied potential up to 1.5 V (Kim and Choi, 2010).

Fig. 4. The conductivity change of the effluent solution in the desorption step at a cell po-tential of 3.5 V obtainedwith a flow rate of 20mL/min and an initial lithium concentrationof 50 mg/L in the range of cell potentials from 0.5V to 2.0 V in the adsorption step.(a) Physisorption, (b) 0.5 V, (c) 1.0 V, (d) 1.5 V, and (e) 2.0 V.

Lee et al. reported that the pH decrease in the effluent solution oc-curs above the cell potential of 1.2 V in the conventional CDI systemdue to the electrochemical reaction (Lee et al., 2010). To check this inour system, the pH change was monitored in the range of applied cellpotential. However, it showed no significant difference compared withthat of physisorption, which implied that there might be no consider-able electrochemical reaction. Meanwhile, the amounts of adsorbedlithium slightly decreased at applied cell potentials above 1.5 V forinitial lithium concentrations of 50 mg/L and 500 mg/L. This slightdrop could not be explained by the electrochemical reaction, consider-ing no significant pH change in applying the cell potential. From thisresult, it seems that the cell potential of 1.0 V in the adsorption step isadequate for higher lithium concentration above 50 mg/L in this work.

The acceleration effect on the adsorption reaction with EFA was notclearly observed within one cycle of tested time (adsorption time of 40min). Thus, to verify this rigorously, we conducted the adsorption reac-tion until it reached an equilibrium state as follows: the LiOH solution(Li+ conc.: 50 mg/L) feeding into the test cell at a flow rate of 40 mL/min was circulated into the influent reservoir, a cell potential of 1.0 Vwas applied with a periodic cycle (on time of 40 min and off time of40 min), and the solution in the influent reservoir was collected whileno cell potential was applied in a periodic cycle. Fig. 6 shows the Li+

Fig. 6. Li+ uptake obtained with EFA and physisorption. (a) Cell potential of 1.0 V and(b) physisorption.

Fig. 8. XRD patterns of (a) adsorbent powder, (b) fabricated adsorbent electrodewith PVAbinder, (c) after adsorption test at a cell potential of 1.0 V, and (d) after repeated cycle testof five timeswith a cell potential of 1.0 V in the adsorption step and 3.5 V in the desorptionstep.

82 T. Ryu et al. / Hydrometallurgy 151 (2015) 78–83

uptake obtained with physisorption and EFA (1.0 V) in which the Li+

uptake (mg/g) was calculated by:

Liþ uptake mgLi=gadsorbentð Þ ¼ Co−Ceð ÞVm

ð1Þ

where Co and Ce are the initial and final Li+ concentrations (mg/L) inthe solution, respectively, V is the solution volume (L), and m is themass of the applied adsorbent (mg).

From the results, the difference in the Li+ uptake betweenphysisorption and adsorption by EFAwas observed until the adsorptionreaction reached an equilibrium state. The adsorption reactionwith EFAreached an equilibrium state faster (approximately 1000min) than thatby physisorption (approximately 3000 min). This implies an accelera-tion effect on the adsorption reaction by EFA. The pH (initial pH: 11.6)after reaching an equilibrium state obtained from the physisorptionand adsorption by EFA decreased to approximately 8.9 and 8.6, respec-tively. The pH decrease after reaching an equilibrium state is due toaccumulation of released H+ ions into the circulated solution, which isgenerally observed in the batch type system (Wang et al., 2009). Afterthe adsorption reaction reached an equilibrium state, the equilibriumlithium uptake of the adsorbent electrode was similar in both cases,which indicates that there was no significant difference in the maxi-mum adsorption capacity. The experiment was repeated three times,and the results were reproducible.

The repeated cycle testwas conducted to confirm the reproducibilityof the adsorption capacity and the durability of the adsorbent electrode,as displayed in Fig. 7. From the results, the conductivity curve in theeffluent solution by the adsorption and desorption processes was simi-larly observed for each repeated cycle. The extracted lithium amountincreased with repetition of up to three cycles and kept an almost con-stant value. The denoted values indicate the extracted lithium amount(mg/g) in the desorption step. The result indicates that the structureof the lithium adsorbent was maintained with good performance inspite of the electrical stimulation during the repeated adsorption anddesorption processes. The adsorbent electrodes before and after the re-peated adsorption/washing/desorption process were also examined byXRD analysis, as shown in Fig. 8. The XRD patterns of the adsorbent elec-trode before the test showed a typical spinel structure of LMO similar tothat of the adsorbent powder [JCPDS card no. 35-0782] and involved nophase transition by the use of the PVA binder, as shown in Fig. 8(a) and(b). The structure was well-maintained after the adsorption test andrepeated cycle test, as shown in Fig. 8(c) and (d). Considering the resultsof the cycle test and XRD analysis, it was revealed that the spinel

Fig. 7. The conductivity change of the effluent solution in the repeated adsorption/washing/desorption process (flow rate of 20 mL/min, applied cell potential of 1.0 V in the adsorptionstep and 3.5 V in the desorption step, and an initial lithium concentration of 50 mg/L ininfluent reservoir).

structure of the adsorbent was not damaged by the electrical stimula-tion in the adsorption and desorption processes by EFA. Fig. 9 showsthe SEM images of the adsorbent electrode before and after the repeatedcycle test of five times. As shown in Fig. 9(a) and (b), the adsorbentparticles were uniformly mixed with the PVA binder, and the three-dimensional pores were distributed throughout the surface of theadsorbent electrode. The morphology of the adsorbent electrode wasnot changed after the cycle test, as shown in Fig. 9(c) and (d). This resultindicates that the adsorbent electrode had good stability againstelectrical stimulation and hydraulic load during the adsorption anddesorption processes.

4. Conclusion

In this study, a novel recovery system for lithium by EFA to improvethe adsorption reactionwas proposed. To verify the feasibility of the im-proved adsorption process, an adsorbent electrode was prepared usinga PVA binder having hydrophilicity, and a test set was fabricated bymodifying the conventionalMCDI system. The adsorption test of lithiumions was conducted to examine the effect of various factors on theadsorption capacity. The results showed that the adsorbed amount oflithium by EFA increased with increasing cell potential up to 1.0 V butslightly decreased at potentials above 1.5 V. The adsorption reactionreaching an equilibrium state by EFAwas relatively improved comparedto that of the conventional adsorption method. The suggested systemalso showed good reproducibility and durability during the repeatedadsorption/desorption cycles. From the results, our suggested systemcould be an alternative recovery system for lithium from aqueous solu-tion containing lithium ions with good economic feasibility due to thereduction of adsorption and desorption times. This process also doesnot require the use of acidic solution in the desorption process. In addi-tion, the less amount of alkali-solution to adjust the pH in the followingseparation step of the lithium ions after the desorption process is con-sumed compared with the desorption process of conventional system.Further studies to improve the adsorption capacity and efficiency, andthe comparative study on the selectivity of lithium ions using salt lakeor Li-spiked seawater are underway in our research group.

Acknowledgments

This research was supported by the National Research Project titled“The Development of Technology for Extraction of Resources Dissolvedin Seawater” of the Korea Institute of Geoscience andMineral Resources(KIGAM), funded by the Ministry of Oceans and Fisheries.

Fig. 9. SEM images of the adsorbent electrode. (a) Cross-sectional image before the test, (b) surface image before the test, (c) cross-sectional image after the 5th cycle test, and (d) surfaceimage after the 5th cycle test.

83T. Ryu et al. / Hydrometallurgy 151 (2015) 78–83

References

Abe, M., Chitrakar, R., Tsuji, M., Fukumoto, K., 1985. Synthetic inorganic ion-exchangematerials XXXIX. Synthesis of titanium (IV) antimonates and their ion exchangeproperties for alkali and alkaline earth metal ions. Solvent Extr. Ion Exch. 3 (1–2),149–172.

Ammundsen, B., Jones, D.J., Roziere, J., Burns, G.R., 1995. Mechanism of proton insertionand characterization of the proton sites in lithium manganate spinels. Chem. Mater.7 (11), 2151–2160.

Biesheuvel, P., Van der Wal, A., 2010. Membrane capacitive deionization. J. Membr. Sci.346 (2), 256–262.

Chitrakar, R., Kanoh, H., Miyai, Y., Ooi, K., 2001. Recovery of lithium from seawater usingmanganese oxide adsorbent (H1.6Mn1.6O4) derived from Li1.6Mn1.6O4. Ind. Eng. Chem.Res. 40 (9), 2054–2058.

Dang, V.-D., Steinberg, M., 1978. Preliminary design and analysis of recovery of lithiumfrom brine with the use of a selective extractant. Energy 3 (3), 325–336.

Epstein, J., Feist, E., Zmora, J., Marcus, Y., 1981. Extraction of lithium from the Dead Sea.Hydrometallurgy 6 (3), 269–275.

Feng, Q., Miyai, Y., Kanoh, H., Ooi, K., 1992. Lithium (1+) extraction/insertion with spinel-type lithium manganese oxides. Characterization of redox-type and ion-exchange-type sites. Langmuir 8 (7), 1861–1867.

Hartley, J., Gore, B., Young, J., 1978. Potential lithium requirements for fusion powerplants. Energy 3 (3), 337–346.

Kim, Y.-J., Choi, J.-H., 2010. Enhanced desalination efficiency in capacitive deionizationwith an ion-selective membrane. Sep. Purif. Technol. 71 (1), 70–75.

Lee, J.-H., Bae, W.-S., Choi, J.-H., 2010. Electrode reactions and adsorption/desorptionperformance related to the applied potential in a capacitive deionization process.Desalination 258 (1), 159–163.

Li, H., Gao, Y., Pan, L., Zhang, Y., Chen, Y., Sun, Z., 2008. Electrosorptive desalination bycarbon nanotubes and nanofibres electrodes and ion-exchange membranes. WaterRes. 42 (20), 4923–4928.

Miyai, Y., Ooi, K., Katoh, S., 1988. Recovery of lithium from seawater using a new type ofion-sieve adsorbent based on MgMn2O4. Sep. Sci. Technol. 23 (1–3), 179–191.

Ooi, K., Miyai, Y., Katoh, S., Maeda, H., Abe, M., 1990. Analysis of pH titration data in a.lambda.-manganese dioxide + lithium hydroxide system on the basis of redoxmechanism. Langmuir 6 (1), 289–291.

Park, B.-H., Choi, J.-H., 2010. Improvement in the capacitance of a carbon electrode pre-pared using water-soluble polymer binder for a capacitive deionization application.Electrochim. Acta 55 (8), 2888–2893.

Ryu, T., Ryu, J.C., Shin, J., Lee, D.H., Kim, Y.H., Chung, K.-S., 2013a. Recovery of lithium byan electrostatic field-assisted desorption process. Ind. Eng. Chem. Res. 52 (38),13738–13742.

Ryu, T., Shin, J., Ryu, J., Park, I., Hong, H., Kim, B.G., Chung, K.-S., 2013b. Preparation andcharacterization of a cylinder-type adsorbent for the recovery of lithium from seawa-ter. Mater. Trans. 54 (06), 1029–1033.

Sagara, F., Ning, W.B., Yoshida, I., Ueno, K., 1989. Preparation and adsorption propertiesof λ-MnO2-cellulose hybrid-type lon-exchanger for lithium ion. Application to theenrichment of lithium ion from seawater. Sep. Sci. Technol. 24 (14), 1227–1243.

Shi, X., Zhou, D., Zhang, Z., Yu, L., Xu, H., Chen, B., Yang, X., 2011. Synthesis and propertiesof Li1.6Mn1.6O4 and its adsorption application. Hydrometallurgy 110 (1), 99–106.

Wang, L., Ma, W., Liu, R., Li, H.Y., Meng, C.G., 2006. Correlation between Li+ adsorptioncapacity and the preparation conditions of spinel lithium manganese precursor.Solid State Ionics 177 (17), 1421–1428.

Wang, L., Meng, C.G., Ma, W., 2009. Study on Li+ uptake by lithium ion-sieve via the pHtechnique. Colloids Surf. A Physicochem. Eng. Asp. 334 (1), 34–39.

Yeom, C.-K., Lee, K.-H., 1996. Pervaporation separation of water-acetic acid mixturesthrough poly (vinyl alcohol) membranes crosslinked with glutaraldehyde. J. Membr.Sci. 109 (2), 257–265.