Chemical Engineering Journal 348 (2018) 1000–1011

Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

Li1−xNi0.33Co1/3Mn1/3O2/Ag for electrochemical lithium recovery from T

brine
⁎
Chosel P. Lawagon1, Grace M. Nisola1, Rosemarie Ann I. Cuevas, Hern Kim, Seong-Poong Lee ,
⁎
Wook-Jin Chung
Energy and Environment Fusion Technology Center (E2FTC), Department of Energy Science and Technology (DEST), Myongji University, Myongjiro-116, Cheoingu, Yongin
City, Gyeonggi Province 17058, Republic of Korea

H I GH L IG H T S

• LiAt reverse
Ni Co Mn O (NCM)/Ag electrochemically captured Li from brine.
1−x 1/3 1/3 1/3 2
+

• NCM/Ag energyapplied current (i), NCM/Ag released LiCl in recovery solution.

• LNCM/Ag operation consumption was minimized at varied i, time, and feed [Li ]. +

• LNCM/Ag can significantly +

can be cycled to produce 96% Li using 2.60 W·h mol Li −1 +
.
• +
reduce operation time for Li mining from brines.

A R T I C LE I N FO A B S T R A C T

Keywords: With the heightened global demand for lithium, more energy-efficient processes with fast Li+ production rates
Brine are essential for sustainable Li+ supply. Electrochemical Li+ recovery is a promising method that could satisfy
Electrochemical these process necessities. However, its success requires highly effective electrodes that could selectively capture
Lithium Li+ at a competitive uptake capacity with minimal energy requirement. Delithiated Li1−xNi1/3Co1/3Mn1/3O2
LiNi1/3Co1/3Mn1/3O2
(NCM) paired with silver (Ag) is introduced as a new electrochemical system for Li+ recovery from brine. NCM is
Recovery
Silver
characterized by its high Li+ selectivity and stability in aqueous environment. At an applied current in brine,
NCM was able to intercalate Li+ into its lattice while the Ag captured the Cl− counter-ion. Reversal of the
current in a receiving solution prompted the release of LiCl. Under optimal conditions (i = ± 0.25 mA cm−2),
NCM can produce 96.4% pure Li+ from brine by expending 2.60 W·h mol−1 Li+. The NCM/Ag was able to
perform consistently and produce an enriched LiCl solution in cycled operations. These promising results in-
dicate that NCM/Ag can be developed as a high-throughput Li+ mining process with low energy requirement.

1. Introduction
Recent technological developments in electric powered vehicles
have heightened the demand for lithium [1–4]. In just over a year, the
Li market for energy related applications has surged from 39% to 46%
[5,6]. This trend raises concern on the sustainability of Li supply in the
future. Approximately 80% of Li comes from lime soda evaporation/
precipitation ponds of brine pools [7]. However, this process is tedious,
requires months to complete, generates secondary waste, and is affected
by many external factors including meteorological patterns [8]. Other
available methods include Li+ adsorption and ion-exchange using li-
thium ion sieves (LIS) [9,10]. Although LIS such as H1.6Mn1.6O4,
H1.33Mn1.67O4, λ-MnO2, and H2TiO3 have moderate to high Li+ uptake
capacities and acceptable selectivities, their fabrication must be well-
designed and executed to prevent capacity, uptake rate, and selectivity
losses [9–16]. With these current limitations, more practical and energy
efficient Li+ recovery approaches must be developed.
Electrochemical techniques have proved to be effective for Li+ se-
questration from brine and seawater as they can significantly increase
Li+ production rate with low expended energy [17–21]. The earliest
report was performed using a λ-MnO2/Pt electrode wherein λ-MnO2
selectively intercalated Li+ at a maximum insertion capacity of
Q = 11 mg g−1 [22]. With the progress in desalination battery and
mixing entropy batteries [23,24], a recently developed LiFePO4/Ag

⁎
Corresponding authors.
E-mail addresses: spleemju2012@gmail.com (S.-P. Lee), wjc0828@gmail.com (W.-J. Chung).
1
Co-first authors.

https://doi.org/10.1016/j.cej.2018.05.030
Received 14 March 2018; Received in revised form 4 May 2018; Accepted 5 May 2018
Available online 07 May 2018
1385-8947/ © 2018 Elsevier B.V. All rights reserved.
C.P. Lawagon et al. Chemical Engineering Journal 348 (2018) 1000–1011

Fig. 1. Conceptual diagram of the NCM/Ag electrochemical process for Li+ recovery. Reactors were not drawn to scale: Vf = 40 mL simulated brine; Vr = 10 mL
30 mM LiCl operated at room temperature.

Fig. 2. Electrode characterization results: (a) LNCM XRD patterns, (b) SEM image (inset: particle size histograms), (c) EDS mapping, and (d) N2 adsorption/
desorption isotherms of LNCM.

(LFP) cell facilitated Li+ insertion in FePO4 via Fe(III) reduction and
counter ion Cl- capture via Ag oxidization [17,18]. By reversing the
applied current, the Li+ and Cl− were released into a fresh receiving
electrolyte solution; repetitive cycles eventually produced a LiCl con-
centrate. Another similar system has used λ-MnO2/Ag [18], whereas a
slightly modified assembly used activated carbon (AC) in lieu of Ag
[25].
Olivine LiFePO4 (LFP), spinel LiMn2O4, and spinel λ-MnO2 have
been routinely reported as Li+ capturing electrodes mainly due to their
Li+ selectivity and compatible operating voltage window in aqueous
environment [7,17–19,21,22,25–27]. However, the long-term use of
these electrodes is still unknown. Meanwhile, λ-MnO2 could be sensi-
tive in different pH environments; it may produce Mn2+ and MnO4- as
disproportionation products, which may leach out of the electrode
[26–28]. These drawbacks highlight the need to find more suitable
electrode materials for selective electrochemical Li+ sequestration.
More importantly, there is still a wide room for system improvements to
achieve lower energy consumption and higher Li+ production rate.
Depending on the Li+ source, electrodes, and operating conditions,
molar energy requirements of 1–33 W·h mol−1 Li+ have so far been

1001
C.P. Lawagon et al. Chemical Engineering Journal 348 (2018) 1000–1011

Fig. 3. Cyclic voltammogram (scan rate 0.5 mV/s, room temperature) of (a) LNCM in 1 M LiCl (Pt as counter, Ag/AgCl as reference); (b) NCM in 1 M LiCl, in mixed
solution (1 M each Mn+ = Na+, Mg2+, K+, Ca2+) with and without 1 M LiCl (Pt as counter, Ag/AgCl as reference); (c) NCM/Ag in 5 M LiCl (Ag as counter, Ag/AgCl
as reference).

reported [17–19,21,22,25,26].
LiNi1/3Co1/3Mn1/3O2 (LNCM) has long been regarded as a promising
material for next-generation lithium ion batteries. It provides high
theoretical capacity (275 mAh g−1) as well as fast and excellent charge/
discharge properties due to its pseudo-zero volume change when cycled
at 2.5–4.4 V (vs. Li/Li+) [29,30]. LNCM has a working potential within
the thermodynamic stability window of aqueous electrolyte (< 1.23 V)
[31]. It has been used in aqueous supercapacitors [31] and aqueous
lithium ion batteries [32–34], but never tested for its ability as Li+
capturing electrode from brines or aqueous streams.
To warrant an excellent electrochemical Li+ recovery performance,
a suitable counter electrode must be paired with LNCM. Materials such
as Ag, Pt, nickel hexacyanoferrate (NiHCF), reversible I-/I3- redox
couple, and polypyrrole reversible Cl- electrodes have been effectively
paired with Li+ capturing cathodes [7,26,27]. But in this report, LNCM
is paired with Ag, due to its well-known selective redox mechanism
with Cl− during Li+ capture and release in aqueous phase [17,18].
Furthermore, Ag reaction with chloride is selective and stable in aqu-
eous solution [17–19].
Herein, a new electrochemical Li+ recovery system is introduced.
As Li+ capturing electrode, LNCM was pre-delithiated as Li1−x[Ni1/
3Co1/3Mn1/3]O2 (NCM) and then paired with Ag electrode to construct
the NCM/Ag cell. To elucidate the mechanisms involved during Li+
insertion and recovery, NCM/Ag was thoroughly characterized in terms
of its textural, surface, morphological, and electrochemical properties.
The performance of the NCM/Ag cell was assessed, optimized, and
compared with other known Li+ capturing electrode assemblies in
terms of Li+ insertion capacity, selectivity, power consumption,

1002
C.P. Lawagon et al.
Fig. 4. NCM/Ag performance at varied i or C-rate (Cf = 0.03 M Li+ with equimolar Mn+, operated for 20 min/step): (a) NCM potential vs Ag/AgCl, (b) Ag potential
vs Ag/AgCl, (c) ΔE time profile, and (d) ΔE vs. q plots for estimation of energy consumption.
recyclability, and product purity.
2. Experimental section
2.1. Materials
Lithium nitrate (+99% LiNO3, Acros Belgium), nickel(II) nitrate
hexahydrate (+97% Ni(NO3)2·6H2O, Daejung Chemicals & Metals Co.,
Ltd. Korea), cobalt(II) nitrate hexahydrate (+97% Co(NO3)2·6H2O,
Daejung Chemicals & Metals Co., Ltd. Korea), and manganese (II) ni-
trate tetrahydrate (+97.5% Mn(NO3)2·4H2O, Acros-Organics USA)
were used as LiNi1/3Co1/3Mn1/3O2 (LNCM) precursors. Polyvinylidene
fluoride (Solef 6010 PVDF, Solvay Korea) was used as a binder and N-
methyl pyrrolidone (99.5% anhydrous NMP, Sigma Aldrich USA) as its
solvent. Super P Carbon (Super C65, Timcal Switzerland) and Titanium
(Ti) foil (99.7% annealed) were used as conducting agent and current
collector for the LNCM electrode, respectively. Silver (99.9% Ag,
0.5–1 µm) was purchased from Alfa Aesar (USA). Nitric acid (60% RHM
HNO3, Junsei Chemical Co., Ltd. Japan) was used in acid digestion
pretreatment for elemental analysis. Lithium chloride (99% LiCl, Acros
Belgium), sodium chloride (99% NaCl, Daejung Chemicals & Metals
Co., Ltd. Korea), magnesium chloride hexahydrate (98% MgCl2·6H2O,
Alfa Aesar USA), and potassium chloride (99.995% KCl, Alfa Aesar
USA) were used for the preparation of simulated brine-like solutions for
electrochemical Li+ recovery experiments. Considering the difficulty of
collecting real samples overseas, simulated brines were prepared based
on the Salar de Uyuni (Bolivia) salt lake containing 234.8 mM Li+,
2566.4 mM Na+, 1208.3 mM Mg2+, 478.3 mM K+, and 5.7 mM Ca2+
[35]. All chemicals were used without further purifications.
2.2. Electrode fabrication
LNCM was synthesized through sol-gel route followed by calcina-
tion as previously reported [36]. Metal nitrate precursors were
1003
Chemical Engineering Journal 348 (2018) 1000–1011
dissolved in deionized (DI) water at stoichiometric proportions. Water
was then slowly evaporated while stirring at 80 °C resulting in a viscous
purple solution. Nitrate ions were partially removed by increasing the
temperature to 120 °C for 8 h followed by drying the sample in an oven
at 120 °C for 12 h. The resulting solid was ground, calcined (5 °C min−1,
900 °C, 12 h), cooled to room temperature, and granulated by slow
mixing in agate mortar and pestle. Carbon black and PVDF (80:10:10 wt
%) were mixed in NMP with either LNCM or Ag to form a slurry. The
mixture was coated on Ti foil with an effective area of 2 cm2. The
electrodes were oven-dried (90 °C, 12 h) to remove NMP, with mass
loadings of c.a. 2–3 mg cm−2 and 3–4 mg cm−2 of active material for
LNCM and Ag electrodes, respectively. Prior use for Li+ capture, LNCM
electrode was oxidized by charging the material at 1C-rate (i.e.
0.3–0.45 mA for 1 h at 93 mA·h g−1 NCM capacity). On the other hand,
Ag was reduced until −0.150 V and then charged with 2 mA for 15 min.
Mass ratio of the prepared LNCM and Ag electrodes was close to 1:1.5
[26].
2.3. Electrode characterization
The textural property of LNCM as powder and electrode (i.e. before
and after use) was examined on an X-ray diffractometer with Cu Kα
source (PANalytical X’pert-Pro, The Netherlands; 40 kV, 30 mA, 0.03
step count−1). Its morphology was observed under Field Emission
Scanning Electron Microscope equipped with Energy Dispersive X-ray
Spectrometer (FESEM-EDX, Hitachi SU-70, Japan). Nitrogen adsorp-
tion/desorption isotherms were carried out in Belsorp-mini II (Bel
Japan, Inc.) at relative pressure range (p/po) between 0.01 and 1.0. The
surface area of the samples was estimated using the Brunauer-Emmett-
Teller (BET) model. Actual LNCM content of 85.2 wt% (ESI Fig. S1) was
confirmed thermo-gravimetrically (Mettler Toledo, DSC823e) by
heating the sample from 30 °C to 1000 °C at 10 °C min−1 in air (flow
rate = 100 mL min−1).
C.P. Lawagon et al. Chemical Engineering Journal 348 (2018) 1000–1011

2.5. Electrochemical lithium recovery

The three-electrode system composed of NCM/Ag with Ag/AgCl as

+
Fig. 5. NCM/Ag performance at varied i or C-rate (Cf = 0.03 M Li with
equimolar Mn+, operated for 20 min/step): (a) Li+ insertion capacity (Q) and
purity of Li+ (% PLi) in the receiving solution and (b) current efficiency (% η)
based on the total q input and q used for Li+ capture.
2.4. Electrochemical characterization
Cyclic voltammetry (CV) and galvanostatic charge-discharge tests
were monitored by a battery cycler (WBCS3000, Won-A-Tech Korea).
Galvanostatic charge and discharge tests were performed in a solution
of 1 M LiCl wherein the LNCM electrode was paired with Ag counter
electrode and Ag/AgCl (saturated KCl) reference electrode. The CV
experiments (25 °C) for LNCM or NCM in pure 1 M LiCl were carried out
using Ag/AgCl (saturated KCl) as the reference- and Pt as the counter
electrodes. Additional CV of NCM was carried out in a mixture of
aqueous electrolytes (i.e. 1 M of each Mn+ = Na+, Mg2+, K+, Ca2+),
with and without 1 M Li+. Finally, the CV of NCM paired with Ag
counter electrode was performed in 5 M LiCl using Ag/AgCl as re-
ference electrode.
reference was assembled (Fig. 1) in a glass reactor with feed volume
Vf = 40 mL, operated at 25 °C. The applied current ( ± i) was regulated
by a battery cycler (WBCS3000, Won-A-Tech Korea) and no voltage
limitation was imposed in the system. A negative current was first ap-
plied in the brine to facilitate Li+ insertion in the NCM lattice and to
capture Cl− at the Ag electrode (step 1). After a period of time, the
electrode assembly was quickly immersed in DI water to remove the
residual brines and then immediately placed in a smaller reactor with
volume Vr = 10 mL containing 30 mM LiCl electrolyte receiving solu-
tion. In LiCl solution, a positive current is applied to extract the Li+
from NCM and desorb the Cl− from Ag (step 2). The applied current in
both steps (i = ± 0.50–1.5 mA) and the operation time
(t = 10–30 min/step) were varied to maximize Li+ recovery at the
lowest energy consumption.
The Li+ insertion capacity (Qf) of the NCM electrode was calculated
using Eq. (1) by measuring its initial (Co) and final (Cf) concentrations
in the brine with a known Vf and mass of NCM (m). For more accurate
values, Li+ insertion capacities (Q) were reported based on the Li+
concentration in the receiving solution (Cr) using Eq. (2) at a known Vr.
The Qf values were used for selectivity calculations (Eqs. (3) and (4))
whereas Q values were reported as actual Li+ insertion capacity. The
Li+ selectivity of the NCM electrode was estimated (Eq. (3)) in terms of
αLi/M which pertains to the ratio of the distribution coefficients (KD) of
Li+ with respect to other cations Mn+. Meanwhile, the KD values of all
cations were determined using Eq. (4).
(Co−Cf )
Qf = × Vf
m (1)
Cr × Vr
Q=
m (2)
KDLi
αMLi =
KDM (3)
Qf
KD =
Cf (4)
The purity of Li+ (% PLi) was calculated as the percentage (%) of Cr
over total cation concentrations (ΣCi) in the receiving solution (Eq. (5))
[21]. Molar energy consumption or W (W·h mol−1 Li+) of recovered
Li+ (Cr × Vr) was estimated from the circular integral of the voltage
profile (ΔE) vs total charge flow, q (Eq. (6)).
Cr
PLi (%) = × 100
∑ Ci (5)
∮ ΔE ·dq
W=
Cr × Vr (6)

The current efficiency (η) of NCM/Ag was calculated using Eq. (7)
where F is Faraday’s constant (F = 96,485 C mol−1) and q is the total
charge flow during Li+ capture [19].

Table 1
NCM/Ag separation performance and molar energy consumption at varied i or C-rate (Cf = 0.03 M Li+ with equimolar Mn+, operated for 20 min/step).
Applied current ± i (mA) 0.25 0.50 0.75 1.0 1.25 1.5

KDLi (mL g−1) 19.42 40.09 48.93 53.81 85.22 95.26

Li
αNa 128.75 159.50 144.33 156.10 42.49 32.11
Li
αMg 79.19 130.80 135.28 125.06 18.71 16.75

αKLi 114.55 189.21 170.34 185.18 100.86 88.44

Li
αCa 186.86 415.73 274.11 387.66 185.57 213.48
W (W·h mol−1 Li+) 10.88 8.22 23.88 29.62 61.38 84.31

1004
C.P. Lawagon et al.
Fig. 6. NCM/Ag performance at varied operation t (Cf = 0.03 M Li+ with equimolar Mn+, i = ± 0.5 mA or 1.13 C): (a) NCM potential vs Ag/AgCl, (b) Ag potential
vs Ag/AgCl, (c) ΔE time profile, and (d) ΔE vs. q plots for estimation of energy consumption.
FCr × Vr
ηLi = × 100%
q (7)
For repetitive Li+ capture and recovery, the separation factors (SF)
of Li+ with respect to other Mn+ were calculated from the ratio of Li+
and Mn+ concentrations (Eq. (8)) at the receiving LiCl solutions (r) and
brine (f) [13].
(C Li+/ C M n + )r
SF =
(C Li+/ C M n + )f (8)
2.6. Analytical methods
Elemental analysis of simulated brine and receiving solution were
carried out using inductively coupled plasma mass spectrometry (ICP-
MS Agilent 7500 series, USA). The samples were pre-treated using
microwave-assisted (MARS-5 CEM, USA) acid digestion. Prior pre-
treatment, the liquids were filtered (0.2 μm Nylon syringe filters) and
3 mL HNO3 was added in each 5 mL sample. Pre-treated samples were
transfer-filtered in polypropylene volumetric flasks (100 mL) and di-
luted with DI water followed by ICP-MS analysis.
3. Results and discussions
3.1. Electrode characterization
The XRD patterns of LNCM powder and LNCM from electrode
(Fig. 2a) can be readily indexed to hexagonal α-NaFeO2 structure with
R3 m space group (JCPDS #09-0063) [36,37]. Both samples had similar
lattice constants a = 2.99 Å and c = 14.2 Å, similar with those reported
in literature [38,39]. The synthesized LNCM is also comparable to
commercially available LNCM showing good splitting of miller indices
at (0 0 6/1 0 2) and (1 0 8/1 1 0), typical for a layered structure [40,41].
Integrated intensity (I) ratio for I(003)/I(104) peaks is > 1.2, confirming
1005
Chemical Engineering Journal 348 (2018) 1000–1011
low cation mixing and good crystalline ordering in the hexagonal layers
[41,42]. This feature indicates excellent electrochemical performance
of LNCM. SEM image (Fig. 2b) reveals that the particle size of LNCM
powder ranged between 200- and 500 nm with an average diameter of
dave = 382.13 nm, similar with the scraped LNCM from electrode (ESI
Fig. S2) with dave = 383.03 nm. SEM-EDS mapping of the LNCM
powder (Fig. 2c) and LNCM electrode (ESI Fig. S3) showed that the
prepared material has uniform distribution of Ni, Co, and Mn. N2 ad-
sorption/desorption isotherms (Fig. 2d) of scraped LNCM exhibits
minor hysteresis loop at p/po = 0.90 in contrast to the pristine powder
[11–14]. The mesoporosity was probably imparted by the PVDF binder
and carbon black. This was reflected by the increased BET surface area
(As) of the LNCM from the electrode (As = 9.29 m2 g−1) as compared to
the pristine sample (As = 1.79 m2 g−1), which has similar value (ESI
Fig. S4) with Ag electrode (As = 1.66 m2 g−1).
3.2. Electrochemical properties LNCM/Ag and NCM/Ag
Cyclic voltammetry (Fig. 3a) in aqueous electrolyte solution (1M
LiCl) of pristine LNCM (vs. Ag/AgCl) reveals anodic (Epa) and cathodic
(Epc) peaks (0.6–0.7 V) that correspond to the reversible redox of Ni2+/
Ni4+ [37] during Li+ intercalation and de-intercalation, respectively
[32,34,41]. The first scan showed a slightly different redox pair (Epa/
Epc = 0.67 V/0.60 V) probably due to the initial restructuring of the
LNCM lattice [43]. Subsequent cycles were stable (Epa/Epc = 0.64 V/
0.61 V), with a very narrow voltage gap < 0.07 V. This result indicates
reversibility [43] and fast Li+ intercalation and de-intercalation process
across the electrode [44]. Galvanostatic charging and discharging test
of LNCM/Ag yielded stable capacities of 93 mA·h g−1 [34], with ∼99%
coulombic efficiency (ESI Fig. S5). This further establishes the struc-
tural stability and reliability of LNCM/Ag system for Li+ intercalation
and de-intercalation [37,44].
As Li+ capturing electrode, LNCM was initially delithiated at 1C-
rate to obtain Li1−x[Ni1/3Co1/3Mn1/3]O2 or NCM prior CV experiment,
C.P. Lawagon et al. Chemical Engineering Journal 348 (2018) 1000–1011

irreversible reduction of Mn4+ [45,46], but this had no effect on the

Li+ capture/release mechanism of NCM. The selective Li+ intercalation
in NCM is primarily attributed to its lower hydration energy and faster
diffusion through the electrode [25]. Moreover, NCM exhibits ion
sieving ability due to Li+ templating effect. Compared to other cations
(ionic radii: Li+ = 76 pm vs. Mg2+ = 72 pm; Ca2+ = 100 pm;
Na+ = 102 pm; K+ = 138 pm), smaller Li+ from the electrolyte solu-
tion can easily insert at the Li+-vacated space of NCM lattice. Larger
cations (Na+, K+, Ca2+) would have slower electrode diffusions [8,19]
than Li+ due to steric hindrance effects. Although Mg2+ has a very
close ionic size to Li+, its higher dehydration energy requirement
makes its NCM insertion more energetically difficult [15,35].
The same CV redox peaks were observed when NCM was paired
with Ag electrode (Fig. 3c). This further demonstrates that Li+ capture
and release in aqueous electrolyte solution can be achieved using NCM/
Ag cell. The absence of peaks from the CV of Ag (Fig. 3c) confirms its
suitability as NCM counter electrode since it selectively reacts with
chloride and is stable in aqueous environment [17–19]. The two-step
Li+ electrochemical recovery process could therefore be described by
Eq. (10), which involves Li+ and Cl− capture (forward) from the brine
and subsequent release of these ions as LiCl as recovery step (backward)
in a receiving electrolyte solution.

Li1 − x [Ni1/3Co1/3 Mn1/3]O2 + x Li+ + x e− + x Ag

+ x Cl− ⇌ Li[Ni1/3Co1/3 Mn1/3]O2 + x AgCl (10)

3.3. Electrochemical Li recovery from brine using NCM/Ag

The electrochemical Li+ recovery process from brine involves the

capture of Li+ by NCM and Cl- by the Ag electrode when a negative
Fig. 7. NCM/Ag performance at varied operation t (Cf = 0.03 M Li+ with
current is applied at the cathode (Fig. 1). Retrieval of Li+ is then carried
equimolar Mn+, i = ± 0.5 mA or 1.13 C): (a) Li+ insertion capacity (Q) and out in a receiving electrolyte, wherein LiCl release is facilitated with
purity of Li+ (% PLi) in the receiving solution and (b) current efficiency (% η) applied reverse current. In this study, the limiting diffusion current (ilim)
based on the total q input and q used for Li+ capture. of Li+ in brine was estimated (Eq. (11)) where n is the number of
electrons, F is Faraday constant, A is the geometric area of the elec-
trode, D is the Li+ diffusion coefficient and δ as the diffusion layer
Table 2
thickness (c.a. 100 μm) [17,24,34]. The obtained ilim = 2.56 mA was
NCM/Ag separation performance and molar energy consumption at varied
operation t (Cf = 0.03 M Li+ with equimolar Mn+, i = ± 0.5 mA or 1.13 C). significantly higher than the tested current range ( ± 0.25–1.5 mA),
which indicates that Li+ intercalation in NCM was not diffusion-limited
Time (min) 10 15 20 25 30 [17,21,24,26]. However, the main challenge is to optimize the process
conditions of the electrochemical system that will lead to the lowest
KDLi (mL g−1) 15.97 23.08 41.60 51.56 52.49
Li 151.72 164.49 212.69 263.58 147.08
possible energy consumption and high Li+ purity in the receiving so-
αNa
Li 65.00 78.30 94.91 94.10 61.77
lution.
αMg
αKLi 149.51 216.13 279.46 346.33 193.25 Cf
ilim = nFAD
Li
αCa 288.71 417.35 674.57 835.98 310.86 δ (11)
W (W·h mol−1 Li+) 9.68 7.82 8.38 9.46 12.11
The process was tuned by controlling the total charge input/output
(q) of the NCM/Ag cell in terms of applied current (i or C-rate) and
where x is the delithiated fraction of Li+. The textural property (ESI Fig. operation time (t). The effect of feed characteristics on the performance
S6) of NCM is characterized by the disappearance of (0 0 6) peak and of NCM/Ag was also investigated to determine its suitability to brine or
the shifted peaks (0 0 3), (1 0 8) and (1 1 0), which caused a slight de- other aqueous Li+ resources.
crease in a and an increase in c values [38,45]. Redox peaks of NCM
(Fig. 3b) in 1 M pure LiCl were slightly broader than those from LNCM 3.3.1. Lithium recovery at varied applied currents
(Fig. 3a) but were still within the expected range (Epc = 0.587, Initial runs were carried out at varied i = ± 0.25–1.5 mA or C-rates
Epa = 0.693). Similar redox peaks were observed in a mixture of elec- 0.56 C–2.88 C in equimolar concentrations of mixed cation feed solu-
trolyte solution containing equimolar concentrations (1 M) of LiCl, tion (30 mM Mn+ = Li+, Na+, K+, Mg2+, Ca2+) for 20 min/step.
NaCl, KCl, MgCl2 and CaCl2. This suggests capture and release of Li+ Individual voltage profiles of NCM and Ag (vs. Ag/AgCl) were used
even in the presence of other cations. This further demonstrates Li+ (Fig. 4a and b) to derive the ΔE time profile (Fig. 4c) and the ΔE vs. q
capture (forward arrow) and release (backward arrow) from the NCM plots (Fig. 4d) of the NCM/Ag cell at each applied i. Increase in i pro-
lattice as expressed in Eq. (9) [33]. vided higher driving force for Li+ capture and release as reflected by
the improvement in Q values (Fig. 5a) and KDLi (Table 1). However, this
Li1 − x [Ni1/3Co1/3 Mn1/3]O2 + x Li+ + x e− ⇌ Li[Ni1/3Co1/3 Mn1/3]O2 (9)
is coupled with the shift of the ΔE curves to more negative (Li+ capture)
+
The selectivity of NCM to Li was confirmed by the absence of CV or positive (Li+ release) values (Fig. 4c), indicating increase in energy
redox couple when LiCl was not added in the mixture of the electrolyte consumption with i [17]. This was confirmed by the obtained in-
solution (Fig. 3b). The observed Epc = 0.324 entails possible tegrated area from ΔE vs. q plots (Fig. 4d) wherein the consumed energy

1006
C.P. Lawagon et al.
Fig. 8. NCM/Ag performance at varied feed concentration (Cf = 0.03 to 0.25 mol Li+ L−1 with equimolar Mn+; i = ± 0.5 mA or 1.13 C for 20 min/step): (a) NCM
potential vs Ag/AgCl, (b) Ag potential vs Ag/AgCl, (c) ΔE time profile, and (d) ΔE vs. q plots for estimation of energy consumption.
increased from 0.11 J to 5.13 J. These results indicate that increase in i
led to higher Li+ capture and release but at higher energy consumption.
Further inspection of ΔE time profiles (Fig. 4c) reveal smoother
curves at lower i but multiple plateaus started to appear at
i ≥ ± 1.0 mA. This can be attributed to the co-insertion and release of
other Mn+ ions with Li+ at higher i or C-rates [19]. It is likely that other
side reactions occurred especially at i = ± 1.5 mA or 2.88C. This is
reflected by the decline in % PLi (Fig. 5a), deviation of calculated q used
for Li+ capture/release from the total q input and decline in % η
(Fig. 5b) [19]. These results confirm the increasing presence of other
Mn+ in the receiving solution with i, as increasing portion of the q input
was expended for their capture and release with Li+.
The αLi/M peaked at i = ± 0.5 mA but started to decline thereafter
(Table 1). This trend indicates that beyond a certain i or C-rate, co-
insertion and release of other Mn+ with Li+ are promoted hence are
detrimental to the selectivity (αLi/M) of the NCM/Ag cell (Table 1).
Thus, fast discharge/charge rates must be limited and must not exceed
1.13 C (i = ± 0.5 mA) to achieve a favorable Li+ capture and release
with minimal recovery of competing Mn+.
The lowest molar energy consumption (Table 1) was achieved at
i = ± 0.5 mA or 1.13 C (8.2 W·h mol−1 Li+). Also, under this condi-
tion, the second highest % PLi was obtained. Hence, applied current
i = ± 0.5 mA (1.13 C-rate) was selected for the subsequent runs as it
has the lowest W and second highest % PLi (97.74%).
3.3.2. Lithium recovery at varied capture and release durations
In the first experiment, under the selected applied current or C-rate
(i = ± 0.5 mA or 1.13 C), the NCM/Ag cell reached only 38% of its
specific capacity (93 mA·h g−1) in 20 min. To observe if more Li+ can
be recovered by NCM/Ag, the total charge flow q was varied while
maintaining the applied i or C-rate by changing the operation t. Thus,
similar runs were carried out at t = 10–30 min using the same type of
mixed feed solution (Section 3.3.1). Similar voltage profiles of NCM
(Fig. 6a) and Ag vs. Ag/AgCl (Fig. 6b) were constructed for the ΔE time
1007
Chemical Engineering Journal 348 (2018) 1000–1011
profile (Fig. 6c) and the ΔE vs. q plots (Fig. 6d) at varied t.
Variation in operation t provided different q for Li+ capture and
release; for t = 10–30 min, q ranged 0.08–0.25 mA·h at constant C-rate.
The improvements in Q (Fig. 7a) values and KDLi (Table 2) were not as
remarkable as in Fig. 5a. This is because the change in t resulted in q
which allowed the NCM/Ag to achieve only 19–59% of its specific ca-
pacity, which was lower than the q range when i was varied (i.e.
19–96% of 93 mA·h g−1). The slopes of the ΔE curves (Fig. 6c) at varied
t were similar considering that they were operated at constant i or C-
rate. The slight variations can be attributed to the differences in elec-
trode samples used in each run [21]. Longer operation t resulted pro-
longed ΔE time profiles (Fig. 6c) hence the energy consumption in-
creased from 0.08 to 0.36 J, estimated from the integrated area ΔE vs. q
plots (Fig. 6d).
At shorter duration (t < 20 min), Li+ capture probably occurred
only at the surface of NCM. Hence, other Mn+ were able to compete
with Li+, which resulted in the relatively lower Q and % PLi values
(Fig. 7a). This was also reflected by the deviation of calculated q used
for Li+ capture from the total q input and % η (Fig. 7b). While Q
steadily increased with t, % PLi and % η peaked at t = 20 min then
declined thereafter. This indicates that selective Li+ capture and release
require careful control of the operation time/step. It is possible that
t = 20 min sufficiently intercalated Li+ into the interior of the electrode
given its faster diffusion than other Mn+. However, competing Mn+
eventually caught up and co-intercalated with Li+ under longer op-
eration hence the decline in % PLi and % η at t > 20 min.
The quality of the receiving solution at varied t reveals that αLi/M
peaked at 20–25 min whereas the lowest molar power consumption
(W = 7.82 W·h·mol−1 Li+) was at t = 15 min (Table 2). The most su-
perior quality of recovered Li+ was achieved at t = 20 min (%
PLi = 97.5%), which likewise exhibited the most efficient power con-
sumption for Li+ capture (% η = 96.2%). Thus, t = 20 min remains the
most suitable condition for selective Li+ capture and release. Overall
results indicate that very short operation t may result in low selectivity
C.P. Lawagon et al. Chemical Engineering Journal 348 (2018) 1000–1011

and quality of recovered Li+ as its capture is only superficial on the

electrode surface along with competing Mn+. Similarly, prolonged op-
eration (c.a. ≥20 min/step) is not advisable due to the eventual co-
insertion and release of competing Mn+.

3.3.3. Lithium recovery at varied feed concentration

Fig. 9. NCM/Ag performance at varied feed concentration (Cf = 0.03 to
0.25 mol Li+ L−1 with equimolar Mn+; i = ± 0.5 mA or 1.13 C for 20 min/
step): (a) Li+ insertion capacity (Q) and purity of Li+ (% PLi) in the receiving
solution and (b) current efficiency (% η) based on the total q input and q used
for Li+ capture.
Under optimal conditions (i = ± 0.5 mA, 20 min/step), Li+ capture
and release were performed at varied Cf. Individual voltages of NCM
(Fig. 8a), Ag (Fig. 8b), and ΔE time profiles reveal gradual shifts to less
negative (Li+ capture) and less positive values (Li+ release) with Cf
increase (Fig. 8c). Increase in Cf of Li+ provides a favorable condition
for its diffusion into the electrode. At higher Cf, the overpotential which
compensates for the possible depletion of analyte in the solid interface
is not a problem. This is in agreement with Fick’s law where bulk
analyte concentration is directly proportional to the rate of diffusion
[47]. Thus, integrated area from ΔE vs. q plots (Fig. 8d) reveal that
energy consumption was reduced from 0.17 to 0.04 J when Cf of Li+
was increased from 0.03 to 0.25 M Li+. This was coupled with slight
increase in Q (Fig. 9a), primarily due to the enhanced mass transport
rate of Li+ from the feed to the NCM electrode during Li+ capture [47].
The decline in KDLi values was purely based on the marginal increase in
Q despite the drastic increase in Cf (Table 3). Meanwhile, the marginal
increase in Q can be related to the limitation imposed on the applied i or
C-rate, total q input fixed for 20 min/step, and amount of active ma-
terial.
The % PLi gradually stabilized > 96% at higher Cf (Fig. 9a). Along
with improved mass transport rate, % η also reached a maximum
value > 96% as Cf was increased. This suggests that at high Cf of Li+,
the NCM/Ag was able to utilize the q input more efficiently for Li+
capture (Fig. 9b). Since KDLi declined with Cf, αLi/M followed similar
trend (Table 3). But as the efficiency of NCM/Ag improved at
Cf = 0.25 M, molar energy consumption was significantly reduced from
8.30 to 2.60 W·h mol−1 Li+. This Cf is close to the Li+ concentration in
Salar de Uyuni salt lake (Bolivia) [35]. This entails that the NCM/Ag
system can be efficiently used for Li+ recovery from brine-like solu-
tions.

3.4. Performance of NCM/Ag for Li+ recovery from brine

Table 3
NCM/Ag separation performance and molar energy consumption at varied feed
A 20 min/step operation with applied current i = ± 0.5 mA
concentration (Cf = 0.03–0.25 mol Li+ L−1 with equimolar Mn+;
( ± 0.25 mA cm−2) will require only 2.60 W·h mol−1 Li+. In terms of
i = ± 0.5 mA or 1.13 C for 20 min/step).
W, NCM/Ag is second to LiFePO4/Ag (1 W·h mol−1 Li+) and λ-MnO2/
Li+ feed Concentration (M) 0.03 0.10 0.25 Ag (1 W·h mol−1 Li+) but more efficient than λ-MnO2/NiHCF
(3.58 W·h mol−1 Li+), λ-MnO2/AC (4.2 W·h mol−1 Li+), LFP/NiHCF
KDLi (mL g−1) 40.87 2.73 1.11
(8.7 W·h mol−1 Li+) and λ-MnO2/Pt (33 W·h mol−1 Li+)
Li
αNa 162.61 28.97 32.18
[17–19,21,22,25,26]. In terms of purity, NCM/Ag produced 96.4% Li+
Li
αMg 66.67 47.20 42.87
which is slightly lower than that of LFP/NiHCF (97.9%), comparable
αKLi 192.90 46.18 52.76
with λ-MnO2/NiHCF (96.2%) but better than λ-MnO2/AC (91.8%), λ-
Li
αCa 393.32 274.86 137.45
MnO2/Ag (91.8%) and LiFePO4/Ag (84.6%) [17,18,21,25,26]. Mean-
W (W·h mol−1 Li+) 8.30 5.36 2.60
while, comparison of Q values was limited since this parameter is rarely
reported in literature. Calculated Q values listed (Table 4) were only

Table 4
Performance of NCM/Ag for electrochemical Li+ recovery compared to other reported electrode materials.
Electrode pair W (W·h mol−1 Li+) Purity (% PLi) Q (mmol Li+ g−1 electrode material) Reference

LiFePO4/Ag 1.00 84.56 1.118 [17]

λ-MnO2/Ag 1.00 91.80 n.a. [18]
NCM/Ag 2.60 96.40 1.560 This study
λ-MnO2/NiHCH 3.58 96.20 2.144 [21]
λ-MnO2/AC 4.20 91.80 n.a. [25]
LiFePO4/NiHCF 8.70 97.86 2.165 [19]
λ-MnO2/Pt 33.0 n.a. 1.585 [22]

n.a. – not available.

1008
C.P. Lawagon et al. Chemical Engineering Journal 348 (2018) 1000–1011

Fig. 10. Long-term performance of NCM/Ag (each cycle i = ± 0.5 mA or 1.13 C, 20 min/step) in simulated Salar de Uyuni brine (composition in Section 2.1): (a)
individual electrode voltages vs Ag/AgCl, (b) ΔE time profile, (c) gradual accumulation of Li+ at the receiving electrolyte solution, and (d) ΔE vs q plots for estimation
of energy consumption.

Table 5 obtained from studies which provided details on electrode material

Long-term performance of NCM/Ag (each cycle i = ± 0.5 mA or 1.13 C, masses. The NCM/Ag has moderate competence in terms of Q
20 min/step) in simulated Salar de Uyuni brine (composition in Section 2.1). (1.56 mmol Li+ g−1), which was lower than λ-MnO2/NiHCH
Cycle Li+ in W (W·h mol−1 SFLi/Na SFLi/Mg SFLi/K SFLi/Ca PLi (%) (2.14 mmol Li+ g−1) and LiFePO4/NiHCF (2.17 mmol Li+ g−1), com-
LiCl (mM) Li+) parable with λ-MnO2/Pt (1.59 mmol Li+ g−1) but better than LiFePO4/
Ag (1.12 mmol Li+ g−1) [17,19,21,22]. Considering W, % PLi, and Q,
1 0.612 1.55 769.33 393.29 464.61 14.83 96.9
NCM/Ag exhibited competitive performance with other reported elec-
5 3.060 1.65 770.54 393.97 460.37 14.68 96.9
10 6.120 2.59 770.64 398.64 441.31 14.89 96.9 trode assemblies for electrochemical Li+ recovery from brines.
15 9.190 3.31 769.81 391.80 467.44 14.98 96.9
20 12.16 4.99 756.98 397.63 468.30 14.71 96.9
3.5. Recyclability of the NCM/Ag for Li+ recovery

The same NCM/Ag cell (i = ± 0.5 mA, 20 min/step) was repeatedly

Fig. 11. Concentration of NCM metal components (Co2+, Mn2+, Ni2+) in the
receiving solution during cycled Li+ capture and release operation (each cycle
i = ± 0.5 mA or 1.13 C, 20 min/step).
used twenty times to demonstrate their reversibility and applicability
for Li+ recovery from simulated brine (Salar de Uyuni composition in
Section 2.1). Time profiles of NCM and Ag (vs. Ag/AgCl in Fig. 10a)
reveal that one complete cycle lasted for 2400 s or 40 min (20 min/
step). The variations in ΔE (Fig. 10b) for every five cycles can be at-
tributed to the slight changes in the electrode arrangement during so-
lution transfer and replacements [21]. Using the same LiCl recovery
solution during Li+ release resulted in its accumulation (Fig. 10c) as
revealed by the steep increase in [Li+] (0.61 mM/cycle, r2 > 0.99)
relative to the competing Mn+ (Na+: 0.009 mM/cycle; Mg2+:
0.008 mM/cycle; K+: 0.003 mM/cycle; Ca2+: 0.001 mM/cycle). This
indicates that Li+ can be enriched in the recovery solution 68–610
times faster than the competing Mn+.
However, a gradual increase in W was observed with increasing
cycle (Table 5). The ΔE vs q plots (Fig. 10d) gradually shifted to either
more negative (Li+ capture) or more positive (Li+ release) values after
repeated cycles, which was probably due to the retardation of Li+
diffusion into the electrode [40,47]. But autopsy of the NCM from the
electrode after the 20th cycle (ESI Fig. S7) reveals negligible change in
its morphology as compared to the pristine sample. Its maintained
structural integrity after several uses suggests other possible cause of

1009
C.P. Lawagon et al.
the gradual W increase. Furthermore, concentrations of NCM electrode
material components such as Co2+, Mn2+ and Ni2+ remained steadily
[9]
low (Fig. 11) and never exceeded trace levels (< 0.012 mg L−1 for
Mn2+, < 0.005 mg L−1 for Co2+, and < 0.08 mg L−1 for Ni2+) during
the 20-cycle run. The absence of electrode component elution in the [10]
receiving solution confirms the stability of the NCM electrode during
cycled operation. Thus, one reason for gradual increase in W could be [11]
the physical loss of electrode material during electrode placements and
solution transfer. Thus, it is recommended for future works to devise
more effective electrode preparation techniques that would minimize [12]
the physical loss of electrode materials.
Nevertheless, at cycle 20, the energy consumption of NCM/Ag
[13]
system is still comparable or remains better than other reported elec-
trode pairs in literature (Section 3.4). Moreover, NCM/Ag performed
consistently as revealed by the similar SF (Eq. (8)) and% PLi in each
[14]
cycle (Table 5). Overall results from cycled Li+ capture and release
demonstrate the potential of NCM/Ag for electrochemical Li+ recovery
applications.
[15]
4. Conclusions
[16]
Electrochemical Li+ recovery using LiNi1/3Co1/3Mn1/3O2/Ag elec- [17]
trodes was successfully demonstrated as the delithiated Li1−xNi1/3Co1/
+ [18]
3Mn1/3O2 or NCM was able to selectively capture the Li from brine
and release them in receiving solution as LiCl (i.e. with Cl− capture/ [19]
release at the Ag electrode), with Q = 1.56 mmol g−1. The NCM/Ag
delivered acceptable performance and consistent product purity after [20]
repeated Li+ intercalation/de-intercalation operation. From its molar
energy consumption of 2.60 W·h mol−1 Li+, Li+ mining from brine can [21]
be carried out at i = ± 0.25 mA cm−2 for 20 min/step to achieve
96.4% PLi. A 1 m2 of NCM electrode (c.a. 30 g NCM at 3 mg NCM cm−2)
[22]
could produce > 5000 mg/L LiCl if released in Vr = 500 mL of recovery
solution for 7.7 cycles. This could significantly reduce the months-long [23]
conventional evaporation/precipitation process to only 5.14 h. Thus,
electrochemical Li+ recovery using NCM/Ag cell is a promising route [24]
towards higher Li+ production rate with low energy consumption.
[25]
Acknowledgements
[26]
This research was supported by the National Research Foundation
[27]
of Korea (NRF) funded by the Ministry of Science and ICT
(2016R1A2B1009221 and 2017R1A2B2002109) and by the Ministry of
Education (2009-0093816). [28]
Appendix A. Supplementary data [29]
Supplementary data associated with this article can be found, in the
[30]
online version, at http://dx.doi.org/10.1016/j.cej.2018.05.030.
References [31]
[1] A. Sonoc, J. Jeswiet, A review of lithium supply and demand and a preliminary [32]
investigation of a room temperature method to recycle lithium ion batteries to re-
cover lithium and other materials, Proc. CIRP 15 (2014) 289–293.
[2] G. Martin, L. Rentsch, M. Höck, M. Bertau, Lithium market research – global supply, [33]
future demand and price development, Energy Storage Mater. 6 (2017) 171–179.
[3] Z.-Y. Ji, F.-J. Yang, Y.-Y. Zhao, J. Liu, J.-S. Yuan, Preparation of titanium-base li-
thium ionic sieve with sodium persulfate as eluent and its performance, Chem. Eng. [34]
J. 328 (2017) 768–775.
[4] J. Hu, J. Zhang, H. Li, Y. Chen, C. Wang, A promising approach for the recovery of
high value-added metals from spent lithium-ion batteries, J. Power Sources 351 [35]
(2017) 192–199.
[5] U.S. Geological Survey, Mineral Commodity Summaries 2017, U.S. Geological [36]
Survey, 2017, p. 202, http://dx.doi.org/10.3133/70180197 (accessed 2017.03.15).
[6] U.S. Geological Survey, Mineral Commodity Summaries 2018, U.S. Geological [37]
Survey, 2018, p. 200, http://dx.doi.org/10.3133/70194932 (accessed 2018.02.25).
[7] J.-S. Kim, Y.-H. Lee, S. Choi, J. Shin, H.-C. Dinh, J.W. Choi, An electrochemical cell
for selective lithium capture from seawater, Environ. Sci. Technol. 49 (2015) [38]
9415–9422.
[8] H.U. Sverdrup, Modelling global extraction, supply, price and depletion of the
1010
Chemical Engineering Journal 348 (2018) 1000–1011
extractable geological resources with the lithium model, Resour. Conserv. Recycl.
114 (2016) 112–129.
C.P. Lawagon, G.M. Nisola, J. Mun, A. Tron, R.E.C. Torrejos, J.G. Seo, H. Kim, W.-
J. Chung, Adsorptive Li+ mining from liquid resources by H2TiO3: equilibrium,
kinetics, thermodynamics, and mechanisms, J. Ind. Eng. Chem. 35 (2016) 347–356.
M.J. Park, G.M. Nisola, A.B. Beltran, R.E.C. Torrejos, J.G. Seo, S.-P. Lee, H. Kim, W.-
J. Chung, Recyclable composite nanofiber adsorbent for Li+ recovery from sea-
water desalination retentate, Chem. Eng. J. 254 (2014) 73–81.
G.M. Nisola, L.A. Limjuco, E.L. Vivas, C.P. Lawagon, M.J. Park, H.K. Shon, N. Mittal,
I.W. Nah, H. Kim, W.-J. Chung, Macroporous flexible polyvinyl alcohol lithium
adsorbent foam composite prepared via surfactant blending and cryo-desiccation,
Chem. Eng. J. 280 (2015) 536–548.
L.A. Limjuco, G.M. Nisola, C.P. Lawagon, S.-P. Lee, J.G. Seo, H. Kim, W.-J. Chung,
H2TiO3 composite adsorbent foam for efficient and continuous recovery of Li+ from
liquid resources, Colloids Surf. A: Physicochem. Eng. Asp. 504 (2016) 267–279.
M.J. Park, G.M. Nisola, E.L. Vivas, L.A. Limjuco, C.P. Lawagon, J.G. Seo, H. Kim,
H.K. Shon, W.-J. Chung, Mixed matrix nanofiber as a flow-through membrane ad-
sorber for continuous Li+ recovery from seawater, J. Membr. Sci. 510 (2016)
141–154.
W.-J. Chung, R.E.C. Torrejos, M.J. Park, E.L. Vivas, L.A. Limjuco, C.P. Lawagon,
K.J. Parohinog, S.-P. Lee, H.K. Shon, H. Kim, G.M. Nisola, Continuous lithium
mining from aqueous resources by an adsorbent filter with a 3D polymeric nano-
fiber network infused with ion sieves, Chem. Eng. J. 309 (2017) 49–62.
A. Umeno, Y. Miyai, N. Takagi, R. Chitrakar, K. Sakane, K. Ooi, Preparation and
adsorptive properties of membrane-type adsorbents for lithium recovery from
seawater, Ind. Eng. Chem. Res. 41 (2002) 4281–4287.
G. Zhu, P. Wang, P. Qi, C. Gao, Adsorption and desorption properties of Li+ on PVC-
H1.6M1.6O4 lithium ion-sieve membrane, Chem. Eng. J. 235 (2014) 340–348.
M. Pasta, A. Battistel, F. La Mantia, Batteries for lithium recovery from brines,
Energy Environ. Sci. 5 (2012) 9487–9491.
J. Lee, S.-H. Yu, C. Kim, Y.-E. Sung, J. Yoon, Highly selective lithium recovery from
brine using a λ-MnO2-Ag battery, Phys. Chem. Chem. Phys. 7690–7695 (2013).
R. Trócoli, A. Battistel, F. La Mantia, Selectivity of a lithium-recovery process based
on LiFePO4, Chem. Eur. J. 20 (2014) 9888–9891.
T. Kim, M. Rahimi, B.E. Logan, C.A. Gorski, Evaluating battery-like reactions to
harvest energy from salinity differences using ammonium bicarbonate salt solu-
tions, ChemSusChem 9 (2016) 981–988.
R. Trócoli, C. Erinmwingbovo, F. La Mantia, Optimized lithium recovery from
brines by using an electrochemical ion-pumping process based on λ-MnO2 and
nickel hexacyanoferrate, ChemElectroChem 4 (2017) 143–149.
H. Kanoh, K. Ooi, Y. Miyai, S. Katoh, Electrochemical recovery of lithium ions in the
aqueous phase, Sep. Sci. Technol. 28 (1993) 643–651.
F. La Mantia, M. Pasta, H.D. Deshazer, B.E. Logan, Y. Cui, Batteries for efficient
energy extraction from a water salinity difference, Nano Lett. 11 (2011)
1810–1813.
M. Pasta, C.D. Wessells, Y. Cui, F. La Mantia, A desalination battery, Nano Lett. 12
(2012) 839–843.
S. Kim, J. Lee, J.S. Kang, K. Jo, S. Kim, Y.-E. Sung, J. Yoon, Lithium recovery from
brine using a λ-MnO2/activated carbon hybrid supercapacitor system,
Chemosphere 125 (2015) 50–56.
R. Trócoli, A. Battistel, F. La Mantia, Nickel hexacyanoferrate as suitable alternative
to Ag for electrochemical lithium recovery, ChemSusChem 8 (2015) 2514–2519.
F. Marchini, D. Rubi, M. del Pozo, F.J. Williams, E.J. Calvo, Surface chemistry and
lithium-ion exchange in LiMn2O4 for the electrochemical selective extraction of LiCl
from natural salt lake Brines, J. Phys. Chem. C 120 (2016) 15875–15883.
E.V. Saenko, G.V. Leont’eva, V.V. Vol’khin, A.S. Kolyshkin, Variation of the com-
position and properties of lithium manganese oxide spinel in lithiation-delithiation
cycles, Russ. J. Inorg. Chem. 52 (2007) 1312–1316.
C. Yang, X. Zhang, M. Huang, J. Huang, Z. Fang, Preparation and rate capability of
carbon coated LiNi1/3Co1/3Mn1/3O2 as cathode material in lithium ion batteries,
ACS Appl. Mater. Interfaces 9 (2017) 12408–12415.
H. Zheng, X. Chen, Y. Yang, L. Li, G. Li, Z. Guo, C. Feng, Self-assembled LiNi1/3Co1/
3Mn1/3O2 nanosheet cathode with high electrochemical performance, ACS Appl.
Mater. Interfaces 9 (2017) 39560–39568.
Y. Zhao, Y.Y. Wang, Q.Y. Lai, L.M. Chen, Y.J. Hao, X.Y. Ji, Pseudocapacitance
properties of AC/LiNi1/3Co1/3Mn1/3O2 asymmetric supercapacitor in aqueous
electrolyte, Synth. Met. 159 (2009) 331–337.
G.J. Wang, L.C. Yang, Q.T. Qu, B. Wang, Y.P. Wu, R. Holze, An aqueous re-
chargeable lithium battery based on doping and intercalation mechanisms, J. Solid
State Electrochem. 14 (2008) 865–869.
F. Wang, Y. Liu, X. Wang, Z. Chang, Y. Wu, R. Holze, Aqueous rechargeable battery
based on zinc and a composite of LiNi1/3Co1/3Mn1/3O2, ChemElectroChem 2 (2015)
1024–1030.
J. Zheng, J. Chen, X. Jia, J. Song, C. Wang, M. Zheng, Q. Dong, Electrochemical
performance of the LiNi1/3Co1/3Mn1/3O2 in aqueous electrolyte, J. Electrochem.
Soc. 157 (2010) A702–A706.
R. Chitrakar, Y. Makita, A. Sonoda, Lithium recovery from salt lake brine by
H2TiO3, Dalton Trans. 43 (2014) (2014) 8933–8939.
S. Patoux, M.M. Doeff, Direct synthesis of LiNi1/3Co1/3Mn1/3O2 from nitrate pre-
cursors, Electrochem. Commun. 6 (2004) 767–772.
H. Zhu, J. Li, Z. Chen, Q. Li, T. Xie, L. Li, Y. Lai, Molten salt synthesis and elec-
trochemical properties of LiNi1/3Co1/3Mn1/3O2 cathode materials, Synth. Met. 187
(2014) 123–129.
Z. Wang, Y. Sun, L. Chen, X. Huang, Electrochemical characterization of positive
electrode material LiNi1/3Co1/3Mn1/3O2 and compatibility with electrolyte for li-
thium-ion batteries, J. Electrochem. Soc. 151 (2004) A914–A921.
C.P. Lawagon et al.
[39] H. Sclar, D. Kovacheva, E. Zhecheva, R. Stoyanova, R. Lavi, G. Kimmel, J. Grinblat,
O. Girshevitz, F. Amalraj, O. Haik, E. Zinigrad, B. Markovsky, D. Aurbach, On the
performance of LiNi1/3Co1/3Mn1/3O2 nanoparticles as a cathode material for li-
thium-ion batteries, J. Electrochem. Soc. 156 (2009) A938–A948.
[40] X.Y. Qiu, Q.C. Zhuang, Q.Q. Zhang, R. Cao, Y.H. Qiang, P.Z. Ying, S.G. Sun, Reprint
of “investigation of layered LiNi1/3Co1/3Mn1/3O2 cathode of lithium ion battery by
electrochemical impedance spectroscopy”, J. Electroanal. Chem. 688 (2013)
393–402.
[41] Y. Li, X. Hou, Y. Zhou, W. Han, C. Liang, X. Wu, S. Wang, Q. Ru, Electrochemical
performance of structure-dependent LiNi1/3Co1/3Mn1/3O2 in aqueous rechargeable
lithium-ion batteries, Energy Technol. 310007 (2018) 391–396.
[42] W. Hua, J. Zhang, Z. Zheng, W. Liu, X. Peng, X.-D. Guo, B. Zhong, Y.-J. Wang,
X. Wang, Na-doped Ni-rich LiNi0.5Co0.2Mn0.3O2 cathode material with both high
rate capability and high tap density for lithium ion batteries, Dalton Trans. 43
(2014) 14824–14832.
[43] Z. Chen, J. Wang, D. Chao, T. Baikie, L. Bai, S. Chen, Y. Zhao, T.C. Sum, J. Lin,
Z. Shen, Hierarchical porous LiNi1/3Co1/3Mn1/3O2 nano-/micro spherical cathode
1011
Chemical Engineering Journal 348 (2018) 1000–1011
material: minimized cation mixing and improved Li+ mobility for enhanced elec-
trochemical performance, Sci. Rep. 6 (2016) 1–10.
[44] J. Li, R. Yao, C. Cao, LiNi1/3Co1/3Mn1/3O2 nanoplates with 0 1 0 active planes ex-
posing prepared in polyol medium as a high-performance cathode for li-ion battery,
ACS Appl. Mater. Interfaces 6 (2014) 5075–5082.
[45] J. Li, Z.R. Zhang, X.J. Guo, Y. Yang, The studies on structural and thermal prop-
erties of delithiated LixNi1/3Co1/3Mn1/3O2(0 < x ≤ 1) as a cathode material in li-
thium ion batteries, Solid State Ionics 177 (2006) 1509–1516.
[46] P. Manikandan, P. Periasamy, R. Jagannathan, Microstructure – twinning and
hexad multiplet(s) in lithium-rich layered cathode materials for lithium-ion bat-
teries, RSC Adv. 4 (2014) 40359–40367.
[47] M. Park, X. Zhang, M. Chung, G.B. Less, A.M. Sastry, A review of conduction
phenomena in Li-ion batteries, J. Power Sources 195 (2010) 7904–7929.
[48] M.S. Palagonia, D. Brogioli, F.L. Mantia, Influence of hydrodynamics on the lithium
recovery efficiency in an electrochemical ion pumping separation process, J.
Electrochem. Soc. 164 (2017) E586–E595.