Adsorption

https://doi.org/10.1007/s10450-020-00208-5

Synergic and competitive adsorption of Li–Na–MgCl2

onto lithium–aluminum hydroxides
Huixiong Jiang1 · Shuiyi Zhang1 · Ying Yang1 · Jianguo Yu1,2

Received: 28 August 2019 / Revised: 19 November 2019 / Accepted: 25 January 2020

© Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract
Layered lithium–aluminum hydroxides have been used for the lithium adsorption for its high selectivity and cyclic stabil-
ity. The adsorption performances of lithium and sodium in the mixing solution with abundant magnesium chloride were
investigated using batch system at 303 K. The lithium adsorption capacity was increased from 5.02 to 5.69 mg/g (initial
lithium concentration is 350 mg/L) with the increasing of initial chloride ions concentration from 200 to 350 mg/L, while
the sodium decreased. Both the adsorption capacities of lithium and sodium were lower in Li–Na–MgCl2 system than that’s
in Li–MgCl2 and Na–MgCl2 system, which indicated the competitive effects between lithium and sodium ions. In addition,
the Langmuir model could describe the isotherms of lithium and sodium well. Then, the homogeneous surface diffusion
model (HSDM) was applied to represent the lithium adsorption kinetics. The surface diffusion coefficients of lithium were
dependent on the initial concentration and affected by sodium ions. The Biot numbers were ranged from 25 to 40 confirm-
ing that the intraparticle diffusion is the limiting step. Breakthrough curves of lithium and sodium demonstrated the lithium
could be enriched and obtained with fixed bed adsorption process, and verified the competitive adsorption phenomenon.

Keywords Lithium adsorption · Competitive effect · Synergic effect · Surface diffusion · Adsorption mechanism

1 Introduction of the lithium resources are found in salt lakes. Therefore,

it is an important way to obtain lithium from salt lake brine
Lithium is one of the most important metal resources and (Ooi et al. 2017; Liu et al. 2018; Hao et al. 2017).
has extensive applications, such as rechargeable batteries, Several technologies, such as calcination, precipitation,
ceramics, alloy, and thermonuclear fusion because of its adsorption, solvent extraction, and electrodialysis, have been
unique electrochemical performance as well other proper- applied in lithium recovery from lithium-bearing brine.
ties. Particularly, lithium-ion batteries are widely developed However, the high magnesium lithium ratio (more than 300)
as efficient storage devices for electronic products and elec- is a great challenge in the lithium recovery from brine in
tric vehicles, which significantly promote the demand of China. Adsorption is an effective and promising process of
the lithium products (Ramakumar et al. 2017; Martin et al. lithium recovery from brines for its selectivity to lithium,
2017; Narins 2017). According to the survey, there are 63 simplicity of the industrial process, and environmental pro-
million tons of lithium resources around the world, and 70% tection (Lawagon et al. 2016; Choubey et al. 2016). Pres-
ently, manganese dioxide (­ MnO2) (Ryu et al. 2013; Zhang
et al. 2007; Hong et al. 2018), titanium dioxide ­(TiO2) (Moa-
* Ying Yang zeni et al. 2015; Zhang et al. 2010; Ramesh et al. 2014)
[email protected]
and layered lithium–aluminum hydroxides (Liu et al. 2018;
* Jianguo Yu Sun et al. 2019a, b) have been widely studied. Although the
[email protected]
lithium adsorption capacity and selectivity of manganese
1
Engineering Research Center of Resources Process and titanium adsorbent are better than those of lithium–alu-
Engineering, Ministry of Education, East China University minum hydroxides, the dissolution losses of manganese
of Science and Technology, Shanghai 200237, China and titanium during the adsorption recycles limited their
2
National Engineering Research Center for Integrated development and industrial application. Owing to the its
Utilization of Salt Lake Resource, East China University cyclic stability, the layered lithium–aluminum hydroxide is
of Science and Technology, Shanghai 200237, China

13
Vol.:(0123456789)
 Adsorption

a promising adsorbent in lithium recovery from high Mg/Li might be explained by the analysis of the adsorbent struc-
ratio brine (Sun et al. 2019b). ture and adsorption mechanism. Therefore, the lithium and
Numerous studies had investigated adsorption perfor- sodium adsorption kinetics were conducted in L–M, N–M
mance systematically for single component system of lith- and L–N–M system, and the kinetic data were analyzed by
ium ions (Wang et al. 2017; Ji et al. 2017; Chitrakar et al. employing homogeneous surface diffusion model. Using
2014; Jiang et al. 2019). In real-life applications, lithium homogeneous surface diffusion model, film and surface
ions commonly coexist with other metal ions, such as diffusion coefficients were evaluated and then the rate con-
sodium, calcium, magnesium ions. In western china, typi- trolling step was identified. Finally, the preliminary column
cal brine contains saturated magnesium chloride, low-grade experiment was conducted in order to evaluate the adsorp-
lithium and other metal ions (Sun et al. 2019b). There may tion performance of column process for lithium recovery.
be interaction effects between coexisting ions during the
adsorption process of lithium ions onto lithium–aluminum
hydroxides. So, in order to figure out the adsorption perfor- 2 Materials and methods
mance of lithium, the competitive and synergic effects of
coexisting ions should be considered. As the concentration 2.1 Materials
of sodium and magnesium ions were higher than other cati-
ons and the chloride ions is the dominant anions, so, in this Commercial lithium–aluminum hydroxides adsorbent con-
work, the effects of sodium magnesium and chloride ions nected with the resin was pre-treated in order to regenerate
on lithium adsorption were mainly investigated. However, and remove the impurities. Resin played the role of car-
the competitive and synergic effects of coexisting ions on rier that provided the pore structure during the formation
lithium adsorption on lithium–aluminum hydroxides had of layered crystals of lithium–aluminum hydroxides and
rarely been reported. enhanced the mechanical performance of the adsorbent. The
For the design of adsorption process, prediction of prepared adsorbent was used for further experimental stud-
adsorption kinetic curve is essential (Ponnusami et al. 2010). ies. The stock solution was prepared according to the real
It is well known that the diffusion models are the most real- brine by dissolving analytical grade of LiCl⋅H2O, NaCl and
istic way to predict the adsorption in porous solids, because ­MgCl2⋅6H2O in deionized water according to the require-
it includes the three simultaneous steps: external mass ments of the experiments. Table 1 listed the properties of
transport, intraparticle diffusion and adsorption on active adsorbent and the brine.
sites (Souza et al. 2017b). The mathematical model is vali-
dated through experimental data of the adsorption kinetics 2.2 Apparatus
and predicted in many conditions (Ponnusami et al. 2010;
Frohlich et al. 2018; Souza et al. 2017a). And, in this work, Inductively Coupled Plasma Optical Emission Spectrometer
the homogeneous surface diffusion model (HSDM) was (ICP-OES, spectra ARCOS, Germany) was used to analyze
used to describe the adsorption kinetics and the identifica- the concentration of metal ions. All the experimental data
tion the limiting mass transfer mechanism (Ramos Suzaki were diluted below 100 mg/L, and the correlation coefficient
et al. 2017; Lin et al. 2017). of the calibration line was above 0.999. A thermostat shaker
As the concentration of magnesium and chloride ions
are much greater than other coexist ions, the magnesium
chloride was represented by the chloride ions for better Table 1  Properties of adsorbent and typical brine
understanding the mechanism. So, in this study, the adsorp- Characteristics Value
tion performance had been studied using the mixture solu-
Adsorbent
tion: Li–MgCl2 (L–M system), Na–MgCl2 (N–M system),
Appearance Light brown, bead
Li–Na–MgCl2 (L–N–M system). The synergic effect of
Mean particle size (cm) 0.065
magnesium chloride on lithium and sodium adsorption
Apparent density (g/L) 1378.7
were investigated by varied the concentration of chloride
Total pore area ­(m2/g) 51.763
ions. In order to determine the competitive effects between
Void of particle 0.2686
lithium and sodium ions, the adsorption performance of
Brine
lithium and sodium ions were investigated in L–M, N–M
Density (mg/L) 1.253
and L–N–M system while the chloride concentration were
Viscosity (cP) 14.04
kept at 280 g/L which were near to the real brine. Two com-
Chloride concentration (g/L) 280
monly used isotherm models were used to fitted the experi-
Lithium concentration (g/L) 0.35
mental data, and the best fitted model would be used in the
Sodium concentration (g/L) 1.4
kinetic studies. The synergic and competitive phenomenon

13
Adsorption

(SW 23, JULABO, Germany) was used for batch adsorption C0 − Ct

experiments. qt = V (1)
m
where qt is the adsorption capacity (mg/g) at given time,
2.3 Experiments C0 is the initial concentration (mg/L) of metal ions in solu-
tion, Ct is the concentration (mg/L) at given time, V is the
Batch experiments were conducted at 303 K for evaluating volume (L) of the solution, and the m is the mass (g) of the
the adsorption performance of lithium and sodium ions onto adsorbent. In order to analyze the synergic and competi-
the lithium–aluminate hydroxides. A series of 250 mL flasks tive effects of sodium and magnesium chloride, the ratio of
with 15 mL wet adsorbent and 150 mL solution at different adsorption capacity (Rq,i) was used, and the parameter was
initial concentration were shaken at thermostat apparatus defined as:
at 150 rpm for sufficient time. Then the supernatants were qe,i
sampled and analyzed after equilibrium. Rq,i =
q0,i (2)
For the synergic effects of magnesium chlorides, the ini-
tial concentration of lithium in L–N–M system (mass ratio where qe,i is the adsorption capacity for metal ions i in the
of Li and Na was 350:1400, according to the real brine) L–N–M component and q0,i is the single component adsorp-
varied from 350, 420, 490, 560, 630, 700, 840, and 1050 tion capacity of that metal ions at the same conditions of
mg/L and the initial concentration of chloride ions (chloride the L–M or N–M system. Parameter Rq,i allows to compare
ions represent magnesium chloride) were 200, 280, 350 g/L the multicomponent adsorbent performance with the results
respectively. obtained in single component systems using the same exper-
For the competitive effects of sodium ions, the initial imental conditions. If Rq,i > 1, the presence of other metal
concentration of chloride ions was kept constant at 280 ions in multicomponent system improves the adsorption of
g/L. The initial lithium concentration varied from 350, 420, metal ion i (i.e., synergic adsorption), if Rq,i = 1, there is
490, 560, 630, 700, 840, and 1050 mg/L in L–M system, no effects on the multicomponent adsorption of metal ion
and the sodium varied from 1400, 1680, 1960, 2240, 252, i, if Rq,i < 1, the adsorption of metal ion i is reduced by the
2800, 3360 and 4200 mg/L in N–M system correspond to presence of other metal ions in the multicomponent system
the lithium concentration. In L–N–M system, the experimen- (i.e., competitive adsorption).
tal conditions were the same with the synergic experiments
when the initial chloride ions were 280 g/L.
The kinetic experiments involved the same methods as 3 Theory
described earlier, except that the samples were collected
at the given time. The initial concentration of lithium and 3.1 Adsorption mechanism
sodium were 350, 700 and 1400, 2800 mg/L both in L–M,
N–M and L–N–M system with the initial concentration of The lithium–aluminum hydroxides could be produced by
chloride kept at 280 g/L. chemical precipitation from the aluminum-lithium system
Fixed bed adsorption experiment was conducted in a ­AlCl3–LiOH–H2O, and Fig. 1 illustrated the structure of
column (internal diameter of 20 mm, bed length of 60 cm) the lithium–aluminum hydroxides. The layered structure of
packed with prepared adsorbent. The simulated brine (Li lithium–aluminum hydroxides mainly consists of gibbsite
350 mg/L, Na 1400 mg/L, Cl 280 g/L) was pumped through (γ-Al(OH)3). The lithium ions occupy the cavity sites at the
column by a peristaltic pump while solid adsorbent kept in center of the hexagon composed of six aluminum atoms
the column. Both the brine and column kept the tempera- which made the layer positively charged, and the chloride
ture at 303 K by using two thermostatic baths. Breakthrough ions are positioned in the middle of two adjacent layers just
curves of lithium and sodium were obtained by plotting the above the lithium metal ions to keep charge balance (Besser-
passing time against with the ratio of metal ions concentra- guenev et al. 1997; Sun et al. 2019a, b; Kozlova et al. 2003).
tion in the effluent and in the feed ­(Ct/C0). The whole adsorption process of lithium is the insertion of
lithium ions to the cavity and the chloride ions lies between
the layers. In the synthesis of the chlorine form double lay-
2.4 Data analysis ered aluminum lithium hydroxides (LiCl⋅2Al(OH)3⋅mH2O)
particles, the intercalation of LiCl precedes disordering of
The amount of adsorbed metal ions on adsorbents was cal- Al-OH layers, the separations between Al-OH layers are at
culated as the difference between the origin concentration the nanoscale, which makes them selective sorbents for LiCl
and the concentration in solution after contact, according to (Kotsupalo et al. 2013). So, it could be reasonable assumed
the mass balance in Eq. (1): that the lithium–aluminum hydroxides show the selectivity

13
 Adsorption

Fig. 1  Structure of the chloride ion intercalated of lithium–aluminum hydroxides [left is the top view; right is the front view]

of lithium toward other metal ions, and the anions would 3.2.2 Freundlich isotherm
promote the insertion of metal ions because the insertion
of cations was accompanied with the anions for the charge Freundlich isotherm is the earlies known relationship
balance. describing the non-ideal and reversible adsorption on hetero-
geneous surfaces as well as multilayer adsorption (Rengaraj
et al. 2003; Foo and Hameed 2010; Chabani et al. 2006). It
3.2 Isotherm models can be given as Eq. (4):

The sorption isotherms express the special relation between qe = kF Ce1∕n (4)
the concentration of the adsorbate and its accumulation onto
the adsorbent when the adsorption process reached to an where kF indicates the adsorption capacity for a unit
equilibrium state at a constant temperature. Therefore, it equilibrium concentration, while 1/n shows the energy or
is important to establish the most appropriate correlation intensity of the reaction and suggests the type of this adsor-
for the equilibrium curves that can be used to optimize the bent–adsorbate system
design of a sorption system. In this study, the two commonly
used adsorption isotherm models, Langmuir and Freundlich 3.2.3 Linear isotherm
models, were applied to fit the equilibrium adsorption data
of lithium and sodium ions. Linear isotherm is the simplest method to represent the
adsorption equilibrium, similar to the Henry’s law in the
gas–liquid equilibrium. The isotherm equation is shown in
3.2.1 Langmuir isotherms Eq. (5):

Langmuir isotherm is probably the most widely used equi- qe = a + bCe (5)
librium equation, it is based on the following assumptions:
where a is the linear equation constant (mg/g), and often
all active sorption site are equivalent, the adsorption is uni-
referred to as the partition coefficient, b is the isotherm con-
form, there is no interaction between a molecule and adja-
stant (L/g). Linear isotherm is widely used in representing
cent sites(Vasiliu et al. 2011; Ramos Suzaki et al. 2017).
the low-concentration adsorption due to its simplicity.
Equation (3) is shown as:
q0 bCe 3.3 Homogeneous surface diffusion model
qe = (3)
1 + bCe
Though adsorption isotherm is fundamentally important
where q0 and b are equation constants related to adsorp- for designing a practical adsorption process, the adsorp-
tion capacity (mg/g) and energy (L/mg), qe and Ce are equi- tion kinetics play an important role in deeply understand-
librium concentration at liquid and solid phase (mg/g) and ing the adsorption mechanism. It is well documented
energy (mg/L) respectively.

13
Adsorption

that the overall rate of adsorption rate on a porous solid 3.4 Error functions
includes the following three simultaneous steps: external
mass transport, intraparticle diffusion, and adsorption on The quality of the fit of several models between the experi-
an active site. mental and the fitting data were evaluated by error functions.
As the low value of void fraction of particle (0.2868) In this study, the fitting conformity was determined by sum
was observed, so, it can be assumed that the surface diffu- square error (SSE) and coefficient of determination (R2). The
sion is the predominant intraparticle mass transfer mecha- lower values of SSE and the higher values of R2, the better
nism (Dotto et al. 2014). In this work, the homogeneous the model fitted to the experimental data.
surface diffusion model (HSDM) was applied on the basis
of assumptions: the surface diffusion was the predominant
intraparticle mass transfer mechanism, adsorption on the 4 Results and discussion
active site was instantaneous, the process was isotherm
(Ponnusami et al. 2010). So, the overall mass balance can 4.1 Synergic effects of magnesium chloride
be expressed:
The adsorption isotherms of lithium and sodium ions as
dC Wa
V = −kf (C − Cp,R=Rp ) (6) affected by the magnesium chloride were investigated,
dt 3Rp 𝜌p and Fig. 2 illustrated the adsorption isotherms of lithium
and sodium at different initial concentration of chloride
where V is the volume of solution (L), C and Cp,R=Rp is
ions. The change of initial chloride concentration ions was
the metal concentration at bulk liquid phase and surface
achieved by adding different amount of magnesium chloride.
of the adsorbent (mg/L), Wa is the mass of adsorbent (g),
Although the mass ratio of Mg/Al, Mg/Li in mixing solution
Rp is the radius of the adsorbent (cm), ρP is the apparent
had remarkable effects on the entering of lithium ions into
density of adsorbent (g/L), kf is the film diffusion coef-
the adsorbent (Hu et al. 2019; Guo et al. 2018) during the
ficient (cm/s).
synthesis of the lithium–aluminum layer double hydroxides
Mass balance in the particle can be expressed:
(LiAl-LDH) or magnesium–aluminum layer double hydrox-
ides (LiAl-LDH). The adsorbent used in this paper were
( )
𝜕q Ds 𝜕 2 𝜕q
= 2 R (7) synthesized in the type of LiAl-LDH with the absence of
𝜕t R 𝜕R 𝜕R
magnesium ions, which indicated that the magnesium ions
where q is the lithium concentration at solid phase can not enter the octahedral vacancies for the lattice selec-
(mg/g), R is the radial coordinate (cm), Ds is the surface tivity during the lithium adsorption process. The magne-
diffusion coefficient (­ cm2/s), εp is the void of the particle. sium ions near saturation and much larger than lithium and
The initial and boundary conditions are given as sodium ions during the experiments. It could be reasonably
follows: considered that magnesium concentration kept constant and
had insignificant effects on the lithium adsorption.
t = 0,0 ≤ R ≤ Rp , C = C0 , q = 0, CP = 0 (8) It showed that the lithium adsorption capacity increased
with the increasing of the chloride ions, while the sodium
𝜕Cp adsorption capacity decreased. The maximum lithium
R = 0, t > 0, =0 (9)
𝜕R adsorption capacity increased from 5.02 to 5.69 mg/g (with
the initial lithium concentration is 350 mg/L) with the initial
𝜕q chloride concentration increasing from 200 to 350 g/L. It
R = Rp , t > 0, 𝜌p Ds = kf (C − Cp,r=Rp ) (10) may be explained that the presence of chloride enhanced the
𝜕R
adsorption lithium which was consistent with the hypothesis
In addition, there is an equilibrium between the lithium ions that the anions promote the adsorption of cations. However,
concentration in particle, Cp, and the concentration in solid the increase of chloride concentration was achieved by add-
phase, q. this equilibrium relationship is represented by the ing the magnesium chloride to the solution, therefore, the
adsorption isotherms based on the fitting of the isotherm increase of chloride means the increase of magnesium ions.
models. And the presence of abundant magnesium ions prohibited
The mathematical model was solved numerically in the adsorption of sodium ions. So, it might be concluded that
gPROMS (Process System Enterprise, UK) using the the adsorbent showed the selectivity of lithium ions toward
orthogonal collocation on finite elements method. Fifty magnesium and sodium ions (Sun et al. 2019a; Kotsupalo
intervals with three order polynomials were used in the et al. 2013). Table 2 listed the Langmuir, Freundlich and
radial domain of adsorbent. The “Parameter Estimation” Linear model parameters and their correlation coefficients.
function in gPROMS was employed for the fitting procedure The Langmuir model can describe the lithium isotherms at

13
 Adsorption

Fig. 2  Metal ions adsorption isotherms at different concentration of magnesium chlorides. a Lithium ions, b sodium ions

Table 2  Isotherm parameters Isotherm models Parameters 350 [g/L (­ Cl−)] 280 [g/L(Cl−)] 200 [g/L(Cl−)]
of metal ions at varied
concentration of magnesium Li Na Li Na Li Na
chloride
Langmuir q0 (mg/g) 5.94 12.1 5.78 29.3 5.52 9.70
b (L/g) 0.0705 4.14E−5 0.0443 1.97E−5 0.0285 7.89E−5
R2 0.964 0.996 0.997 0.997 0.998 0.990
SSE 0.126 3.20E−3 0.0102 2.90E−3 6.20E−3 0.0116
Freundlich kF (mg/g) 2.64 8.44E−4 2.35 7.80E−4 1.81 1.99E−3
n 7.30 1.09 7.00 1.05 5.78 1.17
R2 0.965 0.994 0.921 0.997 0.915 0.985
SSE 0.119 4.30E−3 0.235 3.10E−3 0.279 0.0157
Linear a (mg/g) 4.42 0.0680 4.19 0.0525 3.73 0.189
b (L/g) 0.00407 4.24E−4 0.00308 5.28E−4 0.00313 5.57E−4
R2 0.734 0.992 0.687 0.999 0.703 0.980
SSE 0.906 5.62E−3 0.929 3.22E−3 0.965 0.0210

all the experimental conditions better than the Freundlich lithium adsorption isotherm both in L–M and L–N–M
and Linear models with the higher correlation coefficients system. And three above models can describe the sodium
(R2) lower SSE values. While all the isotherm models could isotherm well. Lithium was much more predominant than
represent the sodium adsorption isotherms well. Therefore, sodium ions as the lithium adsorption could reach 5.1 and
the Langmuir isotherm would be used in the numerical 5.6 mg/g with the equilibrium concentration of 200 mg/L,
simulation of lithium adsorption kinetics and the Linear or while the sodium was 1.0 and 1.5 with the equilibrium con-
Langmuir isotherm model would be used in the describing centration of 2000 mg/L respectively. The lithium adsorp-
of sodium isotherms. tion capacity in L–N–M system (5.42 mg/g) was lower
that’s in L–M system (5.81 mg/g) with the initial lithium
4.2 Competitive effects of sodium concentration is 350 mg/L, and the ratio of adsorption
capacity (Rq,Li) was lower than 1 obviously, which indicted
The experimental data of lithium and sodium isotherm in the presence of sodium ions prohibited the adsorption of
L–M, N–M and L–N–M systems were fitted with Lang- lithium. It could be concluded that there is a competitive
muir, Freundlich and Linear models, as shown in Fig. 3. adsorption effect between lithium and sodium ions on the
And, Table 3 listed the model parameters and their cor- lithium–aluminum hydroxide adsorbent. This competitive
relation coefficients. As can be seen, Langmuir isotherm phenomenon may be caused by following reasons. Firstly,
model was better than Freundlich model in describing the although the LiAl-LDH showed lattice selectivity for

13
Adsorption

Fig. 3  Metal ions adsorption isotherms at L–M, N–M and L–N–M component system. a Lithium ions, b sodium ions

be adsorbed on the resin without selectivity toward lithium

Table 3  Isotherms parameters of metal ions at L–M, N–M and
L–N–M component system. ions, which might lead to the competitive phenomenon.

Isotherm Parameters L–M, N–M system L–N–M system

models
Li Na Li Na 4.3 Effects of contact time
Langmuir q0 (mg/g) 6.11 11.5 5.78 29.3
The effects of contact time on lithium and sodium adsorp-
b (L/g) 0.0569 7.99E−05 0.0443 1.97E−05
tion in L–M, N–M and L–N–M system were investigated
R2 0.994 0.991 0.997 0.997
at 303 K, which were shown in Fig. 4. The equilibrium
SSE 0.0203 0.0152 0.0102 0.00290
adsorption capacities of lithium and sodium in L–M and
Freundlich kF (mg/g) 2.77 2.40E−3 2.35 7.80E−4
N–M system were higher than that on L–N–M system
n 7.84 1.17 7.00 1.05
with the same initial concentration, which was consist-
R2 0.936 0.987 0.921 0.997
ent with the isotherm studies. In addition, the equilibrium
SSE 0.201 0.0216 0.235 3.10E−3
adsorption capacities of lithium were 3.95 and 5.38 mg/g
Linear a (mg/g) 4.61 0.228 4.19 0.0525
approximately in L–N–M system with the initial concen-
b (L/g) 3.16E−3 6.64E−4 3.08E−3 5.28E−4
tration is 350 and 700 mg/L, which were much larger than
R2 0.704 0.982 0.687 0.999
that of sodium (0.70 and 1.38 mg/g respectively) under the
SSE 0.936 0.0294 0.929 3.22E−3
experimental conditions. And, it took long time for lithium
to reach equilibrium, more than 2000 min, while it just
took less than 20 min for sodium to get equilibrium under
lithium cations, it still adsorbed a small amount of other all the experimental conditions, which might because the
metal ions, like sodium ions. Similar phenomenon was adsorption capacity of sodium is much lower than that
found in titanium and manganese ion sieves (Zhang et al. of lithium. The presence of lithium ions decreased the
2010; Xiao et al. 2013), the sodium and lithium adsorp- sodium adsorption with the equilibrium adsorption capac-
tion capacity were 0.044, 0.997 mmol/g ­( Li 2TiO 3) and ity decreased from 1.92 to 1.38 and 1.05 to 0.70 mg/g
0.23, 4.96 mmol/g (­ Li1.6Mn1.6O4), respectively. Under the with the initial concentration of 1400 and 2800 mg/L
experimental conditions of this paper, the concentration of respectively. However, the presence of sodium ions does
lithium was less than sodium concentration, so, the sodium not have significant effects of the lithium adsorption both
ions might occupy the vacancies of octahedral framework in equilibrium and kinetics. So, the study of the lithium
of Al(OH)3 by replacing the lithium ions. Furthermore, the kinetics would neglect the effects of sodium ions due to the
adsorbent consists of LiAl-LDH and resin, the resin played insignificant effects on lithium adsorption and instantane-
the role of the carrier, which enhanced the mechanical per- ous equilibrium of sodium ions.
formance of adsorbent. Meanwhile, some metal ions might

13
 Adsorption

Fig. 4  Effects of contact time on metal ions adsorption capacity at L–M, N–M and L–N–M system with different initial concentration. a Lithium
ions, b sodium ions

4.4 Homogeneous surface diffusion models

4.4.1 Calculation of film diffusion coefficient

From the overall mass balance equation [Eq. (6)], the film
diffusion coefficient can be estimated through the initial
adsorption rate. Initially C = C0 and Cp,R=Rp = 0 (for finite
film mass transfer resistance or finite kf), the initial adsorp-
tion rate r0 is defined:
3kf C0
( )
dq
ro = = (11)
dt t=0 Rp 𝜌p

Equation (11) provides the relationship for calculation

the film diffusion coefficient through the initial adsorption
rate and the initial adsorption rate was determined by using
Fig. 5  Comparison of experimental data and HSDM of lithium kinet-
the bulk phase concentration change between 0 and 10 min
ics at L–M, N–M and L–N–M component system.
in the present study (Yao and Chen 2015). Then the film
diffusion coefficients were calculated and given in Table 4

4.4.2 Estimation of surface diffusion coefficient

Table 4  Diffusion coefficients of lithium at L–M, N–M and L–N–M
The experimental kinetics curves of lithium were obtained component system
under different initial concentration (350 and 700 mg/L) Parameters 350 mg/L 700 mg/L
and system (L–M and L–N–M system). The only unknown
L–M system L–N–M L–M system L–N–M
parameter, Ds, was estimated by fitting the numerical solu-
system system
tion of the HSDM to the experimental kinetic data. And
Table 4 listed the optimal surface diffusion coefficients kf (cm/s) 1.012E−4 1.052E−4 6.482E−5 6.480E−5
and the error functions. Figure 5 depicted the experimental Ds ­(cm2/s) 2.236E−9 2.120E−9 3.301E−9 3.115E−8
kinetic curves of lithium adsorption and the kinetic curves Bis 33.36 39.87 27.47 30.98
predicted with the HSDM model using the optimal values R2 0.9999 0.9991 0.9998 0.9999
of Ds, showing the HSDM fitted the experimental data very SSE 0.009406 0.05073 0.02250 0.4154

13
Adsorption

satisfactorily as the correlation coefficients (R2) were near

to 1 and the low values of SSE under all the experimental
conditions.
As can be seen in the Table 4, the values of Ds increased
from 2.236E−9 to 3.301E−9 ­cm2/s in L-M system with
the initial lithium concentration increased from 350 to 700
mg/L, which showed the surface diffusion appears to depend
on the initial lithium ions concentration. And the same con-
clusion can be found in L–N–M system. This might be due
to the high surface lithium ion loading with high initial con-
centration, hence, the driving force for surface diffusion is
stronger (Choy and McKay 2005; Dotto et al. 2014). As a
result, the fraction uptake of lithium ions onto lithium–alu-
minum hydroxides at high concentration is faster. Regard-
less of the initial concentration of 350 or 700 mg/L, the
surface diffusion coefficients in L–M system were slightly Fig. 6  Breakthrough curves of lithium and sodium with the flow rate
is 3 mL/min temperature is 303 K
larger than that’s in L–N–M systems, which might show the
competitive effects between the lithium and sodium ions.
And the similar conclusion can be found between the dyes concentration apparently. It could be explained that the lith-
adsorption in single and multicomponent system (Choy et al. ium ions compete and replace sodium ions adsorbed in the
2004). adsorbent, then, this part of sodium ions re-entered into the
liquid phase. Consequently, the sodium concentration in the
4.4.3 Rate‑controlling procedure outlet was greater than the feed concentration. In conclusion,
the adsorbent showed the selectivity for lithium than sodium,
The performance of the adsorption kinetics by using HSDM and the fixed bed process could be used in the adsorption
is mainly determined by the following two parameters, kf and and enrichment of lithium effectively.
Ds. The dimensionless Biot number, B ­ is, which reflects the
relative importance of external mass transfer to intraparti-
cle diffusion, was used to verify the influence of each mass 5 Conclusion
transfer step on the adsorption (Souza et al. 2017a; Dotto
et al. 2014). The Biot number is defined by Eq. (12): This work studied the adsorption performance of lithium
and sodium in abundant magnesium chloride solution on
kf Rp C0 lithium–aluminum hydroxides. The lithium adsorption
Bis = (12)
2𝜌p Ds qe capacity increased with the increasing of chloride concen-
tration while the sodium decreased. The lithium adsorp-
According to the literature, a complete dominance of the tion capacity was increased from 5.02 to 5.69 mg/g (initial
external mass transfer exists for ­Bis < 0.5, while a consid- lithium concentration is 350 mg/L) with the increasing of
erable dominance of the intraparticle diffusion exists for initial chloride concentration from 200 to 350 mg/L. So,
­Bis>10. Since the ­Bis values obtained in this work were increasing the concentration of magnesium chloride in real
ranged from 25 to 40 under the experimental conditions, situation is an effective way to improve the lithium adsorp-
which could be seen in Table 4. Consequently, the intrapar- tion performance of lithium–aluminum hydroxides. The
ticle diffusion (surface diffusion) was considered as the rate maximum lithium adsorption capacity in L–N–M system is
limiting procedure. lower than that’s in L–M system, which indicated the com-
petition effects between lithium and sodium ions. Moreover,
4.5 Fixed bed experiment Langmuir models could describe the lithium and sodium
isotherms under all the experimental condition. The analysis
The breakthrough curves of lithium and sodium ions were of the structure might explain the synergic and competitive
plotted with the flow rate of 3 mL/min, which could be seen phenomenon that the anions would promote the adsorption
in Fig. 6. As could be seen in Fig. 6, it took 60 min for of lithium ions while other cations may compete the adsorp-
sodium to reach saturation adsorption in fixed bed, while tion of lithium. It was found that the equilibrium time for
the lithium ions could be detected after 300 min. When lithium was more than 2000 min while for sodium was less
lithium passed through the column (300–400 min), the con- than 20 min, so, the sodium concentration could be consid-
centration of lithium ions in the outlet was greater than feed ered kept constant during the adsorption process of lithium.

13
 Adsorption

HSDM could describe the lithium adsorption kinetic well Guo, X., Hu, S., Wang, C., Duan, H., Xiang, X.: Highly efficient sepa-
both in L–M and L–N–M system with different initial con- ration of magnesium and lithium and high utilization of mag-
nesium from salt lake brine by a reaction-coupled technology.
centration. The values of Ds increased with the increasing Ind. Eng. Chem. Res. 57(19), 6618–6626 (2018). https​://doi.
of the initial lithium concentration, which may be due to the org/10.1021/acs.iecr.8b011​47
stronger diffusion driving force with higher concentration of Hao, H., Liu, Z., Zhao, F., Geng, Y., Sarkis, J.: Material flow analysis
lithium. With the same initial concentration of lithium, the of lithium in China. Resour. Policy 51, 100–106 (2017). https​://
doi.org/10.1016/j.resou​rpol.2016.12.005
surface diffusion coefficients in L–N–M system were lower Hong, H.J., Ryu, T., Park, I.S., Kim, M., Shin, J., Kim, B.G., Chung,
than that’s in L–M system, it might be concluded that the K.S.: Highly porous and surface-expanded spinel hydrogen man-
coexisting ions, such as sodium, magnesium, chloride ions, ganese oxide (HMO)/Al2O3 composite for effective lithium (Li)
would slightly decrease the diffusion rate of lithium ions. recovery from seawater. Chem. Eng. J. 337, 455–461 (2018). https​
://doi.org/10.1016/j.cej.2017.12.130
The Biot number ­(Bis) illustrated that the surface diffusion Hu, S., Sun, Y., Pu, M., Yun, R., Xiang, X.: Determination of bound-
is the rate-controlling step of the whole lithium adsorption ary conditions for highly efficient separation of magnesium and
process. Finally, the breakthrough curves illustrated the fea- lithium from salt lake brine by reaction-coupled separation tech-
sibility of lithium recovery in column adsorption process, nology. Sep. Purif. Technol. (2019). https​://doi.org/10.1016/j.
seppu​r.2019.11581​3
verified the competitive adsorption effects between lithium Ji, Z.Y., Yang, F.J., Zhao, Y.Y., Liu, J., Wang, N., Yuan, J.S.: Prepara-
and sodium ions, and provided the basic data for the further tion of titanium–base lithium ionic sieve with sodium persulfate as
systematic column adsorption experiments. eluent and its performance. Chem. Eng. J. 328, 768–775 (2017).
https​://doi.org/10.1016/j.cej.2017.07.047
Acknowledgement The research was supported by the National Key Jiang, H., Yang, Y., Sun, S., Yu, J.: Adsorption of lithium ions on
R&D Program of China (2017YFB0603104) and NSFC (U1407120, lithium–aluminum hydroxides: equilibrium and kinetics. Can. J.
U1707601, U1607112, 21776089, 51574126). Chem. Eng. (2019). https​://doi.org/10.1002/cjce.23640​
Kotsupalo, N.P., Ryabtsev, A.D., Poroshina, I.A., Kurakov, A.A.,
Mamylova, E.V., Menzheres, L.T., Korchagin, M.A.: Effect of
structure on the sorption properties of chlorine-containing form of
double aluminum lithium hydroxide. Russ. J. Appl. Chem. 86(4),
References 482–487 (2013). https​://doi.org/10.1134/s1070​42721​30400​46
Kozlova, S.G., Gabuda, S.P., Isupov, V.P., Chupakhina, L.: Using NMR
Besserguenev, A.V., Fogg, A.M., Francis, R.J., Price, S.J., O’Hare, in structural studies of aluminum hydroxide intercalation com-
D., Tolochko, V.P.I.: Synthesis and structure of the gibbsite pounds with lithium salts. J. Struct. Chem. 44(2), 198–205 (2003)
intercalation compounds [­ LiAl2(OH)6]X {X = Cl, Br, ­NO3} and Lawagon, C.P., Nisola, G.M., Mun, J., Tron, A., Torrejos, R.E.C., Seo,
­[LiAl2(OH)6]Cl·H2O using synchrotron X-ray and neutron powder J.G., Kim, H., Chung, W.J.: Adsorptive Li + mining from liquid
diffraction. Chem. Mater. 9(1), 241–247 (1997) resources by ­H2TiO3: equilibrium, kinetics, thermodynamics, and
Chabani, M., Amrane, A., Bensmaili, A.: Kinetic modelling of the mechanisms. J. Ind. Eng. Chem. 35, 347–356 (2016). https​://doi.
adsorption of nitrates by ion exchange resin. Chem. Eng. J. 125(2), org/10.1016/j.jiec.2016.01.015
111–117 (2006). https​://doi.org/10.1016/j.cej.2006.08.014 Lin X, Huang Q, Qi G, Xiong L, Huang C, Chen X, Li H, Chen X
Chitrakar, R., Makita, Y., Ooi, K., Sonoda, A.: Synthesis of iron-doped (2017) Adsorption behavior of levulinic acid onto microporous
manganese oxides with an ion-sieve property: lithium adsorption hyper-cross-linked polymers in aqueous solution: Equilibrium,
from bolivian brine. Ind. Eng. Chem. Res. 53(9), 3682–3688 thermodynamic, kinetic simulation and fixed-bed column stud-
(2014). https​://doi.org/10.1021/ie404​3642 ies. Chemosphere 171:231–239. doi:10.1016/j.chemo ​ s pher​
Choubey, P.K., Kim, M-s, Srivastava, R.R., Lee, J-c, Lee, J.-Y.: e.2016.12.084
Advance review on the exploitation of the prominent energy- Liu X, Zhong M, Chen X, Zhao Z (2018) Separating lithium and mag-
storage element: lithium. Part I: from mineral and brine resources. nesium in brine by aluminum-based materials. Hydrometallurgy
Miner. Eng. 89, 119–137 (2016). https​://doi.org/10.1016/j.minen​ 176:73–77. doi:10.1016/j.hydro​met.2018.01.005
g.2016.01.010 Martin, G., Rentsch, L., Höck, M., Bertau, M.: Lithium market
Choy KKH, McKay G (2005) Sorption of cadmium, copper, and zinc research—global supply, future demand and price develop-
ions onto bone char using Crank diffusion model. Chemosphere ment. Energy Storage Mater. 6, 171–179 (2017). https​://doi.
60(8):1141–1150. doi:10.1016/j.chemo​spher​e.2004.12.041 org/10.1016/j.ensm.2016.11.004
Choy, K.K.H., Porter, J.F., McKay, G.: Intraparticle diffusion in single Moazeni, M., Hajipour, H., Askari, M., Nusheh, M.: Hydrothermal
and multicomponent acid dye adsorption from wastewater onto synthesis and characterization of titanium dioxide nanotubes as
carbon. Chem. Eng. J. 103(1–3), 133–145 (2004). https​://doi. novel lithium adsorbents. Mater. Res. Bull. 61, 70–75 (2015).
org/10.1016/j.cej.2004.05.012 https​://doi.org/10.1016/j.mater​resbu​ll.2014.09.069
Dotto, G.L., Buriol, C., Pinto, L.A.A.: Diffusional mass transfer model Narins, T.P.: The battery business: Lithium availability and the growth
for the adsorption of food dyes on chitosan films. Chem. Eng. of the global electric car industry. Extract. Ind. Soc. 4(2), 321–328
Res. Des. 92(11), 2324–2332 (2014). https​://doi.org/10.1016/j. (2017). https​://doi.org/10.1016/j.exis.2017.01.013
cherd​.2014.03.013 Ooi K, Sonoda A, Makita Y, Chitrakar R, Tasaki-Handa Y, Nakazato
Foo, K.Y., Hameed, B.H.: Insights into the modeling of adsorption T (2017) Recovery of lithium from salt-brine eluates by direct
isotherm systems. Chem. Eng. J. 156(1), 2–10 (2010) crystallization as lithium sulfate. Hydrometallurgy 174:123–130.
Frohlich, A.C., Ocampo Perez, R., Diaz Blancas, V., Salau, N.P.G., doi:10.1016/j.hydro​met.2017.10.007
Dotto, G.L.: Three-dimensional mass transfer modeling of ibu- Ponnusami, V., Rajan, K.S., Srivastava, S.N.: Application of film-
profen adsorption on activated carbon prepared by sonication. pore diffusion model for methylene blue adsorption onto plant
Chem. Eng. J. 341, 65–74 (2018). https​://doi.org/10.1016/j. leaf powders. Chem. Eng. J. 163(3), 236–242 (2010). https​://doi.
cej.2018.02.020 org/10.1016/j.cej.2010.07.052

13
Adsorption

Ramakumar, S., Deviannapoorani, C., Dhivya, L., Shankar, L.S., Sun, Y., Yun, R., Zang, Y., Pu, M., Xiang, X.: Highly efficient lithium
Murugan, R.: Lithium garnets: synthesis, structure, Li + con- recovery from pre-synthesized chlorine-ion-intercalated LiAl-
ductivity, Li + dynamics and applications. Prog. Mater. Sci. 88, layered double hydroxides via a mild solution chemistry process.
325–411 (2017). https​://doi.org/10.1016/j.pmats​ci.2017.04.007 Materials (Basel, Switzerland) (2019). https​://doi.org/10.3390/
Ramesh, C., Yoji, M., Kenta, O., Akinari, S.: Lithium recovery from ma121​21968​
salt lake brine by H2TiO3. Dalton Trans. 43(23), 8933–8939 Vasiliu, S., Bunia, I., Racovita, S., Neagu, V.: Adsorption of cefotaxime
(2014) sodium salt on polymer coated ion exchange resin microparticles:
Ramos Suzaki, P.Y., Munaro, M.T., Triques, C.C., Kleinubing, S.J., Kinetics, equilibrium and thermodynamic studies. Carbohydr.
Fagundes Klen, M.R., de Matos Jorge, L.M., Bergamasco, R.: Polym. 85(2), 376–387 (2011). https​://doi.org/10.1016/j.carbp​
Biosorption of binary heavy metal systems: phenomenological ol.2011.02.039
mathematical modeling. Chem. Eng. J. 313, 364–373 (2017). https​ Wang S, Li P, Zhang X, Zheng S, Zhang Y (2017) Selective adsorp-
://doi.org/10.1016/j.cej.2016.12.082 tion of lithium from high Mg-containing brines using ­HxTiO3
Rengaraj, S., Joo, C.K., Kim, Y., Yi, J.: Kinetics of removal of ion sieve. Hydrometallurgy 174:21–28. doi:10.1016/j.hydro​
chromium from water and electronic process wastewater by met.2017.09.009
ion exchange resins: 1200H, 1500H and IRN97H. J. Hazard. Xiao, J.L., Sun, S.Y., Wang, J., Li, P., Yu, J.G.: Synthesis and adsorp-
Mater. 102(2–3), 257–275 (2003). https​://doi.org/10.1016/s0304​ tion properties of ­Li1.6Mn1.6O4 spinel. Ind. Eng. Chem. Res.
-3894(03)00209​-7 52(34), 11967–11973 (2013)
Ryu, T., Ryu, J.C., Shin, J., Lee, D.H., Kim, Y.H., Chung, K.-S.: Recov- Yao, C., Chen, T.: A new simplified method for estimating film mass
ery of lithium by an electrostatic field-assisted desorption process. transfer and surface diffusion coefficients from batch adsorption
Ind. Eng. Chem. Res. 52(38), 13738–13742 (2013). https​://doi. kinetic data. Chem. Eng. J. 265, 93–99 (2015)
org/10.1021/ie401​977s Zhang, Q., Sun, S., Li, S., Jiang, H., Yu, J.: Adsorption of lithium
Souza, P.R., Dotto, G.L., Salau, N.P.G.: Detailed numerical solution ions on novel nanocrystal ­MnO2. Chem. Eng. Sci. 62, 4869–4874
of pore volume and surface diffusion model in adsorption sys- (2007). https​://doi.org/10.1016/j.ces.2007.01.016
tems. Chem. Eng. Res. Des. 122, 298–307 (2017). https​://doi. Zhang, Q.H., Li, S.P., Sun, S.Y., Yin, X.S., Yu, J.G.: Lithium selec-
org/10.1016/j.cherd​.2017.04.021 tive adsorption on low-dimensional titania nanoribbons. Chem.
Souza, P.R., Dotto, G.L., Salau, N.P.G.: Statistical evaluation of pore Eng. Sci. 65(1), 165–168 (2010). https​: //doi.org/10.1016/j.
volume and surface diffusion model in adsorption systems. ces.2009.06.001
J. Environ. Chem. Eng. 5(6), 5293–5297 (2017). https​://doi.
org/10.1016/j.jece.2017.10.012 Publisher’s Note Springer Nature remains neutral with regard to
Sun, Y., Guo, X., Hu, S., Xiang, X.: Highly efficient extraction of jurisdictional claims in published maps and institutional affiliations.
lithium from salt lake brine by LiAl-layered double hydroxides
as lithium-ion-selective capturing material. J. Energy Chem. 34,
80–87 (2019). https​://doi.org/10.1016/j.jeche​m.2018.09.022

13