Magnetically separable magnetite–lithium manganese oxide nanocomposites as reusable lithium adsorbents in aqueous lithium resources

Chemical Engineering Journal 281 (2015) 541–548

Contents lists available at ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Magnetically separable magnetite–lithium manganese oxide nanocomposites as reusable lithium adsorbents in aqueous lithium resources Jihoon Kim, Seunghee Oh, Seung-Yeop Kwak ⇑ Department of Materials Science and Engineering, Research Institute of Advanced Materials (RIAM), Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 151-744, Republic of Korea

h i g h l i g h t s +

 Magnetite–lithium manganese oxides (M–LMOs) were prepared for Li adsorption.  M–LMOs had a spinel structure and contained two crystal phases. +

 Li adsorption efficiency was about 86% after 6 adsorption–desorption cycles.  M–LMOs were conveniently separated from a liquid under an external magnetic field.

a r t i c l e

i n f o

Article history: Received 12 May 2015 Received in revised form 18 June 2015 Accepted 19 June 2015 Available online 25 June 2015 Keywords: Lithium Recovery Lithium-manganese oxide Magnetite Magnetic separation

a b s t r a c t Spinel-structured lithium manganese oxides (LMOs) have generated considerable interest as adsorbents for the recovery of Li ions from aqueous Li resources such as brine, seawater, and concentrated seawater. However, practical applications are limited because powdered adsorbents are hard to handle and separate from a liquid. To overcome this problem, magnetically separable magnetite–LMO composite adsorbents (M–LMOs) were prepared by growing magnetite on LMO. The morphologies, crystal structures, chemical compositions, and magnetic properties of the prepared materials were characterized using various analytical techniques. The results confirmed that M–LMO had a spinel structure and contained two crystal phases. Li+ adsorption experiments were conducted using acid-treated M–LMO (M–HMO). The results confirmed that M–HMO was reusable and selectively adsorbed Li+ in the presence of Na+, K+, and Mg2+; the Li+ adsorption capacity was 6.84 mg/g in LiCl buffer solution and 1.2 mg/g in concentrated seawater, which is a much harsher condition than brine or seawater. M–HMO was conveniently separated from a liquid under an external magnetic field after Li+ adsorption. This is significantly different from conventional Li+ recovery systems such as granulation, foam formation, and membranization. These findings indicate that M–LMO could be used for Li+ recovery from aqueous Li resources and has good potential for practical applications. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction Interest in Li resources has been increasing because of the rapidly increasing demand for Li as a raw material for Li secondary batteries [1]. Li is present in brines and various minerals such as spodumene, petalite, lepidolite, and amblygonite [2]. Seawater is also a potential source of Li. Although the average concentration of Li+ in seawater is very low (0.17 mg/L), the total amount of Li+ in seawater is more than 2.5  1014 kg [2,3]. Li can be recovered from seawater directly or after concentration. Concentrated ⇑ Corresponding author. Tel.: +82 2 880 8365; fax: +82 2 885 9671. E-mail address: [email protected] (S.-Y. Kwak). http://dx.doi.org/10.1016/j.cej.2015.06.090 1385-8947/Ó 2015 Elsevier B.V. All rights reserved.

seawater obtained from desalination plants or salt farms is a good potential source of Li+. In recent years, there has been rapid growth in the construction of desalination plants [4]. Li+ recovery from concentrated seawater is therefore becoming an important challenge for academic researchers and in industry. Lithium manganese oxides (LMOs) with the spinel crystal structure are the most effective materials for Li+ recovery from aqueous Li resources such as seawater, brine, and concentrated seawater. However, powdered adsorbents cannot be used directly in aqueous Li resources because they are difficult to handle and adsorbent recovery from the liquid after Li+ adsorption is difficult, resulting in loss of adsorbent. Granulation, foam formation, and membranization methods have conventionally been used to deal with

542

J. Kim et al. / Chemical Engineering Journal 281 (2015) 541–548

these drawbacks. Macroporous silica beads [5], macroporous cellulose gel beads [6], and poly(vinyl chloride) (PVC) [7] have been used for granulation of Li adsorbents to an appropriate size. Foam adsorbents have been prepared using various binders such as pitch [8] or agar [9], with LMOs as precursors. These granulated and foam adsorbents are used in column systems for Li+ recovery; this method has high energy consumption and a high pressure is needed, leading to adsorbent loss. Other research groups have studied membrane-type Li-adsorbing systems as alternatives to column systems. Membrane-type adsorbents have been prepared by a solvent exchange method using PVC as a binder [10]. Flat sheet [11] and polymeric reservoir ‘‘tea bag’’ configurations [12] have been studied. These membrane-type adsorbents are suitable for Li+ recovery from aqueous Li resources because they can be used in continuous-operation systems, but operation of the pressurized flow component of the membrane system consumes energy [12]. The manufacturing costs are high, and a considerable amount of wastewater containing substances such as N,N-dimethylformamide and N,N-dimethylacetamide [9] is generated. In addition, coverage of the adsorbent surface by polymers decreases the number of active sites for Li+ adsorption. In this study, we investigated a magnetic separation method, using magnetic particles, as an alternative to conventional Li+ recovery systems. Magnetic separation is an effective and simple method for separating particles from a liquid or other fluid. The magnetic particles can be easily separated from solution by a magnetic field and the separated particles can be reused after removal of the adsorbed substance. This method has the advantages of nontoxicity, low cost, and facile synthesis. Magnetic particles are therefore widely used in various fields such as catalysis [13,14], water purification [15–17], and biomedical applications [18–20]. However, magnetic separation in Li+ recovery systems is in its nascent stage and further investigation is necessary to enable the replacement of conventional Li+ recovery systems for aqueous Li resources. A magnetically separable and reusable adsorbent was prepared by growing magnetite (Fe3O4) on LMOs with various Li precursor contents, and used for Li+ recovery. The morphologies, crystal structures, chemical compositions, and magnetic properties of the prepared magnetite–LMO (M–LMO) composite adsorbents were examined using various analytical techniques. The Li+ adsorption capacities of the composites were determined in LiCl buffer solution. Finally, the best composite adsorbent was evaluated for reusability, adsorption selectivity, and Li+ adsorption from concentrated seawater. 2. Experimental 2.1. Materials Manganese sulfate monohydrate (MnSO4H2O, Daejung Chemicals & Metals), lithium hydroxide monohydrate (LiOHH2O, Junsei), hydrogen peroxide (H2O2, 30%, Junsei), iron(II) sulfate heptahydrate (FeSO47H2O, 99%, Sigma–Aldrich), sodium hydroxide pellets (NaOH, Daejung Chemicals & Metals), sodium acetate anhydrous (Sigma–Aldrich), hydrochloric acid (HCl, 35%, Daejung Chemicals & Metals), and potassium permanganate (KMnO4, Junsei) were used as received, without further purification. Highly deionized water (18 MX/cm) was used in all experiments.

method [21]. The Li precursor solutions were prepared by dissolving various amounts of LiOHH2O and 1.2 M H2O2 in deionized water. The LiOH precursor solution (50 mL) was added dropwise to a 0.4 M MnSO4H2O solution (50 mL). The mixture was stirred with a magnetic stirrer for 2 h. The resulting mixture was poured into a Teflon-lined stainless-steel autoclave and crystallized at 110 °C for 8 h. The product was filtered and washed several times with deionized water. The obtained precipitate was dried at 60 °C for at least 12 h and calcined at 400 °C for 4 h. The LiOH precursor concentration was varied from 2.0 to 3.0 M to obtain LMOs with various Li/Mn ratios (denoted by LMO-2.0, LMO-2.5, and LMO-3.0). 2.2.2. Preparation of M–LMO composites The composites were synthesized by growing magnetite crystals on the LMO surface. FeSO47H2O (1.6 g) and sodium acetate (3.0 g) were dissolved in deionized and deoxygenated water in a N2 atmosphere. Then 40% NaOH (3 mL) was added dropwise, with magnetic stirring. The mixture was stirred for 30 min to form a homogeneous Fe(OH)2 solution. An LMO suspension, which was prepared by dispersing the LMO (0.5 g) in deionized water (20 mL) was added to the Fe(OH)2 solution. The mixture was magnetically stirred in a N2 atmosphere at room temperature for 1 h. The mixture was transferred to a Teflon-lined stainless-steel autoclave and reacted at 200 °C for 12 h. The products were separated using a magnet and washed with deionized water. The purification was repeated several times. After purification, the powders were dried at 60 °C for 12 h. M–LMOs prepared from different LMO precursors were denoted by M–LMO-2.0, M–LMO-2.5, and M– LMO-3.0. 2.3. Characterization The LMO and M–LMO crystal structures were examined using X-ray diffraction (XRD) at room temperature, with a Bruker New D8 Advance X-ray diffractometer with Cu Ka radiation (k = 1.5406 Å, 40 kV, 40 mA), with scanning from 10° to 80° at 10°/s. The average crystal diameter was calculated using the Debye–Scherrer formula, D = Kk/bcos h, where D is the crystallite size, K is a constant related to the crystal shape, and usually assumed to be 0.89, k is the wavelength of the X-ray radiation, b is the peak width (full width at half maximum) in radians, and h is the diffraction angle. The morphologies and crystal structures were investigated using high-resolution transmission electron microscopy (HR-TEM), with a JEOL JEM-3000F instrument, at an accelerating voltage of 300 kV. N2 adsorption–desorption isotherms were measured using an ASAP2010 instrument (Micromeritics). The specific surface areas of the samples were calculated, using the multiple-point Brunauer–Emmett–Teller method, from the desorption branch of the isotherm. The magnetic properties of the M–LMOs were investigated by vibrating sample magnetometry (VSM), using the dried powders, at room temperature, with a magnetic field in the range ±10 kOe. The sample compositions were determined using inductively coupled plasma atomic emission spectroscopy (ICP-AES; Varian 730ES). The mean oxidation number of Mn (ZMn) was estimated from the available oxygen, determined by redox titration using KMnO4 solution. X-ray photoelectron spectroscopy (XPS) was performed to determine the electronic and chemical states, using a Kratos Axis-His instrument with monochromatic Mg Ka radiation as the X-ray source.

2.2. Synthesis

2.4. Characterization of adsorption behavior

2.2.1. Preparation of spinel LMO LMO with the spinel structure was synthesized by a hydrothermal method, using a modified version of a previously reported

Li+ adsorption equilibrium experiments were conducted by stirring acid-treated LMO (HMO) or M–LMO (M–HMO) samples (100 mg) in 0.01 M LiCl buffer solution (100 mL) for 72 h at room

543

J. Kim et al. / Chemical Engineering Journal 281 (2015) 541–548

temperature; 0.2 M ammonium hydroxide/ammonium chloride solution was used to adjust the initial pH to 10.10. After adsorption equilibrium was achieved, the adsorbents were separated from the solution by centrifugation or using a magnet. The Li+ concentration in the solution was measured using ICP-AES. The Li+ adsorption capacity per gram of adsorbent (Qe) was calculated from the change in Li+ concentration after adsorption, as follows:

Q e ¼ ðC 0  C e ÞV=W where Qe is the amount of adsorbed Li+ per gram of adsorbent at equilibrium (mg/g), C0 is the initial concentration of the Li solution (mg/L), Ce is the concentration of Li remaining in solution at equilibrium (mg/L), V is the solution volume (mL), and W is the weight of the adsorbent (g). To determine the reusability, the adsorption–desorption cycles were repeated using M–HMO. After Li+ adsorption, M–HMO was regenerated by shaking for 24 h in 0.5 M HCl solution at room temperature. M–HMO was separated using a magnet, washed several times with deionized water, and dried at 60 °C for the next adsorption cycle. Li+ adsorption by the regenerated M– HMO was conducted under the same conditions. This process was repeated 6 times. The selectivity of M–HMO for Li+ over other cations, mainly Na+, K+, and Mg2+, coexisting in aqueous Li resources was evaluated by comparing the distribution coefficients (Kd) of the ions. The Kd value was determined by dispersing M–HMO (0.1 g) in a solution (10 mL, pH 10.10) containing Li+, Na+, K+, and Mg2+ at concentrations of about 0.01 M, and shaking the vial containing the solution to keep the adsorbent dispersed in the solution for 72 h at room temperature. After equilibrium was reached, the concentration of each ion remaining in solution was determined using ICP-AES. The Kd of each cation was determined from the measured cation concentration using the following equation: K d ðmL=gÞ ¼

Adsorbed cation concentration per gram of adsorbent ðmg=gÞ Cation concentration remaining in solution ðmg=mLÞ

Li+ adsorption by M–HMO from concentrated seawater obtained from a salt farm in Shinan, South Korea was performed. M–HMO (0.1 g/L) was immersed in concentrated seawater (Li+: 0.7 ppm) and stirred with a mechanical stirrer for 5 d at room temperature. After separating the M–HMO using a magnet, the seawater was sampled and analyzed using ICP-AES. 3. Results and discussion 3.1. Preparation of LMO The XRD patterns of LMO-2.0, LMO-2.5, and LMO-3.0 in Fig. 1 show that the LMOs have the cubic spinel crystal structure. The positions and relative intensities of all the diffraction peaks corresponding to the (1 1 1), (3 1 1), (2 2 2), (4 0 0), (3 3 1), (5 1 1), (4 4 0), and (5 3 1) planes are in good agreement with those of LMO [Joint Committee on Powder Diffraction Standards (JCPDS) No. 35-0782]. The (4 0 0)/(3 1 1) plane intensity ratio increased with increasing amount of LiOH. It has previously been reported that an increase in the (4 0 0)/(3 1 1) plane intensity ratio indicates that the (4 0 0) plane is growing faster than the (3 0 0) plane [22]. The difference between the growth speeds is caused by suppression of the Jahn–Teller deformation. This is because increasing the LiOH content in the spinel results in more Li ions occupying the 16 d octahedral sites (Mn site). The average crystal sizes of the LMOs, calculated using the Debye–Scherrer equation, were 18.2, 16.9, and 15.7 nm for LMO-2.0, LMO-2.5, and LMO-3.0, respectively. The decreasing crystallite size with increasing amount of LiOH may be caused by the increased concentration of hydroxide ions [23]. The repulsive electrical force on the LMO nanoparticle

Fig. 1. XRD patterns of (a) LMO-2.0, (b) LMO-2.5, and (c) LMO-3.0.

surfaces increases with increasing hydroxide ion concentration, and this suppress further growth of the LMO nanoparticles. The sample structures were further investigated using HR-TEM. A representative image is shown in Fig. 2. As shown in Fig. 2(a), most of the LMO nanoparticles are agglomerated in clusters, because of the high surface energy resulting from their small particle size. The HR-TEM and fast Fourier transform (FFT) images of an LMO particle, shown in Fig. 2(b), indicate that the d-spacing between the lattice fringes is about 4.8 Å, which corresponds to that of the (1 1 1) plane of spinel LMO. The specific surface areas of the samples, calculated from N2 adsorption–desorption isotherms, were 93.95 (LMO-2.0), 113.11 (LMO-2.5), and 150.92 (LMO-3.0) m2/g. The specific surface area increased with increasing LiOH concentration, as expected, because smaller crystallite particles have higher surface areas. These results are consistent with the XRD results. A summary of the physical properties of the LMOs is presented in Table S1. The results of chemical analysis of the LMOs are summarized in Table 1. The Li/Mn ratio in the LMOs increased with increasing LiOH content. The average Mn oxidation numbers (ZMn) of the samples were 3.89 (LMO-2.0) and 4.00 (LMO-2.5, LMO-3.0). Singular and higher oxidation states of metals tend to enhance the chemical stabilities of metal oxides [24,25]. The ZMn value can be used as an indicator of chemical stability in the LMO during Li+ extraction. The Li+ extraction process in acidic solution involves two reactions, i.e., a redox reaction and an ion-exchange reaction [26,27]. The proposed redox mechanism of Li+ extraction is 3þ

þ

2þ 4LiMn Mn4þ O4 þ 8Hþ ! 4Li þ 3Mn4þ þ 4H2 O 2 O4 þ 2Mn

ð1Þ

+

The Li extraction progresses by an ion-exchange mechanism: 4þ

þ

LiMn2 O4 þ Hþ $ HMn4þ 2 O4 þ Li

3+

ð2Þ 4+

2+

The disproportionation of Mn to Mn and Mn is the driving force for the redox-type Li+ extraction. Eq. (1) shows that this reaction is irreversible, and the chemical stability of the LMO may be low as a result of dissolution of Mn2+; this indicates that an LMO containing a large number of Mn3+ ions cannot be used as a reusable Li adsorbent. The proportion of Mn3+ in the LMO could therefore indicate its chemical stability; the chemical stability can be

544

J. Kim et al. / Chemical Engineering Journal 281 (2015) 541–548

Fig. 2. Representative HR-TEM images of (a) LMO cluster and (b) particle in LMO, with corresponding FFT image (inset).

MnZþ þ Fe2þ þ nOH ! MnAþ þ Fe3þ þ nOH ! MnAþ

Table 1 The chemical properties of LMOs. Sample

Li/Mn ratio

ZMn

Chemical formula

LMO-2.0 LMO-2.5 LMO-3.0

0.625 0.825 1.002

3.89 4.00 4.00

Li1.10Mn1.77O4 Li1.37Mn1.66O4 Li1.60Mn1.60O4

þ FeðOHÞ3 ðZ : 3:5—4:0; A :< 3:5Þ

The oxidation of Fe ions results in precipitation of Fe(OH)3 on the LMO surface, and Fe3O4 (magnetite) is formed by coprecipitation of Fe(OH)3 with Fe(OH)2 in the mixture [28]:

Fe2þ þ 2Fe3þ þ 8OH ! Fe3 O4 þ H2 O predicted from the Li/Mn ratio and ZMn value of the LMO [26]. Based on chemical analysis, it was concluded that LMO-2.5 and 3.0 were more stable than LMO-2.0. 3.2. Preparation of M–LMO An outline of the synthetic route to M–LMO is shown in Fig. 3. The reaction scheme is based on the method reported by Krishnamurti et al. [28]. In the reported work, magnetite crystals were easily precipitated on a montmorillonite surface by adding montmorillonite to an alkaline solution of Fe2+. On the basis of this work, we propose a mechanism for the formation of M–LMO. The Mn ions on the LMO surface are reduced and the Fe2+ ions are oxidized, in the following redox reaction [29]:

ð3Þ

2+

ð4Þ

The unreacted Fe(OH)2 can be converted to Fe3O4, via the Schikorr reaction, at high temperature [30,31]:

3FeðOHÞ2 ! Fe3 O4 þ 2H2 O þ H2 ð> 100  CÞ

ð5Þ

The Fe3O4 formed by the Schikorr reaction is dissolved and deposited on the already-formed magnetite crystals on the surface of the LMO, by Ostwald ripening. As a result of these reactions, the Fe sources form magnetite crystals, which grow on the LMO surface. Fig. 4 shows the XRD patterns of a representative M–LMO sample, synthesized from LMO-2.0; two crystal phases are present, i.e., LMO (.) and magnetite (d). Although there are few XRD peaks for the LMO phase, it is easy to differentiate between the two phases because the main peaks for each phase are quite different. The XRD patterns of M–LMO are also different from those of the

Fig. 3. Outline of synthetic route to M–LMO composite adsorbents.

545

J. Kim et al. / Chemical Engineering Journal 281 (2015) 541–548 Table 2 The properties of M–LMOs.

Fig. 4. Representative M–LMO XRD patterns.

mixture of LMO and magnetite (Fig. S1). The morphologies and crystal structures of the M–LMOs were examined using HR-TEM. The results shown in Fig. 5 confirm the presence of two crystal phases. The upper part of the lattice fringes corresponds to the (1 1 1) plane of LMO, with a d-spacing of 4.8 Å. The lattice fringes of M–LMO in the lower part correspond to the (3 1 1) plane of magnetite, with a d-spacing of 2.7 Å. The XRD and HR-TEM results therefore show that magnetite crystals were grown on the LMO. The average crystal size was calculated from the (3 1 1) plane peaks of the LMO and magnetite phase, which overlapped. The M–LMO crystallite size decreased as that of the LMO decreased (Table 2). The magnetite crystals grew on the LMO, therefore the M–LMO crystallite size depended on the LMO crystallite size. The magnetic properties of the M–LMOs at ambient temperature are shown in Fig. 6. The magnetization loops of the M–LMOs were nonlinear with reversible characteristics and no hysteresis, indicating superparamagnetic behavior. The saturated magnetization values (Ms) of the M–LMOs were 44.4, 30.1, and 25.0 emu/g for M–LMO-2.0, M–MO-2.5, and M–LMO-3.0, respectively; these are smaller than that of bulk magnetite. This is attributed to the presence of the LMO phase, which is nonmagnetic, and accounts for part of the total M–LMO mass. However, the Ms values of the M–LMOs were sufficient for them to respond rapidly to an applied

Sample

Crystallite size (nm)

Ms (emu/g)

Li/Mn ratio

Fe/Mn ratio

Mass ratio (LMO:Fe3O4)

M–LMO-2.0 M–LMO-2.5 M–LMO-3.0

26.2 17.5 15.0

44.42 30.09 25.02

0.196 0.214 0.330

1.308 1.377 1.421

0.486:0.514 0.483:0.517 0.482:0.518

magnetic field. The inset photograph shows that the M–LMOs were easily attracted and separated from the aqueous solution by a magnet. M–LMO-2.0 had the highest saturated magnetization value. The magnetic behavior of magnetite is very sensitive to its particle size [32,33]. The higher saturated magnetization value of M– LMO-2.0 shows that it has a larger crystallite size; this is in good agreement with the crystallite size obtained from the XRD data. The Li/Mn and Fe/Mn ratios, and the mass ratios were determined by chemical analysis; the results are listed in Table 2. The mass ratio was calculated from the atomic ratio and chemical formula. The Li/Mn ratios of the M–LMOs were lower than those of the LMOs because magnetite crystal growth destroyed the LMO structure (Fig. S2). The Fe/Mn ratios of the M–LMOs and the mass ratios were similar to each other. This suggests that the Fe sources were almost completely converted to magnetite, and the mass ratio of each sample was determined by the amount of Fe source. XPS was performed to identify the chemical states of the atoms on the M–LMO surfaces. The M–LMO XP spectra had Fe 2p signals at 710 and 723 eV, as shown in Fig. 7(a). These signals are attributed to magnetite, and correspond to the 2p3/2 and 2p1/2 spin–orbital components; this indicates that magnetite crystals were grown on the LMO. Fe 2p signals were not observed for the LMOs. Fig. 7(b) shows the Mn 2p region. The Mn 2p3/2 peak of LMO shifted to a lower binding energy because of a decrease in the Mn oxidation state [34]. This result supports the proposed mechanism for M–LMO synthesis, i.e., growth of magnetite crystals on the LMO starts with a redox reaction between Mn ions and Fe ions. 3.3. Characterization of adsorption behavior Li+ adsorption experiments were performed on the LMOs and M–LMOs after HCl treatment (HMOs and M–HMOs). The Li+ uptakes by the HMOs and M–HMOs are listed in Table 3. The Li+ uptakes by the HMOs increased with increasing Li/Mn ratio, which suggests that larger amounts of Li were inserted into the spinel LMO crystal, and substituted Mn sites, as the Li/Mn ratio increased.

Fig. 5. Representative HR-TEM images of (a) M–LMO cluster and (b) particle in M–LMO, with the corresponding FFT images.

546

J. Kim et al. / Chemical Engineering Journal 281 (2015) 541–548

Fig. 6. Magnetization curves for M–LMOs as function of applied field, measured at room temperature, and photograph of M–LMO in presence of magnetic field (inset).

Fig. 7. XP spectra of LMO and M–LMO: (a) Fe 2p region and (b) Mn 2p region.

Table 3 Li+ adsorption capacity of the samples. Sample

Li+ adsorption capacity (mg/g)

Expected Li+ adsorption capacity (mg/g)

Residual Li+ adsorption capacity ratio

HMO-2.0 HMO-2.5 HMO-3.0 M–HMO-2.0 M–HMO-2.5 M–HMO-3.0

20.84 24.28 30.61 3.97 6.84 7.37

– – – 10.13 11.73 14.75

– – – 39.19% 58.31% 49.94%

The expected Li+ uptakes by the M–HMOs were calculated from the Li+ uptake by the HMOs and the mass ratios of the LMOs in the M– HMOs. The experimental Li+ uptakes by the M–HMOs were lower than the calculated values. M–HMO-2.0 showed the highest loss of Li+ uptake capacity, and the residual capacity ratio was 39%; HMO-2.0 has the lowest chemical stability, based on the Mn oxidation state. This result suggests that the decreased Li+ uptake capacity was caused by LMO destruction during magnetite crystal growth. M–HMO-2.5, which had the highest residual Li+ uptake

capacity, was selected for further investigation of the adsorption behavior. The reusability of M–HMO-2.5 was tested by performing the adsorption–desorption cycles 6 times; the results are shown in Fig. 8. The Li+ adsorption efficiency decreased slightly with increasing cycle number, because the acid treatment damaged the M– HMO structure during Li+ extraction. After reuse 6 times, the Li+ adsorption efficiency was about 86%, which suggests that M– HMO-2.5 has good chemical stability and could be recycled from aqueous Li resources for Li+ recovery. This is because M– HMO-2.5 contains Mn with a singular and high oxidation state. As described in Section 3.1, for an LMO with a low Mn oxidation state, Li+ extraction involves both redox and ion-exchange reactions. The redox reaction destroys the LMO structure, by dissolution of Mn2+, which hinders use of the LMO as a reusable adsorbent. The Li+ extraction and insertion reactions mainly proceed by ion-exchange in M–HMO-2.5, in which the oxidation state of Mn is 4. The adsorption selectivity for Li+ was determined by comparing the equilibrium distribution coefficients (Kd) of Li+, and Na+, K+, and Mg2+, which are the main elements coexisting in seawater

J. Kim et al. / Chemical Engineering Journal 281 (2015) 541–548

547

capacity is therefore more severe in concentrated seawater. Harvianto et al. reported that the Li+ adsorption efficiency for concentrated seawater was lower than that for seawater [36]. Table 5 lists the Li+ adsorption capacities of previously studied manganese oxide composite adsorbents in seawater and brine [6,8–10]. These manganese oxide composite adsorbents were fabricated by granulation or membranization methods to enable the use of powder-type adsorbents in aqueous Li resources for practical industrial applications. Despite the harsh environmental conditions, and much higher amounts of coexisting cations than in seawater and much lower Li+ content than in brine (brine has over 200 ppm Li+), M–HMO-2.5 showed a reasonable Li+ adsorption capacity compared with other adsorbents. Moreover, M–HMO-2.5 is easily separated magnetically and can be reused without loss of adsorbent, unlike other manganese oxide composite adsorbents. M–HMO-2.5 is expected to be an efficient reusable Li adsorbent in aqueous Li resources. 4. Conclusions +

Fig. 8. Li adsorption efficiency of M–HMO-2.5 as a function of cycle number.

Table 4 The distribution coefficient of each metal ions on M–HMO-2.5.

Kd (mL/g)

Li+

Na+

K+

Mg2+

1621.24

3.43

2.89

4.15

Table 5 Li+ adsorption capacity for different manganese oxide composite adsorbent in aqueous lithium resources. Adsorbent

Aqueous lithium resources

Binder content (%)

Li+ adsorption capacity (mg/g)

Refs.

Cellulose Sponge foam Spherical foam Membrane M–HMO-2.5

Seawater Brine Seawater

30 70 2

2.0 1.5 3.4

[6] [8] [9]

Seawater Concentrated seawater

20 50

4.0 1.2

[10] This work

In this study, M–LMO composite adsorbents were prepared, to overcome the shortcomings of powder-type adsorbents in liquids. XRD and HR-TEM showed that the M–LMOs consisted of magnetite and LMO. Chemical analysis and VSM measurements verified that magnetite crystals were grown on the LMO; this changed the Li/Mn ratio, because of LMO destruction. Li+ adsorption experiments on the M–LMOs were performed in LiCl buffer solution. Among the composite adsorbents, M–HMO-2.5 showed the highest residual Li+ adsorption capacity, which means that M–HMO-2.5 had the highest chemical stability. After 6 adsorption–desorption cycles, the Li+ adsorption efficiency was over 86%. For M– HMO-2.5, the order of the equilibrium distribution coefficients (Kd) was Li+  Mg2+ > Na+ > K+, indicating high selectivity for Li+. M–LMO-2.5 displayed a reasonable Li+ adsorption capacity in concentrated seawater, despite the harsh environmental conditions. M–LMO-2.5 was easily separated from a liquid under an external magnetic field within a few minutes. These results show that M– LMOs are promising candidates as convenient, reusable adsorbents for use in aqueous Li resources. Further investigations to decrease LMO destruction by magnetite growth, and adsorption tests using other aqueous Li resources are in progress. Acknowledgement

(Table 4). The order of the Kd values of these metal ions was Li+  Mg2+ > Na+ > K+, indicating high selectivity for Li+, but low selectivities for Na+, K+, and Mg2+. It is difficult for the coexisting metal ions to access the adsorption sites on M–HMO because their ionic radii and free energies of hydration are larger than those of Li+ [35]. The results show that M–HMO-2.5 is an appropriate material for selective adsorption of Li+ from aqueous Li resources in which larger amount of other metal ions are present, such as brine, seawater, and concentrated seawater. Adsorption experiments were performed on M–HMO-2.5 in concentrated seawater that contained 0.7 ppm Li+, 71,750 ppm Na+, 42,370 ppm Mg2+, 8750 ppm K+, and other cations. The Li+ adsorption capacity from concentrated seawater was 1.2 mg/g, which was lower than that from LiCl buffer solution. The Li+ activity in solution decreased with increasing complexity of the solution system, and coexisting cations were adsorbed on the adsorbent surface by electrostatic forces; this resulted in formation of a repulsion region, resulting in a decrease in the Li+ adsorption capacity [8]. Generally, the concentrations in seawater are 0.17 ppm Li+, 10,900 ppm Na+, 1310 ppm Mg2+, and 390 ppm K+. The cation concentrations in concentrated seawater are much higher than those in seawater. The decrease in the Li+ adsorption

This work was supported by ‘‘The GAIA Project’’ of the Korea Ministry of the Environment (RE201402059). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cej.2015.06.090. References [1] A. Ebensperger, P. Maxwell, C. Moscoso, The lithium industry: its recent evolution and future prospects, Resour. Pol. 30 (2005) 218–231. [2] R. Chitrakar, H. Kanoh, Y. Miyai, K. Ooi, Recovery of lithium from seawater using manganese oxide adsorbent (H1.6Mn1.6O4) derived from Li1.6Mn1.6O4, Ind. Eng. Chem. Res. 40 (2001) 2054–2058. [3] Y. Miyai, K. Ooi, S. Katoh, Recovery of lithium from seawater using a new type of ion-sieve adsorbent based on MgMn2O4, Sep. Sci. Technol. 23 (1988) 179– 191. [4] M. Elimelech, W.A. Phillip, The future of seawater desalination: energy, technology, and the environment, Science 333 (2011) 712–717. [5] Y. Onodera, T. Iwasaki, H. Hayashi, K. Torii, Preparation method and lithium adsorption property of k-MnO2-silica composite, Chem. Lett. 19 (1990) 1801– 1804. [6] F. Sagara, W.B. Ning, I. Yoshida, K. Ueno, Preparation and adsorption properties of k-MnO2-cellulose hybrid-type ion-exchanger for lithium ion. Application to

548

[7]

[8]

[9]

[10]

[11]

[12]

[13] [14]

[15]

[16]

[17]

[18]

[19] [20]

J. Kim et al. / Chemical Engineering Journal 281 (2015) 541–548 the enrichment of lithium ion from seawater, Sep. Sci. Technol. 24 (1989) 1227–1243. G. Xiao, K. Tong, L. Zhou, J. Xiao, S. Sun, P. Li, J. Yu, Adsorption and desorption behavior of lithium ion in spherical PVCMnO2 ion sieve, Ind. Eng. Chem. Res. 51 (2012) 10921–10929. L.-W. Ma, B.-Z. Chen, Y. Chen, X.-C. Shi, Preparation, characterization and adsorptive properties of foam-type lithium adsorbent, Microporous Mesoporous Mater. 142 (2011) 147–153. Y. Han, H. Kim, J. Park, Millimeter-sized spherical ion-sieve foams with hierarchical pore structure for recovery of lithium from seawater, Chem. Eng. J. 210 (2012) 482–489. 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. Z. Ji, J. Yuan, X. Guo, J. Wang, L. Li, A preliminary study on preparation of lithium ion-sieve flat sheet membrane, Appl. Mech. Mater. 161 (2012) 144– 147. K.-S. Chung, J.-C. Lee, W.-K. Kim, S.B. Kim, K.Y. Cho, Inorganic adsorbent containing polymeric membrane reservoir for the recovery of lithium from seawater, J. Membr. Sci. 325 (2008) 503–508. U. Jeong, X. Teng, Y. Wang, H. Yang, Y. Xia, Superparamagnetic colloids: controlled synthesis and niche applications, Adv. Mater. 19 (2007) 33–60. P.D. Stevens, J. Fan, H.M.R. Gardimalla, M. Yen, Y. Gao, Superparamagnetic nanoparticle-supported catalysis of Suzuki cross-coupling reactions, Org. Lett. 7 (2005) 2085–2088. X. Zhao, Y. Shi, Y. Cai, S. Mou, Cetyltrimethylammonium bromide-coated magnetic nanoparticles for the preconcentration of phenolic compounds from environmental water samples, Environ. Sci. Technol. 42 (2008) 1201–1206. A.-F. Ngomsik, A. Bee, J.-M. Siaugue, V. Cabuil, G. Cote, Nickel adsorption by magnetic alginate microcapsules containing an extractant, Water Res. 40 (2006) 1848–1856. W. Yantasee, C.L. Warner, T. Sangvanich, R.S. Addleman, T.G. Carter, R.J. Wiacek, G.E. Fryxell, M.G. Warner, Removal of heavy metals from aqueous systems with thiol functionalized superparamagnetic nanoparticles, Environ. Sci. Technol. 41 (2007) 5114–5119. H.M.R. Gardimalla, D. Mandal, P.D. Stevens, M. Yen, Y. Gao, Superparamagnetic nanoparticle-supported enzymatic resolution of racemic carboxylates, Chem. Commun. 35 (2005) 4432–4434. S.S. Banerjee, D.-H. Chen, Magnetic nanoparticles grafted with cyclodextrin for hydrophobic drug delivery, Chem. Mater. 19 (2007) 6345–6349. S. Singh, K.C. Barick, D. Bahadur, Inactivation of bacterial pathogens under magnetic hyperthermia using Fe3O4–ZnO nanocomposite, Powder Technol. 269 (2015) 513–519.

[21] Q.-H. Zhang, S.-P. Li, S.-Y. Sun, X.-S. Yin, J.-G. Yu, LiMn2O4 spinel direct synthesis and lithium ion selective adsorption, Chem. Eng. Sci. 65 (2010) 169– 173. [22] L. Wang, W. Ma, R. Liu, H.Y. Li, C.G. Meng, Correlation between Li+ adsorption capacity and the preparation conditions of spinel lithium manganese precursor, Solid State Ionics 177 (2006) 1421–1428. [23] K. Matsui, M. Ohgai, Formation mechanism of hydrous-zirconia particles produced by hydrolysis of ZrOCl2 solutions, J. Am. Ceram. Soc. 80 (1997) 1949– 1956. [24] K.F. Purcell, J.C. Kotz, Inorganic Chemistry, W.B. Saunders, Philadelphia, 1977. [25] N. Valverde, C. Wagner, Considerations on the kinetics and the mechanism of the dissolution of metal oxides in acidic solutions, Ber. Bunsenges. Phys. Chem. 80 (1976) 330–333. [26] Q. Feng, Y. Miyai, H. Kanoh, K. Ooi, Li+ extraction/insertion with spinel-type lithium manganese oxides. Characterization of redox-type and ion-exchangetype sites, Langmuir 8 (1992) 1861–1867. [27] Q. Feng, H. Kanoh, K. Ooi, Manganese oxide porous crystals, J. Mater. Chem. 9 (1999) 319–333. [28] G.S.R. Krishnamurti, A. Violante, P.M. Huang, Influence of montmorillonite on Fe(II) oxidation products, Clay Miner. 33 (1998) 205–212. [29] A. Navrotsky, Thermodynamics of A3O4–B3O4 spinel solid solutions, J. Inorg. Nucl. Chem. 31 (1969) 59–72. [30] G. Schikorr, Über die reaktionen zwischen eisen, seinen hydroxyden und wasser, Z. Electrochem. 35 (1929) 65–70. [31] P.S. Joshi, G. Venkateswaran, K.S. Venkateswarlu, K.A. Rao, Stimulated decomposition of Fe(OH)2 in the presence of AVT chemicals and metallic surfaces-relevance to low-temperature feedwater line corrosion, Corrosion 49 (1993) 300–309. [32] I. Bica, Nanoparticle production by plasma, Mater. Sci. Eng. B 68 (1999) 5–9. [33] J. Xu, H. Yang, W. Fu, K. Du, Y. Sui, J. Chen, Y. Zeng, M. Li, G. Zou, Preparation and magnetic properties of magnetite nanoparticles by sol–gel method, J. Magn. Magn. Mater. 309 (2007) 307–311. [34] Y. Zhai, J. Zhai, M. Zhou, S. Dong, Ordered magnetic core–manganese oxide shell nanostructures and their application in water treatment, J. Mater. Chem. 19 (2009) 7030–7035. [35] G. Zhu, P. Wang, P. Qi, C. Gao, Adsorption and desorption properties of Li+ on PVC-H1.6Mn1.6O4 lithium ion-sieve membrane, Chem. Eng. J. 235 (2014) 340– 348. [36] G.R. Harvianto, S.-H. Kim, C.-S. Ju, Solvent extraction and stripping of lithium ion from aqueous solution and its application to seawater, Rare Met. (2015), http://dx.doi.org/10.1007/s12598-015-0453-1. in press.