English

• Lithium is known as a new energy metal and drives the world forward^[1]. The rapid development of new energy vehicles leads to the increasing demand for lithium^[2-3]. Lithium are distributed in seawater, salt lake brines and minerals such as spodumene and lepidolite. China has abundant salt lake brines with high mass ratios of Mg to Li.

  Lithium ion-sieve adsorption is the most promising method of recovering lithium from salt lake brines due to its low cost, no pollution and high selectivity. Manganese oxide-based ion sieve H_1.6Mn_1.6O₄ has the highest theoretical adsorption capacity (68 mg·g^-1) and attracts most attention. The most common synthesis process of lithium ion-sieve is hydrothermal formation of LiMnO₂ followed by roasting and acid pickling. Two types of Mn³⁺ resource were adopted during the hydrothermal synthesis of LiMnO₂: One is Mn₂O₃ obtained by roasting electrolytic MnO₂; The other depends on the in-situ formation of MnOOH using oxidant and Mn²⁺-containing solutions^[4-10]. The latter is more convenient compared to the former and widely applied for the synthesis of lithium ion-sieve. However, the actual lithium adsorption capacity in salt lake brine was generally lower than 30 mg·g^-1 and the dissolution percentage of manganese from Li_1.6Mn_1.6O₄ during the acid treatment greater than 3%^[11-14], which limits the industrial application. Many scholars focused on the effect of synthesis parameters on the adsorption capacity and stability of lithium ion-sieve. In fact, Mn²⁺ is easy to oxidize by oxygen in the process of in-situ formation of MnOOH, which makes the synthesis process carry out in an inert atmosphere^[15]. However, keeping an inert atmosphere is inconvenient and high cost. It is seldom reported the effect of oxygen as oxidant on the composition in the process of hydrothermal synthesis of LiMnO₂.

  In this paper, lithium ion-sieve was synthesized by hydrothermal oxidation of MnSO₄ with KMnO₄ to form LiMnO₂ following solid-phase roasting and acid pickling. The effects of hydrothermal temperature, the molar ratio of MnO₄^- to Mn²⁺ and oxygen in the air on the composition of LiMnO₂ were investigated in detail. The adsorption kinetics and separation performance of lithium ion sieve in salt lake brine have been discussed.

  1.   Experimental

  1.1   Materials and reagents

  LiOH·H₂O, KMnO₄ and MnSO₄ produced by Sinopharm Chemical Reagent Co., Ltd. were used as raw materials, and all reagents were of analytical grade. Salt lake brine was from the Lop Nor salt lake and its main composition was listed in Table 1.

  Table 1

  

  Table 1.  Composition of salt lake brine  mg·L^-1

  下载: 导出CSV

  Mg	Na	K	Li	Ca	S	B	Si
  106 100	5 020	2 480	220.2	27.24	6 831	519.0	28.05
  P	Mo	Fe	As	Co	Ga	Cu	V
  5.26	4.50	0.87	0.63	0.53	0.45	0.42	0.32
  Mn	Sn	Cr	Al	Ni	Pb	Y	Cd
  0.28	0.23	0.20	0.13	0.12	0.04	0.04	0.03

  1.2   Pretreatment process

  MnSO₄ solution and KMnO₄ solution were simultaneously added into a LiOH solution through a flowmeter. The mixed solution volume was kept around 500 mL and molar ratio of Li to Mn fixed as 4.5. After mechanically stirring for 2 h, the mixture was transferred into a 1.0 L autoclave.

  1.3   Preparation of LiMnO₂ and precursor Li_1.6Mn_1.6O₄

  LiMnO₂ precipitate was obtained by hydrothermal treating the above mixture for 18 h at a certain temperature. After filtration and washing several times with deionized water, the resulting LiMnO₂ was calcined in a muffle furnace at 350~500 ℃ for 4 h to form the precursor of lithium ion-sieve Li_1.6Mn_1.6O₄.

  1.4   Preparation of lithium ion-sieve

  The precursor (0.6 g) was added to 100 mL of 0.5 mol·L^-1 HCl at 30 ℃ and stirred for 3 hours, centrifuged, washed, and dried to obtain lithium ion sieve. The percentage of dissolved manganese and extraction percentage of Li⁺ from each precursor during acid treatment were calculated by the following equation:

  $ {R_i} = \frac{{{c_i}{v_1}}}{{{m_{\rm{s}}}{w_i}}} $

  (1)

  where R_i is the rate of dissolved metal ions, %; c_i is the concentration of metal ions in the supernatant (mg·L^-1); v₁ is the volume of the solution (L); m_s is the mass of the precursor (mg); w_i is the mass fraction of metal in the precursor (%).

  1.5   Adsorption properties of lithium ion-sieve

  The lithium ion-sieve (0.30 g) was immersed in 100 mL salt lake brine and stirred at 30 ℃ for 20 h. Then the suspension was filtered and the concentra-tions of all the metal ions in the supernatant were determined. The adsorption capacity or the amount of metal ion adsorbed per gram of ion sieve at equili-brium (Q_e), distribution coefficient (K_d), and separation factor (α_Me^Li) were calculated using equations (2~4):

  $ {Q_t} = \frac{{{v_2}\left( {{c_0} - {c_t}} \right)}}{{{m_0}}} $

  (2)

  $ {K_{\rm{d}}} = \frac{{{c_0} - {c_t}}}{{{m_0}{c_t}}} $

  (3)

  $ \alpha _{{\rm{Me}}}^{{\rm{Li}}} = \frac{{{k_{\rm{d}}}({\rm{Li}})}}{{{k_{\rm{d}}}({\rm{Me}})}} $

  (4)

  where Q_t represents the adsorption capacity of Li⁺ (mg·g^-1), c₀ is the concentration of Li⁺ in the salt lake brine (mg·L^-1), c_t is the concentration of Li⁺ in the filtrate after adsorption (mg·L^-1), v₂ is the volume of salt lake brine (L), m₀ is the mass of lithium ion-sieve (g); K_d represents the distribution coefficient of metal ion Meⁿ⁺ in brine, and α_Me^Li is a separation coefficient between lithium and Me (Na⁺, K⁺, Mg²⁺) in the adsorp-tion process.

  1.6   Adsorption kinetics

  The lithium ion-sieve (0.30 g) was added to 100 mL of salt lake brine and stirred at 25 ℃, samples were taken at 1, 2, 6, 12, 20 and 30 h respectively to determine the concentration of Li⁺ in the supernatant to calculate the adsorption capacity of lithium ion sieve. The adsorption process of lithium ion sieve was simulated by pseudo-second-order kinetic model and pseudo-first-order kinetic model, respectively.

  1.7   Circulating property of lithium ion sieve

  The desorption and adsorption of Li⁺ have been repeated 5 times to test the recycling performance of the lithium ion sieve. In each desorption process, the precursor Li_1.6Mn_1.6O₄ were immersed into 0.5 mol·L^-1 HCl solution. The ratio of solid to liquid was controlled at 1/100 (w/V, g·mL^-1). After acid pickling at 30 ℃ for 4 h, the lithium ion-sieve was filtered, washed and dried. Then the ion sieve was added into the salt lake brine to adsorb lithium. The above operations were termed as one cycle. The concentra-tions of manganese and lithium in the effluents were analyzed to calculate the dissolution percentage of manganese and adsorption capacity of lithium of each cycle. The percentage of dissolved manganese and the capacity of lithium adsorption of each cycle were calculated by the equation (1) and the equation (2), respectively.

  1.8   Characterization and chemical analysis

  The crystalline phases were characterized by XRD using a Rigaku Rint-2000 X-ray diffractometer equipped with monochromatized Cu Kα radiation (λ=0.154 056 nm), operating at 40 kV and 100 mA, and a scanning rate of 10°·min^-1 in a 2θ range of 10°~80°. The morphology of the particles was characterized by field emission scanning electron microscopy (MIRA3 TESCAN). The concentrations of manganese and lithium were determined using an atomic absorption spectrophotometer (TAS-999F, Beijing). The manganese valence was characterized after oxidation-reduction titration analyses by using the standard oxalic acid method^[16].

  DTA-TG curve was performed on a MAC science thermal analyzer (system 001 200 TG-DTA) at a heating rate of 5 ℃·min^-1 under an air atmosphere. Open circuit measurement was carried out using a Reference 600+ electrochemical workstation. The working electrode was Pt wire (0.5 mm diameter, 0.5 cm long). The counter electrode was a platinum foil (about 2 cm²). All potentials were measured and reported relative to a saturated Hg/HgO electrode.

  When the MnSO₄ and KMnO₄ solution were added into the LiOH solution, the change of open circuit potential with time was recorded.

  2.   Results and discussion

  2.1   Effect of hydrothermal temperature

  The XRD patterns of LiMnO₂ obtained at different hydrothermal temperatures are shown in Fig. 1. Table 2 shows the extraction percentage of Li⁺ and the percentage of dissolved manganese of the precursor as well as the adsorption capacity of Li⁺ of the corresponding lithium ion sieve.

  Figure 1

  

  Figure 1.  XRD patterns of precipitates obtained at different hydrothermal temperatures

  下载: 全尺寸图片 幻灯片

  Table 2

  

  Table 2.  Performance of precursor and ion-sieve synthesized at different hydrothermal temperatures

  下载: 导出CSV

  Temperature/℃	Dissolution percentage of manganese/%	Extraction percentage of Li+/%	Adsorption capacity of Li+/(mg·g-1)
  140	2.44	84.88	28.35
  160	2.62	86.32	30.29
  180	2.74	84.62	26.56
  200	2.93	85.03	26.46

  It is found from Fig. 1 that all the XRD patterns are indexed to orthogonal LiMnO₂ (PDF No.35-0749). A characteristic impurity peak near 18.7° was observed. A lot of scholars have different opinions about the peak source. Wang et al ascribed the peak to MnO₂ due to Mn(OH)₂ oxidation^[17]. Ji et al. and Liu et al. attributed it to Li₂MnO₃ owing to the oxidation of Mn³⁺ to Mn⁴⁺ by oxygen from the autoclave^[18-19]. Increasing hydrothermal temperature lead to an increase of peak intensity at 18.7°. Supposing that oxygen participates in the reaction to some extent during the pretreatment process and hydrothermal reaction. The impurity peak at 18.7° was attributed to Li₂MnO₃ and LiMn₂O₄. The exact origin of the peak was discussed in the following section.

  To check the effect of hydrothermal temperature on the composition of precursor Li_1.6Mn_1.6O₄, LiMnO₂ was roasted at 450 ℃ for 4 h. Table 2 shows the properties of precursor and adsorption capacity of its ion-sieve H_1.6Mn_1.6O₄. It is observed that the percentage of dissolved manganese increased with increasing hydrothermal temperature. Both the extraction percentage of Li⁺ and adsorption capacity increase slightly with the increasing temperature, reaching a maximum at 160 ℃, and then decreased. The increase in the percentage of dissolved manganese is due to an increase in LiMn₂O₄ content. The extraction percentage of Li⁺ is related to the crystal structure of LiMnO₂. When the hydrothermal temperature reached 140 ℃, the reaction rate became slow and LiMnO₂ growth was imperfect, resulting in a low extraction percentage of Li⁺ and adsorption capacity. When the temperature increased to 160 ℃, LiMnO₂ crystal develops comp-letely, which is beneficial to the extraction and adsorp-tion of Li⁺ ions. Further increasing the temperature leads to the formation of Li₂MnO₃, which in turn reduced Li⁺ extraction and adsorption performance^[20].

  2.2   Effect of molar ratio of MnO₄^-/Mn²⁺

  The formation of orthogonal LiMnO₂ can be expressed by equation (5). The theoretical molar ratio of MnO₄^- to Mn²⁺ was 1:4. However, the presence of either Li₂MnO₃ or LiMn₂O₄ as shown in Fig. 1 indicates that the oxidant was excessive in the hydrothermal reaction. It is assumed that the excessive oxidant originates from oxygen in the air and autoclave.

  To support the assumption, the pretreatment process was carried out in a glove box to completely isolate the oxygen and the molar ratio of MnO₄^-/Mn²⁺ was kept at 1:4. After the pretreatment was completed, the suspension was transferred into the autoclave and N₂ was blown for 30 min. The XRD patterns of the resulting hydrothermal products with and without oxygen isolation are shown in Fig. 2.

  Figure 2

  

  Figure 2.  XRD patterns of hydrothermal product obtained after complete oxygen isolation

  下载: 全尺寸图片 幻灯片

  $ {\rm{MnO}}_4^ - + 4{\rm{M}}{{\rm{n}}^{2 + }} + 12{\rm{O}}{{\rm{H}}^ - } + 5{\rm{L}}{{\rm{i}}^ + } \to 5{\rm{LiMn}}{{\rm{O}}_2} + 6{{\rm{H}}_2}{\rm{O}} $

  (5)

  It can be seen from Fig. 2 that the hydrothermal product is pure LiMnO₂ (PDF No.35-0749) after oxygen was completely isolated. The manganese content was 58.42%, nearly identical to the theoretical content (58.52%) of LiMnO₂, and manganese valence was 3.08. This result indicates that oxygen in the air dissolves into the suspension in the pretreatment process and participates in the oxidation reaction.

  In order to avoid the inconvenience of oxygen removal, an attempt to reduce the amount of oxidant KMnO₄ was made by changing the n_{MnO4^-:nMn²⁺} from 1:4 to 1:5, 1:5.5 even 1:6, the XRD patterns of the resulting precipitates are shown in Fig. 3.

  Figure 3

  

  Figure 3.  XRD patterns of hydrothermal products obtained under different molar ratios of nMnO4-Mn2+:n

  下载: 全尺寸图片 幻灯片

  As shown in Fig. 3, the impurity peak at 18.7° was significantly reduced and even disappeared at a the n_{MnO4^-:nMn²⁺} of 1:6. The results again confirm the fact that oxygen either from air or autoclave participates in the oxidation reaction.

  To further support the fact, the molar ratio of MnO₄^- to Mn²⁺ was kept at 1:6, and N₂ continuously blown into the autoclave for 30 min in order to remove the oxygen in the autoclave and suspension. The XRD patterns of the hydrothermal products are also shown in Fig. 3. Both LiMnO₂ and Mn₃O₄ diffraction peaks were observed. The appearance of Mn₃O₄ indicates that the oxidation reaction was incomplete in the absence of oxygen when the molar ratio of MnO₄^- to Mn²⁺ was 1:6.

  The valence of Mn (Z_Mn) and mass fraction of manganese (w_Mn) in the LiMnO₂ sample obtained at different molar ratios of MnO₄^- to Mn²⁺ were measured. The results are shown in Table 3.

  Table 3

  

  Table 3.  Mn valence in LoMnO₂ at different molar ratios of MnO₄^- to Mn²⁺

  下载: 导出CSV

  nMnO4-Mn2+:n	1:4	1:5	1:5.5	1:6	1:6 (O2 removal)
  wMn/%	55.69	56.62	57.61	58.54	58.70
  ZMn	3.32	3.23	3.15	3.02	3.00

  When the molar ratio of MnO₄^- to Mn²⁺ decreased from 1:4 to 1:6, the Mn valence decreased from 3.32 to 3.02 and the mass fraction of manganese increased from 55.69% to 58.54%. The valence of Mn higher than 3.0 means that the sample contains other compounds besides LiMnO₂. The theoretical mass fraction of manganese in Li₂MnO₃, LiMnO₂, LiMn₂O₄ and Mn₃O₄ is 47.03%, 58.52%, 60.77% and 72.03%, respectively. The sample obtained at a molar ratio of MnO₄^- to Mn²⁺ of 1:6 had a mass fraction of manganese close to LiMnO₂. When the oxygen in the autoclave was removed, a small amount of Mn₃O₄ was formed due to insufficient oxidation, resulting in a higher mass percentage of manganese.

  It is deduced from the mass fraction of manganese and valence in Table 3 that the sample is composed of LiMnO₂, Li₂MnO₃ and LiMn₂O₄. Assuming that the total mass of Mn was 1 mol, LiMnO₂, Li₂MnO₃ and LiMn₂O₄ account for x, y and (1-x-y) mol, respectively. The mass fraction of LiMnO₂ (w), Li₂MnO₃ (w) and LiMn₂O₄ (w) in the sample can be calculated by the following equation (6~10). The results are shown in Table 4.

  Table 4

  

  Table 4.  Hydrothermal product components obtained at different molar ratios of MnO₄^- to Mn²⁺

  下载: 导出CSV

  nMnO4-Mn2+:n	wLiMnO2/%	wLiMnO3/%	wLiMn2O4/%
  1:4	56.54	27.72	15.74
  1:5	67.62	19.13	13.25
  1:5.5	77.23	10.35	12.42
  1:6	98.16	0.16	1.68

  $ 3x + 4y + 3.5(1 - x - y) = {Z_{{\rm{Mn}}}} $

  (6)

  $ \frac{x}{{58.52\% }} + \frac{y}{{47.03\% }} + \frac{{1 - x - y}}{{60.77\% }} = \frac{1}{{{w_{{\rm{Mn}}}}}} $

  (7)

  $ {\omega _{{\rm{LiMn}}{{\rm{O}}_2}}} = \frac{{x{w_{{\rm{Mn}}}}{M_{{\rm{LiMn}}{{\rm{O}}_2}}}}}{{{M_{{\rm{Mn}}}}}} \times 100\% $

  (8)

  $ {\omega _{{\rm{L}}{{\rm{i}}_2}{\rm{Mn}}{{\rm{O}}_3}}} = \frac{{y{w_{{\rm{Mn}}}}{M_{{\rm{L}}{{\rm{i}}_2}{\rm{Mn}}{{\rm{O}}_3}}}}}{{{M_{{\rm{Mn}}}}}} \times 100\% $

  (9)

  $ {\omega _{{\rm{LiM}}{{\rm{n}}_2}{{\rm{O}}_4}}} = \left[ {1 - \left( {{w_{{\rm{LiMn}}{{\rm{O}}_2}}} + {w_{{\rm{L}}{{\rm{i}}_2}{\rm{Mn}}{{\rm{O}}_3}}}} \right)} \right] \times 100\% $

  (10)

  Where M, M and M_Mn represent the molecular weight of LiMnO₂, Li₂MnO₃ and the relative atomic weight of manganese, respectively.

  It can be seen from Table 4 that LiMnO₂ content increased from 56.54% to 98.16% while Li₂MnO₃ decreased from 27.72% to 0.16% when the molar ratio of MnO₄^- to Mn²⁺ changed from 1:4 to 1:6. When the MnO₄^-/Mn²⁺ ratio decreased to 1:6, the product was nearly pure LiMnO₂ with a very small amount of Li₂MnO₃ and LiMn₂O₄. It is deduced that pure LiMnO₂ can be obtained when the amount of oxidant was reduced to a certain extent.

  In order to find the effect of oxygen on the pretreatment process, the change of open circuit potential with pretreatment time was recorded and is shown in Fig. 4. Before the MnO₄^- and Mn²⁺ solutions were added into the LiOH solution, the open circuit potential maintained constant with time. When MnO₄^- and Mn²⁺ were added dropwise into the LiOH solution, redox reaction occurs and forms MnOOH through equation (11) and (12).

  Figure 4

  

  Figure 4.  Open circuit potential with time during pretreatment

  下载: 全尺寸图片 幻灯片

  $ 5{{\rm{H}}^ + } + {\rm{MnO}}_4^ - + 4{\rm{e}} \to {\rm{MnOOH}} + 2{{\rm{H}}_2}{\rm{O}} $

  (11)

  $ 120{{\rm{H}}^ - } + 4{\rm{M}}{{\rm{n}}^{2 + }} - 4{\rm{e}} \to 4{\rm{MnOOH}} + 4{{\rm{H}}_2}{\rm{O}} $

  (12)

  According to the Nernst equation, the open circuit potential depends on the concentration ratio of MnO₄^- to Mn²⁺ in the solution, increasing with increasing concentration of MnO₄^-, and decreasing with increasing concentration of Mn²⁺. Upon MnO₄^- and Mn²⁺ solutions were added into the LiOH solution, the MnSO₄ solution immediately formed Mn(OH)₂ precipitate instead of oxidation to form MnOOH, resulting in the increase of open circuit potential. After a period of time, MnO₄^- and Mn²⁺ ions form MnOOH and the concentration of MnO₄^- decreased, resulting in a sharp drop in the open circuit potential. When the addition rate was equal to the oxidation reaction rate, the concentrations of MnO₄^- and Mn²⁺ remained unchanged and the open circuit potential kept stable.

  When the molar ratio of MnO₄^- to Mn²⁺ changed from 1:4 to 1:6, the maximum open circuit potential decreases from -48.4 to -220 mV due to the decreasing of MnO₄^- concentration. However, the stable open circuit potentials were very close. The results again support the fact that oxygen from the air acts as an oxidant and participates in the oxidation. When the molar ratio of MnO₄^- to Mn²⁺ was high, such as 1:4 or 1:5, excessive oxidant resulted in the oxidation of MnOOH to Mn⁴⁺ and subsequent formation of Li₂MnO₃ and LiMn₂O₄. When the molar ratio of MnO₄^- to Mn²⁺ was relatively low, MnO₄^- combined with oxygen oxidizes all Mn²⁺ ions to produce pure MnOOH followed by conversion to LiMnO₂. As a result, all MnO₄^- and Mn²⁺ ions were transformed to precipitate and made no great difference to the open circuit potential.

  It is deduced that the LiMnO₂ formation reaction proceeds according to equation (13).

  $ \begin{array}{l} {\rm{MnO}}_4^ - + 6{\rm{M}}{{\rm{n}}^{2 + }} + 18{\rm{O}}{{\rm{H}}^ - } + 7{\rm{L}}{{\rm{i}}^ + } + 0.5{{\rm{O}}_2} \to \\ \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;7{\rm{LiMn}}{{\rm{O}}_2} + 9{{\rm{H}}_2}{\rm{O}} \end{array} $

  (13)

  It is calculated that the oxygen from the autoclave only accounts for 23.44% of the theoretical oxygen amount, the other 76.56% is from the dissolved oxygen during the pretreatment process.

  Fig. 5 shows the XRD patterns of Li_1.6Mn_1.6O₄ obtained under different molar ratios of n_{MnO4^-:nMn²⁺}. It is shown that all the XRD patterns are indexed to Li_1.6Mn_1.6O₄ (PDF No.52-1841). A small impurity peak near 20.96° was observed in the range of 1:4, 1:5 and 1:5.5. When the molar radio of MnO₄^- to Mn²⁺ was 1:6, the impurity peak disappeared. In addition, the peak intensity ratio of (400) to(311) plane gradually increased from 1.64 to 2.27 when the molar ratio of MnO₄^- to Mn²⁺ increased from 1:6 to 1:4. Increasing the lithium content in the spinel may lead to growth rate of (400)-plane in the lattice being faster than (311) plane. The structure tend to be regular octahedron^[21]. When molar ratio of MnO₄^- to Mn²⁺ was 1:6, the peak intensity ratio of (400) to(311) was 1.64, close to the theoretical ratio (1.61) denoted by the PDF No.52-1841. This result indicates that reducing KMnO₄ amount can obtain the desirable phase and structure.

  Figure 5

  

  Figure 5.  XRD patterns of Li_1.6Mn_1.6O₄ obtained under different molar ratios of n_{MnO4^-:nMn²⁺}

  下载: 全尺寸图片 幻灯片

  Fig. 6 shows the SEM images of Li_1.6Mn_1.6O₄, lithium-extracted and lithium-reinserted sample. The morphology did not change significantly, the sizes of most particles were in a range of 0.1~0.2 μm, which was small enough and beneficial for ionic diffusion. However, the aggregate phenomenon was reduced after pickling. The looser surface reduces the mass transfer resistance and completes the ion exchange process quickly.

  Figure 6

  

  Figure 6.  SEM images of Li_1.6Mn_1.6O₄ (a), lithium-extracted (b) and lithium-reinserted sample (c) obtained under molar radio of MnO₄^- to Mn²⁺ of 1:6

  下载: 全尺寸图片 幻灯片

  Table 5 shows the properties of Li_1.6Mn_1.6O₄ obtained under different molar ratios of n_{MnO4^-:nMn²⁺} and the adsorption capacity of the corresponding lithium ion-sieve.

  Table 5

  

  Table 5.  Performance of precursor and ion-sieve obtained under different molar ratios of MnO₄^-/Mn²⁺

  下载: 导出CSV

  nMnO4-Mn2+:n	wLiMnO2/%	Extraction percentage of Li+/%	Dissolution percentage of manganese /%	Adsorption capacity of Li+/(mg·g-1)
  1:4	55.85	86.32	2.62	30.29
  1:5	66.70	93.77	2.51	37.80
  1:5.5	77.23	95.48	2.42	39.43
  1:6	98.16	97.82	2.18	42.87

  As can be seen from Table 5, a lower molar ratio lead to a higher adsorption capacity due to smaller amounts of impurities as shown in Table 4. The percentage of dissolved manganese showed a small decrease due to the decrease of LiMn₂O₄ content. It is found that the adsorption capacity of Li⁺ and the extraction percentage of Li⁺ increased with an increase of mass fraction of LiMnO₂.

  2.3   Effect of roasting temperature

  DTA-TG (differential thermal-gravimetric analysis) curve was measured to find the optimum roasting temperature of o-LiMnO₂ obtained at a n_{MnO4^-:nMn²⁺} of 1:6. As shown in Fig. 7, the evaporation of adsorbed water caused nearly 1% weight loss below 200 ℃. When the temperature was in a range from 200 to 480 ℃, LiMnO₂ oxidation occurd through equation 14 and lead to 5.2% weight gain. A weight loss about 0.26% was observed for the temperature between 349 and 390 ℃, which may result from the transformation of Li₂MnO₃ and LiMn₂O₄ to LiMnO₂ accompanied by oxygen release according to reaction (15)^[22].

  Figure 7

  

  Figure 7.  DTA-TG curves of LiMnO2

  下载: 全尺寸图片 幻灯片

  $ 8{\rm{LiMn}}{{\rm{O}}_2} + 2{{\rm{O}}_2} \to 5{\rm{L}}{{\rm{i}}_{1.6}}{\rm{M}}{{\rm{n}}_{1.6}}{{\rm{O}}_4} $

  (14)

  $ {\rm{L}}{{\rm{i}}_2}{\rm{Mn}}{{\rm{O}}_3} + {\rm{LiM}}{{\rm{n}}_2}{{\rm{O}}_4} \to 3{\rm{LiMn}}{{\rm{O}}_2} + 0.5{{\rm{O}}_2} $

  (15)

  The continuous decrease in weight above 480 ℃ was due to the decomposition of Li_1.6Mn_1.6O₄ and the volatilization of Li^{+ [23]}.

  The XRD patterns of Li_1.6Mn_1.6O₄ obtained at different roasting temperatures are shown in Fig. 8.

  Figure 8

  

  Figure 8.  XRD patterns of samples at different roasting temperatures

  下载: 全尺寸图片 幻灯片

  It can be seen from Fig. 8 that pure Li_1.6Mn_1.6O₄ was obtained by roasting LiMnO₂ at 350~450 ℃ for 4 h. The higher the roasting temperature, the better the crystallinity was. When the temperature reaches 500 ℃, some peaks appear at 20.6° and 65.6° and are more likely to be characteristic peaks of Li₂MnO₃ and LiMn₂O₄^[13].

  Fig. 9 shows the extraction percentage of Li⁺ and the percentage of dissolved manganese of the precursor as well as the adsorption capacity of Li⁺ of the corresponding lithium ion sieve at different roasting temperatures. It is observed that the extraction percentage of Li⁺ gradually decreased with an increase of roasting temperature. The percentage of dissolved manganese showed a similar trend but appeared a minimum at 450 ℃. The response can be interpreted from two sides. On the one hand, the extraction rate of Li⁺ was ion exchange between Li⁺ and H⁺ ions. As the roasting temperature increased, the Li_1.6Mn_1.6O₄ crystallinity was well developed and lithium ions were more difficult to escape from the crystal lattice^[24]. On the other hand, Li₂MnO₃ content increased with increasing temperature, while Li₂MnO₃ did not have Li⁺ extraction property in dilute hydrochloric acid. The increase in temperature lead to a more complete oxidation reaction, which in turn decreased the dissolution percentage of manganese. When the temperature exceeded 450 ℃, the sample began to decompose and form Mn³⁺, resulting in an increase in the dissolution rate of manganese.

  Figure 9

  

  Figure 9.  Extraction percentage of Li⁺ and dissolution percentage of manganese (a) and corresponding adsorption capacity of Li⁺ (b) at different roasting temperatures

  下载: 全尺寸图片 幻灯片

  The adsorption capacity of Li⁺ increased with increasing temperature, reaching a maximum of 41.34 mg·g^-1 at 450 ℃. After that, it began to decrease. The crystal structure was not well developed at a low temperature. Correspondingly, the adsorption capacity of Li⁺ was low. When the temperature exceeded 450 ℃, as shown in Fig. 6, the decomposition occured accompanied by Li⁺ volatilization^[22], resulting in the decrease of adsorption capacity of Li⁺.

  2.4   Adsorption kinetics, selective adsorption and cycling performance

  Fig. 10 shows the relationship between the adsorption capacity and time. It is obvious that the adsorption capacity initially increased and then tended to reach equilibrium around 42.76 mg·g^-1 at 20 h.

  Figure 10

  

  Figure 10.  Adsorption capacity aftert different adsorption times

  下载: 全尺寸图片 幻灯片

  The pseudo-first-order kinetic model (equation (16)) and the pseudo-second-order kinetic model (equation (17)) are used to analyze the adsorption process and determine the rate constant of the adsorption process^[24].

  $ \lg \left( {{Q_{\rm{e}}} - {Q_t}} \right) = \lg {Q_{\rm{e}}} - \frac{{t{k_1}}}{{2.303}} $

  (16)

  $ \frac{t}{{{Q_t}}} = \frac{1}{{{k_2}Q_{\rm{e}}^2}} + \frac{t}{{{Q_{\rm{e}}}}} $

  (17)

  where Q_t is adsorption capacity of Li⁺ on ion sieve at adsorption time t, mg·g^-1; Q_e is equilibrium adsorption capacity of Li⁺ on ion sieve, mg·g^-1; k₁ is adsorption rate constant of pseudo-first-order kinetic model, h^-1; k₂ is adsorption rate constant of pseudo-second-order kinetic model, g·mg^-1·h^-1; t is adsorption time, h.

  The pseudo-first-order kinetic model can be transformed into equation (18) and then fitted expo-nentially. The kinetic fitting results are shown in Fig. 11. Table 6 lists the fitted parameters.

  Figure 11

  

  Figure 11.  Pseudo-first-order kinetic (a) and pseudo-second-order kinetic curves (b) fitting of ion sieve adsorption

  下载: 全尺寸图片 幻灯片

  Table 6

  

  Table 6.  Fitting parameters of two kinetic equations

  下载: 导出CSV

  	k1/h	k2/(g·mg-1·h-1)	R2	Qe/(mg·g-1)
  Pseudo-first-order	1.01	—	0.909	41.03
  Pseudo-second-order	—	0.035	0.999	43.60

  The correlation coefficient (R²) of the pseudo-second-order kinetic model for the adsorption process was 0.999, while that of pseudo-first-order kinetic model was 0.909. The result suggests that the adsorption process conformed to the pseudo-second-order kinetic model and was controlled by chemical adsorption process^[25-27].

  The equilibrium adsorption capacity (Q_e) obtained by fitting pseudo-second-order kinetic model was 43.60 mg·g^-1, which was consistent with the maximum adsorption capacity in Fig. 10.

  Table 7 shows the concentration changes of the main metal ions in the salt lake brine before and after adsorption. Only K⁺ ions are adsorbed slightly in salt lake brine, and α_K^Li was 33.38, much larger than that for Li⁺. Indicating that the lithium ion-sieve towards Li⁺ ions had good separation and selectivity performance.

  Table 7

  

  Table 7.  Selective adsorption of the ion sieve in salt lake brine

  下载: 导出CSV

  Metal ion	c0/(mg·L-1)	ct/ (mg:Lmg·L-1)	Kd	αMeLi
  Li+	290	161	0.267	1.0
  Na+	3 020	3 016	0	-
  K+	1 690	1 647	0.008	33.38
  Mg2+	109 820	109 815	0	-

  The cycle performance of lithium ion-sieve is shown in Fig. 12. The initial adsorption capacity of Li⁺ was 42.87 mg·g^-1 and reduced to 37.21 mg·g^-1 after five cycles. The dissolution percentage of manganese decreased from 2.33% to 0.34% and remained stable after 3 cycles.

  Figure 12

  

  Figure 12.  Cyle performance of lithium ion-sieve in salt lake brine

  下载: 全尺寸图片 幻灯片

  The manganese valence increased from 3.94 to 3.97 after one-cycle acid pickling, indicating that a small amount of Mn³⁺ was present in the precursor and disproportionation occured to form Mn⁴⁺ and Mn²⁺. Mn²⁺ entering into the acid solution lead to manganese dissolution. After 2 cycles, the percentage of dissolved manganese tended to be stable, indicating that Mn³⁺ ions had nearly depleted and subsequent manganese dissolution was due to the reduction of MnO₂·0.5H₂O by chloride ions according to equation (19).

  $ {Q_t} = {Q_{\rm{e}}}\left( {1 - {{\rm{e}}^{ - {k_1}t}}} \right) $

  (18)

  The relatively high Li⁺ ion adsorption capacity and selectivity prove that the synthesized lithium ion-sieve is an excellent candidate for Li⁺ recovery from salt lake brines.

  3.   Conclusions

  Li_1.6Mn_1.6O₄ was synthesized through a hydro-thermal reaction to form LiMnO₂ followed by roasting. Acid pickling transforms Li_1.6Mn_1.6O₄ to lithium ion-sieve. The results show that hydrothermal temperature, molar ratio of MnO₄^-/Mn²⁺ have a great influence on LiMnO₂ purity and adsorption performance of the lithium ion sieve. The LiMnO₂ contained a small amount of Li₂MnO₃ and LiMn₂O₄. With the presence of oxygen in pretreatment process or hydrothermal reaction, pure orthogonal LiMnO₂ was obtained at a molar ratio of MnO₄^- to Mn²⁺ of 1:6, deviated from the theoretical ratio of 1:4. Open circuit potential measure-ment confirms that oxygen participates in the oxidation reaction to form LiMnO₂.

  Pure Li_1.6Mn_1.6O₄ can be obtained by roasting pure LiMnO₂ at 450 ℃ for 4 h. The percentage of dissolved manganese was only 2.18% and the extraction percentage of Li⁺ was 97.82%. The adsorption capacity of corresponding lithium ion-sieve in Lop Nor salt lake brine was 42.87 mg·g^-1. Distribution coefficients of metal ions are in the order: Li⁺≫K⁺ > Na⁺ > Mg²⁺. After 5 cycles, the Li⁺ ion adsorption capacity was 37.21 mg·g^-1 and the percentage of dissolved manganese decreased to 0.34%. The adsorption process for the ion-sieve in salt lake brine fits well with a quasi-second-order kinetic model and is controlled by chemical adsorption.

  
  
• 1. [1]

     封国富, 张骁.稀有金属, 2003, 27(1):57-61FENG Guo-Fu, ZHANG Xiao. Chinese Journal of Rare Metals, 2003, 27(1):57-61 doi: 10.3969/j.issn.0258-7076.2003.01.011
     

  2. [2]

     Ebensperger A, Maxwell P, Moscoso C. Resour. Policy, 2005, 30(3):218-231 doi: 10.1016/j.resourpol.2005.09.001
     

  3. [3]

     Scrosati B, Garche J. J. Power Sources, 2010, 195(9):2419-2430 doi: 10.1016/j.jpowsour.2009.11.048
     

  4. [4]

     Banov B, Momchilov A, Trifonova A, et al. J. Power Sources, 1999, 81/82:562-565 doi: 10.1016/S0378-7753(99)00218-9
     

  5. [5]

     Xu H, Sun J, Gao L. Ionics, 2013, 19(1):63-69 doi: 10.1007/s11581-012-0704-1
     

  6. [6]

     Li K Y, Shua F F, Zhang J W, et al. Ceram. Int., 2015, 41(5):6729-6733 doi: 10.1016/j.ceramint.2015.01.116
     

  7. [7]

     Li X H, Su Z, Wang Y B. J. Alloys Compd., 2018, 735:2182-2189 doi: 10.1016/j.jallcom.2017.11.384
     

  8. [8]

     Zhang Q H, Li S P, Sun S Y, et al. Chem. Eng. Sci., 2010, 65(1):169-173 doi: 10.1016/j.ces.2009.06.045
     

  9. [9]

     Xie J L, Huang X, Zhu Z B, et al. Ceram. Int., 2011, 37(1):419-421 doi: 10.1016/j.ceramint.2010.08.045
     

  10. [10]

      Zhao H Y, Chen B, Cheng C, et al. Ceram. Int., 2015, 41(10):15266-15271 doi: 10.1016/j.ceramint.2015.07.213
      

  11. [11]

      Chitrakar R, Kanoh H, Miyai Y, et al. Chem. Mater., 2000, 12(10):3151-3157 doi: 10.1021/cm0000191
      

  12. [12]

      Chitrakar R, Kanoh H, Miyai Y, et al. Ind. Eng. Chem. Res., 2001, 40(9):2054-2058 doi: 10.1021/ie000911h
      

  13. [13]

      Xiao J L, Sun S Y, Wang J, et al. Ind. Eng. Chem. Res., 2013, 52(34):11967-11973 doi: 10.1021/ie400691d
      

  14. [14]

      Shi X C, Zhou D F, Zhang Z B, et al. Hydrometallurgy, 2011, 110(1/2/3/4):99-106 doi: 10.1016/j.hydromet.2011.09.004
      

  15. [15]

      Sun S Y, Xiao J L, Wang J, et al. Ind. Eng. Chem. Res., 2014, 53(40):15517-15521 doi: 10.1021/ie5004625
      

  16. [16]

      Gao A L, Sun Z H, Li S P, et al. Dalton Trans., 2018, 47(11):3864-3871 doi: 10.1039/C8DT00033F
      

  17. [17]

      王禄, 马伟, 韩梅, 等.化学学报, 2007, 65(12):1135-1139WANG Lu, MA Wei, HAN Mei, et al. Acta Chim. Sin., 2007, 65(12):1135-1139 doi: 10.3321/j.issn:0567-7351.2007.12.007
      

  18. [18]

      Ji H M, Yang G, Miao X W, et al. Electrochim. Acta, 2010, 55(9):3392-3397 doi: 10.1016/j.electacta.2010.01.010
      

  19. [19]

      Liu Q, Li Y X, Hu Z L, et al. Electrochim. Acta, 2008, 53(24):7298-7302 doi: 10.1016/j.electacta.2008.04.008
      

  20. [20]

      Wang L, Ma W, Liu R, et al. Solid State Ionics, 2006, 177(17/18):1421-1428 doi: 10.1016/j.ssi.2006.07.019
      

  21. [21]

      Chitrakar R, Kanoh H, Kim Y S, et al. J. Solid State Chem., 2001, 160(1):69-76 doi: 10.1006/jssc.2000.9193
      

  22. [22]

      Chitrakar R, Sakane K, Umeno A, et al. J. Solid State Chem., 2002, 169(1):66-74 doi: 10.1016/S0022-4596(02)00018-X
      

  23. [23]

      Seo H, Na S, Lee B, et al. J. Ind. Eng. Chem., 2018, 64:311-317 doi: 10.1016/j.jiec.2018.03.029
      

  24. [24]

      马立文, 陈白珍, 石西昌, 等.无机化学学报, 2011, 27(4):697-703MA Li-Wen, CHEN Bai-Zhen, SHI Xi-Chang, et al. Chinese J. Inorg. Chem., 2011, 27(4):697-703 
      

  25. [25]

      Langmuir I. J. Am. Chem. Soc., 1918, 40(9):1361-1403 doi: 10.1021/ja02242a004
      

  26. [26]

      Wang L, Meng C G, Han M, et al. J. Colloid Interface Sci., 2008, 325(1):31-40 doi: 10.1016/j.jcis.2008.05.005
      

  27. [27]

      Naiya T K, Bhattacharya A K, Das S K. J. Colloid Interface Sci., 2008, 325(1):48-56 doi: 10.1016/j.jcis.2008.06.003
      

• 

  Figure 1  XRD patterns of precipitates obtained at different hydrothermal temperatures

  下载: 全尺寸图片 幻灯片
  

  Figure 2  XRD patterns of hydrothermal product obtained after complete oxygen isolation

  下载: 全尺寸图片 幻灯片
  

  Figure 3  XRD patterns of hydrothermal products obtained under different molar ratios of nMnO4-Mn2+:n

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Open circuit potential with time during pretreatment

  下载: 全尺寸图片 幻灯片
  

  Figure 5  XRD patterns of Li_1.6Mn_1.6O₄ obtained under different molar ratios of n_{MnO4^-:nMn²⁺}

  下载: 全尺寸图片 幻灯片
  

  Figure 6  SEM images of Li_1.6Mn_1.6O₄ (a), lithium-extracted (b) and lithium-reinserted sample (c) obtained under molar radio of MnO₄^- to Mn²⁺ of 1:6

  下载: 全尺寸图片 幻灯片
  

  Figure 7  DTA-TG curves of LiMnO2

  下载: 全尺寸图片 幻灯片
  

  Figure 8  XRD patterns of samples at different roasting temperatures

  下载: 全尺寸图片 幻灯片
  

  Figure 9  Extraction percentage of Li⁺ and dissolution percentage of manganese (a) and corresponding adsorption capacity of Li⁺ (b) at different roasting temperatures

  下载: 全尺寸图片 幻灯片
  

  Figure 10  Adsorption capacity aftert different adsorption times

  下载: 全尺寸图片 幻灯片
  

  Figure 11  Pseudo-first-order kinetic (a) and pseudo-second-order kinetic curves (b) fitting of ion sieve adsorption

  下载: 全尺寸图片 幻灯片
  

  Figure 12  Cyle performance of lithium ion-sieve in salt lake brine

  下载: 全尺寸图片 幻灯片

  Table 1.  Composition of salt lake brine  mg·L^-1

  Mg	Na	K	Li	Ca	S	B	Si
  106 100	5 020	2 480	220.2	27.24	6 831	519.0	28.05
  P	Mo	Fe	As	Co	Ga	Cu	V
  5.26	4.50	0.87	0.63	0.53	0.45	0.42	0.32
  Mn	Sn	Cr	Al	Ni	Pb	Y	Cd
  0.28	0.23	0.20	0.13	0.12	0.04	0.04	0.03

  下载: 导出CSV

  Table 2.  Performance of precursor and ion-sieve synthesized at different hydrothermal temperatures

  Temperature/℃	Dissolution percentage of manganese/%	Extraction percentage of Li+/%	Adsorption capacity of Li+/(mg·g-1)
  140	2.44	84.88	28.35
  160	2.62	86.32	30.29
  180	2.74	84.62	26.56
  200	2.93	85.03	26.46

  下载: 导出CSV

  Table 3.  Mn valence in LoMnO₂ at different molar ratios of MnO₄^- to Mn²⁺

  nMnO4-Mn2+:n	1:4	1:5	1:5.5	1:6	1:6 (O2 removal)
  wMn/%	55.69	56.62	57.61	58.54	58.70
  ZMn	3.32	3.23	3.15	3.02	3.00

  下载: 导出CSV

  Table 4.  Hydrothermal product components obtained at different molar ratios of MnO₄^- to Mn²⁺

  nMnO4-Mn2+:n	wLiMnO2/%	wLiMnO3/%	wLiMn2O4/%
  1:4	56.54	27.72	15.74
  1:5	67.62	19.13	13.25
  1:5.5	77.23	10.35	12.42
  1:6	98.16	0.16	1.68

  下载: 导出CSV

  Table 5.  Performance of precursor and ion-sieve obtained under different molar ratios of MnO₄^-/Mn²⁺

  nMnO4-Mn2+:n	wLiMnO2/%	Extraction percentage of Li+/%	Dissolution percentage of manganese /%	Adsorption capacity of Li+/(mg·g-1)
  1:4	55.85	86.32	2.62	30.29
  1:5	66.70	93.77	2.51	37.80
  1:5.5	77.23	95.48	2.42	39.43
  1:6	98.16	97.82	2.18	42.87

  下载: 导出CSV

  Table 6.  Fitting parameters of two kinetic equations

  	k1/h	k2/(g·mg-1·h-1)	R2	Qe/(mg·g-1)
  Pseudo-first-order	1.01	—	0.909	41.03
  Pseudo-second-order	—	0.035	0.999	43.60

  下载: 导出CSV

  Table 7.  Selective adsorption of the ion sieve in salt lake brine

  Metal ion	c0/(mg·L-1)	ct/ (mg:Lmg·L-1)	Kd	αMeLi
  Li+	290	161	0.267	1.0
  Na+	3 020	3 016	0	-
  K+	1 690	1 647	0.008	33.38
  Mg2+	109 820	109 815	0	-

  下载: 导出CSV
•