The bauxite used in this study was obtained from an alumina plant in Shanxi Province, China. The chemical composition and loss on ignition (LOI) of the diasporic bauxite are listed in Supplementary Table S1. Before use, it was washed, dried, and smashed to pass through 100 mesh sieves. Deionized (DI) water was used throughout all the experiments. The chemicals used in the experiments were analytical grade and obtained from Sinopharm Chemicals Reagent Co., Ltd. (Shanghai, China).

Lithium/aluminum layered double hydroxides-SiO₂ bauxite (LDH-Si-BX) was prepared as the following procedure illustrated in Figure 1.

1) 50 g natural bauxite was crushed so that we could collect 80–120 um particles. The particles were mixed with 10 g NaOH, placed in a nickel crucible at 500°C with a heating rate of 5 ^°C/min for 2h, and then cooled to room temperature.

The batch test for Li⁺ adsorption was performed in 150 mL conical flasks shaken in a temperature-controlled oscillator. All the Li⁺ solution was collected by filtering through a syringe equipped with a 0.45 μm PES filter head. In the case of a lack of special instructions, the adsorption conditions were controlled at an absorbent dosage of 0.1 g, Li⁺ solution volume of 25.00 mL, room temperature, pH = 7.0, and oscillation rate of 150 rpm. In addition, the pH was adjusted by using 1 mol/L HCl or 1 mol/L NaOH.

To evaluate the kinetics of Li⁺ adsorption onto the absorbents, the absorbents were added to the tapered bottles equipped with 20 mg/L Li⁺ solution. The water samples were collected at 20, 40, 60, 90, 120, 150, 180, and 240 min, and the Li⁺ concentration was analyzed.

The Li⁺ adsorption isotherm was investigated by adding the adsorbents to different concentrations (5, 10, 25, 50, 75, and 100 mg/L) of Li⁺ solution. After adsorption for 24 h, the Li⁺ concentration was tested and calculated.

To further investigate the lithium-ion adsorption performance and selectivity under coexisting ion conditions, the adsorption process was conducted in the solutions of Na⁺, K⁺, Ca^2+, and Mg²⁺ with an initial concentration of 250 mg/L, which is five times that of Li⁺. The absorbents were then additionally oscillated into the above solution and adsorbed to achieve equilibrium, and the Li⁺ content of the remaining solution was detected.

To examine Li⁺ adsorption properties under different pH levels, the adsorbents were mixed with 20 mg/L Li⁺-containing solutions of different pH values ranging from 3.0 to 11.0. After oscillating for 24 h, Li⁺ content and pH value were determined for the remaining solution.

To further study the adsorption-desorption performance of Li⁺ in real brines, 10.0 g LDH-Si-BX was added to 250 mL brines and adsorbed for 6 h to detect the content of lithium ions in brine and calculate the adsorption capacity. During the desorption process, DI water was selected as the desorption solution. For desorption, the adsorbent was loaded into a single glass column with a porous ceramic plug at the bottom and continuously rinsed with DI water until the concentration of lithium in the effluent solution was stabilized.

Surface morphologies of the samples were obtained by field emission scanning electron microscope (FESEM, SUPRA 40, ZEISS, Germany). The physical structure including specific surface areas (S_BET) and pore structure was assessed by using a BET surface area analyzer (TriStar II Plus 2.02, Micro, United States). Zeta Potential Analyzer (Zeta, ZS-90, Malvern, UK) was used to measure the surface charge of absorbents. The Li⁺ concentration was tested by an inductively coupled plasma optical emission spectrometer (ICP-OES, Optima 8000, PerkinElmer, United States). The chemical environment and shift of the absorbents were determined by X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific Instrument Co., United States).

The investigations of the surface morphology of bauxite (BX) and LDH-Si-BX after Li⁺ adsorption were performed by scanning electron microscope (SEM) and transmission electron microscope (TEM). It was found that the BX (Figure 2A) exhibited a typical flat lamellar surface and presented a blocky structure accumulated by a number of lamellas. However, the surface morphology of LDH-Si-BX (Figure 2B) dramatically differed from BX; the main differences were manifested in that the lamellar structure at the edges of the particles was covered and presented some flattened particles and scattered nanosheets, due to the loading of SiO₂ and LDH.

The pore structures of the adsorbents before and after the modification were identified by BET surface area analysis. Figure 3 shows the N₂ adsorption-desorption isotherm profiles for BX(a) and LDH-Si-BX(b), and the inset shows the pore size distribution, specific surface area, average pore size, and total pore capacity as shown in Table 1. As can be seen from Figure 3, for both BX and LDH-Si-BX, the adsorption-desorption isotherms belong to type IV adsorption isotherms in the IUPAC classification, and hysteresis occurs when the relative pressure P/P ₀ > 0.4 for both, indicating the existence of certain mesopores in both BX and LDH-Si-BX. As shown in Table 1, the specific surface area, average pore size, and total pore volume were 13.2 m²/g, 16.6 nm, and 0.05 cm³/g for BX and 152.6 m²/g, 3.1 nm, and 0.11 cm³/g for LDH-Si-BX, respectively. After modification, the average pore size of the adsorbent decreased, while the specific surface area and total pore volume increased significantly. The reason for this change might be the introduction of new substances in the structure of BX after loading SiO₂ and LDH, which caused the structure to become rougher.

The XRD results of the BX and LDH-Si-BX are shown in Figure 4. BX was mainly composed of diaspore (PDF#05–0355) and kaolin (PDF#29–1,488), and the diffraction peaks of the images were narrow and sharp, indicating that BX had high order and good crystallinity. It was noteworthy that the XRD spectra of LDH-Si-BX obtained after the activation modification showed the disappearance of the hydrargyrite diffraction peaks, the decrease of kaolinite intensity, and the appearance of diffraction peaks of LiCl·2Al(OH)₃·nH₂O(PDF#31–0700) and zeolite (PDF#16–0612), which was attributed to the destruction of the crystalline state of kaolin and the conversion of diaspore into new substances under alkali thermal calcination conditions. In the combined effect of acid impregnation, AlCl₃, LiCl, and NaOH, the BX structure was rearranged and partially transformed into a zeolite structure and loaded with LiCl·2Al(OH)₃·nH₂O. It was demonstrated that that LiCl·2Al(OH)₃·nH₂O had the ability to selectively adsorb Li⁺, therefore, the appearance of LiCl·2Al(OH)₃·nH₂O implied that LDH -Si-BX had the potential to adsorb Li⁺.

To study the effect of adsorption time, a kinetic investigation was carried out. As illustrated in Figure 5, the Li⁺ adsorption capacity of LDH-Si-BX increased dramatically and the adsorption rate was fast during the first 125 min. Subsequently, the lithium adsorption capacity of LDH-Si-BX reached more than 80% of its equilibrium adsorption capacity. However, in the subsequent adsorption process, the adsorption rate began to slow down and reached equilibrium. The rapid adsorption phase might be due to the instant contact between the plentiful adsorption sites and adsorbate. When the active sites were gradually occupied, the rate-limiting step of the adsorption process was mainly the diffusion of Li⁺ into the absorbents, which resulted in the plateau. Besides, it is noteworthy that the higher initial concentration exhibited a higher uptake of Li⁺ capacity (1.54 mg/g at 50 mg/L) than that at a lower concentration condition (1.32 mg/g at 25 mg/L and 0.96 mg/g at 10 mg/L), respectively.

To make a further assessment of the adsorption process, the adsorption data derived from the experiment were modeled by using the pseudo-first-order and pseudo-second-order models (Chen et al., 2020; Kamran and Park, 2022). The mathematical expressions are individually shown in Equation 1 and Equation (2): q t = q e 1 − exp − k 1 t (1) q t = k 2 q e 2 t 1 + k 2 q e t (2) where t is the adsorption time (min), q _t is the Li⁺ adsorption capacity at the time of t (mg/g), q _e is the equalized adsorption capacity (mg/g), and k ₁ (1/min) and k ₂ (g/(mgmin)) are considered to be the adsorption rate constant of the pseudo-first-order model and pseudo-second-order model, respectively.

From Figure 5 and Table 2, it was observed that the pseudo-second-order kinetic fitting factor values of R ² were 0.966, 0.972, and 0.974, and higher than those calculated from the pseudo-first-order kinetic model (0.892, 0.884, and 0.943). Furthermore, the calculated saturated adsorption capacity of 10 mg/L, 25 mg/L, and 50 mg/L from the pseudo-second-order kinetic model (1.16, 1.52, and 1.76 mg/g) were closer to the experimental ones of 0.96 mg/g, 1.32 mg/g, and 1.54 mg/g. Therefore, chemisorption might be presumed to be the rate-limiting step (Cheng et al., 2021).

To further describe the Li⁺ adsorption capacity of LDH-Si-BX under various concentrations, the data gained from adsorption experimental procedures were matched by the Langmuir and Freundlich models (Wang et al., 2016; Zhang et al., 2017). The non-linear forms are illustrated by the Equations (3) and (4), respectively. q e = q max K L C e 1 + K L C e (3) q e = K F C e 1 / n (4) where C _e and q _e are the Li⁺ concentration of the solution (mg/L) and the corresponding adsorption capacity when achieving equilibrium (mg/g), respectively, q _max (mg/g) is the saturated adsorption capacity calculated by the Langmuir adsorption model, K _L (L/mg) is the constant of the Langmuir isotherm model, K _F (mg/g) is the Freundlich constant, and n is the Freundlich coefficient.

To determine whether the adsorption process is facilitated or not, the infinitesimal isolation factor R _L was expressed as follows (Li et al., 2014). R L = 1 1 + K L C 0 (5) where C ₀ is the highest initial Li⁺ concentration (mg/L) and K _L is the constant of the Langmuir isotherm model (L/mg).

The fitting results and the Li⁺ adsorption at different concentrations are given in Table 3 and Figure 6. It was shown that with an increase of initial Li⁺ concentration, the equilibrium concentration gradually increased, while the equilibrium adsorption capacity of Li⁺ presented a variation trend of fast increase, then a slight increase, and finally tended to remain constant. It might be because the limited active sites of the adsorbent were occupied by Li⁺. As can be seen from Table 3, the Langmuir model had a higher fitting factor (R ² = 0.991 at 298K, R ² = 0.962 at 308K, and R ² = 0.926 at 318K), which gave a better description of the adsorption process than that of the Freundlich model. This demonstrated a process similar to monolayer adsorption that occurred during the adsorption of Li⁺ by LDH-Si-BX (Huang et al., 2015). In addition, the infinitesimal isolation factor R _L was in the range of 0–1 for both LDH-Si-BX, indicating that these adsorption processes were favorable. Moreover, the calculated saturated Li⁺ adsorption capacities from the Langmuir model by LDH-Si-BX were 1.70 mg/g, 1.88 mg/g, and 2.06 mg/g at 298 K, 308 K, and 318 K, respectively. As can be seen, LDH-Si-BX had a certain adsorption capacity for Li⁺.

In order to identify the feasibility and spontaneity of the Li⁺ adsorption by LDH-Si-BX, different thermodynamic parameters including enthalpy (ΔH ^o ), entropy (ΔS ^o ), and Gibbs free energy (ΔG ^o ) were obtained by linear fit with lnK as a function of 1/T (Figure 7B). The equations are expressed as follows (Chen et al., 2020): K = 1000 M L i L i + Y K L (6) Δ G 0 = − R T ⁡ ln ⁡ K (7) ln ⁡ K = − Δ H 0 R T + Δ S 0 R (8) where K is the dimensionless thermodynamic equilibrium constant, M _Li is the molecular weight of Li (6.941 g/mol), [Li⁺] is the standard concentration of adsorbate and is equal to 1 mol/L, and Y is the activity coefficient (dimensionless). The Li⁺ concentration of experimental brine in this paper is less than 100 mg/L, thus it can be considered that the adsorbate Li⁺ is very diluted and the value of the activity coefficient is assumed to be unitary. R and T denote the distribution, ideal gas constant (8.314 J mol^-1·K⁻¹), and temperature (K), respectively.

Thermodynamic parameters are summarized in Supplementary Table S2. It is observed from the table that the values of ΔG are negative at 298 K, 308 K, and 318 k: 17.826, −18.208, and −18.399 kJ/mol, respectively, In addition, the positive values of enthalpy change ΔH ^o (9.602 kJ/mol) and entropy change ΔS ^o (92.102 J/(molK)) indicated that the adsorption of Li⁺ by LDH-Si-BX was an endothermic and increasing disorder degree process (Chen et al., 2020).

Various cations are always present in the actual brines, which may affect the ability of the adsorbent to remove Li+. In the present study, we independently investigated the effects of Na⁺, Mg²⁺, K^+, and Ca²⁺on Li + adsorption by LDH-Si-BX. As shown in Figure 7, the presence of Na⁺, K⁺, Ca^2+, and Mg²⁺ showed a certain degree of decrease in the adsorption of Li ions, i.e., when the concentration of coexisting ions was five times that of Li⁺, the adsorption capacity of Li⁺, Na⁺, K⁺, Ca²⁺, and Mg²⁺, was 1.53 mg/g, 1.50 mg/g, 1.48 mg/g, 1.49 mg/g, and 1.41 mg/g, respectively, which showed that Mg²⁺ had a more obvious effect on Li⁺ in the adsorption process.

In the co-existing cation effect experiments, the adsorption process was conducted in the solutions with different initial concentrations of Li ⁺ /Me (Me = Na⁺, K⁺, Ca²⁺, and Mg²⁺). The distribution coefficient (K _d ) and separation factor (α _Li / _Me ) were determined by the following equations: K d = Q e / C e (9) α M e L i = K d , L i /   K d , M e (10) where Q _e (mg/g) is the ion equilibrium adsorption capacity of the adsorbent and C _e (mg/L) is the residual cation concentration in the solution.

The relevant parameters are listed in Table 4. The results indicate that the ion-selective sequence of LDH-Si-BX followed the order of Na⁺ > K⁺> Ca²⁺> Mg²⁺ with the α _Li / _Me in a range of 27.98–54.34. It can be seen that LDH-Si-BX had selective adsorption of Li⁺, therefore, the use of LDH-Si-BX for selective absorption of Li⁺ in actual brines possesses some potential applicability.

To investigate the pH sensitivity of LDH-Si-BX in the adsorption of Li⁺, a series of 50 mg/L Li⁺ solutions, which were adjusted to pH levels of 4.0–10.0, were used to test adsorption, as shown in Figure 8A. It was clear to see that the Li⁺ adsorbed by LDH-Si-BX had significant pH sensitivity. The higher adsorption capacity of 1.55 mg/g and 1.50 mg/g were observed at pH = 8 and pH = 7, which was close to the point of zero charges (PZC) of 7.6. When the pH value decreased from 7 to 4, the adsorption amount showed a significantly decreasing trend and dropped to 1.05 mg/g. Moreover, it could be noted that the adsorption decreased to 1.15 mg/g when the pH was increased from 8 to 10. This pH-dependent behavior indicated that the neutral medium was favorable for Li⁺ adsorption by LDH-Si-BX.

The zeta potentials of LDH-Si-BX are shown in Figure 8B. It was found that the isoelectric point (IEP) of LDH-Si-BX was determined as 7.6. The surface of LDH-Si-BX was positively charged by being protonated when the pH < pH _IEP, which was not favorable for lithium-ion adsorption due to charge repulsion. When the pH > pH_IEP, the surface of the adsorbents showed a negative charge, however, the adsorption performance of LDH-Si-BX for Li⁺ still presented a decreasing tendency, which might be attributed to the LiCl·2Al(OH)₃·nH₂O of LDH-Si-BX being converted to the Al(OH)₃ (Zhong et al., 2020). It could be recognized that a neutral environment was conducive to the adsorption of Li⁺.

To further investigate the adsorption performance of LDH-Si-BX on lithium ions in real brines, the study applied LDH-Si-BX to the adsorption of oilfield water. The adsorption kinetics are shown in Supplementary Figure S1 of the Supplementary Material, and the results show that the adsorption reached equilibrium within 180–200 min and was in accordance with the proposed second-order kinetic model (Supplementary Table S3 of the Supplementary Material). The data from the adsorbent/brines ratio study (Figure 9A) showed that the adsorption of lithium ions in oilfield brines by LDH-Si-BX reached 83.3% when the solid-liquid ratio was 1.0 g/mL. In addition, by comparing the adsorption rates for Li⁺, Na⁺, K⁺, Mg²⁺, and Ca²⁺ of LDH-Si-BX, it could be found that the adsorption capacity of LDH-Si-BX for Li⁺ was significantly stronger than that of the other four ions, indicating that LDH-Si-BX also possesses Li⁺ selectivity in the actual brines.

Figure 9B illustrates the changes in the adsorption amount of Li⁺ by LDH-Si-BX after 10 cycles of adsorption-desorption, and the results showed that a decrease in adsorption capacity occurred during the regeneration of the adsorbent, which could be attributed to the adsorption sites being occupied by Mg²⁺ in the brine (Zhong et al., 2020; Zhong et al., 2021). The adsorption amount of Li⁺ by LDH-Si-BX decreased by 16.8% after 10 regenerations, and it was observed that LDH-Si-BX possessed a certain regeneration capability.

To further understand the Li⁺ adsorption mechanism by LDH-Si-BX, x-ray photoelectron spectroscopy (XPS) was used to characterize and analyze the compositional and structural changes of the adsorbents before and after Li⁺ adsorption (LDH-Si-BX-Li).

As seen in Figure 10B, the three peaks of O^2-, -OH, and H₂O (Sleiman et al., 2016) could be combined to fit the high-resolution spectra of O1s for LDH-Si-BX and LDH-Si-BX-Li, and the fitting parameters results (Table 5) demonstrate that the binding energy of -OH, and H₂O did not change significantly before and after adsorption, while the binding energy of O^2- increased from 530.86 eV to 530.94 eV after adsorption. In addition, as can be seen in Figure 10C, the binding energy of Al decreased from 74.44eV to 74.41eV. This phenomenon indicated that the influence of -OH and H₂O in the LDH-Si-BX structure was not significant in the process of Li⁺ adsorption, while the O in the oxide structure played a vital role. Since the LiCl•2Al(OH)₃•nH₂O of LDH-Si-BX was the main substance for Li⁺ adsorption, it could be inferred that the O in the structure of Al-O played an important part in binding with Li⁺.

Moreover, compared to the Li 1s high-resolution scan spectra (Figure 10D) of LDH-Si-BX and LDH-Si-BX-Li, the peak intensity of LDH-Si-BX-Li was significantly greater than that of LDH-Si-BX, which improved the adsorption of Li⁺ by LDH-Si-BX.