Hydrothermal synthesis and adsorption behavior of H₄Ti₅O₁₂ nanorods along [100] as lithium ion-sieves

Abstract

The adsorption method is a promising route to recover Li⁺ from waste lithium batteries and lithium-containing brines. To achieve this goal, it is vital to synthesize a stable and high adsorption capacity adsorbent. In this work, Li₄Ti₅O₁₂ nanorods are prepared by two hydrothermal processes followed by a calcination process. Then the prepared Li₄Ti₅O₁₂ nanorods are treated with different HCl concentrations to obtain a H₄Ti₅O₁₂ adsorbent with 5 μm length along the [100] direction. The maximum amount of extracted lithium can reach 90% and the extracted titanium only 2.5%. The batch adsorption experiments indicate that the H₄Ti₅O₁₂ nanorod maximum adsorption capacity can reach 23.20 mg g⁻¹ in 24 mM LiCl solution. The adsorption isotherms and kinetics fit a Langmuir model and pseudo-second-order model, respectively. Meanwhile, the real adsorption selectivity experiments show that the maximum Li⁺ adsorption capacity reaches 1.99 mmol g⁻¹, which is far higher than Mg²⁺ (0.03 mmol g⁻¹) and Ca²⁺ (0.02 mmol g⁻¹), implying these nanorods have higher adsorption selectivity for Li⁺ from Lagoco Salt Lake brine. The adsorption capacity for Li⁺ remains 91% after five cycles. With the help of XPS analyses, the adsorption mechanism of Li⁺ on the H₄Ti₅O₁₂ nanorods is an ion exchange reaction. Therefore, this nanorod adsorbent has a potential application for Li⁺ recovery from aqueous lithium resources.

---

1. Introduction

Lithium, an indispensable metal, is widely used in medicines, spacecraft, ceramics, thermonuclear reaction, metallurgical engineering and so on.^1–3 In recent years, the lithium supply has not satisfied the demands of the lithium battery market especially in battery consumption.^4,5 Lithium resources mainly exist in ores and brines. According to a United States Geological Survey report, lithium resources in salt lakes account for 69% of the world's lithium reserves. So, recovery of lithium from brine has attracted great interest from researchers in recent years. However, it is a challenge to seek an environment-friendly, costless method to extract lithium from salt lakes.

Some methods including precipitation, solvent extraction,⁶ membrane separation,⁷ and adsorption,^8–10 have been reported to recover lithium. Among these methods, adsorption has been recognized as one of the most promising methods for Li⁺ recovery owing to its high ion selectivity and environment-friendly properties.^11,12 Usually, spinel lithium manganese oxides (LiMn₂O₄,¹³ Li_1.33Mn_1.67O₄ ¹⁴ and Li_1.6Mn_1.6O₄,¹⁵ lithium titanium oxides (Li₂TiO₃ and Li₄Ti₅O₁₂) and LiCl·2Al(OH)₃·xH₂O have been used to recover lithium. Spinel lithium manganese oxides have been synthesized as the precursor for preparation lithium ion-sieves to recover Li⁺ in sea water and brines. However, serious Mn loss of lithium manganese oxides during the acid process impede their industrial application. Compared with lithium manganese oxides, lithium titanium oxides have good adsorption capacities and slight Ti loss in acid process. Li₄Ti₅O₁₂ has the same spinel structure like Li_1.33Mn_1.67O₄, which also has good adsorption capacity.¹⁶ The reason for low dissolution is that the titanium valence remains stably +4 during leaching and adsorption process. Furthermore, Li₄Ti₅O₁₂ has well anti-acidic property and is suitable to extract lithium due to the Ti–O bond. Kaneko et al. investigated the effect of LiCl·2Al(OH)₃·xH₂O morphology on the Li⁺ adsorption behavior and presented the surface area was an important factor for adsorption process.¹⁷ Li et al. prepared yolk–shell structured Li₄Ti₅O₁₂ to increase the surface area and the adsorption uptake can reached to 28.46 mg g⁻¹ in 50 mM LiCl solution.¹⁸ Wei et al. recovered about 59.1 mg g⁻¹ of lithium in 144 mM LiCl solution by using H₄Ti₅O₁₂ as adsorbent.¹⁹ Li et al. also synthesized three-dimensionally H₄Ti₅O₁₂ exhibited the high adsorption performance.¹⁶ So, the morphology and surface area have important factors to affect the adsorption behavior. According to the above H₄Ti₅O₁₂ adsorbents, H₄Ti₅O₁₂ has good adsorption performance and recycle stability. Considering that the nanowires and nanorods have similar one-dimensional structures, we believe that nanorods adsorbents have the similar surface area. However, research about the adsorption behavior of H₄Ti₅O₁₂ nanorods has still not been reported.

In this study, the Li₄Ti₅O₁₂ nanorods was prepared by hydrothermal reaction followed by heat treatment in the air, and the H₄Ti₅O₁₂ along [100] direction growth was obtained by acid treatment. The white samples were characterized by XRD, SEM, TEM, XPS, FTIR and TG. The batch adsorption experiments (such as pH value, adsorption temperatures, lithium concentrations, adsorption selectivity and recycle stability) of H₄Ti₅O₁₂ nanorods were investigated. In addition, the H₄Ti₅O₁₂ nanosheets was used to compare the influence of morphology on the Li⁺ adsorption performance. Meanwhile, the adsorption behaviors and ions selective properties from Lagoco Salt Lake brine were obtained. Meanwhile, the adsorption behaviors and ions selective properties from Lagoco Salt Lake brine were obtained. The possible adsorption mechanism of H₄Ti₅O₁₂ nanorods was elaborated.

2. Experimental

2.1 Preparation of H₄Ti₅O₁₂ nanorods

The Li₄Ti₅O₁₂ nanorods was fabricated using two hydrothermal process followed by the calcined treatment. The synthesis process is modified from Li report.²⁰ First, the 3.0 g anatase TiO₂ was added to the 150 mL 10 M NaOH solution by stirring 1 h and sonicated for another 0.5 h. Then the above solution was transformed to Teflon-lined autoclave in 200 mL volume, and maintained for 48 h at 150 °C. When the autoclave was cooled to the room temperature, the white powder was collected by centrifugation, and washed with deionized water and 0.1 M HCl solution until the pH reached to neutral. After acid treatment, the 0.5 g titanium oxides were dispersed in a 40 mL 0.8 M LiOH aqueous solution. Then, the above solution was transformed to the autoclave and kept at 95 °C for 24 h. The obtained sample was washed with deionized water and ethanol for several times, the white sample was dried at 60 °C and calcined at 550 °C for 6 h to obtain Li₄Ti₅O₁₂ nanorods. The H₄Ti₅O₁₂ nanorods was prepared by the ion-exchanged methods. Briefly, the 0.1 g Li₄Ti₅O₁₂ nanorods were added to the 10 mL 0.2 M HCl solution for 48 h to obtain the H₄Ti₅O₁₂ nanorods. Then, the white sample washed by deionized water and ethanol for several times, and dried at 60 °C for 8 h. This acid process can be expresses by reaction (1) and the amount of extracted Li⁺ (a_Li⁺) and the amount of extracted Ti⁴⁺ (a_Ti⁴⁺) are used to determine the extraction efficiency for lithium and the stability of lithium ionic sieve, which are calculated by eqn (2). The synthesized process of H₄Ti₅O₁₂ nanorods is depicted in Scheme 1.

Li₄Ti₅O₁₂ + 4H⁺ = H₄Ti₅O₁₂ + 4Li⁺ (1)

where C_t (mg L⁻¹) is the concentration of Li⁺ (or Ti⁴⁺) in solution at random time, V (L) is the volume of the solution, and S (mg) is the mass of lithium or titanium in Li₄Ti₅O₁₂ nanorods.

Scheme 1 Illustration of the synthesis procedure of H₄Ti₅O₁₂ nanorods lithium ion sieves.

2.2 Li⁺ adsorption on the H₄Ti₅O₁₂ nanorods

2.2.1 Li⁺ adsorption of H₄Ti₅O₁₂ nanorods. The batch adsorption experiments were carried by adding 0.1 g adsorbent into 50 mL of different concentrations of Li⁺ solutions with a shaking speed of 150 rpm. The pH was adjusted by HCl (0.6 M) and KOH (5.0 M). The adsorption isotherms were investigated in 24 mM LiCl solutions at 25 °C, 35 °C and 45 °C for 4 h at pH 13. The adsorption kinetics were obtained in 12, 24 and 36 mM LiCl solutions at 25 °C for different times at pH 13. The adsorption capacity was calculated by eqn (3).

2.2.2 Selectivity of H₄Ti₅O₁₂ nanorods. The selectivity of H₄Ti₅O₁₂ nanorods was carried by dispersing 0.1 g H₄Ti₅O₁₂ into 50 mL 24 mM simulation solutions and real brine (Lagoco Salt Lake, Tibet) containing Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, Ca²⁺ and Mg²⁺. The concentrations of all metal ions were tested by ICP after the adsorption equilibrium. The Li⁺ selectivity is characterized by distribution coefficient (K_d) and separation factor (α^Li_M) as shown in eqn (4) and (5).

The initial and equilibrium Li⁺ concentration is defined as C₀ (mg L⁻¹) and C_e (mg L⁻¹), the volume of solution names V (L), m (g) is mass of adsorbent, α^Li_M is separation factors that means the Li to M ion (Na, K, Rb, Cs), respectively.

2.2.3 Recyclability of adsorbents. The reuse of H₄Ti₅O₁₂ nanorods was performed on desorption and regeneration processes. Firstly, 0.8 g H₄Ti₅O₁₂ were added into 400 mL LiCl solution with initial concentration of 24 mM at pH 13, and equilibrated for 4 h. Afterwards, the H₄Ti₅O₁₂ after Li⁺ adsorption was immerged into 0.2 M HCl aqueous solution and stirred for 24 h to carry out the regeneration process. Then, white precipitates were washed by ethanol and deionized water for three times, respectively. The precipitates were dried in an oven at 60 °C for 12 h. In the next reutilization experiments, the volume of LiCl solution was obtained according to adsorbent dosage (in this study, H₄Ti₅O₁₂ dosage was 2 g L⁻¹). The reuse of H₄Ti₅O₁₂ was lasted for five cycles following the above steps.

2.3 Characterization of adsorbents

X-ray diffraction (XRD) (XPert PRO, PANalytical, Netherlands) was used to characterize the crystalline phase of samples. The morphology and size of samples were observed by SEM (SU8010, HITACHI, JAPAN) and TEM (Titan G2 60-300, FEI, USA). The concentrations of ions were measured with Inductively Coupled Plasma (Optima 7000DV, PerkinElmer, USA) and Ion Chromatography System (ICS-1100). The BET surface area and pore size distributions of H₄Ti₅O₁₂ nanorods was measured through N₂ adsorption–desorption isotherms on a TriStar II 3020 nitrogen adsorption apparatus (Micromeritics, USA). X-ray photoelectron spectroscopy (XPS) analysis was recorded on a PHI 5300x multi-technique system with a Mg-K_α X-ray source. FT-IR (Tensor 27, Bruker, Germany) patterns were collected from 400 cm⁻¹ to 4000 cm⁻¹ with a resolution of 2 cm⁻¹. DSC-TG curves of material was performed on a Netzsch Leading thermal analyzer (STA449F3, NETZSCH, Germany) at a heating rate of 10 °C min⁻¹ in nitrogen atmosphere.

3. Results and discussion

3.1 Material characterization

Fig. 1 presents the XRD patterns of titanium oxides obtained from different process. In the XRD spectrum of first hydrothermal process, the diffraction peaks at 8.96°, 18.01°, 28.0°, 38.7° and 59.1° were assigned to the reflections of (200), (400), (110), (501) and (020) planes, respectively, which were correlated with H₂Ti₂O₅·xH₂O (PDF no. 47-0124).²¹ Through second hydrothermal treatment, the synthesized sample can be perfectly indexed to the layer-structured orthorhombic Li_1.81H_0.19Ti₂O₅·xH₂O unit cell with a = 16.66 nm, b = 3.979 nm, and c = 3.007 nm (PDF no. 47-0123) (Fig. 1a).²² After Li_1.81H_0.19Ti₂O₅·xH₂O was calcined at 550 °C for 6 h, the peaks at 18.33°, 35.57° and 43.24° appear, which correspond to the spinel Li₄Ti₅O₁₂ (PDF no. 49-0207), implying the layer-structured of Li_1.81H_0.19Ti₂O₅·xH₂O was majorly transformed to spinel Li₄Ti₅O₁₂ with only minor phase of anatase TiO₂ (Fig. 1b).²³ After acid treatment, the H₄Ti₅O₁₂ maintains the spinel structure and the intensity become weakly (Fig. S1†). In addition, all peaks are slightly shifted, implying the unit cell dimension is changed by the ion exchange between H⁺ and Li⁺.

Fig. 1 The XRD patterns of titanium oxides. (a) H₂Ti₂O₅·xH₂O and Li_1.81H_0.19Ti₂O₅·xH₂O, (b) Li₄Ti₅O₁₂ and H₄Ti₅O₁₂.

TG-DSC curves were applied to investigate the phase transformation of the Li_1.81H_0.19Ti₂O₅·xH₂O upon heating and curves were shown in Fig. 2. A total weight loss of about 12.71% was observed between 40–800 °C. The first weight loss region is 40–200 °C and the corresponding weight loss is about 7.60%. Basing on the literatures, the first weight loss region corresponded to the phase transformation from Li_1.81H_0.19Ti₂O₅·xH₂O to Li_1.81H_0.19Ti₂O₅ by an endothermal peak appeared about 193.4 °C.²⁴ The second weight loss region is 200–400 °C and the weight loss is about 4.5%, which corresponds to the phase transformation from Li_1.81H_0.19Ti₂O₅ to Li₄Ti₅O₁₂.²⁵ Heating to 200–400 °C results in the collapse of the layered structure and the formation of Li₄Ti₅O₁₂. The weakly broad peak appears at 600 °C to 800 °C and the weight loss is about 0.6%, which is ascribed to the Li evaporation.²⁶

Fig. 2 The DSC-TG curves of Li_1.81H_0.19Ti₂O₅·xH₂O.

The size and morphology of titanium oxides were observed by SEM and images were showed in Fig. 3. The image of H₂Ti₂O₅·xH₂O shows one-dimensional nanorods with the average length around 300 nm (Fig. 3a). After second hydrothermal process, the Li_1.81H_0.19Ti₂O₅·xH₂O still retains the one-dimensional morphology with the length around 5 μm at 95 °C for 24 h (Fig. 3b). However, when treated temperature increasing to 120 °C for 24 h in Fig. S2b,† the nanorods morphology disappears and is completely converted to particles with the size from 50 nm to 100 nm (Fig. S1†), indicating that the relatively low temperature is beneficial for keeping the nanorods morphology. After calcined process at 550 °C for 6 h (Fig. 3c), the Li₄Ti₅O₁₂ possess a rod-like morphology with largely rough surface. After acid treatment, the Li₄Ti₅O₁₂ nanorods is converted to H₄Ti₅O₁₂ nanorods without the morphology change (Fig. 3d).

Fig. 3 The SEM images of (a) H₂Ti₂O₅·xH₂O, (b) Li_1.81H_0.19Ti₂O₅·xH₂O, (c) Li₄Ti₅O₁₂ and (d) H₄Ti₅O₁₂.

The TEM technology was used to further characterize the Li₄Ti₅O₁₂ nanorods structure and the images were shown in Fig. 4. It is easily observed that the nanorods structure is very stable even after ultrasonic processing and electron beam attacking during TEM test in Fig. 4a and b. The spacing lattices were 0.48 nm and 0.21 nm, which match well with the (111) and (−400) planes or equivalent facets of spinel Li₄Ti₅O₁₂. Therefore, it can conclude that the Li₄Ti₅O₁₂ nanorods is growth along [100] direction with rough surface (Fig. 4c and d).²⁷ From the above results, the Li₄Ti₅O₁₂ nanorods has a high crystallinity and the rough surface would be benefit for Li⁺ and H⁺ ion exchange.

Fig. 4 (a) and (b) TEM images and (c) and (d) high resolution TEM images of Li₄Ti₅O₁₂. The inset of (c) is corresponding FFT pattern, and inset of (d) is the Li₄Ti₅O₁₂ (111) top view crystal structure from the [100] direction.

The precursor Li₄Ti₅O₁₂ and the final adsorbent H₄Ti₅O₁₂ were identified by typical FTIR spectra and presented in Fig. 5. In the spectrum of Li₄Ti₅O₁₂ and H₄Ti₅O₁₂, the peaks at 3442.17 cm⁻¹ and 1632.82 cm⁻¹ are ascribed to the O–H vibrations of adsorbed water in the sample surface,²⁸ while the peaks at 1075.87 cm⁻¹ and 459.89 cm⁻¹ are resulted from Ti–O and Li–O vibrations.²⁹ After acid treatment, the new peak appears at 504.36 cm⁻¹ was imputed to the vibration of Ti–O vibrations, which indicates that the Li⁺/H⁺ ion exchange reaction present in acid treatment process.²⁷

Fig. 5 The FTIR spectra of Li₄Ti₅O₁₂ nanorods and H₄Ti₅O₁₂ nanorods.

In order to further understand the impact of nanorods structure on the adsorption process, the surface area and pore distribution were investigated by BET and the results were shown in Fig. 6. The adsorption/desorption isotherm is type IV with slight hysteresis according to the IUPAC classification and the pore size is almost 25 nm, which will be benefit to the H⁺ and Li⁺ exchange. The BET surface area of H₄Ti₅O₁₂ nanorods is calculated using BET equation to be 84.34 m² g⁻¹.

Fig. 6 N₂ adsorption–desorption isotherms of H₄Ti₅O₁₂ nanorods.

3.2 Adsorption studies

3.2.1 Acid treatment. In order to investigate the optimal HCl concentrations for the lithium and titanium extraction, the Li₄Ti₅O₁₂ nanorods were treated in different HCl concentrations and the results were shown in Fig. 7. As shown in Fig. 7a, the lithium can easily leach from Li₄Ti₅O₁₂ nanorods and reach the equilibrium almost 10 h. The amount of extracted lithium increases from 48% to 89% when the HCl concentration increases from 0.05 M to 0.20 M. When the HCl concentration increases from 0.50 M to 1.00 M, the amount of extracted lithium changes slightly. In contrast, the amount of extracted titanium increases slightly from 0.1% to 2.0% when HCl concentration changes from 0.05 M to 0.20 M. When the HCl concentration increases from 0.50 M to 1.00 M, the amount of extracted titanium changes dramatically from 11% to 19% (Fig. 7b). Generally speaking, lithium ion sieves have to possess the little dissolution of titanium and large extracted of lithium to increase the adsorption performance. Therefore, 0.20 M HCl solution is adopted in the subsequent study to maximize the extraction efficiency for lithium ions and minimize the dissolution of titanium ions from Li₄Ti₅O₁₂ nanorods.

Fig. 7 The amount of extracted (a) lithium and (b) titanium from Li₄Ti₅O₁₂ nanorods in different HCl concentrations.

3.2.2 Effect of pH value on the adsorption capacity. The pH value of the initial solution has an important impact on the Li⁺ adsorption and the Li⁺ adsorption capacity of H₄Ti₅O₁₂ nanorods was shown in Fig. 8. In this study, the adsorption capacity of nanorods adsorbent increased when the pH value changed from 4 to 13. Especially, when the pH value is changed from 12 to 13, the adsorption capacity is increased from 16.49 mg g⁻¹ to 22.06 mg g⁻¹. So, the pH value of solution at 13 is optimized to investigate the adsorption behavior in our adsorption experiments. The influence of pH on lithium uptake capacity is acquired by empirical and mechanistic models. The mechanism can be explained by ion-exchange behavior, which H⁺ ions exchanged by Li⁺ ions. It is indicated that the high pH is benefit to the Li⁺ adsorption and this process can be explained by reaction (6).

H₄Ti₅O₁₂ + 4Li⁺ = Li₄Ti₅O₁₂ + 4H⁺ (6)

Fig. 8 The effect of pH value on Li⁺ adsorption (C₀: 24 mM LiCl, adsorbent: 0.1 g, volume: 50 mL, T: 25 °C).

3.2.3 The influence of morphologies on the Li⁺ adsorption. To investigate the influence of different morphologies on the Li⁺ adsorption, the H₄Ti₅O₁₂ nanosheets was derived from Li₄Ti₅O₁₂ nanosheets and the adsorption behaviors and adsorption capacities were investigated between H₄Ti₅O₁₂ nanorods and H₄Ti₅O₁₂ nanosheet. The XRD patterns show that the peaks of white sample obtained from the hydrothermal process can be assigned to Li_1.81H_0.19Ti₂O₅·xH₂O and the peaks can be indexed to spinel Li₄Ti₅O₁₂ after heat treatment (Fig. S2†). The SEM image of Li_1.81H_0.19Ti₂O₅·xH₂O nanosheets displays nanosheets structure with length close to 1 μm (Fig. S3a†). The morphologies and sizes of Li₄Ti₅O₁₂ and H₄Ti₅O₁₂ have little changed after heating and acid leaching treatments (Fig. S3b and c†).

The adsorption capacities of H₄Ti₅O₁₂ with different shapes entirely increase with the increasing of LiCl concentrations (Fig. S4†). Especially, the maximum adsorption capacities of H₄Ti₅O₁₂ nanosheets and nanorods are 18.8 mg g⁻¹ and 21.8 mg g⁻¹ in 24 mM LiCl solutions, respectively (Fig. S4a and b†). Meanwhile, the impact of adsorption temperatures on the Li⁺ adsorption process was systematically investigated. It is easily found that the higher temperatures of LiCl solutions are benefit to increase the adsorption capacities of different shapes in Fig. S4c and d.† Meanwhile, the results of adsorption temperatures on adsorption capacities of H₄Ti₅O₁₂ nanorods shows that the higher temperatures are benefit to increase the adsorption capacities and the adsorption capacities of H₄Ti₅O₁₂ nanorods are higher than H₄Ti₅O₁₂ nanosheets. From the above results, it is easily indicated that the H₄Ti₅O₁₂ nanorods has good adsorption capacity and the H₄Ti₅O₁₂ nanorods was used to perform the following adsorption experiments.

Besides, considering different adsorbents have different adsorption capacities, we added the comparison of the adsorption performance with other reported Li-adsorbents in Table 1. From this table, the manganese oxides (HMn₂O₄, H_1.33Mn_1.67O₄ and H_1.66Mn_1.67O₄) have larger adsorption capacities and the adsorption uptakes can reach to 10–40 mg g⁻¹. However, the dissolution of manganese is an inevitable problem (the loss of Mn >3%). Although, some researchers have been devoted to decrease the loss of manganese in acid leaching process through coated or doped methods.^30,31 The loss of manganese decreases obviously and the dissolution of manganese are still larger than the dissolution of titanium of titanium oxides (H₂TiO₃ and H₄Ti₅O₁₂).^32,33 LiCl·2Al(OH)₃·xH₂O adsorbents possess lower cost, simple preparation and adaptable advantages in real brines. But the adsorption capacities (5–8 mg g⁻¹) need to improve in industrial application.³⁴ Compared to manganese oxides and LiCl·2Al(OH)₃·xH₂O, titanium oxides have high adsorption capacities and lower dissolution of titanium. These results imply that this H₄Ti₅O₁₂ nanorods adsorbent has a potential application in Li⁺ recovery.

Table 1 Comparison of the adsorption performance with other reported Li-adsorbents

Materials	qe (mg g^−1)	Te (h)	Loss of Ti/Mn	Adsorption environment			Ref.
				pH	T (°C)	CLiCl (g L^−1)
HMn2O4	11.8	4	11%	12	25	0.02	31
H1.33Mn1.67O4	12.1	4	10%	12	45	0.20	31
H1.66Mn1.67O4-doped	30.8	48	3.9–4.42%	12	25	0.08	30
H1.66Mn1.67O4 nanorods	35.0	24	5.2%	11	25	0.15	35
H2TiO3	16.6	22	0.75%	14	25	1.00	32
H2TiO3 nanoparticles	34.2	24	—	12	25	0.5	33
Yolk–shell H4Ti5O12	28.5	8	1%	13	25	0.35	36
H4Ti5O12 nanosheets	20.2	4	1.8%	13	45	0.17	This work
H4Ti5O12 nanorods	31.4	4	2%	13	45	0.25	This work

---

3.2.4 Adsorption isotherms. In order to understand Li⁺ adsorption behavior on H₄Ti₅O₁₂ nanorods, the Langmuir (eqn (7)) and Freundlich models (eqn (8)) were used to analysis adsorption process:³⁷where q_m (mg g⁻¹) is the maximum uptake capacity in theory, b (L mg⁻¹) is the Langmuir constant, k_F (mg g⁻¹) and n are the Freundlich constant.

Fig. 9 and Table 2 show the fitting dates of the Langmuir and Freundlich isotherms models, which are usually used to understand the behavior of monolayer and multilayer adsorption on the adsorbent surface. The Li⁺ adsorption process is fitted better with the Langmuir isotherms models due to higher correlation R² at different reaction temperatures, indicating that the ions exchanged process determine by Li⁺ adsorbed on the surface of H₄Ti₅O₁₂ nanorods with a monolayer adsorption.

Fig. 9 (a) Langmuir and (b) Freundlich isotherms models for Li⁺ adsorbed on the H₄Ti₅O₁₂ nanorods at different adsorption temperatures. (adsorbent: 0.1 g, volume: 50 mL, shaking speed: 150 rpm).

Table 2 The fitting results of the Langmuir and Freundlich isotherms models for Li⁺ adsorbed on the H₄Ti₅O₁₂ nanorods at different reaction temperature

T (°C)	Langmuir model			Freundlich model
	qm (mg g^−1)	b (L mg^−1)	R^2	kF (mg g^−1)(L mg^−1)	1/n	R^2
25	25.99	0.24	0.99	10.59	4.67	0.98
35	26.33	0.28	0.99	10.75	4.53	0.96
45	29.74	0.21	0.94	11.95	4.73	0.90

---

The Li⁺ adsorption behaviors of H₄Ti₅O₁₂ nanorods at different temperature (25 °C, 35 °C, 45 °C) are depicted in ESI (Fig. S5†). The Li⁺ adsorption uptake of H₄Ti₅O₁₂ nanorods is enhanced with increase of temperature, indicating the adsorption is endothermic reaction.³⁸

3.2.5 Adsorption kinetics. The Li⁺ adsorption behavior on H₄Ti₅O₁₂ nanorods was fitted by the pseudo-first-order and pseudo-second-order kinetic models and their mathematical form is expressed by the eqn (9) and (10):^39,40

ln(q_e − q_t) = ln q_e − K₁t (9)

where q_e (mg g⁻¹) and q_t (mg g⁻¹) are the adsorption capacity at equilibrium and a random time t (min); the pseudo first-order and pseudo-second-order adsorption constants denote K₁ (min⁻¹) and K₂ (g (mg min)⁻¹), respectively.

The fitting results of relevant kinetic parameters of pseudo-first-order and pseudo-second-order models were displayed in Fig. 10. From the fitting dates, it is easily observed that the R² values of pseudo-second-order are much higher than the R² values of pseudo-first-order. Moreover, the experimental adsorption capacity (q_e,exp) values (Fig. S6†) match well with the theoretical adsorption capacity (q_e,cal) values calculated from the pseudo-second-order model rather than the pseudo-first-order model (Table 3). These results indicate that the Li⁺ adsorption process can be well described by the pseudo-second-order model, and this suggests that the Li⁺ adsorption on H₄Ti₅O₁₂ nanorods is controlled by the chemical exchanged process.

Fig. 10 (a) Pseudo-first-order and (b) pseudo-second-order kinetics plots for the Li⁺ adsorption on the H₄Ti₅O₁₂ nanorods at different LiCl concentrations (adsorbent: 0.1 g, volume: 50 mL, shaking speed: 150 rpm, temperature: 25 °C).

Table 3 The pseudo-first-order and pseudo-second-order kinetics models match for the Li⁺ adsorption behavior on the H₄Ti₅O₁₂ nanorods at different Li⁺ concentrations

C0 (mg g^−1)	qe,exp (mg g^−1)	Pseudo-first-order			Pseudo-second-order
		qe,cal (mg g^−1)	K1 (min^−1)	R^2	qe,cal (mg g^−1)	K2 (min^−1)	R^2
12	17.86	3.40	0.01	0.63	18.07	5.03 × 10^−4	0.99
24	20.81	9.20	0.01	0.51	21.03	2.53 × 10^−4	0.99
36	23.05	8.59	0.02	0.89	23.20	2.47 × 10^−4	0.99

---

3.2.6 Adsorption selectivity. Considering that the concentration of Li⁺ is very low while the coexist ions (Na⁺, K⁺, Rb⁺, Cs⁺) are higher in the brines, the adsorption selectivity experiments of nanorods adsorbent was furtherly investigated in other alkali metal ions and coexist ions.

The results of Li⁺ adsorption selectivity in non-competitive and competitive solution were illustrated in Fig. 11. It is easily found that this adsorbent has good Li⁺ adsorption selectivity, both in non-competitive and competitive solutions. The maximum adsorption abilities can reach to 3.05 mmol g⁻¹. The distribution coefficient (K_d) and separation factor (α^Li_M) of various cations were shown in Table 4. The order of K_d values of H₄Ti₅O₁₂ nanorods for these cations is Li⁺ ≫ K⁺ > Rb⁺ ≈ Cs⁺ > Na⁺, which may be caused by the ion radius of Li⁺ (0.059 nm) can migrate to the pore channel of nanorods adsorbent while Na⁺, K⁺, Rb⁺, Cs⁺ with lager ion radius are only can uptake on the surface of adsorbent without enter into it.⁴¹ The real adsorption selectivity experiments show that the maximum Li⁺ adsorption capacity reach to 1.99 mmol g⁻¹, which was far higher than Mg²⁺ (0.03 mmol g⁻¹) and Ca²⁺ (0.02 mmol g⁻¹), indicating that the prepared materials have excellent separation performance of Li⁺ and Mg²⁺ in Lagoco Salt Lake brine in Table 5. Consequently, the nanorods adsorbent has a high capacity for Li⁺ ions and well ion selective property for Li⁺ in aqueous solution.

Fig. 11 Equilibrium adsorption capacities of H₄Ti₅O₁₂ nanorods adsorbent in the (a) non-competitive solution with 24 mM cation and (b) a competitive solution with 24 mM for each cation.

Table 4 Selective adsorption of H₄Ti₅O₁₂ nanorods in solutions containing Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺

Metal ions	qe (mmol g^−1)	Kd (mL g^−1)	α^LiM
Li^+	2.81	800.28	1
Na^+	0.06	2.73	296.51
K^+	0.39	12.13	66.78
Rb^+	0.16	8.54	94.77
Cs^+	0.14	7.57	106.83

---

Table 5 Adsorption selectivity of H₄Ti₅O₁₂ nanorods in Lagoco Salt Lake

Metal ions	qe (mg g^−1)	qe (mmol g^−1)	Kd (mL g^−1)	α^LiM
Li^+	13.80	1.99	690.00	1.00
Na^+	12.50	0.54	50.00	13.80
K^+	1.00	0.03	20.00	34.50
Mg^2+	0.90	0.03	45.00	15.33
Ca^2+	0.91	0.02	45.65	15.12

---

3.2.7 Adsorbent recycling. The regeneration capabilities of H₄Ti₅O₁₂ nanorods were evaluated by five sequential cycles of adsorption/desorption process and the results including adsorption capacities and XRD patterns were shown in Fig. 12 and S6.† It can be easily found that the adsorption capacity for Li⁺ decreases slightly after five adsorption experiments. In addition, the adsorption capacity of Li⁺ still maintains 91% of the first adsorption capacity in the fifth cycle (Fig. 12). The XRD pattern in Fig. S7(a)† shows that the peaks of H₄Ti₅O₁₂ in each cycle are similar to the original pattern of H₄Ti₅O₁₂, which indicates the well stability of H₄Ti₅O₁₂. After the Li⁺ adsorption, the diffraction peaks of samples can be indexed to the spinel Li₄Ti₅O₁₂ unit cell with a = 8.3588, b = 8.3588 and c = 8.3588 nm (PDF no. 49-0207) in Fig. S7(b),† which is indicated that the H⁺ ions are exchanged by the Li⁺ ions and the structure is transforming from the H₄Ti₅O₁₂ to Li₄Ti₅O₁₂ without structure collapsing. These results indicate that the H₄Ti₅O₁₂ nanorods has high adsorption capacity in inorganic Li⁺ ion exchange materials and well chemical stability.

Fig. 12 The Li⁺ adsorption capacities of H₄Ti₅O₁₂ nanorods in each recycling process. (adsorbent: 0.1 g, C₀: 24 mM LiCl, T: 25 °C, time: 4 h).

3.2.8 Adsorption mechanism. The element compositions of H₄Ti₅O₁₂ and Li₄Ti₅O₁₂ nanorods were investigated by XPS in Fig. 13 and the mechanism of Li⁺ adsorbed on H₄Ti₅O₁₂ surface was elaborated. Strong peaks of O 1s and Ti 2p can be clearly observed in Fig. 13a. The intensity of Li 1s on Li₄Ti₅O₁₂ increases obviously than H₄Ti₅O₁₂ (Fig. 13b), probably due to the H–O transformation to Li–O bonds. Before the Li⁺ adsorption, the high-resolution XPS spectrum of Ti 2p shows the two peaks at 458.81 eV and 464.46 eV in Fig. 13c, which correspond well with characteristic Ti 2p_3/2 and Ti 2p_1/2 peaks of Ti⁴⁺.⁴² After Li⁺ adsorption, there is 0.34 eV deviation of Ti 2p_3/2, which can be ascribed to H–O–Ti transformation to Li–O–Ti. From the image of Fig. 13d, there is 0.35 eV deviation of O–Ti and the intensity of –OH increases after the Li⁺ adsorption, which may ascribe to the H–O–Ti transformed to Li–O–Ti. Based on the literatures, adsorption mechanism is consisted of redox reactions^43,44 and ion exchange.⁴⁵ There is no valence change of Ti after Li⁺ adsorption (Fig. 13c), indicating that the Li⁺ adsorption on H₄Ti₅O₁₂ is an ion exchange reaction which matches well with the Dubinin–Radushkevich isotherm model.

Fig. 13 The XPS analyses of H₄Ti₅O₁₂ and Li₄Ti₅O₁₂ nanorods. (a) Survey spectra of H₄Ti₅O₁₂ and Li₄Ti₅O₁₂ and (b) Li 1s; (c) the high-resolution spectra of Ti 2p and (d) the high-resolution spectra of O 1s.

4. Conclusions

In this study, the Li₄Ti₅O₁₂ nanorod were prepared by hydrothermal methods through two steps followed by a calcination process. Then, the H₄Ti₅O₁₂ nanorods along the [100] direction were obtained by HCl acid treatment. Batch adsorption experiments were conducted to investigate the adsorption behavior of the nanorod adsorbent. Acid treatment experiments show the maximum amount of extracted lithium can reach 90% and the amount of extracted titanium only reaches 2.5%. The adsorption behavior of H₄Ti₅O₁₂ is monolayer adsorption and the adsorption process is a chemical exchange process. Adsorption selectivity experiments show that nanorod adsorbent has high selectivity for Li⁺ ions in simulated brine and real brines. In addition, the Li⁺ adsorption capacity still remains 16.83 mg g⁻¹ even after five recycling experiments. The Li⁺ adsorption on H₄Ti₅O₁₂ nanorods is an ion exchange reaction by adsorption mechanism analysis. Furthermore, this nanorod adsorbent is a promising candidate as an environment friendly Li⁺ adsorbent for successful application to exploit liquid lithium resources.

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

This study was supported by the NSFC (No. U1607105, No. 51302280). Natural Science Foundation in Qinghai Province (No. 2018-GX-101, No. 2019-HZ-808, No. 2018-ZJ-722). The Thousand Talents Plan in Qinghai Province and Youth Innovation Promotion Association of Chinese Academy of Sciences (No. 2016377).

Notes and references

1. J. W. An, D. J. Kang, K. T. Tran, M. J. Kim, T. Lim and T. Tran, Hydrometallurgy, 2012, 117–118, 64–70.
2. L. Y. Zhang, D. L. Zhou, G. He, F. H. Wang and J. B. Zhou, Mater. Lett., 2014, 135, 206–209.
3. S. Ziemanna, M. Weil and L. Schebeka, Resour., Conserv. Recycl., 2012, 63, 26–34.
4. B. Swain, Sep. Purif. Technol., 2017, 172, 388–403.
5. S. Hu, Y. Li, F. Lai, X. Zhang, Q. Li, Y. Huang, X. Yuan, J. Chen and H. Wang, RSC Adv., 2015, 5, 17592–17600.
6. J. Nan, D. Han and X. Zuo, J. Power Sources, 2005, 152, 278–284.
7. H. B. Li and L. D. Zou, Desalination, 2011, 275, 62–66.
8. I. H. Chowdhury, A. H. Chowdhury, P. Bose, S. Mandal and M. K. Naskar, RSC Adv., 2016, 6, 6038–6047.
9. I. H. Chowdhury and M. K. Naskar, RSC Adv., 2016, 6, 67136–67142.
10. H. N. M. E. Mahmud, A. K. O. Huq and R. b. Yahya, RSC Adv., 2016, 6, 14778–14791.
11. S. Kaneko and W. Takahashi, Colloids Surf., 1990, 47, 69–79.
12. W. J. Chung, R. E. C. Torrejos, M. J. Park, E. L. Vivas, L. A. Limjuco, C. P. Lawagon, K. J. Parohinog, S.-P. Lee, H. K. Shon, H. Kim and G. M. Nisola, Chem. Eng. J., 2017, 309, 49–62.
13. C. Özgür, Solid State Ionics, 2010, 181, 1425–1428.
14. J. L. Xiao, X. Y. Nie, S. Y. Sun, X. F. Song, P. Li and J. G. Yu, Adv. Powder Technol., 2015, 26, 589–594.
15. F. R. Qian, M. Guo, Z. Q. Qian, Q. Li, Z. J. Wu and Z. Liu, ChemistrySelect, 2019, 4, 10157–10163.
16. N. Li, K. F. Gan, D. Lu, J. L. Zhang and L. Z. Wang, Res. Chem. Intermed., 2017, 44, 1105–1117.
17. S. Kaneko and W. Takahashi, Colloids Surf., 1990, 47, 69–79.
18. N. Li, D. Lu, J. Zhang and L. Wang, J. Colloid Interface Sci., 2018, 520, 33–40.
19. S. D. Wei, Y. F. Wei, T. Chen, C. B. Liu and Y. H. Tang, Chem. Eng. J., 2020, 379, 122407.
20. Y. Li, G. L. Pan, J. W. Liu and X. P. Gao, J. Electrochem. Soc., 2009, 156, A495.
21. H. Zhang, G. R. Li, L. P. An, T. Y. Yan, X. P. Gao and H. Y. Zhu, J. Phys. Chem. C, 2017, 111, 6143–6148.
22. K. Zhu, H. Gao and G. Hu, J. Power Sources, 2018, 375, 59–67.
23. Y. Sha, B. Zhao, R. Ran, R. Cai and Z. Shao, J. Mater. Chem. A, 2013, 1, 13233.
24. T. Gao, H. Fjellvåg and P. Norby, J. Phys. Chem. B, 2008, 112, 9400–9405.
25. N. Li, T. Mei, Y. Zhu, L. Wang, J. Liang, X. Zhang, Y. Qian and K. Tang, CrystEngComm, 2012, 14, 6435.
26. S. Wang, W. Quan, Z. Zhu, Y. Yang, Q. Liu, Y. Ren, X. Zhang, R. Xu, Y. Hong, Z. Zhang, K. Amine, Z. Tang, J. Lu and J. Li, Nat. Commun., 2017, 8, 627.
27. J. Shu, Electrochem. Solid-State Lett., 2008, 11, A238.
28. M. R. Mohammadi and D. J. Fray, J. Sol-Gel Sci. Technol., 2010, 55, 19–35.
29. J. H. Kim, S. y. Bae, J. H. Min, S. W. Song and D. W. Kim, Electrochim. Acta, 2012, 78, 11–16.
30. F. R. Qian, B. Zhao, M. Guo, Z. Q. Qian, N. C. Xu, Z. J. Wu and Z. Liu, Hydrometallurgy, 2020, 193, 105291–105300.
31. H. S. Wang, J. J. Cui, M. L. Li, Y. F. Guo, T. L. Deng and X. P. Yu, Chem. Eng. J., 2020, 389, 124410–124419.
32. L. Y. Zhang, D. L. Zhou, Q. Q. Yao and J. B. Zhou, Appl. Surf. Sci., 2016, 368, 82–87.
33. R. Chitrakar, Y. Makita, K. Ooi and A. Sonoda, Dalton Trans., 2014, 43, 8933–8939.
34. J. Zhong, S. Lin and J. G. Yu, J. Colloid Interface Sci., 2020, 572, 107–113.
35. F. R. Qian, M. Guo, Z. Q. Qian, Q. Li, Z. Wu and Z. Liu, ChemistrySelect, 2019, 4, 10157–10163.
36. N. Li, D. L. Lu, J. L. Zhang and L. Z. Wang, J. Colloid Interface Sci., 2018, 520, 33–40.
37. Z. Liu, Y. Q. Zhou, M. Guo, B. L. Lv, Z. J. Wu and W. Z. Zhou, J. Hazard. Mater., 2019, 371, 712–720.
38. X. C. Shi, D. F. Zhou, Z. B. Zhang, L. L. Yu, H. Xu, B. Z. Chen and X. Y. Yang, Hydrometallurgy, 2011, 110, 99–106.
39. D. Ding, K. X. Li, D. Z. Fang, X. S. Ye, Y. Q. Hu, X. L. Tan, H. N. Liu and Z. J. Wu, ChemistrySelect, 2019, 4, 13630–13637.
40. Z. Liu, R. Yu, Y. P. Dong, W. Li and B. L. Lv, Chem. Eng. J., 2017, 309, 815–823.
41. Q. Feng, H. Kanoh, Y. Miyai and K. Ooi, Chem. Mater., 1995, 7, 1226–1232.
42. H. Li, L. Shen, X. Zhang, J. Wang, P. Nie, Q. Che and B. Ding, J. Power Sources, 2013, 221, 122–127.
43. X. Xu, Y. Chen, P. Wan, K. Gasem, K. Wang, T. He, H. Adidharma and M. Fan, Prog. Mater. Sci., 2016, 84, 276–313.
44. J. C. Hunter, J. Solid State Chem., 1981, 3, 142–147.
45. X. Luo, K. Zhang, J. Luo, S. Luo and J. Crittenden, Environ. Sci. Technol., 2016, 50, 13002–13012.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0ra05094f

---