1. Introduction

Emissions of greenhouse gases cause severe global warming, and consequently, major climate change. Reduction on the dependence on fossil fuels and finding alternative sources of clean energy is extremely essential. Rechargeable batteries are one such alternative sources for generating energy. Currently the lithium ion battery (LIB) is a very successful form of batteries although there is much research being conducted to substitute conventional LIBs with more efficient and less costly ones. With the growing demand on lithium, drastic price jumps can be prevented by mining the metal or extracting it from salt-water sources. Since mining is both dangerous and nonrenewable, extracting lithium from natural water sources is sustainable and necessary. Disposing of briny waste after desalination is an immediate issue. The recovery of lithium ions from brine will be a win-win situation as it reduces polluted water while recovering precious metals. Recycling LIBs by recovering their lithium content is another valuable process which will prevent further environmental pollution. Different processes are used for the extraction of the metals, but currently membrane separation is one of the most efficient methods of extraction [1-5].

There are many methods for the desalination of seawater using nanofiltration (NF) membranes [6-10]. Reverse osmosis (RO) is a well-established process for desalination as it simply uses applied pressure to forcibly filter the feed through the NF membrane to produce a purer solution. The selectivity in RO is mostly determined by pore size of the membrane and its chemical affinity toward water and lithium ions. Electrodialysis is another effective process in which charged membranes, such as the cation and anion exchange membrane (CEM and AEM), are used to selectively filter certain ions when voltage is applied. Similar to electrodialysis, electrolysis is an electrochemical process that selectively mobilizes ions for effective separation and collection. Furthermore, fabricating composite membranes is a favorable method to optimize lithium recovery in any separation process. Today, many studies dedicate their research to modifying NF membranes in order to collect pure lithium solutions by overcoming the challenge of rejecting magnesium and other salts in the seawater that interfere greatly during the filtration process. For instance, ion-imprinted membranes are highly selective of lithium ions and easily reject other ions.

Many recent studies that explain novel composite membranes or analyze the effectiveness of RO NF, electrodialysis, electrolysis, or other methods show great potential for the separation of lithium ions from brine, especially for large-scale recovery from salt-water sources. The synthesis, characterization, and results of some of these studies on lithium recovery are discussed in this review.

2. Membrane separation process

2.1. Nanofiltration

S. H. Park et al. introduced an efficient two-step process for Li recovery from brine via nanofiltration (NF) and membrane distillation (MD) [11]. Commercially available polyamide (PA) based NF membranes and a hydrophobic and porous polyvinylidene fluoride (PVDF) membrane were purchased for NF and MB, respectively. Artificial brine samples with varying ratios of eight different salts were created to imitate Li-containing natural water sources around the world. Ion-coupled plasma optical emission spectrometry (ICP-OES) was used to determine cation concentrations before and after Li recovery. Field-emission scanning electron microscopy (FE-SEM) visualized the NF membrane surface morphologies after larger salts were filtered out and crystalized. To compare the efficiency of the novel two-step membrane process with naturally-occurring salt precipitation in water sources, a solar-based evaporation method for brine was deployed. This involved solar evaporation, solvent extraction, and chemical treatment, which proved to be time consuming and cost-intensive due to high usage of energy and chemicals. Moreover, the final Li concentration extracted from the brine is low due to poor separation from other remain- ing salts. The two-step membrane separation process improved upon these flaws and consequently reduced recovery time and total cost. First, the brine is filtered through the NF membranes to filter out divalent cations, such as Ca²⁺ and Mg²⁺, which tend to crystallize (this was to prevent formation of crystals on the MB membrane surface in the next step). Then the thinned brine was passed through the MD membrane to further separate out the remaining salts and produce a Li-rich solution. The Li concentration (1,200 ppm) after the two-step membrane process resulted in up to twelve times the concentration of the solar-evaporated solution (100 ppm), which was highly desirable for efficient Li collection afterwards. In addition, the membranes exhibited a water flux of 22.5 L m^-2 h^-1 which was approximately sixty times greater than the evaporation method, significantly decreasing operation costs and time. This simple two-step membrane process has great potential for Li recovery in real large-scale water sources.

L. Wang et al. proposed a non-PA, metal-coordinated NF membrane for separation processes in brines [12]. Using polyethersulfone (PES) as the substrate, m-phenylenediamine (MPD) ligand solution, CuCl₂ metal solution, and NaIO₄ oxidant solution were added in respective order on top of the PES substrate and shaken for 5 h to react and polymerize, as illustrated in Fig. 2. Afterwards, the Cu-MPD NF membrane was obtained and completed with posttreatment. Transmission electron microscopy (TEM) characterized the cross-section and surface layers of the NF membrane. Quartz crystal microbalance with dissipation (QCM-D) analyzed the chemical properties of the Cu-MPD complexes that were formed during the reaction step. In lithium recovery processes, current commercialized NF membranes (based on PA) have shown to be ineffective due to poor Li⁺/Mg²⁺ selectivity. They also possess an inherent water-permeance/salt-selectivity trade-off, which means increasing permeance usually decreases the effects of selectivity, and vice versa. This study incorporated a complex of copper metal and organic MPD, where the Cu²⁺ acted as both a crosslinking agent for MPD chains as well as positively charged sites. Performancewise, the novel Cu-MPD NF membrane maintained an excellent water permeance of 16.2 ± 2.7 LMH bar^-1, rivaling the performance of high water permeance (conventional) NF membranes. For Li⁺/Mg²⁺ selectivity, the membrane exhibited 8.0±1.0, which was a significant improvement from the poor selectivities of the same NF membranes. In addition to overcoming the inherent trade-off performance of conventional NF membranes, the Cu-MPD complex showed properties of pHresponse. It shows better overall performance of the membrane at lower pH—and anti-biofouling, proving the long-term stability of the novel membrane even in complex brine mixtures. Other morphological characterizations (Fig. 3) determined the average pore radius (1.25 nm) and porosity (21%) and revealed a pore tunability factor by adjusting monomers or cations, extending the possible applications of the Cu-MPD NF membrane in the separation field.

P. Xu et al. focused on improving the concentration of Li⁺ and rejection of Mg²⁺ by synthesizing a PA-based layer of nano-functionalized networks on NF membranes [13]. Using PES as the substrate, a PA layer—doped with potassium carboxylate functionalized multi-wall carbon nanotubes (MWCNTs-COOK)—was synthesized on top of the PES substrate via the interfacial polymerization of aqueous solution (containing MWCNTs-COOK) and organic phase (trimesoyl chloride (TMC) in n-hexane). Dopant concentrations included 0, 30, 100, 150, 300, and 1000 ppm and labeled NF1 to NF6, respectively. FE-SEM provided magnified images of the membrane surface morphologies, mostly focusing on uniform distribution of dopants and the intactness of the PA layer. Water contact angle (WCA) measured the hydrophilicity of each synthesized membrane. With the COOK portion providing high positive charges on the membrane surface, the MWCNTs would enhance the hydrophilic properties; combined, a tight network of cross-linked structure is created, allowing highly efficient water channels. More specifically, the novel MWCNTs-COOK NF membrane in artificial brine exhibited an outstanding performance where the highest water flux was 12.23 L m^-1 h^-1 bar^-1 and an Li⁺/Mg²⁺ separation factor of 58. To emphasize this significance, the water flux is on par with some of the highest performing NF membranes reported, but its separation factor is approximately triple the average separation factor. It is good to note that the 150 ppm concentration performed the best, and membrane characterizations indicating that higher concentrations of MWCNTs-COOK led to interferences in the polymerization step and consequently formed loose layers in the final membrane. Moreover, hydrophilicity increased with increasing dopant concentrations, but this is essentially useless if the membrane is not tightly intact and cannot filter efficiently. Nonetheless, the MWCNTs- COOK novel NF membrane has proven its superiority over other reported NF membranes and shows its potential in future lithium recovery technologies in larger brine sources.

2.2. Electrodialysis

M. Bazrgar et al. fabricated a CEM to separate Li⁺ from Na⁺ using electrodialysis [14]. After purchasing commercially available CEM resin and powdering into microparticles, the CEM was dissolved in dimethyl formamide (DMF) and well combined with another prepared solution of PES/DMF. The homogenized mixture was then casted and dried into a membrane and further modified with a thin surface coating of spinel-type lithium selective adsorbent particles in poly(acrylic acid) (this membrane was called M-CEM). An alternative membrane with the selective particles inserted directly into the CEM matrix was fabricate (C-CEM). Fourier transform infrared (FTIR) analyzed the chemical interactions between functional groups in the CEMs, which confirmed the peaked presence of the adsorbent material in the novel membranes where it was absent in the pure CEMs. Ion exchange capacity (IEC) characterization measured the electrochemical properties and Li⁺ mobility of the novel membranes. Generally, the performance of the C-CEM was slightly worse than M-CEM, but they were overall similar in effect. Although unmodified CEMs have high cation flux, their cation selectivity is low, especially in the cases where cations are similar in properties such as Li⁺/Na⁺. By incorporating a layer of lithium-selective adsorbent particles, the modified CEMs would improve the cation selectivity; however, ion flux would be hindered due to pore blockages. In detail, the M-CEM in artificial brine showed a molar Li⁺ flux of 0.149 mol min^-1 m^-2 and molar selectivity of 32.2, which was an 18.58% decrease and 62.3% increase, respectively, compared to pure CEM. Despite the decrease in flux, the more pressing issue of selectivity was significantly improved in comparison. The drawback in flux, however, cannot be ignored, which is why the critical concentration of the selective adsorbents was determined to be 0.6 wt%, as any higher particle loadings would form large agglomerations and considerably effect the flow efficiency. As long as critical loadings are managed, modified CEMs to improve cation selectivity shows promising results for lithium ion recovery.

M. Grageda et al. studied the effects of changing electrodialysis parameters to determine ways to produce higher concentrations of LiOH—a highly pure cathode precursor desirable in batteries—without producing energy-intensive lithium carbonates in between [15]. As seen in Fig. 4 and Fig. 5, an acrylic cell is created where the anode (graphite with impure LiCl solution as the anolyte) and cathode (stainless steel or nickel with LiOH solution as the catholyte) compartments are strictly separated by a membrane. The membrane used were commercially available copolymer perfluorosulfonic acid/polytetrafluoroethylene. X-ray diffraction (XRD) measured the crystal structures in the collected solid from the final catholyte, verifying the major presence of LiOH. In this study, temperature, current density, membrane thickness, impurity concentration, and electrode material were varied to optimize operating conditions as well as LiOH recovery. In terms of energy efficiency, literature suggests that reported specific electricity consumptions (SECs) ranged between 5 to 15 kW h per kg of LiOH. The optimal conditions, therefore, should produce high purity LiOH (above 99%) while consuming close to 5 kW h kg^-1 LiOH. First, temperatures exceeding 80°C generally decreased purity most likely due to the thermal expansion of membrane pores and decrease in ion selectivity. The highest purity of 99.93% was obtained below 75°C, but electrolytic conductivity was lowered, causing SEC to increase (i.e., over 11 kW h kg^-1 LiOH). Second, current densities exceeding 2400 A m^-2 would generally overproduce Li⁺ at a rate higher than the ions could migrate through the membrane, decreasing the overall LiOH generation per energy consumption. The lowest SEC (7.25 kW h kg^-1 LiOH) was measured when the current density was set to 1200 A m^-2 at temperatures at 85°C, although the purity decreased to 93.65%. Third, thicker membranes, lower concentrations of impurities in the initial anolyte, and nickel as the cathode improved purity due to improved selectivity and ion flow rate. Depending on the desired result either high LiOH purity or low SEC can be achieved, but it is difficult to say whether further adjustments on only experimental parameters could allow simultaneous improvement in lithium production and energy consumption.

T. Ounissi et al. synthesized a composite membrane —or mixed matrix membrane (MMMs)—to improve upon commercialized NF membranes flexibility and selectivity for lithium ions [16]. To begin, polyepichlorohydrin (PECH) and 1,4-diazabicyclo[2.2.2]octane (DABCO) were dissolved in dimethylacetamide (DMAc) to form a polymer binder solution. Lithium ion conducting glass ceramic powder (LICGC) was mixed with DMAc and BRIJ76 (C₁₈H₃₇[C₂H₄O]₁₀OH) surfactant before it was dispersed (“blended”) uniformly into the PECH-DABCO. Finally, PES base solution was combined to the dispersion, and the resultant mixture was casted and dried to form the novel lithium selective composite membrane (LCM). SEM characterized the membrane surface and cross-sectional morphologies. An INSTRON traction machine was used to test the tensile strength of the LCM. The LICGC membrane alone may inherently possess good lithium selectivity, but its high rigidity risks brittleness and cracking under pressure. On the other hand, most polymer membranes are highly flexible, yet they possess poor selectivity. The hope was to combine these two materials to compensate for each drawback, fabricating a strong and flexible film that is highly selective of lithium. For the composite membrane to work, however, the LICGC inorganic particles must be homogenous with the polymer binder; this was verified during membrane characterization where microscopic images showed even dispersions of LICGC particles and no agglomeration formation. Moreover, the mechanical properties were greatly affected, confirming that a higher polymer content resulted in a higher rupture point (meaning it can withstand greater stress/ strain). The LCM was tested under dialysis (not to be confused electrodialysis) conditions using an impure Li⁺ feed containing similar cations such as Na⁺ and K⁺, which would be filtered into distilled water. It is rare for the Li⁺ selectivity to surpass 100 for both commercialized or novel NF membranes, but the LCM exhibited superior lithium ion selectivity of 363 for Li⁺/Na⁺ and 278 for Li⁺/K⁺. Although flexibility is indeed critical for a long lasting performance, it is crucial to incorporate LICGC—or any lithium selective additive— in lithium recovery technologies. Despite the inorganic nature of LICGC, this study has proven that blending/dispersion method can be used to create a homogenous and working membrane.

P. P. Sharma et al. doped sulfonated poly(ether ether ketone) (SPEEK) polymer with nanomaterials highly selective in lithium in order to synthesize a cation exchange membrane to be applied in electrodialysis [17]. In detail, lithium manganese oxide nanoparticles were dispersed in DMAc solvent and then later combined with a prepared SPEEK/DMAc solution; the resulting mixture was casted and dried to form a thin composite membrane. The universal testing machine (UTM) characterized the synthesized membranes’ mechanical stability in terms of stress/strain and elasticity. Ionic conductivity of membranes was determined by AC impedance spectroscopy. Because of the high impurity factors of brines extracted from natural resources, such as salt lakes, lithium recovery membranes that are highly selective and durable for multiple cycles is greatly desired. In terms of durability, the nanomaterial/ SPEEK composite membrane (nano-SPEEK) showed decrease in elasticity when increasing the loading of nanomaterials and decreasing polymer; however, the overall strength of the membrane increased due to the introduction of crystalline nanomaterials in the matrix. The ionic conductivity also increased when introducing nanomaterials because the transport of ions became smoother due to increase in hydrophilicity of the membrane (i.e., wider water channels). Consequently, the nano-SPEEK in artificial brine showed a natural lithium adsorption of 15.2 mg g^-1, which was five times the intake of Mg. In addition, a respectable 64% of lithium was recovered during electrodialysis, where selectivities for Li/Mg, Li/K, and Li/Na were 4.82, 3.0, and 2.17, respectively. Although this novel membrane does not extract the purest lithium solution through dialysis, it was determined that the nano-SPEEK is suitable for recovering Li⁺ from larger scale brine sources with good efficiency.

2.3. Electrolysis

C.H. Díaz Nieto et al. proposed a new methodology of concentrating Li⁺ in brines by coupling membrane electrolysis with crystallization to remove large concentrations of Na⁺ [18]. In this study, a three-compartment acrylic electrochemical cell was applied. Commercially available AEM was used to separate the anode and middle compartments whereas CEM was used to separate the cathode and middle compartments; titanium mesh and stainless steel mesh were the anode and cathode, respectively. In terms of electrolytes, the anode compartment contained (bi)carbonate buffer, the cathode contained NaHCO₃, and the middle contained the brine (real brine with/without Mg²⁺ and Ca²⁺ removed). XRD was used to measure the crystal structures of the precipitated solids. In general, the lithium recovery process was structured in three main steps: first was removing Mg and Ca ions from real brine in the form of hydroxides, second is the crystallization and removal of Na⁺ in the form of NaHCO₃, and third is the crystallization and collection of Li⁺ in the form of Li₂CO₃. (These carbonates are formed when the cations are selectively passed through the CEM into the cathode and reacted with CO₂ during water electrolysis to precipitate pure carbonates.) Through characterizations, it was confirmed that the precipitated NaHCO₃ had an excellent purity of 99.5%, highly concentrating the remaining brine with Li⁺ for more efficient recovery in the following step. In addition, the resultant product was pure water. To be specific, the water salinity was about two orders of magnitude lower than the initial brine and the SEC to treat 1 m³ of brine was measured to be 331.3 kW h. Although this three-step membrane electrolysis procedure was both energy and time consuming, the benefits of extracting pure salts was obtaining a by-product of fresh water, which is arguable worth the cost in environments where natural fresh water sources are depleting at alarming rates.

W.R. Torres et al. focused on improving the lithium carbonate recovery efficiency in the previously mentioned three-stage membrane electrolysis methodology [19]. The acrylic based electrochemical cell was divided into three compartments. The anode (containing a titanium current collector in a buffer of carbonates) and middle (where brine is introduced) were separated by commercially available AEM; likewise, the middle and cathode (containing stainless steel current collector) were separated by commercially available CEM. The first stage of electrolysis removed any large divalent salts from a prepared artificial brines. The second stage removed Na⁺ to further concentrate the resulting brine with lithium. The third stage, the focus of this study, would then crystallize the lithium in the form of lithium carbonate (Li₂CO₃) by introducing carbon dioxide. XRD characterization determined the crystalline structure of the resulting Li₂CO₃ crystal powders, comparing the purity of the powders produced from brines of different initial ion content. The brine samples included Li⁺/K⁺ solution, Li⁺/Na⁺ solution, and Li⁺/K⁺/Na⁺ solution, and the diffraction peaks of the powdered sample from each brine were identical to pure Li₂CO₃ (from reference), indicating a successful lithium carbonate recovery regardless of brine content or concentration. SEM images showed the surface morphologies of the anodic and cathodic exchange membranes before and after each stage of electrolysis. Interestingly, the untouched membranes already had micro holes and cracks; however, only the AEM seemed to increase in defects considerably after each stage of electrolysis. Otherwise, both AEM and CEM showed increase in salt deposits that remained imbedded in the membrane even after thorough cleaning. In terms of efficiency, the purity of the collected Li₂CO₃ salt range between 93.8 and 97.5 wt% with an energy consumption of 70.6 kWh m^-3. In addition, a by-product of low-saline water was obtained. Although increasing the energy input certainly improves the efficiency of lithium recovery, the financial consequences are an issue. It is, therefore, imperative to find alternative solutions to reduce electrochemical resistance or raise current densities, perhaps starting with innovations for exchange membranes.

2.4. Others

J. Cui et al. introduced a facile membrane fabrication for lithium adsorption and recovery by polymerizing biodegradable hybrids of cellulose acetate (CA) and chitosan (CS) [20]. The CA/CS membrane was synthesized by evenly blending dimethyl sulfoxide, CS, CA, PVA, PVP, and glutaraldehyde and casting the resulting solution into a film. This was then doped with dopamine and then polymerized with TiO₂ (obtained from hydrolyzed tetrabutyl titanate). Finally, the modified CA/CS membrane was combined with LiCl and other crosslinking agents to polymerize the novel lithium ion imprinted composite membrane (LIICM). The incorporation of lithium ions would create pore pockets highly tuned for only Li⁺, essentially leaving “imprints” when the ions were later rinsed and removed before testing. Atomic force microscopy (AFM) observed the surface roughness of the membranes, which showed a higher roughness for LIICM than unmodified CA/CS membranes, increasing specific surface area for adsorption. FTIR measured the characteristics peaks of the respective functional groups in the LIICM, which indicated a successful synthesis as all the appropriate peaks were shown. In the adsorption experiments, the novel LIICM exhibited an excellent maximum Li⁺ adsorption capacity of 20.08 mg m^-1 when immersed in an LiCl solution. The selectivity was then tested with the LIICM immersed in solution of Li⁺ and one other type of ion. In detail, the selectivity ratios of Li⁺/Na⁺, /K⁺, /Mg²⁺, and /Ca²⁺ were respectively 1.78, 2.43, 2.60, and 3.61. As for reusability, the LIICM had an Li⁺ loss of only about 7% and negligible membrane/imprint damage even after six full cycles, confirming the sustainability of using LIICMs.

M. Mohammad et al. introduced a composite of metal organic framework (MOF) and metal phenolic network (MPN) into membranes to advance lithium ion extraction processes [21]. Using polypropylene (PP) as the substrate, two coatings were applied. The first coating was the MPN layer, which was synthesized by casting an aqueous solution of tannic acid and FeCl3 hydrate to form TA-Fe^III particles. The second coating was the MOF layer, which was a result of casting an aqueous solution of Zn²⁺ and 2 methylimidazole to form zeolitic imidazolate framework particles (ZIF-8). Zeta potential characterization determined the surface charge of the novel TA-Fe^III/ZIF-8 membrane, which indicated a shift to a negative potential upon the incorporation of TA-Fe^III particles in the MPN layer. The increase in negative charge would consequently attract Zn²⁺ and promote the nucleation of ZIF-8 particles in the MOF layer. SEM images showed the cross-section of the composite membrane, showing smooth interface adhesion and confirming strong chemical interaction between the two active layers. Because MOFs are known for their advanced tunability in pore size and chemical selection, it was considered a desirable component to filter mixtures consisting of chemicals with different sizes and chemical properties. The TA-Fe^III/ZIF-8 membrane was used as a separator between the anode and cathode (both containing the same salt solution) of a simple electrochemical cell to test its ion transport properties. Various selectivity factors of monovalent ion/Mg²⁺ concentrations were calculated, including 4.49 for K⁺/Mg²⁺, 4.0 for Na⁺/Mg²⁺, and 3.87 Li⁺/Mg²⁺; the selectivity factor for Ca²⁺/Mg²⁺ was 1.1 (negligible difference). Although membrane was highly selective of the monovalent ions compared to the divalent ones due to size disparity, there were negligible differences in selectivity between the monovalent ions themselves, similar to the divalent ions (Ca²⁺/Mg²⁺). In fact, K⁺ seemed to have greater in ion mobility than Li⁺ through the membrane, which was most likely a chemical interaction issue. Indeed, the Ta-Fe^III/ZIF-8 membrane is not suitable for Li⁺ extraction alone; however, it is certainly a highly efficient membrane to remove divalent ions to promote concentrating lithium in brines for later extraction processes.

A. Razmjou et al. studied the fundamental elements of nanostructured membranes that are crucial for ion selective properties, especially lithium ions [22]. By collecting and reviewing various works on Li⁺ selective membranes, this study was able to narrow down key aspects in the nanostructural design, including pore channels and Li⁺ selective chemicals. In terms of ion channels, considering the continuity of the tunnel, the selective pore size, and the negative surface charge are significant factors during the membrane synthesis step. Firstly, discontinuous channels—usually occurring when uniformity is disturbed by heterogeneously mixing or aggregation of additives—physically hinder ion mobility through membranes. These are best tackled during the synthesis process by either templating to create artificial and uniform tunnels or evenly dispersing ion-selective additives to form additional bridges or pathways. Secondly, pore size is important as a sieving component to quickly remove any large impurities; however, it is not so simple to synthesize pore sizes smaller than the hydration ionic diameter of lithium ions. A solution to this is to allow larger pore sizes up to the hydration diameter of lithium ions but to also ensure a negative surface charge through the channel to selectively pull the ions and reject water. As seen in Figure 9, the side view of the protein channel illustrates the efficient transport of the monovalent cation through the negatively charged tunnel, while rejecting much of the H₂O molecules (dehydration). Surface charge was determined to be one of the factors of nanostructures that is both effective in selectivity as well as tunable. One simple means to increase the negative charge of the inner linings of the nanochannels is to attach functional groups. For instance, incorporating chemicals rich in carbonyl groups (C=O) may introduce plenty of oxygen sites that provide negative charges to attract positive cations, as in Fig. 6. All in all, membranes with nanostructures that possess nanochannels that maintain uniform and continuous pathways, encourage selective dehydration of Li⁺ ions, and provide plenty of negative surface charge were observed to show the best results in terms of lithium ion selectivity.

S. Roobavannan et al. designed a process involving membrane distillation and chemical ion sieving to recover both pure water and lithium salt from real seawater [23]. Commercially available polytetrafluoroethylene (PTFE) hydrophobic membrane was used during the distillation step. Acid manganese oxide (HMO) —synthesized by reacting in situ prepared LiMnO₂ with HCl—was used as the ion-sieving solution during the final lithium extraction step. Both untreated and pretreated seawater were tested to study the effects inorganic species concentration as well as divalent ion concentration. WCA measurements determined the hydrophobicity and surface morphology of the membrane prior and after direct contact membrane distillation (DCMD). XRD characterization analyzed the crystalline properties of the HMO as a powder before and after its intake of Li⁺ from the resultant seawater after DCMD. In general, the extraction procedure is as follows: (1) either untreated or pretreated seawater samples were allowed to pass through the PTFE membrane, and water extraction was calculated (2) after DCMD the highly concentrated solution was mixed with HMO, and ion concentrations were calculated. In actuality, the DCMD step with PTFE membrane showed a significant increase in water recovery for seawater pretreated with oxalic acid compared to untreated (i.e., ~50% increased to ~90%). This was most likely due to the initial removal of Ca ions from the seawater with oxalic acid pretreatment. The water recovery results effected the following ion sieving process. To explain, after DCMD, the pretreated solution had 7 times the ion concentration of the untreated one, resulting in a greater adsorption of Li⁺ in HMO (17.8 mg g^-1). All in all, this study concluded that multiple stages of the lithium recovery process are necessary, especially when using real saline sources; these include pretreating (i.e., removing divalent species), concentrating lithium ions, and finally extracting (in this study, adsorption and desorption).

C. Yu et al. developed a novel composite membrane, ester-functionalized and lithium ion-imprinted, for recovering lithium ions from seawater [24]. Applying sugar-templated polydimethylsiloxane (PDMS) as the substrate, the PDMS base was copolymerized with polydopamine (PDA). The PDMS-PDA was then reacted with succinic anhydride to obtain PDMS-COOH, then reacted with thionyl chloride to obtain PDMS-COCl. The PDMS-COCl substrate was then imprinted by mixing with a prepared solution of ligand calix [4] arene, which resulted in the novel Li ion-imprinted membrane (Li-IIM). FTIR analysis detected functional groups of the synthesized membranes. Basically, Si-O-Si was detected for pristine PDMS, additional N-H for PDMSPDA, carboxyl groups for PDMS-COOH, and C-Cl for the final PDMSl-COCl membrane, verifying a successful synthesis of a functionalized membrane. Energy dispersive X-ray (EDX) spectroscopy revealed nearly 50% of lithium content in the Li-IIM, indicating a successful synthesis, but no lithium content in the PDMSCOCl membrane due to the lack of ligand ion templating. To emphasize, the significance of the novel Li-IIM was the high Li⁺ adsorption capacity resulting from the abundance of lithium-suitable cavities and functional groups. Specifically, the Li-IIM possessed a high rebinding capacity of 50.872 mg g^-1, and good selectivities of Na⁺, K⁺, and Rb+ (respectively 1.71, 4.56, 3.80) were approximately triple the results of PDMSCOCl. Moreover, the ester group structure exhibited regeneration abilities after several cycles of adsorption/ desorption, indicating a long-term stability for the novel composite membrane.

3. Conclusions

Lithium recovery from natural salt water sources is beneficial for LIB production, water purification, and brine waste reduction. Reverse osmosis nanofiltration, electrodialysis, and electrolysis using membranes are ion-separation processes that are effective for lithium extraction from mixtures of monovalent and multivalent metal ion solutions. Membrane separation processes are desirable as they can be highly selective of lithium ions and cost effective. For more advanced separation performance, some studies have modified these membranes into composite membranes in order to increase lithium ion permeation while maintaining high rejection of other ions. This review has discussed the successful synthesis of these novel membranes and their characterizations, including morphological (structural) analysis and ionic conductivity. Although these studies have shown mixed results, some better than others, there is no doubt that lithium recovery, especially with membrane filtration, is a field that requires further investment. Nonetheless, the various combinations of composite membranes and separation processes reveal the vast potential for innovative lithium recovery in the near future.