Stress-homogenized spatial architectures via entropy-driven self-assembly enabling high-performance and durable lithium extraction

Significance

The growing global demand for lithium necessitates more efficient and sustainable selective electrodes for electrochemical extraction from salt lakes. Spinel-type LiMn₂O₄ is highly promising, but its poor cycling performance is intrinsically limited by its inherent Jahn–Teller effect and lattice distortion issues. This paper proposes an entropy-driven amphiphilic self-assembly strategy for constructing stress-homogenized multilayer core–shell structures. By precisely tuning the internal geometric structure to optimize stress–strain distribution, our approach effectively mitigates stress concentration. This material achieved a record lithium capacity and exceptional cycling stability, outperforming current state-of-the-art materials. Ultimately, our work provides an engineering paradigm for creating advanced materials that could enable next-generation, high-performance water treatment and resource recovery systems.

Abstract

The global transition to sustainable energy demands efficient lithium extraction from brines. While electrochemical lithium extraction using LiMn₂O₄ (LMO) holds great promise, its practical application is hindered by mechanical degradation caused by anisotropic volume changes and stress accumulation during cycling. Herein, we present an entropy-driven amphiphilic self-assembly strategy that engineers stress-homogenized multilayer core–shell architectures, which innovatively mitigates stress accumulation by tuning the internal geometric structure to optimize stress–strain behavior, thereby synergistically enhancing ion distribution, transport kinetics, and electrochemical stability. This hierarchical interlayer architecture ensures uniform Li⁺ distribution and redistributes internal stresses, mitigating localized stress concentrations and lattice expansion to preserve structural integrity throughout cycling. The optimized LMO establishes a dual benchmark for both capacity and cycling stability in hybrid capacitive deionization, achieving a remarkable lithium extraction capacity of 4.78 mmol g⁻¹ with 96% retention over 100 cycles, outperforming both its unoptimized counterpart and other reported materials of the same type. Finite element simulations further elucidate a 48% reduction in maximum stress compared to disordered counterparts, underscoring the critical coupling between ion diffusion and stress evolution. This paradigm provides a pathway for developing advanced materials with intrinsically stable architectures for sustainable lithium extraction.

The accelerating global transition toward sustainable energy and electrification is underpinned by a secure supply of lithium resources (1). The environmental footprints and geopolitical bottlenecks of hard-rock mining necessitate the exploitation of unconventional brines via advanced, eco-benign technologies (2, 3). Electrochemical lithium extraction (ELE) has emerged as a pivotal solution, leveraging intrinsic redox mechanisms to circumvent chemical reagent consumption and ensure specific Li⁺ selectivity compared to traditional separation methods (4, 5). Critically, the industrial scalability of ELE technologies is predicated on the kinetic properties of the host electrode materials. Among candidates, LiMn₂O₄ (LMO) distinguishes itself owing to its cost-effectiveness, earth abundance, and unique three-dimensional spinel framework (comprising 8a-16c-8a pathways) (6, 7). In stark contrast to the kinetically constrained one-dimensional olivine channels of its primary competitor, LiFePO₄ (LFP), the isotropic architecture of LMO not only supports a high theoretical capacity of 148 mAh g⁻¹ but also facilitates Li⁺ diffusion coefficients 2 to 5 orders of magnitude higher (4, 8). This kinetic superiority underpins the high-rate capability and ion-sieving efficiency requisite for industrial-scale deployment.

Despite these intrinsic merits, the deployment of LMO in long-cycling ELE processes is severely impeded by its inherent structural instability (9–12). While earlier investigations primarily scrutinized the Jahn–Teller distortion itself, critical attention has increasingly shifted toward the downstream consequence: distortion-induced stress accumulation (13, 14). Specifically, during repeated Li⁺ de/intercalation cycles, LMO undergoes a reversible cubic-to-tetragonal phase transition accompanied by a substantial ~7.7% anisotropic volume expansion (notably, a 16% increase in the c/a ratio) (15). These transitions often propagate heterogeneously across individual nanoparticles, precipitating nonuniform stress distributions and localized strain concentrations (16, 17). Such cumulative mechanical stress exacerbates microcrack formation and particle pulverization, ultimately compromising both lithium extraction efficiency and long-term cycling stability. To mitigate such structural degradation, recent advances in the broader field of lithium-ion batteries (LIBs) have pivoted toward interface engineering and lattice stress regulation (18, 19). For instance, strategies such as reinforcing the cathode–electrolyte interface in LiCoO₂ (LCO) and employing high-entropy doping in ultrahigh-Ni Co-free single-crystalline cathodes have proven effective in suppressing stress-induced microcracking and phase-transition fatigue (20, 21).

Prevailing optimization strategies for ELE are predominantly predicated on surface passivation, typified by protective layers such as metal oxides and phosphates (9, 22–25). While recent studies demonstrate that Li_1.5Al_0.5Ge_1.5(PO₄)₃(LAGP) coatings can enhance the stability of LMO particles in brine environments, boosting its capacity retention from ~60 to ~88% after 30 cycles (26), this “passive defense” incurs an intrinsic penalty: The dense interfaces required to exclude water molecules inevitably erect additional transport barriers, severely compromising the rapid kinetics essential for capturing Li⁺ from dilute sources (27, 28). More fundamentally, rigid exterior coatings fail to accommodate the volumetric strain arising from intrinsic lattice breathing, leaving the active material susceptible to chemo-mechanical degradation via stress accumulation (29). Resolving this dichotomy necessitates a paradigm shift from superficial modification to bulk architecture engineering. Here, inspired by stress-relief mechanics, we propose an intrinsic geometric regulation strategy that engineers lattice-adaptive buffer spaces to facilitate stress delocalization, thereby achieving optimal ion diffusion kinetics and mechanical stability.

In this work, we present a strategy to construct a stress-homogenized multilayer core–shell LiMn₂O₄ architecture, achieved via entropy-driven self-assembly to enable uniform, synchronized lattice strain and avert stress accumulation, effectively addressing mechanical degradation in electrochemical lithium extraction from brines. By finely tuning the entropy-driven self-assembly of amphiphilic precursors, where their hydrophobic chains aggregate to spontaneously form micelle templates for the subsequent layer-by-layer process, we achieve a meticulously ordered internal structure with graded interlayer channels. This facilitates uniform Li⁺ distribution and optimizes ion transport kinetics, effectively mitigating localized stress concentrations and ensuring superior electrochemical performance and long-term structural integrity. Notably, LMO-MCS-4 demonstrates exceptional Li⁺ selectivity in Mg²⁺/Li⁺ systems (separation factor = 1,514.35) and stable performance in simulated Salar de Atacama brines, as revealed by a three-electrode setup. In the continuous-flow hybrid capacitive deionization (HCDI) system, the optimized LMO-MCS-4 achieves a remarkable Li⁺ extraction capacity of 4.78 mmol g⁻¹ with 96% retention over 100 cycles, outperforming the unoptimized material, which has a capacity of 3.14 mmol g⁻¹ and suffers from a 70% capacity decay. Finite element simulations show a 48% reduction in maximum stress for the multilayer core–shell structure compared to the disordered dense one, further revealing the synergistic coupling between ion diffusion and stress evolution. This work represents a paradigm shift in electrode material design, offering a strategic blueprint for high-performance, durable materials for sustainable lithium extraction technologies.

Results

Characterization of Multilayer Core–Shell LMO-MCS Electrode.

The fabrication of the LMO-MCS-n series (n = 1 to 5) exploits an entropy-driven self-assembly strategy coupled with controlled coprecipitation (Fig. 1A). Systematic optimization demonstrates that precursor architecture is dictated by surfactant concentration (SI Appendix, Fig. S1), with the LMO-MCS-4 regime (7% CTAB) surpassing the critical micelle concentration (CMC) to establish stable supramolecular templates (30). Thermodynamically, this assembly is propelled by the hydrophobic effect: Before micellization, water molecules are constrained into highly ordered, clathrate-like cages surrounding the hydrophobic alkyl tails (31). The aggregation of these chains into micellar cores triggers the liberation of these confined water molecules into the bulk phase. The resulting surge in translational solvent entropy (∆S_solve ≥ 0) serves as the dominant energetic driver, effectively offsetting the enthalpic penalty to minimize the system’s Gibbs free energy (∆G < 0) (31). This entropically stabilized interface orchestrates material growth: The cationic Stern layer of the micelles electrostatically recruits HCO₃⁻/CO₃²⁻ anions, which subsequently coordinate Mn²⁺ ions to direct a precise, layer-by-layer nucleation of MnCO₃ (SI Appendix, Fig. S3) (32, 33). This templating mechanism yields hierarchical core–shell spheres, strictly distinguishing them from the disordered, stochastic agglomerates formed in the absence of the entropic template (LMO-MCS-1, SI Appendix, Fig. S2). Subsequent calcination converts the precursor into porous Mn₂O₃ while faithfully inheriting the entropy-derived morphology. The final LMO-MCS exhibits a distinct particle size expansion relative to the Mn₂O₃ intermediate, corroborating the successful high-temperature intercalation of Li⁺ ion into the host lattice.

Fig. 1.

Synthesis and characterization of LMO-MCS-4. (A) Schematic illustration of the synthetic route for multilayered core–shell LMO-MCS-4. (B and C) SEM images of LMO-MCS-1 (control group) nanospheres and internal structure. (D and E) SEM images of LMO-MCS-4 nanospheres and internal structure. (F–H) HRTEM images of LMO-MCS-4. (I) SAED pattern of LMO-MCS-4. (J) XRD patterns of MnCO₃, Mn₂O₃, and LMO-MCS-4. (K) High-resolution Mn 2p XPS spectrum of LMO-MCS-4. (L) Nitrogen adsorption–desorption isotherms and corresponding pore diameter distribution curves of LMO-MCS-4.

SEM analysis (Fig. 1 B–E) reveals that while the series shares a uniform ~1 μm spherical morphology enveloped by a ~150 nm thick shell of aggregated nanocrystallites, the internal architectures exhibit a distinct entropy-driven evolution. Unlike the dense LMO-MCS-1 or the quasi-disordered intermediates LMO-MCS-2/3 (SI Appendix, Figs. S1 and S2), which lack geometric periodicity, LMO-MCS-4 achieves a thermodynamically optimal multilayer topology characterized by ~80 to 90 nm ordered gaps (Fig. 1 D and E and SI Appendix, Fig. S3). This specific architecture represents the ideal equilibrium between structural integrity and synthetic economy, as higher surfactant loadings (LMO-MCS-5) yield diminishing morphological returns. Crucially, these ordered interspaces function as “strain-buffering reservoirs” to accommodate lattice expansion while establishing rapid channels for uniform lithium transport.

High-resolution transmission electron microscopy (HRTEM) images of LMO-MCS-4 (Fig. 1F) reveal well-defined lattice fringes, indicating high crystallinity with discernible (111) and (311) planes (34), corresponding to interplanar spacings of 0.48 nm and 0.24 nm, respectively (Fig. 1 G and H). Selected area electron diffraction (SAED) patterns (Fig. 1I) show rings indexable to the (111), (311), (400), (511), and (440) planes of spinel LiMn₂O₄. X-ray diffraction (XRD) patterns (Fig. 1J and SI Appendix, Fig. S4) confirm the phase evolution from MnCO₃ (JCPDS 44-1472) to Mn₂O₃ (JCPDS 41-1442) and the final LiMn₂O₄ product (JCPDS 35-0782), exhibiting pure phases without peak splitting or impurities. Structural characterization of LMO-MCS-4 confirms the cubic spinel structure with space group Fd-3m and a lattice parameter of 8.2476 Å (35), where Li⁺ occupies 8a tetrahedral sites, and Mn³⁺/Mn⁴⁺ reside in 16d octahedral sites within the O²⁻ 32e framework (SI Appendix, Fig. S5). This spinel structure facilitates 3D Li⁺ diffusion pathways, crucial for high-rate performance. High-resolution X-ray photoelectron spectroscopy (XPS) (Fig. 1K and SI Appendix, Fig. S6) confirms the presence of mixed-valent Mn³⁺/Mn⁴⁺ states with a near 1:1 ratio in the Mn 2p_3/2 and Mn 2p_1/2 spectra, consistent with Li⁺ intercalation and the redox-active Mn³⁺/Mn⁴⁺ couple during electrochemical cycling (36).

Fig. 1L shows the N₂ adsorption/desorption isotherms and Barrett–Joyner–Halenda (BJH) pore size distribution for LMO-MCS-4. The IV-type isotherm with an H3-type hysteresis loop indicates mesoporous characteristics (37), with a specific surface area of 4.27 m² g⁻¹ and a pore size distribution predominantly in the mesoporous and macroporous range (10 to 50 nm). In contrast, LMO-MCS-1 exhibits a lower surface area of 3.56 m² g⁻¹ and fewer pores (SI Appendix, Fig. S7). The relatively favorable surface area and pore structure of LMO-MCS-4 are expected to facilitate Li⁺ transport. Notably, the differences in surface area and pore size are relatively minor, suggesting that the performance disparity is likely attributed to the structural arrangement of the particles rather than just the textural properties.

Electrochemical Properties of LMO-MCS Series Electrodes.

Electrochemical measurements revealed that the multilayer core–shell microstructure of LMO-MCS-4 significantly enhances Li⁺ transport and (de)intercalation processes compared to the disordered LMO-MCS-1 electrodes. The electrochemical properties of various LMO-MCS-n electrodes were first investigated in 1 M LiCl solution. As shown in Fig. 2A, the CV curves for the LMO-MCS-n series, measured at a scan rate of 1 mV s⁻¹ within 0.4 to 1.1 V, displayed two pairs of distinct oxidation-reduction peaks. These peaks indicate the reversible Li⁺ intercalation and deintercalation processes, corresponding to the two distinct redox processes of Mn (e.g., Mn³⁺/Mn⁴⁺) in the spinel structure (34). Notably, LMO-MCS-4 achieves the highest peak current of 3.43 mA cm⁻² at 0.93 V, significantly outperforming LMO-MCS-1, which reaches only 1.59 mA cm⁻² (SI Appendix, Table S4). This higher peak current indicates faster reaction kinetics and enhanced electrochemical performance, likely due to reduced charge-transfer resistance and improved ion diffusion, as supported by earlier structural analysis.

Fig. 2.

Electrochemical properties of LMO-MCS-n electrodes. (A) CV curves of LMO-MCS-n electrodes at a scan rate of 1 mV s⁻¹ in 1 M LiCl solution. (B) GCD curves of different LMO-MCS-n electrodes at a current density of 1 A g⁻¹ in 1 M LiCl solution. (C) Discharge capacity of LMO electrodes at a discharge rate of 0.5 to 10 A g⁻¹. (D) CV curves of LMO-MCS-4 at scan rates ranging from 0.1 to 1.0 mV s⁻¹. (E) Linear correlation between logarithmic oxidation/reduction peak currents and logarithmic scan rates for LMO-MCS-4. (F) Specific capacitance of the LMO-MCS-1 and LMO-MCS-4 electrodes at various scan rates based on the CV curves. (G) Cycle performance for LMO-MCS-n electrodes at 1 A g⁻¹. (H) Discharge curves of LMO-MCS-4 in 1 M MCl solution (M = Li⁺, Na⁺, K⁺, Mg²⁺, Ca²⁺) at a current density of 1 A g⁻¹, and its CV curves at a scan rate of 1 mV s⁻¹.

Further confirmation of these electrochemical improvements was provided by GCD measurements (Fig. 2B). All LMO-MCS electrodes showed two distinct plateaus during charge/discharge, confirming a two-stage Li⁺ intercalation/deintercalation process, which is consistent with the reversible redox reactions observed in the CV curves. These processes correspond to the following reactions:

$${\mathrm{LiMn}}_2{\mathrm{O}}_4-0.5{\mathrm{e}}^{-}-0.5\mathrm{Li}\rightleftharpoons {\mathrm{Li}}_{0.5}{\mathrm{Mn}}_2{\mathrm{O}}_4,$$

$${\text{Li}}_{0.5}{\text{Mn}}_2{\mathrm{O}}_4-0.5{\mathrm{e}}^{-}-0.5\text{Li}\rightleftharpoons \lambda -\text{MnO}2.$$

LMO-MCS-4 demonstrates a notably longer discharge time compared to the other electrodes, indicating a higher specific capacity. As shown in Fig. 2 B and C, LMO-MCS-4 achieves an impressive capacity of 138.5 mAh g⁻¹ at the discharge rate of 0.5 A g⁻¹, outperforming LMO-MCS-3 (134.8 mAh g⁻¹), LMO-MCS-2 (131.4 mAh g⁻¹), LMO-MCS-5 (122.1 mAh g⁻¹), and LMO-MCS-1 (116.3 mAh g⁻¹). Particularly at higher discharge rates, LMO-MCS-4 exhibits significantly enhanced rate performance compared to LMO-MCS-1 (Fig. 2C and SI Appendix, Fig. S8). The reversible discharge capacity of LMO-MCS-4 at 0.5, 1, 3, 5, 8, and 10 A g⁻¹ is 138.5, 127.9, 100.1, 80.1, 61.6, and 53.4 mAh g⁻¹, respectively, while LMO-MCS-1 shows a lower capacity of 116.3, 104.6, 75.6, 55.8, 23.2, and 17.4 mAh g⁻¹ at the same rates. This significant improvement in rate performance is attributed to the unique microstructure of LMO-MCS-4, which facilitates more efficient ion diffusion channels and enhances ion transport during the charge-discharge cycles.

Fig. 2D presents the CV curves of LMO-MCS-4 at scan rates ranging from 0.1 to 1.0 mV s⁻¹, accompanied by linear correlation analysis in Fig. 2E. The corresponding data for LMO-MCS-1 highlight differences in electrochemical kinetics (SI Appendix, Fig. S9). According to the Bruce-Dunn theory (38), expressed by the power-law equation I(V) = aV^b, electrochemical kinetics can be quantified by analyzing the relationship between the scan rate (v) and the response current (i). A b-value of 0.5 indicates a diffusion-limited process, while a value approaching 1.0 suggests capacitive-controlled behavior. The calculated b-values for the cathodic and anodic peaks in LMO-MCS-4 were 0.968 (peak 1), 0.922 (peak 2), 0.952 (peak 3), and 1.022 (peak 4), respectively (Fig. 2E). This result suggests that the (de)lithiation process in the modified nanolayered core–shell LMO-MCS-4 is primarily controlled by surface kinetics rather than limited by bulk Li⁺ diffusion. This indicates that Li⁺ transport within LMO-MCS-4 is rapid, forecasting excellent rate capability and fast charging/discharging performance for the modified material.

The specific capacitance is calculated from the CV curves at varying scan rates for LMO-MCS-1 and LMO-MCS-4 (Fig. 2F). At a scan rate of 1.0 mV s⁻¹, LMO-MCS-4 achieves a specific capacitance of 492.41 F g⁻¹, which closely matches its theoretical value of 532.8 F g⁻¹and significantly surpasses that of LMO-MCS-1 (373.68 F g⁻¹). This indicates a higher charge storage capacity and superior electrochemical performance for LMO-MCS-4. Fig. 2G further presents the 100-cycle GCD profiles of LMO-MCS-n electrodes at a current density of 1 A g⁻¹. LMO-MCS-4 exhibits an outstanding capacity retention of 91.7%, a significant improvement over LMO-MCS-1, which retains 72.1% of its initial capacity. Notably, this enhanced stability was achieved without the use of dopants or changes in crystal structure, highlighting that the primary cause of improved cycling stability is the structural difference, with the multilayer core–shell architecture significantly outperforming the dense, disordered configuration.

Fig. 2H illustrates the CV and GCD curves of LMO-MCS-4 in competitive ion solutions. In contrast to the distinct, reversible redox peaks observed in the LiCl electrolyte, no significant electrochemical signals were detected in Na⁺, K⁺, Mg²⁺, and Ca²⁺ containing solutions, indicating negligible insertion/extraction processes for these ions. Additionally, discharge curves in the other competitive ion solutions exhibit weak responses with no voltage plateaus, indicating that larger or higher-valent ions cannot effectively intercalate into the material’s lattice, with any minimal current likely arising from nonspecific adsorption or minor side reactions, further highlighting the exceptional Li⁺ selectivity of LMO-MCS-4. Collectively, these electrochemical results highlight that the structural modifications in LMO-MCS-4 significantly boost its lithium extraction efficiency in aquatic environments. With high specific capacity, superior rate performance, excellent cycling stability, and Li⁺ selectivity, LMO-MCS-4 emerges as a promising material for advanced energy storage applications.

Specific Lithium Extraction in the Three-Electrode Setup.

The electrochemical performance of LMO-MCS-n electrodes for lithium extraction from brines was evaluated using an enlarged 2 × 2 cm² working electrode in a sequential batch three-electrode configuration. LMO-MCS-4 demonstrated longer charge-discharge cycles with 10 to 70 mM LiCl (Fig. 3A), reflecting a higher Li⁺ intercalation capacity than LMO-MCS-1 (SI Appendix, Fig. S10). According to Fig. 3D and SI Appendix, Table S5, LMO-MCS-4 achieved a maximum extraction capacity of 4.05 mmol/g in 70 mM LiCl, a 33% increase over LMO-MCS-1 (3.04 mmol/g), quantitatively highlighting its superior Li⁺ extraction capability.

Fig. 3.

The lithium extraction performance of LMO-MCS-4 in the three-electrode intercalation/deintercalation setup. (A) Inset: Discharge profiles of LMO-MCS-4 in LiCl solutions of varying concentrations (10, 30, 50, and 70 mM). Main Panel: Corresponding charge profiles in a 10 mM NaCl receiving solution. (B) Inset: Discharge profiles of LMO-MCS-4 in binary solutions with different molar ratios of Mg: Li. Main Panel: Corresponding charge profiles in a 10 mM NaCl receiving solution. (C) Inset: Discharge profiles of LMO-MCS-4 during 5 consecutive cycles in brine. Main Panel: Corresponding charge profiles in a 10 mM NaCl receiving solution. (D) Li⁺ extraction capacity of LMO-MCS-1 and LMO-MCS-4 in LiCl solutions of varying concentrations (10, 30, 50, and 70 mM). (E and F) Extraction and separation performances of LMO-MCS-4 and LMO-MCS-1 in binary solutions with different molar ratios of Mg: Li. (G) Properties of competing ions and electrochemical mechanism for selective lithium extraction in brine. (H) Schematic illustration of the Li⁺ intercalation-deintercalation mechanism in the three-electrode system. (I) Comparative separation performance of LMO-MCS-1 and LMO-MCS-4 for simulated brine. (J) Evolution of receiving solution concentration during 5 extraction cycles of LMO-MCS-4 in simulated brine.

Extracting Li⁺ from complex salt lake brines is challenging due to the similar ionic radii of Mg²⁺ (72 pm) and Li⁺ (76 pm) (SI Appendix, Table S2). To evaluate separation efficiency, binary solutions with Mg²⁺: Li⁺ molar ratios of 2:1, 5:1, 10:1, and 20:1 (with [Li⁺] = 5 mmol L⁻¹) were tested. As shown in Fig. 3B, charge-discharge plateaus shorten in binary solutions compared to single LiCl or simulated brine, attributed to the extremely low Li⁺ concentration (5 mmol/g). The Li⁺/Mg²⁺ selectivity coefficients for LMO-MCS-4 (Fig. 3E) were determined to be 281.47, 557.17, 1092.35, and reached a maximum of 1514.35 (achieved under the most competitive conditions of Mg²⁺/Li⁺=20:1 at a current density of 0.1A g⁻¹), with corresponding Li⁺ extraction capacity of 2.99,2.45, 2.39, and 2.17 mmol/g, respectively (SI Appendix, Table S5). Although Li⁺ extraction decreases with increasing Mg²⁺ concentration, the separation coefficients rise significantly, reflecting minimal Mg²⁺ intercalation and exceptional Li⁺ selectivity of LMO-MCS-4. Fig. 3F shows that LMO-MCS-1 exhibits the same trend. The observed selectivity is attributed to the significantly higher dehydration energy barrier of Mg²⁺ (1,921 kJ/mol) compared to Li⁺ (519 kJ/mol) (Fig. 3G and SI Appendix, Table S3), which kinetically hinders Mg²⁺ intercalation despite their similar ionic radii. Additionally, the higher charge density of Mg²⁺ leads to stronger polarization and elevated migration energy barriers within the spinel structure.

The Li⁺ recovery potential of LMO-MCS-4 was further evaluated in the simulated Chilean Salar de Atacama brine (SI Appendix, Table S1). Fig. 3I shows that the selectivity coefficients of LMO-MCS-4 for Li⁺ are 106.30 ${\alpha }_{{\mathrm{Ca}}^{2+}}^{{\mathrm{Li}}^{+}}$, 128.98 ${\alpha }_{{\mathrm{K}}^{2+}}^{{\mathrm{Li}}^{+}}$, 107.50 ${\alpha }_{{\mathrm{Mg}}^{2+}}^{{\mathrm{Li}}^{+}}$, and 140.5 ${\alpha }_{{\mathrm{Na}}^{2+}}^{{\mathrm{Li}}^{+}}$, all superior to LMO-MCS-1. Brine deintercalation cycling (Fig. 3 H and J) revealed consistent Li⁺ enrichment in the receiving solution (~0.4 mmol/L per cycle, ~3 mmol/g capacity), with negligible extraction of Na⁺, K⁺, Mg²⁺, and Ca²⁺. This high Li⁺ selectivity under galvanostatic conditions in the three-electrode system aligns with electrochemical observations, which showed exclusive Li⁺ redox peaks and plateaus.

The selectivity for Li⁺ during electrochemical intercalation arises from two primary mechanisms (Fig. 3G). First, ion dehydration and charge transfer occur at the electrical double layer (EDL), which is essential for distinguishing Li⁺ from other ions (39). The marked differences in hydration enthalpy between Li⁺ and ions like Mg²⁺ and Ca²⁺ significantly contribute to this selectivity. Second, the migration of ions into the crystal structure and their occupation of interstitial sites is facilitated by their size (11). The smaller ionic radius of Li⁺ allows for more effective integration into the spinel’s 3D channel system, thus enhancing the efficiency of intercalation. On the other hand, larger ions such as Na⁺, K⁺, and Ca²⁺ encounter considerable steric hindrance (40), obstructing their insertion into the lattice. Additionally, the engineered imprinting vacancies within the spinel manganese oxide framework likely provide specific coordination sites and favorable electronic structures that preferentially interact with Li⁺, thereby increasing its extraction capacity and selectivity. Moreover, LMO-MCS-4 consistently outperformed LMO-MCS-1 and other LMO-MCS variants (LMO-MCS-2, -3, -5) in Li⁺ extraction and separation under identical conditions (SI Appendix, Figs. S11–S13 and Table S5). These findings highlight that the micelle-assisted self-assembled LMO multilayer core–shell nanostructure significantly enhances both Li⁺ selectivity and extraction capacity, demonstrating more favorable Li⁺ transport and extraction performance.

Continuous Lithium Recovery in the Constructed HCDI System.

The present study comparatively assessed the lithium ion extraction performance of the LMO-MCS-4 electrode in a constructed HCDI system, with the aim of further enhancing the efficiency and applicability of lithium extraction. The results demonstrated that LMO-MCS-4 exhibited superior performance at 1.2 V in 10 mM LiCl, achieving a high Li⁺ extraction capacity of 3.14 mmol g⁻¹ and a maximum extraction rate of 0.14 mmol g⁻¹ min⁻¹. This was a significant improvement on the performance of other LMO-MCS electrodes (Fig. 4A). The Ragone plot (Fig. 4B) further confirmed the highest lithium ion extraction capacity and rate for LMO-MCS-4, establishing its efficiency for lithium capture.

Fig. 4.

Continuous lithium recovery performance of LMO-MCS-4 in the HCDI system. (A and B) Li⁺ extraction capacity of LMO-MCS-n electrodes in 10 mM LiCl at 1.2 V and the corresponding Ragone plot. (C and D) Li⁺ extraction capacity of LMO-MCS-4 in 10 to 70 mM LiCl at 1.2 V and the corresponding Ragone plot. (E and F) Li⁺ extraction capacity of the LMO-MCS-4 in 10 mM LiCl at 0.8 to 1.2 V and the corresponding Ragone plot. (G) Schematic of the HCDI device for lithium recovery. Components: 1. Acrylic Plate, 2. Silicone Gasket, 3. Graphite Plate, 4. Acrylic Spacer. (H) Separation performance of LMO-MCS-4 in Mg²⁺/Li⁺ binary solutions with molar ratios from 2:1 to 20:1. (I) Selectivity coefficients of LMO-MCS-4 for various ions in simulated salt lake brine. (J) Cycling performance of LMO-MCS-4 in HCDI at 1.2 V with 96% capacity retention over 100 cycles (Inset: concentration-time profiles). (K) Comparison of Li⁺ extraction capacity, rate, and stability from this work with previous reports.

Fig. 4 C and E systematically examined the impact of feed concentration and applied voltage on the Li⁺ extraction performance of LMO-MCS-4. As the concentration of LiCl was increased from 10 to 70 mM, a significant enhancement in both lithium ion extraction capacity (up to 4.78 mmol g⁻¹) and rate (up to 0.16 mmol g⁻¹ min⁻¹) was observed (Fig. 4C). However, when LiCl concentration exceeds 50 mM, capacity growth follows an asymptotic trajectory, primarily dictated by the spinel’s stoichiometric ceiling as crystallographic vacancies are depleted (5) Concurrently, elevated ionic strength compromises EDL permselectivity at the counter electrode, triggering parasitic co-ion fluxes that attenuate charge efficiency (41). A similar trend appeared in the Ragone plot (Fig. 4D), showing performance improved with increasing concentration. As Fig. 4E demonstrates, raising the applied voltage from 0.8 to 1.2 V in 10 mM LiCl substantially enhanced both the lithium ion extraction capacity (2.09 to 3.14 mmol g⁻¹) and rate (0.10 to 0.14 mmol g⁻¹ min⁻¹). The Ragone plot (Fig. 4F) shifted toward higher power with increasing voltage, indicating that stronger electric fields favor lithium ion extraction.

Fig. 4H presents the Li⁺ selectivity coefficients (${\alpha }_{{\mathrm{M}\mathrm{g}}^{2+}}^{{\mathrm{L}\mathrm{i}}^{+}}$) for LMO-MCS-4, showing values of 2.5, 5, 20, and 33 at corresponding Mg²⁺: Li⁺ molar ratios. These yielded Li⁺ extraction capacities of 1.62, 0.95, 0.62, and 0.4 mmol/g, respectively. While demonstrating robust separation, HCDI selectivity is slightly attenuated by electrochemical coupling effects inherent to the two-electrode design. Unlike the strictly potential-clamped three-electrode setup, the HCDI cell operates under a floating interfacial potential governed by dynamic capacitive matching with the carbon counter electrode. Furthermore, the intensified electric field from the narrow interelectrode spacing (0.5 mm) induces overpolarization, triggering parasitic nonspecific electrosorption of competing cations within the electric double layer that dilutes the intrinsic Faradaic selectivity (42, 43). In simulated brine (Fig. 4I), LMO-MCS-4 displayed Li⁺ selectivity coefficients of 66.64 $({\alpha }_{{\mathrm{C}\mathrm{a}}^{2+}}^{{\mathrm{L}\mathrm{i}}^{+}})$, 69.25 $({\alpha }_{{\mathrm{K}}^{2+}}^{{\mathrm{L}\mathrm{i}}^{+}})$, 44.65 $({\alpha }_{{\mathrm{M}\mathrm{g}}^{2+}}^{{\mathrm{L}\mathrm{i}}^{+}})$, and 57.31 (${\alpha }_{{\mathrm{N}\mathrm{a}}^{2+}}^{{\mathrm{L}\mathrm{i}}^{+}})$, demonstrating its potential for effective Li⁺ separation in complex matrices.

Cycling stability plays a critical role in the practical application of materials, directly influencing their long-term performance and reliability. The robustness of LMO-MCS-4 was further validated under practical HCDI operating conditions. Remarkably, LMO-MCS-4 retains 96% of its initial capacity after 100 cycles (Fig. 4J), significantly surpassing the 70% retention of the disordered LMO-MCS-1 (SI Appendix, Fig. S15). While most reported LMO-based counterparts suffer from capacity fading—typically retaining less than 90% of their initial capacity after 50 to 100 cycles (SI Appendix, Table S6)—LMO-MCS-4 achieves a breakthrough in durability. As illustrated in Fig. 4K, it stands as a benchmark with both superior extraction capacity (4.78 mmol g⁻¹) and cycling stability. This exceptional performance is attributed to the optimized multilayer core–shell structure, which facilitates continuous and uniform ion diffusion pathways and contributes to outstanding cycling stability. This enhanced stability likely stems from the structural integrity maintained during repeated cycling, ensuring consistent performance over time.

Stress Homogenization Mechanisms for Sustainable Lithium Extraction.

Density Functional Theory (DFT) calculations were performed to evaluate the binding energies of various ions within the LMO-MCS-4 structure (Fig. 5A). The calculated binding energies were listed in increasing order as Li⁺(−5.70 eV), Mg²⁺(−5.38 eV), Na⁺(−3.64 eV), Ca²⁺(−1.09 eV), and K⁺(−0.82 eV). This trend demonstrates the enhanced Li⁺ selectivity of the LMO-MCS material. Electrochemical impedance spectroscopy (EIS) of LMO materials (0.1 to 10⁵ Hz, Fig. 5 B and C) reveals that LMO-MCS-4 exhibits lower charge transfer resistance (Rct) and Li⁺ diffusion resistance (R_Li+) compared to other LMO materials, indicating enhanced charge transfer kinetics and Li⁺ diffusion rates. Quantitative analysis of Li⁺ chemical diffusion coefficients (D_Li⁺) was performed by fitting impedance data to ω^-1/2 and Z’ using the equation:

$$D=\frac{R^2{T}^2}{2n^4{F}^4{A}^2{\sigma }^2{c}_0^2},$$

Fig. 5.

Mechanism of enhanced lithium extraction performance and cycling stability in multilayer core–shell structures of LMO-MCS-4. (A) Comparison of binding energies and intercalation schematics for competing coexisting ions. (B) EIS results for LMO-MCS materials. (C) Relationship between Z’ and ω−1/2 corresponding to the EIS curves. (D) Evolution of Li⁺ concentration profiles within LMO-MCS-1 over time. (E) Evolution of Li⁺ concentration profiles within LMO-MCS-4 over time. (F and G) Absolute von Mises stress contour maps corresponding to the concentration distribution. (H) Mechanistic comparison of the structural advantages and cycle stability of LMO-MCS-1 and LMO-MCS-4.

where R is the gas constant (8.314 J mol⁻¹ K⁻¹), T is the absolute temperature (298.15 K), A is the effective surface area of the electrode (1 cm²), n is the number of electrons transferred in the electrode reaction (n = 1 for Li⁺ intercalation/deintercalation), F is the Faraday constant (96,485 C mol⁻¹), C is the lithium-ion concentration in the electrode material (C_Li⁺ = 0.022 mol cm⁻³), and σ is the slope of the Z’ vs. ω^−1/2 line obtained from Fig. 5C. Detailed calculations are provided in the SI. The calculated diffusion coefficient for the multilayered core–shell spheres of LMO-MCS-4 (D_Li+ = 7.15 × 10⁻¹⁴ cm² s⁻¹) is higher than that of the disordered spheres of LMO-MCS-1 (D_Li+ = 3.28 × 10⁻¹⁴ cm² s⁻¹), highlighting the significant positive impact of the multilayered core–shell structure on facilitating Li⁺ diffusion.

To systematically investigate the influence of material microstructure on ion diffusion and the resulting mechanical responses, we constructed models of the multilayer core–shell structure (LMO-MCS-4) and the disordered dense structure (LMO-MCS-1) using COMSOL Multiphysics. Detailed modeling parameters and morphological settings are provided in SI Appendix. Fig. 5 D and E illustrate the Li⁺ distribution evolution within both structures at different times (t = 0 s, 5 s, and 10 s). Initially (t = 0 s), both models are in a steady state before diffusion, with the electrolyte providing the same boundary concentration. As time progresses (at t = 5 s and 10 s), significant differences in diffusion dynamics emerge. In the disordered dense structure, Li⁺ accumulates near the surface and diffuses slowly into the interior, resulting in a pronounced concentration gradient. By contrast, the core–shell structure exhibits a more uniform and continuous concentration distribution, indicating superior diffusion efficiency and structural connectivity.

The inhomogeneous concentration further induces differential stress fields (Fig. 5 F and G). In concentrated regions (e.g., the surface layer of LMO-MCS-1), Li⁺ intercalation causes volume expansion, resulting in pronounced local absolute stress concentrations, consistent with the diffusion-induced stress mechanism (44). In sharp contrast, the multilayer core–shell architecture facilitates a more uniform Li⁺ concentration distribution, transforming the sharp stress peaks into a homogenized stress field. Consequently, the maximum von Mises stress is dampened by 48%, while the stress differential is drastically narrowed (SI Appendix, Fig. S18), highlighting the effectiveness of the multilayer structure in mitigating mechanical stress through ionic homogenization. Additionally, the layered structure provides buffer space for the anisotropic volumetric changes that occur during lithium extraction. Overall, the multilayer core–shell nanosphere structure minimizes stress concentration and ensures uniform stress/strain distribution during extensive lithiation/delithiation cycles. This is crucial for achieving superior structural integrity and long-term cycling stability. These results are in strong agreement with the enhanced extraction capacity and structural stability that have been observed in experimental studies.

As conceptually visualized in Fig. 5H, this study establishes a synergistic “structure-interface” modulation paradigm that effectively decouples the intrinsic trade-off between high adsorption capacity and long-term cycling stability. The multilayer core–shell geometry achieves spatiotemporal stress homogenization, dissipating lattice distortion energy to suppress crack formation and ensure intrinsic mechanical robustness. Complementing this, the hierarchical structure eliminates kinetic “dead zones” and creates open channels for the dual acceleration of mass transport, fully maximizing active site utilization. This combined strategy of stress alleviation and kinetic boosting establishes a rational blueprint for engineering high-performance electrochemical extraction materials.

The electrochemical resilience and structural fidelity of the engineered materials were rigorously evaluated under a prolonged 400-cycle (~400 h) regime (Fig. 6A). LMO-MCS-4 exhibited exceptional durability, sustaining ~80% capacity, whereas the pristine LMO-MCS-1 suffered a rapid decay to 39%. This performance divergence is corroborated by electrolyte analysis (Fig. 6B), which revealed a ~69% suppression in manganese dissolution for LMO-MCS-4 (3.62 µg/L) compared to LMO-MCS-1 (11.77 µg/L), confirming the efficacy of the modification strategy in mitigating active material loss. Postmortem characterizations elucidate the underlying failure mechanisms. SEM imaging (Fig. 6C) unveils catastrophic pulverization and structural collapse in LMO-MCS-1—symptomatic of volume breathing and localized stress accumulation—while LMO-MCS-4 preserved a pristine, crack-free spherical morphology (Fig. 6D). XRD analysis (Fig. 6E) further identifies parasitic phase evolution in LMO-MCS-1, evidenced by the emergence of MnOOH (−111) and Mn₂O₃ (222/440) impurity peaks, signifying proton-induced phase transition and surface hydroxylation. In contrast, LMO-MCS-4 retained a highly crystalline spinel framework. Fig. 6 F–H substantiates this degradation pathway. LMO-MCS-1 displayed intensified signals for oxygenated defects (C-O/C = O), substantial lattice oxygen loss, and an elevated Mn⁴⁺/Mn³⁺ ratio (1.3), indicative of accelerated Mn³⁺ disproportionation. Conversely, LMO-MCS-4 maintained a stable valence ratio (1.1) and robust lattice oxygen integrity, confirming the effective suppression of parasitic phase transitions and surface corrosion. Conclusively, the entropy-driven multilayer core–shell architecture establishes a stress-homogenized system that effectively arrests mechanical degradation and chemical corrosion, providing a robust foundation for durable lithium extraction from brine.

Fig. 6.

Performance and characterization of LMO-MCS-1 and LMO-MCS-4 after 400 cycles in the HCDI system at 1.2 V. (A) Capacity retention profile during cycling. (B) Comparison of dissolved Mn ion concentration. SEM images of LMO-MCS-1 (C) and LMO-MCS-4 (D). (E) XRD patterns. High-resolution XPS spectra of (F) C1s, (G) O1s, and (H) Mn 2p.

Industrial Scalability and Future Perspectives.

Electrochemical lithium extraction technologies remain predominantly in the nascent stages of research and development (R&D), with large-scale commercialization representing a shared frontier for concerted academic and industrial efforts. While the LMO-MCS-4 electrode demonstrates exceptional intrinsic properties, its industrial translation necessitates addressing critical scale-up challenges spanning manufacturing, operation, and sustainability. Regarding synthesis, the nonlinear amplification of fluid dynamics in large-scale reactors poses a risk to batch consistency (45); future upscaling must therefore leverage Computational Fluid Dynamics (CFD) to ensure the uniform shear environments required for entropy-driven self-assembly (46). Equally critical is thermal regulation across the lifecycle: Precise temperature control is imperative to direct entropy-driven assembly during synthesis, while active thermal management systems are requisite in industrial facilities to maintain the optimal 15 to 25 °C extraction window against the severe ambient fluctuations of high-altitude brines. Beyond these physical controls, the operational challenge extends to system-level resilience against interfacial fouling by brine impurities (e.g., Ca²⁺, Mg²⁺, organics), demanding a systems engineering approach that integrates advanced pretreatment (47). Economically, despite higher initial capital expenditures compared to solar evaporation (48), the technology offers competitive lifecycle costs driven by inexpensive precursors and reduced operational expenditures, particularly when synergized with geothermal energy (49). Ultimately, the transition to a truly green industry will hinge on establishing closed-loop recycling protocols to enable the circular regeneration of these robust electrode materials (50).

Discussion

In this work, we have successfully developed a strategy employing entropy-driven supramolecular self-assembly to engineer a stress-homogenized multilayer core–shell LiMn₂O₄ structure for robust electrochemical lithium extraction. The ordered structure promotes efficient Li⁺ transport and ensures uniform concentration distribution, effectively mitigating high stress concentrations and alleviating diffusion-induced stress. This design significantly enhances both the lithium extraction efficiency and cycling stability of the material in electrochemical lithium extraction from brines. In the three-electrode deintercalation setup, the optimized LMO-MCS-4 exhibits a Li⁺ extraction capacity of 4.05 mmol g⁻¹, along with a high Mg²⁺:Li⁺ separation factor of 1,514.35 in a Mg²⁺:Li⁺ = 20:1 binary system. Furthermore, in simulated Salar de Atacama brine, the selectivity coefficients for Ca²⁺, K⁺, and Na⁺ all exceed 100, indicating strong anti-interference capability in complex ionic environments. In the continuous-flow HCDI system, LMO-MCS-4 exhibited a maximum Li⁺ extraction capacity of 4.78 mmol g⁻¹ and a remarkable capacity retention of 96% after 100 cycles and sustaining 85% after 400 cycles, representing the best cycling stability reported to date. DFT calculations reveal that the superior Li⁺ selectivity results from a lower binding energy relative to competing ions. COMSOL simulations demonstrate that the multilayer core–shell architecture significantly enhances Li⁺ concentration uniformity and reduces maximum von Mises stress by 48% compared to disordered structures, validating a synergistic mechanism of ionic homogenization–induced stress mitigation. Comparative postcycling characterization after 400 cycles confirms that the superior stability of LMO-MCS-4 is attributed to its robust structural integrity and enhanced interfacial chemical stability maintained throughout the long-term operation. This work offers a unique perspective for designing materials crucial to lithium extraction from salt lakes.

Materials and Methods

Material Synthesis and Electrode Fabrication.

The LMO-MCS-n samples (n = 1-5) were synthesized by adding a MnSO₄ solution (1.081 g in 100 mL water) dropwise into an NH₄HCO₃/CTAB mixed solution at 30 °C (600 rpm). The receiving mixture contained 4.744 g NH₄HCO₃ (establishing a 10:1 molar ratio to Mn) and varying CTAB amounts (0%, 2%, 5%, 7%, 10% relative to MnSO₄·H₂O) in 100 mL of deionized water. After aging (24 h) and drying (60 °C), the MnCO₃ precipitate was calcined at 600 °C for 8 h to yield Mn₂O₃. Finally, the precursor was lithiated with LiOH·H₂O (1:1.06 molar ratio) at 700 °C for 10 h to synthesize LMO-MCS-n. The indices n = 1-5 correspond to the ascending CTAB percentages. For electrochemical evaluation, electrodes were fabricated by coating a homogeneous slurry of the active material, Super P Li, and PVDF (8:1:1 mass ratio) onto graphite substrates. Specifically, drop-casting was employed for the three-electrode system, while a film applicator was used to prepare electrodes for the HCDI system. Detailed chemical dosages and specific fabrication procedures for both electrode types are provided in SI Appendix.

Material Characterization and Electrochemical Measurements.

The morphology, crystal structure, and surface properties of the samples were systematically characterized using SEM, HRTEM, XRD, XPS, and BET analyses. Electrochemical performance was evaluated via CV, GCD, and EIS in a three-electrode configuration employing a Pt counter electrode and Ag/AgCl reference electrode. Specific instrument models and detailed operating parameters are provided in SI Appendix.

Li+ Extraction Experiments in Electrochemical (De)intercalation Setup and HCDI System.

Three-electrode setup for lithium recovery assessment.

A three-electrode system was established to investigate the electrochemical lithium deintercalation properties of LMO-MCS-n. Working electrodes were prepared by coating delithiated LMO-MCS-n onto 2 × 2 cm² graphite sheets, with a Pt foil counter electrode and an Ag/AgCl reference electrode. Lithium recovery experiments involved capture and release steps at a constant temperature (20 to 25 °C) to ensure consistent kinetic baselines. For the capture step, lithium extraction was conducted at 100 mA g⁻¹ within 0.4 to 1.1 V in 60 mL of feed solution under three conditions: 1) 10 to 70 mM LiCl to evaluate Li⁺ capture capacity; 2) binary Mg²⁺: Li⁺ mixtures (2:1 to 20:1, with [Li⁺] fixed at 5 mM) to assess Li⁺ selectivity; and 3) simulated Salar de Atacama brine (SI Appendix, Table S1) to assess practical performance. Following extraction, electrodes were rinsed with ultrapure water. For the release step, the LMO-MCS-n electrode served as the cathode in 60 mL of 10 mM NaCl receiving solution, with Li⁺ released at 100 mA g⁻¹ over the same voltage range. Ion concentrations were analyzed by ICP-OES (ICP 5800, Thermo Scientific) before and after each capture–release cycle.

HCDI system for continuous lithium recovery.

A custom flow-through HCDI device was constructed, based on the desalination system developed in our previous work (34), to assess continuous lithium recovery under practical conditions. The system employed LMO-MCS-n as the anode and activated carbon (AC) as the cathode, separated by a 1 mm silicone gasket to form a flow channel. A constant voltage (0.8 to 1.2 V) was applied using a potentiostat (IT-E122, ITECH), while 50 mL of feed solution was circulated at 10 mL min⁻¹. The same LiCl, Mg²⁺: Li⁺ binary mixtures, and synthetic brine solutions were used to enable direct comparison with the three-electrode setup. ICP-OES quantified ion concentrations at fixed minute intervals throughout the recovery process. Detailed calculations for Li+ intercalation capacity, separation coefficient, and extraction efficiency are provided in SI Appendix.

Theoretical Calculations.

DFT calculations were performed to analyze the intercalation thermodynamics of competing ions in LMO-MCS-4. Comparative simulations of Li+ diffusion kinetics in LMO-MCS-1 and LMO-MCS-4 were conducted using the finite element method. Detailed experimental procedures are documented in SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

Study data are included in the article and/or SI Appendix.