Pulsed Charging‐Induced Interfacial Stabilization Enables Highly Stable Iron Phosphate Electrode for Ultrafast Electrochemical Lithium Extraction

ABSTRACT

Amid the soaring global demand for lithium, electrochemical lithium extraction using iron phosphate (FePO₄) electrodes is promising, yet its practical application in salt lake brines is severely constrained by high Mg²⁺/Li⁺ ratios. Herein, we unravel the intrinsic degradation mechanism by demonstrating that Mg²⁺ intercalation weakens the Fe─O bonds in FePO₄, triggering Fe²⁺ dissolution, and inducing the formation of Fe₂O₃ surface species that deteriorate electrode performance. To tackle this critical challenge, we propose a novel pulsed charging protocol where reverse potentials not only efficiently flush out surface‐intercalated Mg²⁺ but also enrich Cl⁻ on the electrode surface, thereby modulating the electronic structure and lowering the Mg²⁺ deintercalation barrier. The FePO₄||Ag cell integrated with this protocol exhibits superior Li⁺ extraction kinetics (∼20 mg g⁻¹ h⁻¹), high LiCl product purity (99.95%), and exceptional cycling stability (800 h operation with ∼80% performance retention) in real salt lake brines. This work establishes a theoretical basis for electrode protection and provides a robust strategy to advance high‐efficiency electrochemical lithium extraction.

FePO₄ is prone to degrade when working in salt lake brines with high Mg²⁺/Li⁺ ratios for electrochemical lithium extraction. This work unravels the mechanism for Mg²⁺‐induced FePO₄ degradation. A novel pulsed‐charging protocol is implemented to induce surface Cl⁻ enrichment, which effectively modulates the local electronic structure and lowers the Mg²⁺ deintercalation barriers, enabling highly stable and ultrafast lithium extraction.

1. Introduction

With the booming development of energy storage technologies, the global demand for lithium (Li) has surged exponentially, rendering the efficient extraction of Li resources a critical priority in both academia and industry [1, 2]. Lithium extraction from brines has become a core strategy to secure Li supply, given that brines account for ∼64% of global Li reserves, far exceeding the reserves of Li‐bearing ores [3]. Among various lithium extraction techniques, electrochemical lithium extraction using Li⁺‐selective electrodes has garnered considerable attention, thanks to its inherent advantages of high techno‐economic efficiency and environmental friendliness [4]. Similar to aqueous lithium‐ion batteries, electrochemical lithium extraction relies on ion intercalation/deintercalation reactions, where the selection of suitable host materials is a key factor determining Li⁺ extraction performance. Numerous recent studies have identified olivine iron phosphate (FePO₄) as a highly promising host material, primarily due to its suitable working potential and intrinsic thermodynamic preference for Li⁺ intercalation [5]. However, the Li⁺ extraction rate of FePO₄ electrodes is rather poor compared to other Li⁺ intercalation materials (e.g., LiMn₂O₄ with three‐dimensional ion pathways) due to its one‐dimensional ion diffusion channels [6]. This is made even worse when dealing with real salt lake brines containing low Li⁺ concentrations and high concentrations of competitive ions (e.g., Mg²⁺). Although increasing the applied voltage can enhance Li⁺ intercalation kinetics and thus improve the Li⁺ extraction rate to some extent, it inevitably induces Mg²⁺ intercalation. Owing to a higher diffusion barrier for Mg²⁺ deintercalation than that for intercalation (1.03 eV vs. 0.58 eV), Mg²⁺ intercalation into FePO₄ seems irreversible, which potentially blocks Li⁺ transport channels and degrades the Li⁺ extraction rate [7]. Furthermore, Mg²⁺‐intercalated FePO₄ may undergo structural degradation, compromising the electrode stability, which critically hinders the practical application of FePO₄ for electrochemical lithium extraction [7, 8, 9]. Although extensive efforts have been made to mitigate this challenge (e.g., tailoring crystal structures and electrode/electrolyte interfaces) [10, 11], the intrinsic degradation mechanisms of FePO₄ electrodes in brines with high Mg²⁺/Li⁺ ratios remain elusive. Furthermore, developing an effective strategy capable of breaking through the inherent trade‐off between Li⁺ extraction rate and cycling stability is highly imperative.

In this work, a comprehensive analysis of the Mg²⁺ intercalation chemistry of olivine FePO₄ in aqueous electrolytes was performed for the first time. Leveraging the distinct electrochemical behaviors of Li⁺ and Mg²⁺, a pulsed charging protocol was designed and developed. As illustrated in Scheme 1, the brief reversed potential pulse flushes Mg²⁺ out of the FePO₄ and enriches Cl⁻ on the electrode surface to modulate the electronic structure and lower the Mg²⁺ deintercalation barrier, thus achieving efficient Mg²⁺ deintercalation. This multifaceted mechanism effectively inhibits Mg²⁺‐induced interfacial degradation and stabilizes the FePO₄ electrode/electrolyte interface. As a result, the FePO₄||Ag electrochemical lithium extraction cell operated under the pulsed charging protocol exhibits superior lithium extraction kinetics and cycling stability in both simulated and real salt lake brines. Overall, this work uncovers the intrinsic degradation mechanism of FePO₄ in brines with high Mg²⁺/Li⁺ ratios and provides a theoretical foundation for electrode protection strategies, thereby enabling long‐term operation of electrochemical lithium extraction systems.

SCHEME 1.

Mechanism of the pulsed charging protocol proposed in this work.

2. Results and Discussion

2.1. Mg²⁺ Intercalation Chemistry and Induced Degradation of FePO₄

Li⁺ and Mg²⁺ exhibit distinct intercalation behaviors in FePO₄ owing to their inherent differences in ionic radius, charge density, and dehydration energy [12]. As shown in Figure 1a, linear sweep voltammetry (LSV) curves of FePO₄ obtained in 1 M LiCl (blue line) and 1 M MgCl₂ (red line) aqueous electrolytes clearly demonstrate that the cathodic peak potential of Mg²⁺ (−0.50 V vs. Ag/AgCl) is markedly lower than that of Li⁺ (0.14 V vs. Ag/AgCl). While reducing the intercalation potential to accelerate Li⁺ intercalation kinetics is desirable for efficient lithium extraction, it inevitably induces the co‐intercalation of Mg²⁺ into the FePO₄ lattice, which poses a critical challenge for FePO₄‐based lithium extraction systems operating in brines with high Mg²⁺/Li⁺ ratios.

FIGURE 1.

(a) Linear sweep voltammetry (LSV) curves of FePO₄ obtained in 1 M LiCl (blue line) and 1 M MgCl₂ (red line) aqueous electrolytes at a scan rate of 0.03 mV s⁻¹. (b) Chronopotentiometry (CP) curves of various samples obtained in 1 M LiCl aqueous electrolyte at 1 C, 5 C, and 10 C (1 C = 147 mA g⁻¹). (c) Specific capacities and half‐capacity voltage differences (V _half) of various samples, derived from (b). (d) Nyquist plots, (e) distribution of relaxation times (DRT) plots, and (f₁–f₄) local electrochemical impedance mapping (LEIM) results of pristine FePO₄ during 24 h soaking in 1 M LiCl aqueous electrolyte. (g₁) Atomic force microscopy (AFM), (g₂) conductive AFM (C‐AFM), (g₃) scanning electron microscopy (SEM), and (g₄) transmission electron microscopy (TEM) images of pristine FePO₄ after 24 h soaking in 1 M LiCl aqueous electrolyte. (h) Nyquist plots, (i) DRT plots, and (j₁–j₄) LEIM results of Mg²⁺‐intercalated FePO₄ during 24 h soaking in 1 M LiCl aqueous electrolyte. (k₁) AFM, (k₂) C‐AFM, (k₃) SEM, and (k₄) TEM images of Mg²⁺‐intercalated FePO₄ after 24 h soaking in 1 M LiCl aqueous electrolyte. (l) Grazing incidence x‐ray diffraction (GIXRD) patterns and (m) Raman spectra of various samples. (n) X‐ray photoelectron spectroscopy (XPS) depth profiling results of Mg²⁺‐intercalated FePO₄ after 24 h soaking in 1 M LiCl aqueous electrolyte. Scale bars: LEIM (1 mm × 1 mm scan area), AFM and C‐AFM (400 nm), SEM (100 nm), TEM (5 nm). Red and blue contours represent high and low intensity, respectively.

To elucidate the degradation mechanism of FePO₄ induced by Mg²⁺ intercalation, a series of control experiments were conducted: the pristine FePO₄ electrode (Pristine) was first held at −0.6 V versus Ag/AgCl in 1 M MgCl₂ aqueous electrolyte for 1 h to allow the intercalation of Mg²⁺ (Mg²⁺‐intercalated electrode, denoted as 0 h), which was then soaked in 1 M LiCl aqueous electrolyte for 24 h (denoted as 24 h). Chronopotentiometry (CP) tests at 1 C (147 mA g⁻¹), 5 C, and 10 C (Figure 1b,c) reveal that the 0 and 24 h samples delivered drastically reduced capacities (from 125 mAh g⁻¹ for Pristine to 61 and 85 mAh g⁻¹ for 0 and 24 h at 1 C) and enlarged half‐capacity voltage differences (V _half, from 112 mV to 427 and 599 mV for 0 and 24 h at 1 C). These results demonstrate that Mg²⁺ intercalation significantly degrades the Li⁺ intercalation kinetics of FePO₄ electrodes.

To unravel the origin of this degradation, in situ electrochemical impedance spectroscopy (in situ EIS) and distribution of relaxation times (DRT) analysis were performed on the pristine FePO₄ electrode in 1 M LiCl electrolyte. No significant changes in the Warburg coefficient (C _w, a parameter reflecting Li⁺ diffusion kinetics [13], calculated from the Nyquist plots in Figure 1d) or charge transfer impedance were observed over 24 h of soaking. The peak highlighted by the white box in Figure 1e, derived from DRT analysis of the in situ EIS data with a time constant of approximately 0.03 s, corresponds to the resistance associated with the charge transfer process [14]. Local electrochemical impedance mapping (LEIM, Figure 1f₁–f₄ , measurement details and reliability verification are provided in the Supporting Information) further confirmed negligible variations in electrode conductivity over the soaking period. Complementary conductive atomic force microscopy (C‐AFM) measurements (Figure 1g₁–g₂ ) identified high‐conductivity regions corresponding to conductive carbon black, which act as electron transport bridges connecting low‐conductivity FePO₄ particles. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images of the electrode after 24 h of soaking (Figure 1g₃–g₄ ) revealed well‐defined and sharp boundaries of FePO₄ particles after 24 h of soaking, highly consistent with the pristine state (Figure S1 in the Supporting Information).

In sharp contrast, the Mg²⁺‐intercalated electrode exhibited drastic impedance evolution upon immersion in 1 M LiCl electrolyte (Figure 1h). The C _w gradually decreased, indicating accelerated Li⁺ diffusion kinetics [15]. However, DRT analysis (Figure 1i) uncovered a significant increase in charge transfer impedance with prolonged soaking time. LEIM measurements (Figure 1j₁–j₄ ) demonstrated a gradual reduction in electrode conductivity during soaking. The AFM image (Figure 1k₁ ) revealed no significant change in the surface roughness of the Mg²⁺‐intercalated electrode after 24 h. However, according to the current mapping measured by C‐AFM (Figure 1k₂ ), the high‐conductivity regions corresponding to conductive carbon black were no longer observable, indicating that the electronic conductive network was covered or encapsulated by the surface reaction products. Further morphological characterization via SEM and TEM (Figure 1k₃–k₄ ) clearly visualized the growth of these flake‐like species on the surfaces of FePO₄ particles after 24 h of soaking.

To determine the chemical composition of these flake‐like species formed on the Mg²⁺‐intercalated FePO₄ electrode following 24 h of soaking in 1 M LiCl electrolyte, complementary characterization techniques were employed. Grazing incidence x‐ray diffraction measurements (GIXRD, Figure 1l) revealed distinct diffraction peaks corresponding to the (110) plane of α‐Fe₂O₃ in the soaked Mg²⁺‐intercalated sample [16]. Notably, these peaks were absent in both the pristine FePO₄ electrode and the unsoaked Mg²⁺‐intercalated electrode, confirming that Fe₂O₃ formation was triggered by both Mg²⁺ intercalation and subsequent LiCl electrolyte immersion. To corroborate the GIXRD identification, Raman spectroscopy was conducted (Figure 1m), with reference spectra of Fe₂O₃ and LiFePO₄ provided for comparison in Figure S2, further validating the presence of the Fe₂O₃ phase on the soaked sample. To further investigate the spatial distribution of the Fe₂O₃ layer relative to the FePO₄ bulk, x‐ray photoelectron spectroscopy (XPS) depth profiling was performed (Figure 1n). Results showed that phosphorus (P) signals attributed to FePO₄ were nearly undetectable on the surface of the Mg²⁺‐intercalated electrode after 24 h of soaking but gradually intensified with increasing etching depth. Collectively, these characterizations confirm the formation of Fe₂O₃ on the surface of the Mg²⁺‐intercalated electrode after soaking.

Theoretical calculations were conducted to further explain the underlying mechanism of the observed surface evolution. To this end, crystal orbital Hamiltonian population (COHP) and integrated COHP (ICOHP) analyses were performed on FePO₄ and Mg_0.5FePO₄ (Figure 2a–c). The results revealed that Mg²⁺ intercalation weakens the Fe─O bond strength in Mg_0.5FePO₄, with the bond length increasing from 2.16 to 2.26 Å. Notably, the negative ICOHP (─ICOHP) value serves as a quantitative descriptor of bond strength, with lower values denoting weaker covalent interactions [17]. Specifically, the Fe─O spin‐up channel exhibited only a marginal reduction from 0.34 to 0.32. In stark contrast, the ─ICOHP value for the Fe─O spin‐down channel underwent a substantial decline, plummeting from 0.79 to 0.57. Since the total bonding interaction of Fe─O bonds arises from the combined contribution of both spin channels, the pronounced loss of covalent bonding in the spin‐down configuration directly undermines the structural stability of Mg_0.5FePO₄ relative to pristine FePO₄. This reduced interatomic interaction within the crystal lattice renders the framework thermodynamically susceptible to solvation‐induced dissolution and subsequent phase decomposition into Fe₂O₃ upon exposure to water molecules.

FIGURE 2.

Crystal orbital Hamilton population (COHP) analysis of (a) FePO₄ and (b) Mg_0.5FePO₄. The red, yellow, orange, and purple spheres represent O, Fe, Mg, and P atoms. (c) Negative ICOHP (─ICOHP) values of FePO₄ and Mg_0.5FePO₄. (d) Open circuit potential (OCP) variations of various samples during 24 h of soaking. (e) Pourbaix diagram for iron at a concentration of 10⁻⁷ M. (f) Gibbs free energy changes of the reactions involved in the degradation mechanism. (g) Schematic illustration of the proposed Mg²⁺‐induced degradation mechanism.

The time‐dependent variations in open‐circuit potential (OCP) of the pristine FePO₄ electrode and the Mg²⁺‐intercalated FePO₄ electrode during 24 h of soaking in 1 M LiCl are displayed in Figure 2d. Consistent with the stability predictions from the COHP calculations, the Mg²⁺‐intercalated FePO₄ electrode exhibits a more rapid OCP decline than the pristine FePO₄ electrode, demonstrating an enhanced propensity for spontaneous decomposition in water [18]. By correlating the OCP profiles with the Pourbaix diagram of iron (Figure 2e) [19], we further elucidate the mechanism underlying the transformation of leached Fe²⁺ into the surface α‐Fe₂O₃ phase. Nernstian analysis indicates that the equilibrium potential for this oxidative conversion is ∼0.314 V versus RHE at pH = 7. Crucially, the experimental OCP (∼0.741 V vs. RHE, corresponding to ∼0.130 V vs. Ag/AgCl at pH = 7) is far more anodic, creating a considerable 427 mV overpotential that drives interfacial Fe²⁺ oxidation. Under open‐circuit conditions, the conductive carbon network acts as a capacitive reservoir, continuously accommodating the generated electrons [20]. This charge injection directly causes the dynamic downward relaxation of the OCP over 24 h while sustaining the continuous interfacial phase transition. Details of the above analysis are provided in the Discussion 5.1 in the Supporting Information.

Herein, we propose the following multistep reaction mechanism for the Mg²⁺‐induced degradation of FePO₄: (I) Mg²⁺ intercalation into FePO₄ generates Mg_0.5FePO₄ (Equation 1); (II) Mg_0.5FePO₄ decomposes into free Mg²⁺, Fe²⁺, and H₂PO₄ ⁻ ions facilitated by water molecules (Equation 2), where H₂PO₄ ⁻ is the dominant phosphate speciation at neutral pH, as validated by the solution equilibrium analysis in Supporting Information: Discussion 5.2; (III) Fe²⁺ ions transform to Fe₂O₃ (Equation 3). Thermodynamic calculations (Figure 2f, details are provided in the Supporting Information) corroborate that while step (I) is thermodynamically unfavorable (Δ_r G _I  = +29.2 kJ mol⁻¹), steps (II) and (III) are spontaneous (Δ_r G _II  = −10.6 kJ mol⁻¹ and Δ_r G _III  = −31.6 kJ mol⁻¹, respectively). Crucially, the H⁺ generated from step (III) dynamically fuels the decomposition in step (II), sustaining the dominance of H₂PO₄ ⁻. As a result, the net reaction (Equation 4) exhibits an overall negative Gibbs free energy change of −13.0 kJ mol⁻¹. This dictates that although step (I) requires an external potential to overcome its initial energy barrier, the entire coupled process is thermodynamically spontaneous, with Mg²⁺ ions serving as pivotal reaction mediators (Figure 2g).

$$\begin{array}{c}\mathrm{FeP}{\mathrm{O}}_4\left(s\right)+0.5\mathrm{M}{\mathrm{g}}^{2+}\left(aq\right)+{\mathrm{e}}^{-}\to \mathrm{M}{\mathrm{g}}_{0.5}\mathrm{FeP}{\mathrm{O}}_4\left(s\right)\end{array}$$ (1)

$$\begin{array}{cc}\hfill \mathrm{M}{\mathrm{g}}_{0.5}\mathrm{FeP}{\mathrm{O}}_4\left(s\right)& +2{\mathrm{H}}^{+}\left(aq\right)\to 0.5\mathrm{M}{\mathrm{g}}^{2+}\left(aq\right)\hfill \\ & +\mathrm{F}{\mathrm{e}}^{2+}\left(aq\right)+{\mathrm{H}}_{2}P{\mathrm{O}}_4^{-}\left(aq\right)\hfill \end{array}$$ (2)

$$\begin{array}{c}\mathrm{F}{\mathrm{e}}^{2+}\left(aq\right)+1.5{\mathrm{H}}_{2}O\left(l\right)\to 0.5\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3\left(s\right)+3{\mathrm{H}}^{+}\left(aq\right)+{\mathrm{e}}^{-}\end{array}$$ (3)

$$\begin{array}{c}\mathrm{FeP}{\mathrm{O}}_4\left(s\right)+1.5{\mathrm{H}}_{2}O\left(l\right)\to 0.5\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3\left(s\right)+{\mathrm{H}}^{+}\left(aq\right)+{\mathrm{H}}_{2}P{\mathrm{O}}_4^{-}\left(aq\right)\end{array}$$ (4)

Additionally, COHP analysis reveals that Li⁺ intercalation into FePO₄ also weakens the Fe─O bonding interaction to some extent. However, thermodynamic evaluations demonstrate a fundamental difference in their aqueous stabilities: Unlike Mg_0.5FePO₄, which undergoes spontaneous interfacial dissolution at low Mg²⁺ concentrations (< 0.2 mmol L⁻¹), the Δ_r G for LiFePO₄ dissolution remains strictly positive across all possible concentration ranges (> 10⁻⁷ mol L⁻¹). This confirms that LiFePO₄ is thermodynamically immune to aqueous decomposition, validating that the structural deterioration is uniquely driven by Mg²⁺ intercalation. Details of the above analysis are provided in the Supporting Information: Discussion 5.3.

In summary, within Mg²⁺‐rich salt lake environments, the irreversible retention of co‐intercalated Mg²⁺ within the FePO₄ framework drives the spontaneous phase decomposition of the transient Mg_0.5FePO₄ intermediate into parasitic, flake‐like α‐Fe₂O₃ aggregates. These insulative species degrade the electrode/electrolyte interface and disrupt the continuous electronic transport pathways between FePO₄ particles and the conductive carbon matrix. The resulting severe electrode polarization fundamentally undermines the Li⁺ extraction kinetics and cycling stability of the lithium extraction system.

2.2. Pulsed Charging Protocol

The aforementioned results illustrate a dilemma: Mg²⁺ intercalation is inevitable at low intercalation potentials when rapid Li⁺ intercalation kinetics are desired. This trade‐off effect has long puzzled the field of electrochemical lithium extraction. It is vital to develop a strategy that enables rapid lithium extraction while minimizing Mg²⁺ co‐intercalation. To this end, it is necessary to analyze the distinct electrochemical behaviors of Li⁺ and Mg²⁺ intercalation in FePO₄. To clarify the kinetic differences, we performed differential voltage analysis (dU/dQ) on data obtained from galvanostatic intercalation experiments in 1 M LiCl and MgCl₂ electrolytes (Figure 3a,b). Here, U denotes the electrode potential (V vs. Ag/AgCl) and Q represents the accumulated specific capacity (mAh g⁻¹) during intercalation. Results reveal that FePO₄ exhibits the electrochemical response of a classic battery‐type material during Li⁺ intercalation, whereas it behaves as a pseudocapacitance‐type material during Mg²⁺ intercalation [21, 22, 23]. Specifically, the intrinsic solid‐state ion diffusion coefficient dictates the diffusion depth of guest ions, which in turn governs the spatial distribution of ion storage. The diffusion energy barrier of Li⁺ in FePO₄ is considerably lower than that of Mg²⁺, enabling Li⁺ to fully access the entire FePO₄ particle [24, 25]. In contrast, the intercalation range of Mg²⁺ into FePO₄ is highly restricted within a limited time due to its high diffusion energy barrier (Figure S3).

FIGURE 3.

Differential voltage analysis (dU/dQ) curves derived from galvanostatic intercalation profiles of FePO₄ in (a) 1 M LiCl and (b) 1 M MgCl₂ electrolytes. (c) Theoretical diffusion distances of Li⁺ and Mg²⁺ within the FePO₄ lattice plotted as a function of time. Representative (d) potential and (e) current profiles of the pulsed charging protocol. (f) Heat map of specific capacity as a function of the charging rate and reverse potential. (g) High‐resolution Fe 2p x‐ray photoelectron spectroscopy (XPS) spectra of electrodes after 1 h of pulsed charging under various reverse potentials. (h) XPS‐quantified surface Mg atomic content of the electrodes after pulsed charging under various reverse potentials and differential capacitance curve of the FePO₄ electrode measured in 1 M MgCl₂ electrolyte. (i) Heat map of specific capacity as a function of the charging rate and duty cycle. (j) Ex situ x‐ray diffraction (XRD) patterns of the electrodes after pulsed charging with various duty cycles. (k) Rietveld‐refined XRD patterns of the electrode after pulsed charging with a duty cycle of 0.95 in 1 M LiCl electrolyte for 1 h. Red and blue contours represent the high and low value, respectively.

Based on the distinct ion storage behaviors of Li⁺ and Mg²⁺ within the FePO₄, we rationally designed a pulsed charging protocol to resolve this dilemma: By applying pulsed potentials (the representative potential and current profiles are shown in Figure 3d,e), Li⁺ intercalation and surface‐limited Mg²⁺ intercalation occur at cathodic potentials. Subsequently, the surface‐intercalated Mg²⁺ ions are stripped off by switching to an anodic potential during the electrochemical rinsing process, whereas the Li⁺ ions that have penetrated deeper into the bulk crystal lattice remain largely confined within the core [26].

First, we systematically evaluated the effect of reverse potentials on capacity retention. The intercalation potential was fixed at −0.6 V versus Ag/AgCl, and the experimental procedures were identical to that in Figure 1c except for the variation of reverse potentials. As shown in Figure 3f, the capacity gradually recovers with increasing reverse potentials, whereas an excessive increase in reverse potentials leads to a slight capacity decay. This decay can be attributed to water electrolysis and chlorine evolution reactions at high anodic potentials, where the by‐products attack FePO₄ and deteriorate its performance according to previous reports [27]. Figure 3g presents the high‐resolution Fe 2p XPS spectra of the electrodes after 1 h of pulsed charging. The characteristic peak at 710.8 eV is assigned to Fe (II), and the peak at 711.9 eV corresponds to Fe (III) (reference spectra of FePO₄ and LiFePO₄ are provided in Figure S4). With increasing reverse potentials, the Fe valence state gradually evolves from Fe (II) to Fe (III). This observation confirms the deintercalation of surface Mg²⁺ and the formation of a pristine FePO₄ shell at sufficiently high reverse potentials. As shown in Figure 3h, the XPS‐derived Mg content for the samples with different reverse potentials (Figure 3g) further corroborates this conclusion. Additionally, the point of zero charge (PZC) of FePO₄ is determined to be ca. 0.6 V versus Ag/AgCl, which corresponds to the potential at which the measured specific capacitance C reaches its minimum value C ₀​ [28]. Accordingly, setting the reverse potential to this value enables Mg²⁺ deintercalation while modulating the competitive adsorption of Cl⁻, thereby optimizing the performance of electrochemical lithium extraction (a detailed analysis is provided in the Supporting Information: Discussion 5.4. Notably, the adsorption and accumulation of Cl⁻ exert a promotional effect on the Mg²⁺ deintercalation process, which will be discussed in detail later).

We further investigated the influence of different duty cycles on capacity retention (Figure 3i). A lower duty cycle corresponds to a smaller proportion of intercalation time, yet it provides a more prolonged interval that facilitates a more thorough deintercalation of surface Mg²⁺ (e.g., symmetric pulses of 100 s at −0.6 V vs. Ag/AgCl and 100 s at +0.6 V vs. Ag/AgCl correspond to a duty cycle of 0.5). However, this is unfavorable for achieving a high Li⁺ extraction rate due to the curtailed Li⁺ intercalation time. We gradually increased the duty cycle and found that the capacity retention remained satisfactory at an optimized duty cycle of 0.95 (i.e., 190 s at −0.6 V vs. Ag/AgCl + 10 s at +0.6 V vs. Ag/AgCl). Ex situ x‐ray diffraction (XRD) patterns of the electrodes subjected to pulsed charging with various duty cycles (Figure 3j) confirm that the crystal structure gradually transforms from FePO₄ (characterized by the diffraction peak at 17.1° corresponding to the (200) plane) to LiFePO₄ (characteristic diffraction peak at 17.9° corresponding to the (200) plane) with increasing duty cycles [3]. Using the optimized pulsed charging parameters (190 s intercalation at −0.6 V vs. Ag/AgCl + 10 s deintercalation at +0.6 V vs. Ag/AgCl), FePO₄ was completely converted into LiFePO₄ within 1 h in 1 M LiCl electrolyte (Rietveld‐refined XRD structure of the electrode after pulsed charging is shown in Figure 3k). This demonstrates the feasibility of electrochemical lithium extraction via the pulsed charging protocol.

Furthermore, when MgCl₂ was replaced with other magnesium salts, including Mg(NO₃)₂, MgSO₄ and Mg(TFSI)₂, comparable capacity retention and robust interfacial stability were consistently achieved (Figure S5). This observation further verifies the broad universality of the proposed pulsed charging protocol across diverse electrolyte configurations. Given that Cl⁻ serves as the predominant anion species in practical, high‐Mg²⁺/Li⁺ ratio salt lake brines [29], MgCl₂ was strategically selected as the model electrolyte to systematically elucidate the underlying mechanisms in the following section.

2.3. Mechanistic Study

To verify that pulsed charging can effectively deintercalate Mg²⁺ ions, various surface analysis techniques were employed to characterize the surface chemical composition of the samples. To minimize interference from the residual electrolyte on the results, the samples were thoroughly rinsed with deionized water prior to characterization (detailed rinsing protocols are provided in the Supporting Information) [30]. Energy‐dispersive x‐ray spectroscopy (EDX) spectra (Figure 4a) were first conducted. For the pristine sample directly soaked in 1 M MgCl₂ electrolyte for 1 h without subsequent rinsing or applying any electrochemical processes (denoted as the soaked sample), Mg and Cl signals attributed to MgCl₂ residues were clearly observed. In contrast, no Mg or Cl signals could be detected after thorough rinsing. The surface Mg content under the pulsed charging protocol was significantly lower than that under the constant‐potential charging protocol, which was further corroborated by the high‐resolution Mg 1s XPS spectra (Figure 4b) and time‐of‐flight secondary ion mass spectrometry (ToF‐SIMS) results (Figure 4c).

FIGURE 4.

(a) Energy‐dispersive x‐ray spectroscopy (EDX) spectra of the electrode surfaces under four conditions: washed, soaked, constant‐potential charging, and pulsed charging. (b) High‐resolution Mg 1s x‐ray photoelectron spectroscopy (XPS) spectra and (c) time‐of‐flight secondary ion mass spectrometry (ToF‐SIMS) depth profiles of the electrodes after various charging protocols. (d) Electrochemical quartz crystal microbalance (EQCM) results for various charging protocols. (e) Mg²⁺/Cl⁻ density ratio profiles as a function of the distance from the electrode surface at various applied voltages. Inset displays a representative snapshot of the molecular dynamics (MD) simulation cell (red, green, blue, and white spheres represent Mg, Cl, O, and H atoms). (f) Mg²⁺ and (g) Cl⁻ density profiles as a function of the distance from the electrode surface at various applied voltages. (h) Mg─O and (i) Mg─Cl radial distribution functions (RDFs) at various voltages. In situ ATR‐FTIR spectra of the electrode/electrolyte interface collected at (j) various stepped potentials and (k) dynamic pulsed potentials. Red and blue contours represent high and low absorbances, respectively. Schematic crystal structures of (l) Mg_0.5FePO₄ and (m) Cl⁻‐adsorbed Mg_0.5FePO₄. Total density of states (DOS) and partial density of states (PDOS) plots for (n) Mg_0.5FePO₄ and (o) Cl⁻‐adsorbed Mg_0.5FePO₄. (p) Energy profiles and the corresponding intermediate configurations at each reaction step for Mg²⁺ transport across the electrode/electrolyte interface during the deintercalation process. The red, yellow, orange, blue, green, and grey spheres represent O, Fe, Mg, P, Cl, and H atoms.

Electrochemical quartz crystal microbalance (EQCM) was further utilized to monitor the real‐time electrode mass changes during different charging protocols. As shown in Figure 4d, under the constant‐potential charging protocol, the electrode mass gradually increases, corresponding to Mg²⁺ intercalation and water molecule co‐adsorption. However, the calculated Δ mass/Δ electron ratio after 300 s is considerably lower than 12 g mol⁻¹ e⁻ (i.e., the theoretical value corresponding to the ideal formation of Mg_0.5FePO₄). Subsequent calculations reveal the actual stoichiometry of the newly formed species to be Mg_0.016Fe(II)_0.016Fe(III)_0.984PO₄. This finding further proves that Mg²⁺ intercalation is highly limited and strictly confined to the electrode surface, exhibiting pseudocapacitive behavior. In contrast, under the pulsed charging protocol (100 s pulses at −0.6 V vs. Ag/AgCl and 100 s pulses at +0.6 V vs. Ag/AgCl over a total duration of 3600 s, a model pulsed charging protocol utilized for clear signal capturing), the electrode mass did not decline during the high‐potential intervals as would be expected for only Mg²⁺ deintercalation, but instead increased slowly. Crucially, given that the overall deintercalation of Mg²⁺ was confirmed by the reduced surface Mg²⁺ content after employing the pulsed charging protocol, this unexpected slow mass increase is hypothesized to result from the combined effects of Mg²⁺ deintercalation and chloride ion (Cl⁻) adsorption [31].

EQCM experiments reveal that the application of reverse pulses promotes the surface enrichment of Cl⁻. This phenomenon was further validated through molecular dynamics (MD) simulations by modeling the electric double layer (EDL) at the electrode/electrolyte interface under −0.6 V and +0.6 V bias conditions. As shown in Figure 4e, Mg²⁺ ions are enriched while Cl⁻ ions are depleted in the EDL at −0.6 V, whereas a sharp inversion occurs at +0.6 V. The relationships between the local ion densities (Mg²⁺ and Cl⁻) and the distance from the electrode surface are shown in Figure 4f,g, further validating the potential‐induced EDL reconstruction [32]. The radial distribution functions (RDFs) of Mg─O and Mg─Cl were also analyzed (Figure 4h,i). The results show that the external electric field exerts a negligible effect on the coordination between water molecules and Mg²⁺ but significantly influences the coordination between Cl⁻ and Mg²⁺ in the first hydration shell of Mg²⁺ [33]. Specifically, at +0.6 V, more Cl⁻ ions entered the first hydration shell of Mg²⁺, indicating that a positive electrode potential effectively drives the accumulation and enrichment of Cl⁻ ions at the electrode surface.

Electrochemical attenuated total reflection surface‐enhanced infrared absorption spectroscopy (EC‐ATR‐SEIRAS) serves as a powerful characterization tool for probing the dynamic electrode/electrolyte interface [34]. Infrared spectra of FePO₄ were collected in 1 M MgCl₂ electrolyte at various potentials gradually stepped from −0.6 to +0.6 V versus Ag/AgCl. Using the absorbance spectrum acquired at −0.6 V as the baseline, the differential infrared spectra were calculated and presented in Figure 4j. With increasing electrode potential, the intensity of the absorption band at 1240 cm⁻¹ gradually increases. The potential‐driven specific adsorption of Cl⁻ ions and the resulting Fe─Cl coordination trigger local surface polarization that shortens the adjacent P─O bond lengths and elevates their bond strengths. As a result, the classic asymmetric stretching modes of the PO₄ ³⁻ tetrahedra are blue‐shifted and prominently resolve at 1240 cm⁻¹ [35], displaying a reversible, periodic evolution in tandem with the imposed electrochemical potential profiles (Figure 4k).

Crystal structures along with corresponding density of states (DOS) and partial density of states (PDOS) profiles for Mg_0.5FePO₄ and Cl⁻‐adsorbed Mg_0.5FePO₄ (denoted as Cl‐Mg_0.5FePO₄) are presented in Figure 4l–o, respectively. DFT calculations reveal that the Cl⁻ adsorption induces new electronic states in the band gap of the Mg_0.5FePO₄ (010) surface, reducing the band gap from 0.31 to 0 eV. Such gap states originate from Fe─Cl bonding interactions, where the orbital hybridization between Fe 3d and Cl 3p gives rise to band gap narrowing [36]. This enhances the local electronic conductivity, and decreases charge‐transfer resistance, thereby facilitating the rapid deintercalation of Mg²⁺ [37].

According to the previous literature, anion adsorption on the FePO₄ surface can significantly modulate the energy barrier for ion deintercalation [38]. As shown in Figure 4p, we calculated the Gibbs free energy profiles for the deintercalation of Mg²⁺ from surface Mg_0.5FePO₄ in the presence and absence of surface‐adsorbed Cl⁻. The results demonstrate that Mg²⁺ deintercalation from FePO₄ is more favorable (exhibiting a lower energy barrier) when Cl⁻ ions are adsorbed at the electrode surface. These results further explain why the pulsed charging protocol can effectively suppress interfacial degradation. Additionally, our calculation reveals that the lattice substitution of bulk‐coordinated PO₄ ³⁻ by Cl⁻ requires a positive formation energy (0.27 eV, Figure S6), demonstrating that such a lattice substitution is thermodynamically unfavorable. This result ensures that the pulsed charging protocol will not trigger structural collapse.

2.4. Lithium Extraction Performance

To evaluate the Li⁺ extraction efficiency of the pulsed charging protocol, a two‐electrode electrochemical lithium extraction cell featuring a FePO₄||Ag configuration was assembled. Within this device, the FePO₄ electrode was employed for reversible Li⁺ intercalation/deintercalation, while the Ag electrode served as the counter electrode to sustain charge balance via reversible Cl⁻ trapping/releasing through the Ag/AgCl redox conversion (the intercalation potentials for both protocols are the same as −0.6 V vs. Ag counter electrode, experimental details are provided in the Supporting Information) [2]. First, the lithium extraction performances under the constant‐voltage charging protocol and pulsed charging protocol were evaluated in simulated salt lake brines with a fixed Li⁺ concentration of 200 ppm and varied Mg²⁺/Li⁺ molar ratios ranging from 0 to 100. As shown in Figure 5a, the Li⁺ extraction rates (calculated based on the intercalation duration) under the pulsed charging protocol are significantly higher than those achieved by the conventional constant‐voltage method. With an increasing Mg²⁺/Li⁺ molar ratio, the Li⁺ extraction rates decrease due to the selective polarization effect induced by Mg²⁺ ions. Notably, the Li⁺ extraction rates attenuate to a lesser extent under the pulsed charging protocol compared to the constant‐voltage method with the increase of Mg²⁺/Li⁺ molar ratio, indicating that the pulsed charging protocol can effectively reconstruct the surface EDL, thereby mitigating the electrostatic charge‐shielding effect caused by the accumulation of Mg²⁺ at the electrode/electrolyte interface [12, 39].

FIGURE 5.

(a) Li⁺ extraction rate of various charging protocols in simulated salt lake brines (Li⁺ concentration: 200 ppm, Mg²⁺/Li⁺ ratio varied from 0 to 100). (b) Li⁺ extraction rate of various charging protocols in real salt lake brines and the purity of LiCl product obtained by pulsed charging protocol. (c) Benchmarking of this work against previous studies in the field. (d) Long‐term stability of various charging protocols in real salt lake brines. (e) Contour plot of initial capital investment for electrodes (in USD, logarithmic scale) as a function of Li⁺ extraction rate and daily Li₂CO₃ production. (f) Contour plot of total capital investment for electrodes (in USD, logarithmic scale) as a function of Li⁺ extraction rate and electrode service life. The annual Li₂CO₃ production capacity is assumed as 10 kt and the project lifetime is assumed as 15 years. Red and blue contours represent the high and low expenditures, respectively.

To demonstrate its practical viability, the electrochemical lithium extraction performance was further validated in real salt lake brines: brine sourced from the West Taijinar Salt Lake, Qinghai, China (Li⁺ concentration: 386 ppm, Mg²⁺/Li⁺ molar ratio: 19) and brine sourced from the Salar del Rincón, Salta, Argentina (Li⁺ concentration: 587 ppm, Mg²⁺/Li⁺ molar ratio: 13). The results shown in Figure 5b consistently demonstrate that the pulsed protocol delivers superior Li⁺ extraction rates compared to the conventional constant‐voltage charging protocol. Furthermore, the final LiCl product obtained by the pulsed charging protocol exhibits a high purity of 99.95%, fully satisfying the quality criteria for battery‐grade materials. As shown in Figure 5c, the Li⁺ extraction rate achieved by the proposed pulsed charging protocol (∼20 mg g⁻¹ h⁻¹) positions this work at the forefront of the field, highlighting the efficacy of this novel strategy for advancing lithium extraction technology (detailed comparison data are provided in Table S1 in the Supporting Information). More importantly, the pulsed charging protocol significantly enhances the stability of the FePO₄ electrode. As shown in Figure 5d, it achieved 800 h of stable operation in real salt lake brines, with its performance retaining approximately 80% of the initial value. This represents one of the best reported stability performances (Figure S7) [40]. In contrast, the lithium extraction performance of the constant‐voltage method decayed rapidly within 400 h. This superior longevity is attributed to the periodic, potential‐driven electrochemical rinsing of Mg²⁺ from the FePO₄ crystal lattice during the reverse pulse phases, which successfully prevents lattice distortion and eliminates the interfacial degradation triggered by irreversible Mg²⁺ intercalation.

To validate the engineering importance of achieving a high Li⁺ extraction rate, techno‐economic analysis (TEA) was conducted (detailed parameters and calculation methods are provided in the Supporting Information). Figure 5e displays the contour plot of the initial electrode‐related capital investment (in USD, logarithmic scale) as a function of the Li⁺ extraction rate and daily Li₂CO₃ production capacity. As the daily Li₂CO₃ production escalates, elevating the Li⁺ extraction rate yields a considerable reduction in the initial electrode‐related capital investment. Specifically, for a target daily Li₂CO₃ production of 29 t d⁻¹ (corresponding to an annual capacity of 10 kt), accelerating the Li⁺ extraction rate from a conventional benchmark of 5 to 20 mg g⁻¹ h⁻¹ accomplished in this work lowers the electrode‐related capital investment from $625 999 to $156 500. In a volatile global lithium market with fluctuating Li₂CO₃ prices [41], this technical advantage not only alleviates upfront capital intensity and investment risks, but also substantially curtails the investment payback period, thereby rendering electrochemical lithium extraction more economically attractive for industrial commercialization [42]. Furthermore, our analysis reveals a critical, previously overlooked paradigm that Li⁺ extraction rate plays an equally pivotal role in electrode service life. While prior studies have emphasized extending electrode service life to reduce costs [6], our finding suggests that for a fixed annual Li₂CO₃ production capacity (as shown in Figure 5f, with an annual Li₂CO₃ production of 10 kt and a project lifetime of 15 years), the total electrode‐related cost associated with an Li⁺ extraction rate of 20 mg g⁻¹ h⁻¹ with an electrode lifetime of 150 days is identical to that of 5 mg g⁻¹ h⁻¹ with an electrode lifetime of 600 days. Crucially, our optimized pulsed‐charging protocol simultaneously addresses both bottlenecks by delivering a 2.2‐fold boost in extraction kinetics and a 5‐fold extension in operational lifespan, thereby substantially reducing the recurrent stack‐manufacturing and replacement operational overheads (Figure S8). The underlying reason is that a sluggish Li⁺ extraction rate fundamentally necessitates a massive required active electrode mass during plant operation to meet the production quota. This inflated baseline mass directly multiplies the economic penalty incurred during each replacement cycle, thereby driving up cumulative expenditures. Consequently, maximizing the kinetic extraction rate represents a key factor in achieving economic viability in large‐scale brine operations.

3. Conclusion

This study addresses the long‐standing issue of FePO₄ electrode degradation during electrochemical lithium extraction from salt lake brines with high Mg²⁺/Li⁺ ratios and develops a novel pulsed charging protocol to enable efficient and stable Li⁺ extraction. We reveal that Mg²⁺ intercalation at low potentials induces severe structural instability in FePO₄, which further leads to the formation of Fe₂O₃ on the electrode surface under the combined effects of aqueous medium and electrode potential. The newly formed Fe₂O₃ layer isolates FePO₄ particles from the conductive carbon black, consequently causing a severe decay in the electrochemical performance of the electrode. To circumvent this bottleneck, a pulsed charging protocol was rationally designed based on the distinct spatial distributions and electrochemical kinetics of Li⁺ and Mg²⁺ within FePO₄. Specifically, Mg²⁺ ions are predominantly confined to the near‐surface region due to severe diffusion limitations, whereas Li⁺ ions diffuse deeply into the bulk phase. The applied potential pulses efficiently strip off surface‐intercalated Mg²⁺ and drive the dynamic enrichment of adsorbed Cl⁻ at the electrode/electrolyte interface. The adsorbed Cl⁻ modulates the surface electronic structure of Mg_0.5FePO₄, narrowing its band gap and thus enhancing the intrinsic electronic conductivity of the electrode interface. Concurrently, Cl⁻ adsorption directly lowers the energy barrier for Mg²⁺ deintercalation at the electrode/electrolyte interface, synergistically accelerating the interfacial charge‐transfer dynamics and ion‐transport kinetics for efficient Mg²⁺ deintercalation. Notably, bulk Li⁺ diffusion is barely affected by these interfacial modulation processes, while surface Mg²⁺ deintercalation is significantly promoted. Incorporating this optimized protocol into a two‐electrode FePO₄||Ag cell delivers a high Li⁺ extraction rate of ∼20 mg g⁻¹ h⁻¹ and yields LiCl with a purity of 99.95% in real salt lake brines. Furthermore, the system demonstrates promising durability, maintaining 80% of its initial performance after 800 h of continuous operation. In summary, this study not only provides a fundamental understanding of Mg²⁺‐induced FePO₄ degradation but also establishes a versatile strategy for tailoring the electrode/electrolyte interface via pulsed charging. Looking forward, these findings pave the way for the practical application of FePO₄‐based systems in lithium extraction from high Mg²⁺/Li⁺ brines and inspire the rational design of efficient electrode materials and operating protocols for other electrochemical separation and purification technologies.

Author Contributions

Ziquan Wang: methodology, data curation, investigation, conceptualization, writing – original draft. Zhujie Liang: methodology, validation. Yongbing Tang: formal analysis, validation. Xuehui Li: validation, formal analysis. Minjun Kim: writing – review and editing, formal analysis. Yusuke Yamauchi: writing – review and editing, formal analysis, validation, visualization. Libo Deng: conceptualization, methodology, supervision, writing – review and editing, funding acquisition, investigation.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: The authors have cited additional references within the Supporting Information [1–124].

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.