A surface chemistry-regulated gradient multi-component solid electrolyte interphase for a 460 W h kg⁻¹ lithium metal pouch cell

Abstract

Lithium (Li) metal is an ideal anode for high energy density rechargeable Li batteries. However, parasitic reactions and an uneven native oxide layer on the surface lead to uncontrollable Li deposition and dendrite growth, significantly restricting its practical application. Here, we introduce a simple and scalable surface chemical approach involving spray casting of dilute 2,2-difluoro-2-(fluorosulfonyl)acetic acid (DFFSA) solution onto the Li surface, meticulously regulating ion transfer and improving interface stability to achieve stable cycling of the Li anode. The spontaneous in situ reaction between Li and DFFSA eliminates the uneven native oxide layer, forming an organic fluorinated carboxylate lithium salt on the outermost surface and a graded inorganic layer composed of LiF, Li₂S, and Li₂SO₃ inside, resulting in a multi-component artificial solid electrolyte interphase (SEI). This multi-component SEI, as evidenced by visualization techniques and computational methods, exhibits enhanced Li affinity and wettability, enabling rapid lithium-ion transport and dendrite-free, uniform lithium deposition. Consequently, the modified Li||LiCoO₂ full cell retains 77.85% capacity after 1200 cycles in carbonate-based electrolytes. A 461.6 W h kg⁻¹ pouch cell with a capacity of 5.49 A h, a low N/P ratio of 1.28 and a lean electrolyte of 1.6 g A h⁻¹, demonstrates an impressive capacity retention of 84.7% after 100 cycles at 0.5C. This work provides a simple and promising surface engineering strategy and enlightens the multi-component SEI design for promoting the practical application of high energy density Li metal batteries.

---

Broader context

To address the challenges of uneven Li deposition and dendrite formation in carbonate-based electrolytes, conventional strategies such as electrolyte design, 3D current collectors, and artificial SEI coatings are constrained by the native passivation layer and issues of scalability and cost-effectiveness. A simple and scalable surface chemistry approach is proposed, involving spray casting of a dilute solution of fluorosulfonyl organic carboxylic acid on the Li surface. This technique not only obviates the uneven native passivation layer but also fosters the formation of a multi-component artificial SEI on the Li surface. Featuring a lithium fluoro-organic carboxylate outer layer and a gradient inorganic layer internally enriched with LiF, Li₂S, and Li₂SO₃, this interphase offers excellent Li affinity, good wettability and rapid lithium-ion transport, thereby promoting uniform, dendrite-free Li deposition. This SEI design ensures enduring stability of the Li anode. Consequently, the Li||LiCoO₂ cell achieves a notable capacity retention of 77.85% after 1200 cycles and a 5.49 A h Li||LiNi_0.8Co_0.1Mn_0.1O₂ pouch cell demonstrates an outstanding energy density of 461.6 W h kg⁻¹ with an 84.7% capacity retention after 100 cycles, outperforming most recently reported Li metal batteries. This work introduces a promising approach to advancing the practical application of high-energy-density Li metal batteries.

1. Introduction

The lithium metal anode (LMA), esteemed as the “Holy Grail” of lithium (Li) batteries, exhibits a remarkable theoretical specific capacity of 3860 mA h g⁻¹ and possesses the lowest electrochemical potential of −3.04 V (versus the standard hydrogen electrode (SHE)).^1–3 Consequently, the Li metal anode is envisioned as a linchpin for achieving the next generation of high-energy-density rechargeable batteries. However, the challenge of achieving reversible cycling of Li metal during repeated charging and discharging remains unresolved and urgently demands attention.⁴ Li metal exhibits high reactivity with most organic electrolytes, especially with carbonate-based electrolytes that are currently the most commonly used. During the cycling process in carbonate-based electrolytes, side reactions occur at the interface of the Li anode, resulting in the formation of porous Li dendrites as well as significant volume changes.⁵ These issues severely limit the cycle life and safety of Li metal batteries (LMBs).

To maintain the interfacial stability of LMAs and suppress excessive side reactions, the construction of a stable solid electrolyte interphase (SEI) between the Li metal and the electrolyte represents an effective strategy. This can be achieved through various approaches such as the application of artificial SEI coatings,^6–10 electrolyte design,¹¹ ionic liquid electrolytes,¹² and electrolyte additives.^13–15 However, parasitic reactions still occur on the surface of LMAs when carbonate-based electrolytes are used. Moreover, the formation of an interphase with LiF-rich components continuously consumes fluorine-containing components in the electrolyte, leading to the formation of a severely inhomogeneous SEI that compromises the ability to protect the LMAs. Finally, the native passivation layer on the Li surface can cause inhomogeneous lithium-ion flux and promote the growth of Li dendrites. Even with the most effective electrolyte design, these passivation layers could not be completely removed from the Li surface.

Eliminating the native oxide layer and constructing a uniform and dense artificial SEI on the Li surface are thus ideal alternatives. For instance, Choi et al.¹⁶ demonstrated the promotion of homoepitaxial Li deposition by removing the native oxidation layer using simple bromine-based acid–base chemistry. Cui et al.¹⁷ proposed a new strategy for stabilizing the SEI and LMAs by employing sulfur vapor upon contact with LMA to form a highly ionic conductive and uniform Li₂S interfacial layer. Sun et al.⁶ utilized the spontaneous in situ reaction between lithium and heptafluorobutyric acid (HFA) to eliminate the surface oxide layer and generate a lithiophilic interphase of HFA-Li, promoting dendrite-free uniform Li deposition and significantly improving the cycle stability.

Despite the significant advances, eliminating the native oxide layer and constructing a multicomponent artificial SEI with high ionic conductivity through simple surface chemistry still remain challenging. Here, we demonstrate a facile surface chemistry approach that involves spraying a diluted solution of 2,2-difluoro-2-(fluorosulfonyl) acetic acid (DFFSA) onto the Li surface, reacting with the native oxide layer and forming a multicomponent, highly ion-conductive artificial SEI with a thickness of 1–2 μm to achieve an efficient and stable cycling DFFSA–Li anode. The lithium organic fluoro-carboxylate layer formed on the outermost layer enhances the wettability between Li metal and electrolyte, effectively reducing the interface impedance. The gradient artificial SEI formed on the inner layer, which is rich in LiF, Li₂S, and Li₂SO₃, enhances the Li⁺ diffusion within the interphase, reduces polarization, and promotes uniform Li⁺ deposition. As a result, when coupled with the LiCoO₂(LCO) cathode with an areal density of 10.7 mg cm⁻², the full coin cell delivers a high-capacity retention of 77.85% after 1200 cycles. Furthermore, the Li||LCO and Li||LiNi_0.8Co_0.1Mn_0.1O₂(NCM811) pouch cells show stable cycling with high-capacity retentions even under high cathode loading (39–49.7 mg cm⁻²) and thin Li foil (50 μm) with limited Li excess. A 5.49 A h Li||NCM811 pouch cell achieves an impressive energy density of 461.6 W h kg⁻¹ and delivers a lifespan of 100 cycles under a low negative/positive capacity ratio of 1.28 and lean electrolyte of 1.6 g A h⁻¹. In this study, the simple surface engineering strategy is developed to form a multicomponent, highly ion-conductive artificial SEI with a lithiophilic interface, thereby enhancing the stability of the anode and promoting the practical application of high-energy-density Li metal batteries.

2. Results and discussion

2.1. Design principle of surface chemistry with dilute DFFSA solution treatment

Here, we choose a dilute DFFSA in tetrahydrofuran (THF) to treat the Li foil surface to produce a protective multicomponent artificial SEI via a spontaneous chemical reaction, as shown in Fig. 1a. DFFSA possesses multi-polar functional groups such as –FSO₂, C–F, and –COOH, which can be readily transformed into LiF, Li₂S, –COOLi, and other organic components during chemical or electrochemical reactions. The Bare-Li surface contains a native oxidation layer mainly composed of Li₂CO₃ and Li₂O,^18,19 which are naturally formed in a drying chamber or glove box environment due to trace amounts of CO₂ or oxygen. The Li oxide layer natively formed on the Li foil is known to impose substantially high resistance for Li-ion diffusion and uneven Li deposition compared to the SEI layer produced by the (electro)chemical reaction with the electrolyte.¹⁶ Therefore, the highly active DFFSA can directly react with Li metal and the native oxide layer on the Li surface, forming a multicomponent artificial SEI composed of inorganic substances such as LiF, Li₂S, and Li₂SO₃ and organic components such as C–F and C₂SO₄F₃Li. Compared with the native passivation layer on Bare-Li, the multicompetent artificial SEI constructed by DFFSA enables rapid longitudinal Li ion transportation and promotes dendrite-free, uniform Li deposition to improve the cycling stability of LMAs.

Fig. 1 Schematic illustration of the in situ formation process and protection mechanism of the organic lithium fluorocarboxylate and inorganic LiF/Li₂S/Li₂SO₃ layers.

Characterization of the multicomponent SEI protective interphase

A 0.5 wt% DFFSA solution in THF was sprayed onto the surface of the bare Li, and then the THF solvent was evaporated in a glove box to obtain the DFFSA–Li anode. The use of low-concentration DFFSA solutions for surface treatment, on the one hand, can reduce the consumption of cost-intensive DFFSA, which is beneficial for reducing the cost of large-scale Li surface modification. On the other hand, it can prevent excessive corrosion reactions between the strongly acidic DFFSA and Li, which is conducive to the formation of a thinner multi-component SEI. The surface morphology of Bare-Li and DFFSA–Li after fluorosulfonylcarboxylic acid treatment was determined by field emission scanning electron microscopy (FE-SEM), as shown in Fig. 2a and b, respectively. Notably, the treated DFFSA–Li in Fig. 2b exhibited a flat and smooth surface morphology, with no observable pores, cracks, or other irregularities similar to those found on the Bare-Li surface (Fig. 2a). This indicates that during the DFFSA treatment process, chemical reactions occurred between DFFSA and the Bare-Li surface layer, eliminating the inhomogeneous surface oxide layers, impurities and defects. As can be seen from the cross-sectional SEM images (Fig. S1, ESI†), the thickness of the artificial SEI is about 1–2 μm. Fig. 2c and d show the optical images of LMAs before and after treatment. It can be observed that the surface of DFFSA–Li is cleaner and smoother, indicating that a significant interfacial reaction has occurred.

Fig. 2 SEM images of (a) Bare-Li and (b) DFFSA–Li surfaces. Optical images of the (c) Bare-Li and (d) DFFSA–Li anodes. Depth-dependent XPS spectra of DFFSA–Li (e) F 1s and (f) S 2p. The 3D overlays of the ion fragments of (g) F⁻, (h) S²⁻, (i) SO₃⁻, and (j) C₂SO₄F₃⁻. (k) DFT calculations of the reaction pathways of DFFSA and Li.

To clarify the chemical reaction mechanism and interfacial composition between DFFSA and the Li foil surface, the XRD patterns of the reaction products of the main native oxides (Li₂CO₃ and Li₂O) on the Li surface and excess DFFSA were preliminarily characterized (Fig. S2 and S4, ESI†). After excess DFFSA reacted with Li, the Li diffraction peak disappeared, and a complex derivative peak emerged, indicating the possible formation of fluorinated carboxylate Li salts.⁶ Interestingly, the reaction of DFFSA with Li₂O and Li₂CO₃ yielded a product with prominent LiF diffraction peaks rather than the expected lithium carboxylate-like reactant with Li. This is atypical for typical acid–base neutralization reactions. This abnormal reaction behavior may be attributed to the special properties of the –FSO₂ group,^20,21 the low energy of the S–F bond, which may break first during the reaction, leading to the formation of inorganic substances such as LiF or Li₂S. To confirm this speculation, we conducted density functional theory (DFT) calculations of the possible reaction pathways. According to the reaction shown in the ESI† (Fig. S3), DFFSA can react with Li₂CO₃ and Li₂O to produce LiF.

In order to further explore the reaction mechanism of DFFSA and the surface composition of DFFSA–Li, X-ray photoelectron spectroscopy (XPS), infrared spectroscopy (IR), and secondary time-of-flight mass spectrometry (TOF-SIMS) were carried out. In the XPS depth profiling experiments, the XPS spectra at different etching stages revealed the intensity characteristics of LiF, Li₂S, and C₂SO₄F₃⁻ within the SEI as they varied with depth (Fig. 2e and f). At the initial etching stage, the F 1s spectrum exhibited peaks corresponding to C–F (686.4 eV) bonds (primarily originating from C₂SO₄F₃⁻) and LiF (684.9 eV).⁶ The strongest signal of LiF indicates its dominant position within the SEI. The LiF signal strengthens as the etching progresses deeper, indicating that its concentration gradually increases from the surface of the SEI towards the inner layer. Concurrently, the C–F signal rapidly diminished, revealing that C₂SO₄F₃⁻ is predominantly located in the outermost layer of the SEI. In the S 2p spectrum, peaks associated with Li₂SO₃ and Li₂S (160.9 eV) (ref. 8) were identified, and the proportion of the Li₂S signal increased with the depth. The 3D ion fragment maps from TOF-SIMS in Fig. 2g–i and the depth sputter curves (Fig. S6 and S7, ESI†) align with the XPS results, jointly revealing the gradient distribution characteristics of the multicomponents within the SEI. The inner inorganic SEI components rich in LiF and Li₂S facilitate rapid Li⁺ transport, leading to the homogenization of the Li-ion flux and inhibition of the growth of lithium dendrites. Meanwhile, the presence of organic components such as C₂SO₄F₃⁻ and C₂SO₃F₃⁻ (Fig. 2j and Fig. S5, ESI†) are primarily distributed in the outer layer of the SEI, exhibiting good affinity towards lithium.⁶ These organic components also contribute to isolating side reactions between lithium and the electrolyte.⁶ Furthermore, IR was employed to study the changes in the bond information on the Li surface (Fig. S8, ESI†). The IR signals were collected from the Li surface in diffuse reflectance mode. The results showed that the peak corresponding to C=O vibration shifted from 1773.2 cm⁻¹ (characteristic of carboxylic acid) to 1673.9 cm⁻¹ (–COOLi), indicating the formation of Li carboxylates after treatment. These results further proved the successful formation of organic–inorganic composite SEI on the Li surface after DFFSA treatment.

To clarify the rationality of the interfacial reaction products, DFT calculations regarding the reaction pathway of DFFSA with Li (Fig. 2l) were performed. The results of the possible reduction pathway indicate that DFFSA first reacts with the alkali metal Li to form lithium carboxylate. Subsequently, the –SO₂F group detaches and undergoes bond cleavage, generating LiF and LiSO₂. Under conditions of excess Li, LiSO₂ is unstable and decomposes into Li₂S and Li₂SO₃. These results suggest that surface treatment with a dilute solution of DFFSA can not only eliminate the native oxide layer on the Li surface but also form a surface layer enriched in LiF, Li₂S, Li₂SO₃ and Li organic carboxylate, which is expected to improve the lithiophilicity and cycling stability of the Li metal anode interface.

Effect of DFFSA–Li on suppressing Li dendrite growth

A preliminary evaluation of the effect of DFFSA concentration on the stability of the Li anode was conducted. Fig. S9 (ESI†) presents the rate performance of Li||LCO batteries after treatment with DFFSA at concentrations of 0.2, 0.5, and 1.0 wt%. The DFFSA–Li anodes treated with dilute DFFSA solutions all exhibit significantly enhanced cycling stability. Among them, the DFFSA–Li anode treated with 0.5 wt% DFFSA dilute solution demonstrates the most outstanding rate performance, achieving a capacity of 86.67 mA h g⁻¹ even at a high rate of 5C. Furthermore, electrochemical impedance spectroscopy (EIS) measurements were conducted to assess the interfacial impedance of DFFSA–Li treated at different concentrations (Fig. S10a, ESI†). The impedance spectra consist of a semicircle at high frequencies and a diagonal line at low frequencies, corresponding to the charge transfer resistance (R_ct) and Warburg impedance, respectively. It can be observed that the R_ct of the batteries assembled with DFFSA–Li anodes is significantly reduced and exhibits a trend of initial decrease followed by an increase with increasing DFFSA concentration (Fig. S10b, ESI†). Notably, the R_ct value of DFFSA–Li treated with 0.5 wt% concentration was the lowest at 89.6 Ω, which is about one-fourth of that of Bare-Li with 388.4 Ω. These preliminary results suggest that it is necessary to treat metal Li with an appropriate concentration of DFFSA to form a well-structured and significantly stable multi-component SEI film. Optimizing the concentration of DFFSA is crucial for achieving a balanced SEI on LMAs. A low concentration of DFFSA, such as 0.05 wt%, results in a thin SEI with limited inorganic components, hindering ionic conductivity and mechanical stability. Conversely, high concentrations (100 wt%) lead to excessive organic residue accumulation,¹⁰ impeding ion transport and enhancing brittleness. XPS analysis (Fig. S11, ESI†) validates these contrasting effects, emphasizing the need to find the optimal DFFSA concentration to maximize battery performance. Considering these results, the DFFSA–Li anodes treated with a 0.5 wt% DFFSA dilute solution were employed in subsequent tests and descriptions.

Coulombic efficiency is an important metric for evaluating the cycling reversibility of LMAs. The influence of DFFSA–Li on the CE was determined using two methods: the “reservoir” method and the asymmetric Li||Cu cell.²²Fig. 3a shows the test curves for DFFSA–Li and Bare-Li anodes utilizing a method where 50 μm Li is used as the “reservoir”. The CE of the DFFSA–Li anode, with 99.04%, is significantly higher than that of Bare-Li, with only 93.15%. The asymmetric Li||Cu cell assembled with DFFSA–Li exhibits enhanced cycling stability over 50 cycles, while the CE of Bare-Li begins to decline significantly after 25 cycles (Fig. S12, ESI†). The improvement in CE and the increase in stability indicate an enhancement in the reversibility of the Li anode reaction. The high coulombic efficiency further contributes to the superior interfacial stability of the Li anodes.

Fig. 3 (a) Voltage profiles of the DFFSA–Li and Bare-Li anodes by the “reservoir” method for the coulombic efficiency test. (b) Voltage profiles of the symmetric cells with a cycling capacity of 1 mA h cm⁻² at current densities of 1, 3, 5, 8 and 10 mA cm⁻². (c) Voltage profiles of the symmetric cells with a cycling capacity of 3 mA h cm⁻² at 3 mA cm⁻². (d), (e) and (f) The enlarged voltage–time curves of the Li||Li symmetric cells in (c) EIS of the (g) Bare-Li and (h) DFFSA–Li cells after 5, 20, and 50 cycles of the cell. (i) EIS of the Li||LCO cells with Bare-Li and DFFSA–Li before cycling.

To evaluate the interfacial transport rate and stability of the Li surface in DFFSA–Li, Li||Li batteries were constructed. Rate capability tests were conducted on symmetric batteries (Fig. 3b and Fig. S14a, ESI†) using a carbonate electrolyte. In the rate capability test, the cell using DFFSA–Li exhibited a more stable voltage curve and lower polarization. The polarization gap increased as the current density increased, indicating that DFFSA–Li has a strong ability to adapt to high-flux Li⁺ transport. Meanwhile, galvanostatic charge–discharge tests were performed using Li||Li symmetric cells (Fig. 3c–f). The symmetric DFFSA–Li battery exhibited a stable voltage distribution for over 300 hours at 3 mA cm⁻², with a significant reduction in the polarization voltage, while the Bare-Li only cycled for 135 hours. Moreover, the results of symmetric cells using an ether-based electrolyte of 1.0 M LiTFSI in DOL and DME (v/v = 1 : 1) with 1.0 wt% LiNO₃ in Fig. S14b (ESI†) also indicate that DFFSA–Li significantly reduce the polarization voltage and enhance the cycling stability, which suggests the effectiveness of DFFSA modification across different electrolyte systems due to its ability to form a stable and uniform SEI on the surface of the Li anode. These results indicate that the dilute DFFSA modification is a versatile approach and demonstrates great potential in improving the stability of LMAs.

To further verify the advantages of the artificial SEI formed on the surface of the DFFSA–Li in symmetric cells, EIS measurements were conducted. The impedance of the symmetric batteries was measured after the 5th, 20th, and 50th cycles, and the results are shown in Fig. 3g and h. The impedance of the DFFSA–Li symmetric batteries was lower than that of the Bare-Li symmetric cells at various cycles. EIS measurements were also performed on Li||LCO cells, and the impedance of the DFFSA–Li battery was 89.6 Ω, which was only about 1/4 of that of Bare-Li (388.4 Ω). The EIS results confirm that the artificial SEI formed by DFFSA can accelerate Li⁺ and charge transfer, thereby promoting uniform Li deposition and faster interfacial reaction kinetics.²³

To directly observe the changes in deposition behavior on the surface of the Li anodes before and after DFFSA treatment, symmetric batteries were assembled, and an in situ optical microscope was used to observe the Li deposition behavior. Fig. 4a displays real-time optical images of in situ Li deposition in DFFSA–Li and Bare-Li symmetric batteries at 0, 2, 4, 6, 10, and 20 minutes, respectively, at a current density of 2 mA cm⁻². After 2 minutes of deposition on Bare-Li, Li dendrites had already grown noticeably. As the deposition process progressed, the dendrite growth intensified, forming distinct dendritic structures on the Bare-Li surface. In contrast, the growth of Li dendrites on the DFFSA–Li surface was effectively suppressed, and no dendrite formation was observed even after 20 minutes. Morphological changes in the Li||Li symmetric cells after 50 cycles were observed using SEM (Fig. 4b–g). During cycling, the Li deposition capacity was 1 mA h cm⁻² at a current density of 1 mA cm⁻². The uneven morphology on the Bare-Li surface was evident, and the dendrites exhibited a loose and porous structure with cracks, indicating uneven Li deposition (Fig. 4b and c). In contrast, the surface of DFFSA–Li was remarkably flat and dense, and no obvious dendrites were observed (Fig. 4e and f). Cross-sectional SEM images after cycling revealed that the thickness of the DFFSA–Li reaction surface layer was only 8.6 μm (Fig. 4g), obviously thinner than that of Bare-Li (37.5 μm) (Fig. 4d), further confirming that the multicomponent SEI could promote uniform Li deposition and suppress the Li dendrites growth effectively.

Fig. 4 (a) Optical images of the in situ Li deposition on Bare-Li and DFFSA–Li at a deposition current density of 2 mA cm⁻². SEM images of the Li||Li symmetric cells containing (b)–(d) Bare-Li and (e)–(g) DFFSA–Li after 50 cycles of plating/stripping processes. (h)–(j) Depth-dependent XPS spectra of the symmetric cells after 10 cycles. (k) Changes in the LiF composition with etching depth after 10 cycles of Bare-Li and DFFSA–Li.

To investigate the chemical composition changes of the DFFSA–Li surface during cycling, depth-dependent XPS analysis was conducted on Li anodes after 10 cycles at a current density of 1 mA cm⁻² with a Li deposition capacity of 0.5 mA h cm⁻² per cycle (Fig. 4c–f). In the F 1s spectra, peaks of C–F (686.4 eV) and LiF (684.9 eV) were observed on both Bare-Li and DFFSA–Li surfaces. However, the LiF content on the DFFSA–Li surface was significantly higher than that on the Bare-Li surface. Moreover, with increasing etching depth, the LiF content sharply increased while the proportion of organic components gradually decreased, indicating a higher inorganic component content in the inner layers of the SEI for DFFSA–Li. On the contrary, the Bare-Li surface exhibited an opposite trend. In the C 1s spectrum (Fig. S15, ESI†), Li₂CO₃ (288.2 eV) and C–O (286.2 eV)²⁴ were observed on both Bare-Li and DFFSA–Li, which mainly originated from the decomposition of the carbonate electrolyte in Bare-Li.²⁵ However, the peaks from Li₂CO₃ and C–O in DFFSA–Li were weaker than those in Bare-Li. Peaks attributed to sulfur were also observed on the DFFSA–Li surface, including Li₂S (160.9 eV) and Li₂SO₃ (167.3 eV), and the proportion of Li₂S components increased with increasing etching depth. Based on a comparative analysis of elemental contents obtained by XPS (Fig. S16, ESI†), the surface contents of C and O on DFFSA–Li decreased while the content of the F element increased compared to Bare-Li. This indicates that DFFSA treatment can effectively increase the contents of LiF and Li₂S in the SEI, reduce the decomposition of the electrolyte, and inhibit the formation of Li₂CO₃. The XPS results further demonstrate that DFFSA treatment could form a multi-component SEI rich in LiF and Li₂S, maintaining good stability during cycling. TOF-SIMS analysis of Li metal anodes cycled in symmetric cells (Fig. S17, ESI†) indicates that there was significant inhomogeneity in the LiF distribution within the SEI of Bare-Li. Additionally, electrolyte decomposition results in the accumulation of phosphorus-containing species in deeper layers, hindering the stable cycling of the cells. In contrast, the SEI of DFFSA–Li exhibits a uniform and thin structure and contains a Li₂S layer with high ionic conductivity, effectively preventing direct contact with the electrolyte and promoting rapid Li-ion transport. This reduction in direct contact led to decreased generation of P-containing species and enhanced battery performance and stability (Fig. S18, ESI†).

To investigate the wettability of the Bare-Li and DFFSA–Li surfaces in carbonate-based electrolytes, contact angle measurements were conducted (Fig. 5a). The contact angle measuring 38.9° on the Bare-Li surface was significantly higher compared to that on the DFFSA–Li surface, which was 28.2°. This suggests that DFFSA–Li exhibits better wettability than Bare-Li. This can be attributed to the presence of polar groups containing organic lithium carboxylates on the surface of DFFSA–Li, which facilitates interaction with polar electrolytes.²⁶ Therefore, DFFSA treatment enhances the affinity between the Li surface and the ester-based electrolyte. Higher affinity and wettability enable efficient Li-ion transport and uniform distribution near the Li anode, thereby reducing the interfacial impedance of the batteries. The Tafel curves of Bare-Li and DFFSA–Li symmetric cells were also measured, and the exchange current density (j₀) was calculated based on the Tafel curves (Fig. 5b). The j₀ value of DFFSA–Li (0.23 mA cm⁻²) was significantly higher than that of Bare-Li (0.0258 mA cm⁻²). A higher j₀ value indicates faster Li-ion conduction at the electrode–electrolyte interface,^27,28 further confirming the advantages of the artificial SEI formed by the DFFSA. EIS tests of symmetric batteries were conducted at different temperatures to calculate the activation energy for Li deposition on the DFFSA–Li and Bare-Li surfaces (Fig. 5c–e). The activation energy on the DFFSA–Li surface (53.18 kJ mol⁻¹) was lower than that of the Bare-Li (59.11 kJ mol⁻¹), reflecting a lower energy barrier for Li deposition on the DFFSA–Li surface, enabling faster and more uniform Li deposition.

Fig. 5 (a) Contact angles of Bare-Li and DFFSA–Li. (b) Tafel plots of DFFSA–Li and Bare-Li. EIS plots of the Li||Li symmetric cells containing (c) Bare-Li and (d) DFFSA–Li at different temperatures before cycling. (e) Activation energy (E_a) of DFFSA–Li and Bare-Li. (f) Comparison of Li-ion diffusion energy barriers for different SEI components. Top view of the Li migration pathways on the (g) LiF(001), (h) Li₂S(111), (i) Li₂O(111) and (j) Li₂CO₃(001) surfaces. The energy profile for Li ion diffusion on (k) LiF (001), (l) Li₂S(111), (m) Li₂O(111) and (n) Li₂CO₃(001) surfaces using the CI-NEB method.

DFT calculations were performed to gain insights into the underlying mechanism behind the fast kinetics enabled by the multi-component SEI on the DFFSA–Li anode. The main SEI components on the DFFSA–Li surface are LiF and Li₂S, while those on the Bare-Li surface are primarily Li₂CO₃ and Li₂O. Therefore, DFT calculations were focused on the diffusion pathways (Fig. 5g–j) and diffusion energy barriers of Li⁺ on the surfaces of LiF(001), Li₂S(111), Li₂O(111) and Li₂CO₃(001) (Fig. 5k–n) (the crystal plane selected is the dominant crystal plane). As shown in Fig. 5f, Li₂S(111) exhibited the lowest diffusion energy barrier of 0.1202 eV, followed by LiF(001) at 0.1303 eV. The diffusion energy barriers on the surfaces of Li₂O(111) and Li₂CO₃(001) were relatively high, at 0.229 eV and 0.3988 eV, respectively. The lower diffusion energy barrier enables the SEI rich in LiF/Li₂S to not only provide sufficiently fast Li-ion transport kinetics, enabling the Li anode to maintain good electrochemical performance even under higher current densities, thereby improving the battery life and rate performance, but also promote more uniform dispersion of Li⁺ on the Li anode, inhibiting the growth of Li dendrites. This is consistent with the experimental results.

Electrochemical performance of DFFSA–Li

To investigate the improvement of DFFSA treatment on the cycling stability of Li anodes in full batteries, DFFSA–Li was coupled with the LCO cathode (10.7 mg cm⁻²) to assemble Li||LCO coin cells and cycled under conditions of 0.3C charging and 1C discharging (1C = 178 mA g⁻¹), as shown in Fig. 6a–c. The initial discharge capacity of the DFFSA–Li||LCO cell was 153.98 mA h g⁻¹, and it remained at 119.88 mA h g⁻¹ over 1200 cycles, achieving a high-capacity retention rate of 77.85%. Compared to recent publications on the Li anode, the DFFSA–Li anode exhibits a significant advantage in terms of long-term cycle retention (Fig. 6b). In contrast, the Bare-Li||LCO cell experienced a rapid decline in capacity from 145.7 mA h g⁻¹ to 79.7 mA h g⁻¹ after only 530 cycles, and the battery suffered a sudden failure during the 532 cycle. Moreover, the DFFSA–Li cell demonstrated a superior ability to maintain its original potential distribution in the charge–discharge curves (Fig. 6d) compared to the Bare-Li (Fig. 6c). This was primarily attributed to the improved interfacial stability of the DFFSA–Li anode containing a multicomponent SEI. Conversely, the unstable interface of Bare-Li accelerated the growth of Li dendrites, leading to a rapid decline in battery performance. Additionally, the rate performance of the Li||LCO cells was tested. As shown in Fig. 6e, the DFFSA–Li||LCO cell exhibited superior rate performance compared to the Bare-Li||LCO cell at rates ranging from 0.2C to 5C. Even under the high-rate charging and discharging conditions of 5C, the DFFSA–Li cell maintained a capacity of 86.67 mA h g⁻¹. This indicates that the DFFSA–Li anode exhibits a higher rate performance, enabling rapid Li-ion transport and diffusion across the multi-component SEI even under conditions of high current density.

Fig. 6 (a) Long-term cycling performance of the coin cells with Bare-Li and DFFSA–Li at 1/3C charge and 1C discharge. (b) Comparison of the lifespans and area capacity retention of the coin cells with those reported in recent studies.^23,29–37 Charge–discharge curves of the coin cell taken at the 100th, 200th, 300th, 400th, and 500th cycles with (c) Bare-Li and (d) DFFSA–Li. (e) Rate capability of coin cells with DFFSA–Li and Bare-Li for cycling from 0.2 to 5C. (f) and (g) Cycling performance of the Li||LCO pouch cells under high cathode loading conditions. (h) A schematic illustration of fabricated 5.49 A h pouch cell under practical conditions. (i) The pie chart of the weight distribution of all components in the pouch cell. (j) Cycling performance of the 5.49 A h Li||NCM811 pouch cell at 0.2C charge/0.5C discharge. (k) Comparison of Li||NCM811 pouch cell performance with recently published studies.^32,38–47

The dilute DFFSA treatment boasts the advantages of simplicity and scalability. To further evaluate the application potential of DFFSA–Li anodes in high-capacity Li metal pouch batteries, pouch cells were fabricated using ultra-high areal loading LCO cathodes (49.7 mg cm⁻²) and thinner Li anodes (100 μm, 50 μm). The detailed pouch cell parameters are listed in Table S1 (ESI†). Compared to Bare-Li, the cycling stability of the pouch batteries assembled with DFFSA–Li was significantly improved. The pouch cell with 100 μm DFFSA–Li demonstrated a stable cycling performance over 110 cycles with 83.1% capacity retention. In contrast, the capacity of the pouch cell with 100 μm Bare-Li began to rapidly decay after only 19 cycles, and decreased to 23 mA h g⁻¹ after 65 cycles (Fig. 6f). Similarly, the pouch cell with 50 μm DFFSA–Li maintained capacity retention of 85.3% after 70 cycles, whereas the pouch cell with 50 μm Bare-Li suffered rapid capacity decay after only 40 cycles (Fig. 6g). These results demonstrate that the application of dilute DFFSA to metallic Li can effectively improve the cycling stability of practical Li batteries.

To further validate the cycling stability of DFFSA–Li anodes in high-energy-density Li metal batteries, pouch cells with a capacity of 5.49 A h were fabricated using high-specific-capacity NCM811 (39 mg cm⁻², 3.9 mA h cm⁻² for each side) and ultra-thin DFFSA–Li foils (50 μm, 5 mA h cm⁻² for each side) (Fig. 6h), giving an N/P ratio of only 1.28. Meanwhile, the electrolyte to capacity ratio was controlled at 1.6 g A h⁻¹. Detailed pouch cell parameters, including the Al foil, separator, package, and taps, are presented in Fig. 6i and Table S2 (ESI†). The pouch cells were initially activated at 0.1C for two cycles and then cycled at 0.2C charging and 0.5C discharging; they exhibited an initial discharge capacity of 21.84 W h, with an overall weight (including the aluminum-plastic film packaging and tabs) of 47.29 g (Fig. S19, ESI†), resulting in a practical energy density of 461.6 W h kg⁻¹ (specific calculation details are provided in Fig. 6j and Table S2, ESI†). After 90 cycles, the discharge capacity was 4.9 A h with a capacity retention rate of 92.3%, corresponding to a capacity decay rate of 0.086% per cycle. After 100 cycles, the capacity retention was 84.7%. Notably, the high energy density (461.6 W h kg⁻¹) and excellent stability (0.086% decay rate) of our DFFSA–Li||NCM811 pouch cell outperform most of the recently reported Li metal pouch cells (Fig. 6k) under harsh testing conditions. The above findings demonstrated the tremendous potential of economical and simple DFFSA dilute solution treatment for achieving stable LMBs for practical Li metal pouch cells.

3. Conclusion

In conclusion, our research validated the ability of a dilute DFFSA solution to remove the passivating layer from the Li surface while concurrently constructing an organic–inorganic hybrid SEI rich in LiF, Li₂S and Li₂SO₃, effectively enhancing the cycling stability of LMAs. The protective interphase generated after diluted DFFSA treatment improves the chemical affinity between the Li surface and electrolyte, regulates the Li deposition behavior, and suppresses the formation of Li dendrites. Comprehensive theoretical calculations and experimental validation confirmed the effective enhancement of Li⁺ transport kinetics facilitated by the multicomponent gradient artificial SEI. Notably, DFFSA–Li exhibits excellent electrochemical performance, with a capacity retention of 77.85% after 1200 cycles in the Li||LCO full cell using ester-based electrolyte. When coupled with the 39 mg cm⁻² NCM811 cathode, the 5.49 A h DFFSA–Li||NCM811 pouch cell delivers an impressive energy density of 461.6 W h kg⁻¹ with a capacity decay rate of only 0.086% per cycle. This simple and feasible approach provides a promising strategy for improving the cycling stability of practical high-energy LMBs.

4. Experimental section

Preparation of the DFFSA–Li anode

DFFSA (50 mg) was stirred overnight in tetrahydrofuran (THF) (10 g). Subsequently, 50 μL of the above DFFSA solution was dropped on the surface of Li foil (15.6 mm diameter) and left at room temperature for 12 hours to evaporate the THF to obtain the DFFSA–Li anode. The preparation method of the DFFSA–Li electrode in the half-cell is the same as that of the whole cell test, but the former uses 450 μm thick Li foil, and the latter uses 50 μm or 100 μm thick Li foil, and the solution amount is amplified according to the area ratio. The entire process is carried out in a glove box (O₂ < 0.01 ppm, H₂O < 0.01 ppm).

Battery assembly and testing

Coin cells are assembled in a high-purity argon glove box. The LCO and NCM811 cathodes used in the experiment were purchased from Guangdong Canrd New Energy Technology Co., Ltd. All coin cells were CR2016 with the prepared DFFSA–Li (15.6 mm diameter) as the anode, 1M LiPF₆ in EC and DEC (v/v = 3 : 7) with 10% FEC and Celgard separator (19 mm diameter) as the electrolyte and separator, respectively, with a coin cell electrolyte dosage of 50 μL. The constant current charge/discharge test is performed within the voltage window of 3.0–4.4 V. The pouch cell is assembled in a dry room with a dew point temperature below −56 °C. The cathode, anode and separator are assembled using a lamination method. The manufacturing process includes ultrasonic welding, packaging, and vacuum sealing with electrolyte injection. The rest of the conditions are the same as for the coin cell; the cycle rate is 0.2C charge/0.5C discharge.

Asymmetric Li||Cu battery method: The Li||Cu battery uses Cu foil with a diameter of 19 mm as the cathode. It is charged at 1 mA cm⁻² to a capacity of 1 mA h cm⁻², and then discharged to a cut-off voltage of 1 V. The “reservoir” method: a button battery is assembled with 50 μm Li foil (Q_Li = 10 mA h cm⁻²) as the positive electrode and a 450 μm Li sheet as the negative electrode. Ten cycles are conducted at a current density of 1 mA cm⁻² and a capacity of 3 mA h cm⁻² (Q_C). Finally, it is charged to a cut-off voltage of 1.5 V, and the remaining capacity after cycling is Q_S. The formula for calculating CE is as follows:

Tafel curve. The Tafel curve is measured by an electrochemical workstation with a voltage range from −0.16 V to 0.16 V and a scan rate of 1.0 mV S⁻¹. The exchange current density was obtained using the Tafel equation

η = A(log j − log j₀)

where η and A are the overpotential and the kinetic constant, respectively. j₀ represents the exchange current density, the value of which can be obtained from the intersection of the extrapolated linear part of the log j versus η plot with the η = 0 line.

Li deposition activation energy. Assemble Li||Li symmetric cells were assembled, and EIS tests were conducted at different temperatures. The formula for calculating the activation energy is:
where R_ct, A₀, R, and E_a represent the charge-transfer resistance, pre-exponential constant, the standard gas constant and the activation energy, respectively. Therefore, E_a can be extracted from the slope of the plot of log R_ctvs. 1/T.

Material characterization

Infrared spectra were recorded on a Fourier transform infrared spectroscope (FTIR, Nicolet IS50) at a wavelength of 4000–400 cm⁻¹. JSM-7610FPlus(Jeol) SEM was used to characterize the morphology of Li. Time-of-flight secondary ion mass spectrometry was conducted on an ION-TOF TOF.SIMS5 (30 keV, 0.75 pA, ion species: Bi³⁺⁺). An in situ optical microscope (Olympus, DSX1000) was used to observe and record the Li deposition process. The contact angles of DFFSA–Li and Bare-Li were determined using a SINDIN SDC-350 fully automatic contact angle measuring instrument. The X-ray diffractometer (XPS) was a PHI Versa Probe 4. X-ray diffraction (XRD) analysis was conducted on a Rigaku X-ray diffractometer (XRD, Cu Kα radiation). Li₂O and Li₂CO₃ were individually dispersed in minimal quantities of tetrahydrofuran (THF), with the subsequent addition of stoichiometric DFFSA to facilitate the reaction. The dried reaction products were then subjected to XRD analysis, scanning at 10° min⁻¹ with the 2θ range of 10–80° to elucidate their crystalline structures.

DFT calculations

Reaction pathway calculations. The reaction pathways calculations for the reduction of DFFSA were carried out in the ORCA program⁴⁸ using Becke's three-parameter hybrid method and the Lee–Yang–Parr correlation functional (B3LYP)⁴⁹ at the 6-311+G(d,p) level along with the D3 dispersion correction. The solvent–solute interaction was considered using the universal solvation model of SMD.⁵⁰ Frequency analysis was performed to ensure the ground state of molecular structures.

Diffusion energy barrier calculations. The calculations of Li ion diffusion behaviors on the surface were carried out using spin-polarized DFT with the generalized gradient approximation(GGA) and Perdew–Burke–Ernzerhof (PBE) as implemented in the Vienna ab initio simulation package (VASP).⁵¹ The plane-wave energy cutoff was set to 500 eV for all calculations. The convergence threshold was set at 10⁻⁵ eV for the iteration in the self-consistent field (SCF) and 0.02 eV Å⁻¹ for the maximum force component. The van der Waals interactions were described using empirical corrections in the DFT-D3. The supercell sizes were set to , 1 × 2 × 2, 1 × 2 × 2, and 1 × 2 × 2 for LiF(001), Li₂S(111), Li₂O(111) and Li₂CO₃(001) slab models, respectively. We utilized a 3 × 3 × 1 Monkhorst–Pack k-point mesh for the LiF(001), Li₂S(111), Li₂O(111) and Li₂CO₃(001) surface diffusions. To avoid articular interaction between neighboring images, a vacuum spacing of more than 25 Å was introduced in the surface diffusion calculations. The climbing image-nudged elastic band (CI-NEB) method⁵² was employed to determine the energy barrier for Li ion diffusion.

Author contributions

S. Liu conceived and designed this work and wrote the paper. M. Pang conducted all the experiments and co-wrote the paper. Y. Li, W. Sun and G. Zhou participated in the analysis of the experimental data, theoretical calculations, discussions of the results and the editing of the manuscript. Z. Jiang and C. Luo conducted part of the battery assembly and testing studies. Z. Yao, T. Fu, T. Pan, Q. Guo, S. Xiong and C. Zheng participated in editing of the manuscript. The manuscript was written through contributions from all authors. All authors have given approval to the final version of the manuscript.

Data availability

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

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

This work was financially supported by the Hunan Provincial Natural Science Foundation (Grant No. 2022JJ30663, 2022JJ40551) and the National Natural Science Foundation of China (Grant No. 51702362).

References

1. X. Shen, H. Liu, X.-B. Cheng, C. Yan and J.-Q. Huang, Energy Stor. Mater., 2018, 12, 161–175.
2. F. Shi, A. Pei, A. Vailionis, J. Xie, B. Liu, J. Zhao, Y. Gong and Y. Cui, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, 12138–12143.
3. W. Sun, S. Liu, Y. Li, D. Wang, Q. Guo, X. Hong, K. Xie, Z. Ma, C. Zheng and S. Xiong, Adv. Funct. Mater., 2022, 32, 2205471.
4. X. Zheng, L. Huang, X. Ye, J. Zhang, F. Min, W. Luo and Y. Huang, Chem, 2021, 7, 2312–2346.
5. Y. Gu, W.-W. Wang, Y.-J. Li, Q.-H. Wu, S. Tang, J.-W. Yan, M.-S. Zheng, D.-Y. Wu, C.-H. Fan, W.-Q. Hu, Z.-B. Chen, Y. Fang, Q.-H. Zhang, Q.-F. Dong and B.-W. Mao, Nat. Commun., 2018, 9, 1339.
6. Y. Xie, Y. Huang, Y. Zhang, T. Wu, S. Liu, M. Sun, B. Lee, Z. Lin, H. Chen, P. Dai, Z. Huang, J. Yang, C. Shi, D. Wu, L. Huang, Y. Hua, C. Wang and S. Sun, Nat. Commun., 2023, 14, 2883.
7. X. Yao, J. Wang, S. Lin, C. Tao, X. Zhang, W. Wang, C. Zhao, L. Wang, J. L. Bao, Y. Wang and T. Liu, Adv. Energy Mater., 2023, 13, 2203233.
8. H. Li, F. Zhang, W. Wei, X. Zhao, H. Dong, C. Yan, H. Jiang, Y. Sang, H. Chen, H. Liu and S. Wang, Adv. Energy Mater., 2023, 13, 2301023.
9. X. Liu, J. Liu, T. Qian, H. Chen and C. Yan, Adv. Mater., 2020, 32, 1902724.
10. J. Yang, J. Hou, Z. Fang, K. Kashif, C. Chen, X. Li, H. Zhou, S. Zhang, T. Feng, Z. Xu and M. Wu, Chem. Eng. J., 2022, 433, 133193.
11. Y. Li, M. Liu, K. Wang, C. Li, Y. Lu, A. Choudhary, T. Ottley, D. Bedrov, L. Xing and W. Li, Adv. Energy Mater., 2023, 13, 2300918.
12. X. Ma, J. Yu, X. Zou, Y. Hu, M. Yang, F. Zhang and F. Yan, Cell Rep. Phys. Sci., 2023, 4, 101379.
13. J. Zhang, X. Yue, Z. Wu, Y. Chen, Y. Bai, K. Sun, Z. Wang and Z. Liang, Nano Lett., 2023, 23, 9609–9617.
14. P. Zhou, Y. Xia, W.-H. Hou, S. Yan, H.-Y. Zhou, W. Zhang, Y. Lu, P. Wang and K. Liu, Nano Lett., 2022, 22, 5936–5943.
15. S. Yan, F. Liu, Y. Ou, H.-Y. Zhou, Y. Lu, W. Hou, Q. Cao, H. Liu, P. Zhou and K. Liu, ACS Nano, 2023, 17, 19398–19409.
16. M. Baek, J. Kim, K. Jeong, S. Yang, H. Kim, J. Lee, M. Kim, K. J. Kim and J. W. Choi, Nat. Commun., 2023, 14, 1296.
17. H. Chen, A. Pei, D. Lin, J. Xie, A. Yang, J. Xu, K. Lin, J. Wang, H. Wang, F. Shi, D. Boyle and Y. Cui, Adv. Energy Mater., 2019, 9, 1900858.
18. W. Xu, J. Wang, F. Ding, X. Chen, E. Nasybulin, Y. Zhang and J.-G. Zhang, Energy Environ. Sci., 2014, 7, 513–537.
19. D. Lin, Y. Liu and Y. Cui, Nat. Nanotechnol., 2017, 12, 194–206.
20. W. Yu, Z. Yu, Y. Cui and Z. Bao, ACS Energy Lett., 2022, 7, 3270–3275.
21. K. S. Jiang, G. M. Hobold, R. Guo, K.-H. Kim, A. M. Melemed, D. Wang, L. Zuin and B. M. Gallant, ACS Energy Lett., 2022, 7, 3378–3385.
22. A. Mohammadi, S. Djafer, S. Sayegh, A. J. Naylor, M. Bechelany, R. Younesi, L. Monconduit and L. Stievano, Chem. Mater., 2023, 35, 2381–2393.
23. S. Fang, F. Wu, S. Zhao, M. Zarrabeitia, G. T. Kim, J. K. Kim, N. Zhou and S. Passerini, Adv. Energy Mater., 2023, 13, 2302577.
24. H. Chen, J. Liu, X. Zhou, H. Ji, S. Liu, M. Wang, T. Qian and C. Yan, Chem. Eng. J., 2021, 404, 126470.
25. X. Li, J. Liu, J. He, H. Wang, S. Qi, D. Wu, J. Huang, F. Li, W. Hu and J. Ma, Adv. Funct. Mater., 2021, 31, 2104395.
26. R. Pathak, K. Chen, A. Gurung, K. M. Reza, B. Bahrami, F. Wu, A. Chaudhary, N. Ghimire, B. Zhou, W. H. Zhang, Y. Zhou and Q. Qiao, Adv. Energy Mater., 2019, 9, 1901486.
27. S. Huang, K. Long, Y. Chen, T. Naren, P. Qing, X. Ji, W. Wei, Z. Wu and L. Chen, Nano-Micro Lett., 2023, 15, 235.
28. X. He, Z. Liu, Y. Yang, Z. Wang, Y. Chen, Q. Zhang, Z. Shi, Y. Tan, X. Yue and Z. Liang, J. Mater. Chem. A, 2023, 11, 10155–10163.
29. K. H. Han, J. Y. Seok, I. H. Kim, K. Woo, J. H. Kim, G. G. Yang, H. J. Choi, S. Kwon, E. I. Jung and S. O. Kim, Adv. Mater., 2022, 34, e2203992.
30. Y. Shi, Q. Li, X. Hu, Y. Liao, W. Li, Y. Xu, B. Zhao, J. Zhang and Y. Jiang, Nano Energy, 2024, 120, 109150.
31. Z. Sun, Y. Wang, S. Shen, X. Li, X. Hu, M. Hu, Y. Su, S. Ding and C. Xiao, Angew. Chem., Int. Ed., 2023, 62, e202309622.
32. Y. Zhang, P. Zhao, Q. Nie, Y. Li, R. Guo, Y. Hong, J. Deng and J. Song, Adv. Mater., 2023, 35, e2211032.
33. Y. Gao, Z. Yan, J. L. Gray, X. He, D. Wang, T. Chen, Q. Huang, Y. C. Li, H. Wang, S. H. Kim, T. E. Mallouk and D. Wang, Nat. Mater., 2019, 18, 384–389.
34. M. H. Kim, T. U. Wi, J. Seo, A. Choi, S. Ko, J. Kim, U. Jung, M. S. Kim, C. Park, S. Jin and H. W. Lee, Nano Lett., 2023, 23, 3582–3591.
35. A. Bhargav, H. Y. Asl and A. Manthiram, J. Mater. Chem. A, 2023, 11, 9772–9783.
36. Y. Liang, C. Shen, H. Liu, C. Wang, D. Li, X. Zhao and L. Z. Fan, Adv. Sci., 2023, 10, 2300985.
37. X.-B. Cheng, S.-J. Yang, Z. Liu, J.-X. Guo, F.-N. Jiang, F. Jiang, X. Xiong, W. Bo Tang, H. Yuan, J.-Q. Huang, Y. Wu and Q. Zhang, Adv. Mater., 2023, 36, e2307370.
38. Q. Kang, Y. Li, Z. Zhuang, H. Yang, L. Luo, J. Xu, J. Wang, Q. Guan, H. Zhu, Y. Zuo, D. Wang, F. Pei, L. Ma, J. Zhao, P. Li, Y. Lin, Y. Liu, K. Shi, H. Li, Y. Zhu, J. Chen, F. Liu, G. Wu, J. Yang, P. Jiang and X. Huang, Adv. Mater., 2024, 36, 2308799.
39. W. Chen, Y. Hu, Y. Liu, S. Wang, A. Hu, T. Lei, Y. Li, P. Li, D. Chen, L. Xia, L. Xue, Y. Yan, G. Lu, M. Zhou, Y. Fan, H. Yang, X. Tao, X. Wang, Y. Li and J. Xiong, Adv. Mater., 2024, 36, 2312880.
40. X. Xie, L. Wei, J. Lu, A. Xu, B. Wang, X. Xiao, A. Wang, Z. Jin, Z. Shi and W. Wang, Energy Storage Mater., 2024, 67, 103323.
41. L.-P. Hou, X.-Y. Li, C.-X. Bi, Z.-X. Chen, Z. Li, L.-L. Su, P. Shi, C.-B. Jin, B.-Q. Li, J.-Q. Huang, X.-Q. Zhang and Q. Zhang, J. Power Sources, 2022, 550, 232144.
42. H. Y. Zhi Chang, Anqiang Pan, Ping He and Haoshen Zhou, Nat. Commun., 2022, 13, 6788.
43. D. Li, H. Liu, C. Wang, C. Yan, Q. Zhang, C. W. Nan and L. Z. Fan, Adv. Funct. Mater., 2024, 34, 2315555.
44. Y. H. Tan, Z. Liu, J. H. Zheng, Z. J. Ju, X. Y. He, W. Hao, Y. C. Wu, W. S. Xu, H. J. Zhang, G. Q. Li, L. S. Zhou, F. Zhou, X. Tao, H. B. Yao and Z. Liang, Adv. Mater., 2024, 2404815.
45. A. Wang, Y. Nie, Y. Zhao, D. Xu, L. Zhang, Z. Zhao, L. Ren, S. Zhou, X. Liu and J. Luo, Adv. Funct. Mater., 2024, 2401462.
46. Z. Wu, C. Wang, Z. Hui, H. Liu, S. Wang, S. Yu, X. Xing, J. Holoubek, Q. Miao, H. L. Xin and P. Liu, Nat. Energy, 2023, 8, 340–350.
47. P. Zhao, Y. Li, S. Chen, H. Fan, Y. Feng, L. Hu, Y. Zhang, Q. Nie, H. Pei, C. Yang, J. Deng, C. Bao and J. Song, Adv. Energy Mater., 2022, 12, 2200568.
48. F. Neese, Wires Comput. Mol. Sci., 2011, 2, 73–78.
49. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
50. C. J. C. A. D. G. T. Aleksandr and V. Marenich, J. Phys. Chem. B, 2009, 113, 6378–6396.
51. G. Kresse, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
52. G. Henkelman, B. P. Uberuaga and H. Jónsson, J. Chem. Phys., 2000, 113, 9901–9904.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ee02311k

---