Synthesis and characterization
At low SEM magnification, the q2D-cMOF self-assembles into uniform two-dimensional sheet morphologies (Fig. 1b). This quasi-two-dimensional structure effectively exposes the surface of the nanocrystals, which is beneficial for providing more channels for charge transport and adsorbing additional ions, thereby enhancing the transmission of fluoride and Li+ due to the larger contact area. MOF nanoparticles are tightly bound to each other through van der Waals forces. Notably, the crystal size of q2D-cZ8 is a regular dodecahedron with a particle size of about approximately 100 nm (Fig. 1c). The crystal size of q2D-cZ67 and q2D-cU66 are hexahedrons and octahedrons, measuring 150 nm and 200 nm, respectively (Fig. S2a-d). q2D-cNiMOF displays a lamellar microstructure with a diameter of 300 nm (Fig. S2e-f). Despite the diversity in coordination monomers and metal nodes, and the variability in particle size, the q2D-cMOF orderly self-assembles into sheet-like quasi-two-dimensional structures under consistent freeze-drying conditions. The growth mechanism of q2D-MOF is characterized as the nanoparticle extrusion process driven by the ice template, which is elaborated on in the supporting information. TEM images further confirmed the regular hexagonal structure of q2D-cZ8 under plane projection (Fig. 1d). Complete quadrilaterals, diamonds, and rectangles under the plane projection were observed for q2D-cZ67, q2D-cU66, and q2D-cNiMOF (Fig. S3). High-resolution TEM images reveal nanocrystalline regions with distinct lattice fringes, corresponding well with the distribution of the respective metal elements. This indicates that metal atoms are interlinked with the carbon framework, preserving the quasi-two-dimensional structure (Fig. S4).
The morphology of q2D-FcMOF undergoes noticeable alterations due to corrosion by F2. The surface becomes rougher with folds emerging. The inherent regular dodecahedral and hexahedral structures of q2D-FcZ8 and q2D-FcZ67 are disrupted by fluoridation, leading to the etching of clear edges (Fig. 1e, S5a). The etched edges and corners exhibit high curvature, increasing the area available for contact with active substances. The particles are adhered to each other, with the interfaces between them being occupied. q2D-cZ8 and q2D-cZ67 is lack of stability, and F2 reacts with Zn and Co as metal nodes to form metal fluoride, resulting in the contraction of the hollow structure. Nonetheless, these materials still retain the characteristics of the quasi-two-dimensional structure and the inherent microscopic size of q2D-cZ8 and q2D-cZ67. q2D-FcU66 also contracted to a certain extent, and the particle size was reduced to approximately 150 nm, but an obvious interface between particles is still retained (Fig. S5b). The q2D-FcNiMOF maintains its lamellar stacking structure, with no significant changes in the secondary lamellar size (Fig. S5c). TEM images of q2D-FcZ8 indicate an uneven distribution of particles, with layers being stacked upon each other (Fig. S6a). In high-resolution TEM images, distinct lattice fringes corresponding to the (110) and (101) crystal faces of ZnF2 are apparent. This suggests that q2D-FcZ8 is essentially a heterojunction composite structure comprising ZnF2 and a fluorinated organic carbon framework (Fig. 1f). Elemental analysis of q2D-FcZ8 reveals a uniform distribution of C, N, F, and Zn (Fig. 1g, h). TEM images of q2D-FcU66 display the retention of the boundary and appearance of secondary particles post-fluoridation, accompanied by the emergence of numerous strip structures. Lattice stripes with obvious strip structures can be observed corresponding to the (211) crystal face of ZrF4, indicating that fluorinated q2D-FcU66 is essentially a heterojunction composite structure (Fig. S6b, c).
The porous textures were measured by N2 isothermal adsorption to analyze the pore size and distribution. Fig. S7 describes the pore size distribution. q2D-cMOF has plentiful micropores, as well as pores across the range of mesopores to micropores, which are mainly derived from the cavities of carbon NPs. The specific surface area and pore size of the fluorinated products are greater than those of the precursors because F2 etched the carbon skeleton at the pore edge; thus, expanding the pore structure and providing an ordered channel for the transport of Li+ (Fig. 1i, S8). In addition, the outward contraction of quasi-two-dimensional structures during pyrolysis can prevent a decrease in microporosity and inhibit the collapse of the organic framework.
The XRD patterns of q2D-cZ8 show a broad peak at approximately 22.5°, corresponding to the diffraction of the (002) lattice planes of low-graphitic carbon. Conversely, broad reflection peaks corresponding to related metal compounds can be clearly observed in the XRD patterns of q2D-cZ67 (CoO), q2D-cU66 (ZrO2), and q2D-NiMOF (Ni2S3) owing to the peak corresponding to high-intensity metal compounds masking the carbon peak of the lower graphitic carbon (Fig. S9). The hyperactive fluorine atoms substitute for the heteroatom, and a broad peak emerges for the corresponding fluorinated metal (ZnF2, CoF2, ZrF4, and NiF2), indicating that the fluorinated metal is dispersed in the amorphous carbon skeleton and effectively hybridized (Fig. 1j, S10).
The Raman spectrum exhibits the proportions of defects, structural distortions, and edge planes in the D-band, while the G-band represents the number of layers and the integrity of the lattice [38]. Compared to those of q2D-cMOF (Fig. S11) and q2D-FcMOF (Fig. 1k), the increase in the calculated ID/IG ratio and broadening of the G-band reveal that the chemical bond formed between the C and F atoms destroys the original arrangement and inherent aromaticity with increasing disorder.
The fluorine species are further characterized by Fourier transform infrared spectroscopy. q2D-cMOF exhibited a significant signal at 1620 cm–1, corresponding to the characteristic peak of the C=C bond on the aromatic ring (Fig. S12). Further fluorination does not completely demolish the aromatic conjugated structure, but more prominent peaks appeared at 1212 cm–1, corresponding to the peak attributed to the stretching vibration of covalent C-F bonds (Fig. 2a). The bands observed near 1385 cm–1 are characteristic peaks of the -CF2 group. These perfluorinated groups are formed by the destruction of aromatic structures by fluorine radicals.
X-ray photoelectron spectroscopy survey demonstrated that the q2D-cMOF had similar concentrations of metal atoms in the range of 7–12 at% (Fig. S13a, S14). The pronounced F signal in the analysis indicates substantial F-doping, with the concentration of metal atoms decreasing to 4–7 at% (Fig. S13b, 2b). The characteristic metal peaks of the fluorinated products exhibit a significant shift towards higher binding energy. This shift occurs because the chemical environment of metal elements is altered by the transformation of M-C (M = Zn, Co, Zr, Ni) bonds into metal fluorides, which possess higher binding energy (Fig. S15). The presence of metal ions coordinated with F leads to a stronger electron absorption effect. This strong M-F bond interaction facilitates charge transfer from the metal to the matrix, creating an electron-enriched environment around adjacent C atoms. Such an environment can appropriately regulate the adsorption of Li on q2D-cMOF. The C1s spectrum displays an asymmetric peak (Fig. 2c, S16). The emergence of C-O is probably attributed to the slight oxidation of carbon materials at room temperature. The appearance of peaks corresponding to various fluorine-containing groups in q2D-FcMOF indicates an effective reaction between q2D-cMOF and F2 along with the transformation of conjugated sp2 hybrid C atoms into sp3 hybrid C atoms that bonded with F atoms. Ulteriorly, the F1s spectrum of q2D-FcMOF can be deconvoluted into semi-ionic C-F bonds (688.0 eV), covalent C-F bonds (689.1 eV) and metal-F bonds, reflecting the hybrid coexistence of ionic, semi-ionic and covalent F bonds, respectively (Fig. 2d). The N1s spectrum of q2D-cZ8, q2D-cZ67, and q2D-cNiMOF can be deconvoluted into peaks corresponding to pyridinic N (398.2 eV), pyrrolic N (399.5 eV), and graphitic N (401.0 eV). Among these, pyridinic nitrogen contributes the predominant component, aligning with the structure of the organic monomer used in the MOF synthesis. The relative contents of pyridinic N, pyrrolic N, and graphitic N in the fluorinated products remain similar to their precursors. However, a slight shift of these N peaks towards higher binding energy by ~2 eV is attributed to local environmental changes due to the presence of F atoms, which further evidence that fluorination significantly impacts the valence bond binding energy (Fig. S17).
X-ray Absorption Fine Structure (XAFS) was utilized to investigate the structural and coordination morphology [39]. The k absorption edge of q2D-cZ8 and q2D-FcZ8 lies between those of Zn and ZnO, indicating a mixed valence state of Zn elements between 0 and +2. The valence of fluorinated q2D-FcZ8 appears higher (Fig. 2e). The extended X-ray absorption fine structure (EXAFS) illustrates substantial changes in the coordination environment of Zn after extensive fluoridation (Fig. 2f). The dominant peaks of q2D-cZ8 at 1.6 Å and 2.2 Å correspond to the coordination peaks of Zn-O and Zn-Zn, respectively. In contrast, the dominant peak of q2D-FcZ8 at about 1.7 Å relates to the coordination peak of Zn-F, confirming successful fluorination [40-42]. Wavelet transform (WT) of Zn, ZnO, q2D-cZ8, and q2D-FcZ8 presented in Fig. 2g, further supports these findings. The WT contour map peak of q2D-cZ8 around 7 Å corresponds to the Zn-Zn coordination bond. Meanwhile, the prominent peak in the WT contour map of q2D-FcZ8 around 9 Å corresponds to the Zn-F bond, providing additional proof that gas-phase fluoridation alters the coordination properties of Zn [43].
Deposition-dissolution of q2D-FcMOF@Li
To explore the interaction between Li+ and q2D-FcMOF, the chemical composition and structure of the ASEI layer were determined by time-of-flight secondary ion mass spectrometry (TOF-SIMS) (Fig. 3a, S18) [44]. q2D-FcZ8@Li contains various groups according to 3D rendering. Zn, CF2–, CF3–, and ZnF2– groups, which have higher surface contents, form the outer layer of the organometallic frame. With increasing sputtering time, the contents of CF2–, CF3– and ZnF2– decreased. More LiF appears as the main component of the inner SEI layer, which contains small amounts of C2HO– and CH–.
XPS depth analysis on the surface of q2D-FcZ8@Li was conducted to scrutinize the composition and state of ASEI in detail (Fig. 3b, d) [45]. Prior to Ar+ sputtering, signals for Li-F and Li-O bonds in Li1s, C=C, semi-ionic and covalent C-F bonds in C1s, Zn-F, semi-ionic and covalent C-F bonds in F1s, and Zn2p are predominantly observed, confirming q2D-FcZ8 effective reaction with Li. As sputtering progresses, component proportions vary with the depth of sputtering etching. Initially (3–9 min), prominent signs of C=C, C-F, and Zn-F bonds are key components of the outer layer of the organometallic frame. With increased sputtering depth (12–15 min), the strength of the C-F bond peak decreases, while Li-F content sharply rises, and Zn signals gradually fade, indicating that inorganic LiF becomes the dominant component of the inner SEI layer. In the late sputtering phase (18 min), the emergence of Li0 signals suggests that Ar+ has nearly penetrated the SEI layer, exposing fresh Li metal in some areas. This trend implies that the double-layer ASEI main structural components are an outer organometallic frame layer and an inner inorganic LiF layer with a stronger affinity for Li. Surface element compositions of q2D-FcZ8@Li after 50 cycles (1 mA cm–2 @ 1 mAh cm–2) were also analyzed (Fig. 3c, e). Before Ar+ sputtering, q2D-FcZ8@Li mainly comprises Li2S, derived from slight electrolyte decomposition, with weak signals of Li-O, and stronger signals of Li-F, Zn, and C-F. As sputtering progresses, signals of LiF, Zn, and C quickly recover. The Li-F signal notably surpasses Li-O at the 9th min, and Zn is consistently present, indicating that the double-layer ASEI remains intact with excellent cyclic and structural stability. In contrast, traditional SEI layers, composed of disorganized distributions of Li-O, LiF, and Li2S, are observed at all depths, dominated by oxygenated organic components produced by severe electrolyte decomposition, generating the inevitably huge consumption of active Li (Fig. S19, S20). The results demonstrate that the organometallic frame/inorganic LiF hybrid double-layer ASEI uniformly adheres to the Li surface, enhancing the mechanical strength of the SEI layer. The outer Zn layer provides uniform Li nucleation sites, enabling smooth and reversible Li movement. The F-rich inner layer acts as an exceptional electronic insulator, creating a strong e-blocking environment that inhibits side reactions and Li dendrite formation, effectively preventing electrolyte decomposition. LiF is also a Li+ conductor [46], which can regulate the flux, facilitating rapid Li+ transfer.
Li deposition morphology is significantly influenced by varying energy barriers during nucleation and transport (Fig. 3f). After 4 h of deposition at 1 mA cm–2, bare Li exhibits coarse, unevenly distributed Li dendrites and island-like Li structures on the surface (Fig. 3f1). A loose porous layer of freshly deposited Li is confirmed by cross-sectional views (Fig. 3f4). Under identical deposition conditions, q2D-FcZ8@Li maintains a smooth, dendrite-free surface post-deposition. At higher current densities (3 mA cm–2 and 5 mA cm–2), the quantity and size of dendrites increase, showing moss-like Li deposits with uneven edges and exposing severely damaged structures and dead Li, indicating poor electrochemical stability (Fig. 3f2, f3). The absence of obvious dendrite growth on q2D-FcZ8@Li surfaces suggests that the organometallic frame/inorganic LiF hybrid double-layer ASEI effectively inhibits Li dendrite growth.
Based on the foregone electrodeposition model, the process of Li deposition and the evolution of Li+ concentration were simulated using the finite element method (Fig. 3g). Severe concentration polarization and irreversible depletion of Li+ occur in the Li deposition simulation behavior at 8 mA cm–2 for 120 min. Large dendrite Li growth starts at nucleation sites due to the tip effect, and the spontaneously formed fragile SEI layer with poor Li+ conduction ability causes an uneven electric field. The increase in the local concentration of Li+ near the deposition site is significantly driven by the accumulated negative charge, generating the formation of dead Li and capacity loss (Fig. 3g1, g2). In contrast, Li deposition appears flat deposited on q2D-FcZ8@Li. Since the diffusion coefficient of the double-layer ASEI is significantly greater than the traditional spontaneous SEI, Li+ can be rapidly conducted during continuous plating/stripping to avoid excessive local Li+ concentrations and maintain a uniform electric field at the interface, achieving uniform deposition of Li+ (Fig. 3g3, g4).
The electrochemical performances of the Li||Cu half-cells assembled by pairing Cu foils with Li are demonstrated in Fig. 4a. The overpotentials of the Li||Cu, q2D-FcZ8@Li||Cu, q2D-FcZ67@Li||Cu, q2D-FcU66@Li||Cu, q2D-FcNiMOF@Li||Cu and CFx@Li||Cu cells were 31, 10, 11, 13, 12, and 21 mV, respectively (Fig. 4a, S21). q2D-FcZ8@Li has the lowest Li deposition barrier and achieves a coulombic efficiency (CE) of 99.12% after 500 cycles. The CE of q2D-FcZ67@Li, q2D-FcU66@Li, and q2D-FcNiMOF@Li decreased to less than 95.0% after 461, 305, and 381 cycles, while that of CFx@Li and bare Li decreased to less than 95.0% after 270 and 142 cycles, revealing that the inorganic LiF layer can control the growth of Li dendrites to some extent, but the influence of the metal site and quasi-two-dimensional carbon skeleton is more remarkable. Variations in metal nodes significantly influence the activity of metal sites on Li+, thereby affecting electrochemical performance. q2D-FcZ8@Li showed the lowest polarization voltage (38 mV) during the 50th cycle (Fig. 4b), in comparison to Li||Cu at 72 mV. This is attributed to the controlled homogeneous growth of Li in the double-layer ASEI, leading to extremely flat deposition under a uniform local electric field. The voltage profiles of q2D-FcZ8@Li||Cu and q2D-FcZ67@Li||Cu remained nearly unchanged across different cycles (Fig. S22). Electrochemical impedance spectroscopy (EIS) revealed a more stable interface for q2D-FcZ8@Li, with lower impedance (Fig. S23). In contrast to the precarious interface of bare Li, the interface charge transfer resistance (Rct) of q2D-FcZ8@Li||Cu remain stable over 100 cycles, consistent with rapid Li+ migration in q2D-FcZ8@Li. When the deposition rate and capacity increased to 2 mAh cm–2 at 2 mA cm–2, q2D-FcZ8@Li||Cu demonstrated an outstanding average CE of 98.7 % after 400 cycles, markedly superior to bare Li, which fell below 95% after the 142nd cycle (Fig. 4c, S25a, b). At even higher rates of 4 mAh cm–2 at 2 mA cm–2, q2D-FcZ8@Li||Cu maintained the longest cycles and highest average CE of 98.3%, whereas the CE of bare Li dropped to 85% after merely 41 cycles (Fig. S24, S25c, d). Additionally, q2D-FcZ8@Li||Cu showed excellent electrochemical properties at 1 mAh cm–2@1 mA cm–2 using a standard carbonate electrolyte (1 M LiPF6 EC/DEC/DMC) (Fig. 4d), achieving a high coulomb efficiency of 97.9% after 300 cycles, due to the insulating LiF intermediate layer, which significantly avoid the side reactions between Li and the carbonate electrolyte. This suggests that the robust, lithiophilic, high Li+ flux ASEI can uniformly adsorb Li+ on the surface.
Long-term cyclic stability and rate performance were assessed by Li||Li symmetric cells (Fig. 4e, S26). The symmetric cell with q2D-FcZ8@Li, operating at 1 mAh cm–2 at 1 mA cm–2, showed a low hysteretic voltage of 44 mV after 3600 h, indicating low and smooth polarization. Conversely, the hysteretic voltage of bare Li showed a trend of enlargement gradually, the hysteretic voltage increases reaching 88 mV after 800 h and the cell failure occurs after 1052 h. The symmetric q2D-FcZ67@Li, q2D-FcU66@Li, and q2D-FcNiMOF@Li cells exhibited stable cycling performance, lasting for 1200 h. In contrast, the CFx@Li symmetric cell showed increasing polarization voltage after 685 h and failed after 1363 h. When the Li deposition capacity was raised to 2 mAh cm–2 at 10 mA cm–2 (Fig. S27), the symmetric Li cell fails after 147 h. The symmetric q2D-FcZ67@Li maintained excellent cycle stability with a relatively low hysteresis voltage of 398 mV, lasting up to 250 h. The hysteretic voltage of symmetric q2D-FcZ67@Li, q2D-FcU66@Li, and q2D-FcNiMOF@Li after 100 h is 504, 792, and 586 mV, respectively. The rate performance of symmetric q2D-FcZ8@Li, q2D-FcZ67@Li, and bare Li is carried out under current densities ranging from 1 to 8 mA cm–2 with a capacity of 1 mAh cm–2 (Fig. 4f). The symmetric q2D-FcZ8@Li presents inconspicuous hysteresis, with steady voltage polarization of 48, 98, 127, 191 mV at current density of 1, 2, 4, 8 mA cm–2. When the current density returns to 1 mA cm–2, the hysteresis voltage reaches 53 mV. The symmetric q2D-FcZ67@Li also has the ability to cope with step changes in the rate current. In contrast, when the current density is returned to 1 mA cm–2, the bare Li has severe voltage polarization, and the polarization voltage increases by 36%. These results imply that the double-layer ASEI can improve the deposition behavior of Li+, which effectively disperses the current density in the battery. Adequate Li+ supplementation and continued transport activate the internal active site and prevent the intensification of polarization.
To ascertain the distribution of "dead Li" and "SEI-Li+" during the long-term cycle, Li after 200 cycles was verified by titration gas chromatography (TGC) to assess the residue of active "reversible Li" (Fig. 4g) [47]. Compared to bare Li, q2D-FcZ8@Li and q2D-FcZ67@Li exhibited higher reversible capacities of 99.09% and 98.84%, respectively, and lower amounts of "dead Li" and "SEI-Li+." This indicates that the double-layer ASEI effectively enhances the active "reversible Li" in prolonged cycles and curbs side reactions. To assess the impact of the double-layer ASEI on charge transfer kinetics, symmetric cells of bare Li, q2D-FcZ8@Li, and q2D-FcZ67@Li were assembled and evaluated using Tafel plots for exchange current density (Fig. S28). q2D-FcZ8@Li displayed an enhanced current density of 0.48 mA cm–2, which is 539.3% higher than bare Li, indicating sluggish kinetics in the latter. Cyclic voltammetry (CV) tests show that q2D-FcZ8@Li and q2D-FcZ67@Li have higher current response values (Fig. S29), reflecting improved electrochemical reaction kinetics between the electrode and electrolyte, and has a reduced Li nucleation barrier and low overpotential during plating/stripping, while the high ionic conductivity accelerated Li+ transfer kinetics. The activation energy (Ea) for charge transport on double-layer ASEI was calculated by fitting data to the Arrhenius equation [48], based on electrochemical impedance spectra measured at various temperatures (Fig. 4h, S30). Temperature-dependent EIS from 298.15 K to 338.15 K revealed that both q2D-FcZ8@Li and q2D-FcZ67@Li had significantly lower interface process resistance (Ri), which includes charge transfer and SEI resistance. The Ea values for q2D-FcZ8@Li and q2D-FcZ67@Li were 0.48 eV and 0.52 eV, respectively, lower than bare Li (0.58 eV), supporting the benefits of double-layer ASEI in stabilizing the Li/electrolyte interface and enhancing Li+ transport; thus, improving electrode stability and Li stripping/plating kinetics.
Electrochemical performance of high-energy LMBs with q2D-FcMOF@Li
The q2D-FcMOF@Li was coupled with LFP cathode to assemble the LMB, evaluating its potential in full batteries (Fig. 5a, S31). The q2D-FcZ8@Li||LFP achieved excellent cycle performance with an initial capacity of 143.24 mAh g–1 at 1 C and maintains a capacity retention of 86.37% over 600 cycles with ultrahigh CE (99.7%). Under similar circumstances, the q2D-FcZ67@Li||LFP cell provides an initial capacity of 140.83 mAh g-–1 and maintains a capacity retention of 74.64% over 600 cycles with an ultrahigh CE of 99.3%. In contrast, the specific capacity of Li||LFP, q2D-FcU66@Li||LFP, and q2D-FcNiMOF@Li||LFP decreased to 100 mAh g–1 after 288, 396, and 470 cycles, respectively. Polarization comparisons of various electrodes at the 50th and 200th cycles (Fig. S32) showed that q2D-FcZ8@Li||LFP and q2D-FcZ67@Li||LFP remained stable over long cycles, with minimal polarization increase, unlike Li||LFP and CFx@Li||LFP, which exhibited significant polarization increases after prolonged cycling (Fig. S33). The practical application rate performance of q2D-FcZ8@Li||LFP was demonstrated in Fig. 5b and S34, showing a high discharge capacity of 155.56 mAh g–1 at 0.1 C, and maintaining 115.65 mAh g–1 at a high rate of 5 C, with only a slight increase in voltage polarization (Fig. S35). The Li||LFP displayed a discharge capacity of 150.01 mAh g–1 at 0.1 C, but significantly decreased at 5 C (54.05 mAh g–1), aligning with the poor morphological evolution of Li deposition.
In situ infrared spectroscopy was utilized to monitor dynamic changes in electrolyte composition (Fig. 5c). As charging progressed, distinct double peaks emerged in the 2300 ~ 2387 cm–1 band, associated with the -O-C=O group formed from electrolyte decomposition (Fig. 5d). The peak intensity eventually stabilized, indicating effective inhibition of electrolyte decomposition in subsequent electrochemical processes, thereby protecting the Li/electrolyte interface via the double-layer ASEI.
CV curves are employed to confirm that the SEI layer enhances the kinetics of Li+ transfer at the interface (Fig. 5e, S36). Li||LFP exhibits a pair of redox peaks corresponding to the Fe2+/Fe3+ redox couple. In comparison to pure Li, q2D-FcZ8@Li||LFP displays a stronger redox peak and a smaller voltage difference between the reduction peak and the oxidation peak (ΔV=0.152 V). Conversely, the CFx@Li||LFP interface demonstrates the poorest Li+ transfer kinetics due to CFx's limited electrical conductivity, which impedes the Li+ transfer rate. The galvanostatic intermittent titration technique (GITT) was used to determine the Li+ diffusion coefficients (Fig. S37). The functional relationship between the electrochemical diffusion coefficient of Li+ and the voltage during the charge/discharge process was calculated. The DLi+ value gradually increases during the initial charging stage. However, after reaching the order of 10-9, DLi+ begins to decrease due to the accumulation of abundant polyphase mixed crystals in the positive layer, which hinders Li+ migration. During the discharge process, DLi+ continues to rise as the cathode continuously lithiates, forming a phase that facilitates Li+ conduction with a shorter Li+ diffusion path. Overall, the diffusion coefficient of Li+ in q2D-FcZ8@Li||LFP is greater than that in Li||LFP, demonstrating that q2D-FcZ8@Li||LFP possesses better reaction kinetics and accelerates the diffusion of Li+.
To further examine the impact of cathode mass loading and electrolyte volume on cycle efficiency, cells were constructed and tested under conditions of high cathode mass loading (~15 mg cm–2) and lean electrolyte (E/C = 18 g Ah–1). The q2D-FcZ8@Li||LFP and q2D-FcZ67Li||LFP cells demonstrated 200 stable cycles, as depicted in Fig. 5f, with capacity retention rates of 94.78% and 81.07%, respectively. Conversely, the capacity of bare Li declined rapidly after 88 cycles, with a capacity retention rate of only 5.26% after 100 cycles. When q2D-FcZ8@Li was paired with an NCM811 cathode, the q2D-FcZ8@Li||NCM811 exhibited superior electrochemical performance, as shown in Fig. 5g, S38. Over 150 cycles at 0.5 C, q2D-FcZ8@Li maintained an initial capacity of 169.47 mAh g–1, a capacity retention rate of 81.88%, and an average CE of 99.17%.
Air stability of q2D-FcMOF@Li
The double-layer ASEI on Li also contributes to preventing air penetration. Optical photographs presented in Fig. S39 reveal the corrosion degree of q2D-FcZ8@Li and Li when exposed to air (average humidity ≈ 30%). Bare Li gradually lost metallic luster after 30 min of exposure from inert gas to air and turned completely gray and black with a rough texture after 2 hours, whereas q2D-FcZ8@Li showed no significant change in color, shape, or texture after being exposed to air for 24 hours. XPS analysis illustrated that each peak of bare Li exposed to air underwent displacement and intensity changes to varying extents. Unprotected Li reacted with air to form LiOH, leading to a significant increase in oxygen content and further reacting with N2 to form Li3N, while the composition of q2D-FcZ8@Li exhibited remarkable air stability (Fig. 5h, S40–43).
Additionally, the electrochemical properties of Li after air exposure were examined. Fig. 5i displays typical electrochemical stripping curves at 1 mA cm–2 for bare Li and q2D-FcZ8@Li. The initial specific capacity of bare Li was 3439.90 mAh g–2, which decreased to 2345.11 mAh g–2 after 1 hour, indicating almost no electrochemical activity after 2 hours. In contrast, q2D-FcZ8@Li maintained an inherent specific capacity of 3274.34 mAh g–2 and a capacity retention rate of 91.69% after 2 h. The charge transfer resistance of q2D-FcZ8@Li increased slightly from 113 to 2098 Ω after 2 h of exposure to air, while that of bare Li increased from 79 to 223 Ω (Fig. 5j). The plating/stripping process of symmetric Li is proposed in Fig. S44, showing that bare Li-1h maintained cycle stability for only 231 hours, whereas q2D-FcZ8@Li-2h has cycle stably for 600 hours. When paired with LFP, the capacities of Li-1h||LFP significantly degraded, while q2D-FcZ8@Li-2h achieved a stable 400 cycles with a capacity retention rate of 85.51% (Fig. S45). These results indicate that exposure to air generates abundant corrosion products, hindering electron and Li+ transport and reducing the capacity of Li. Unprotected Li is unable to effectively resist corrosion; thus, significantly increasing the risk of Li storage in an aerobic state, but the double-layer ASEI effectively serves as a barrier against air to prevent the reaction of active Li.