MOF liquid-infusion preparing MOF Glass infused high-voltage cathode
In their previous studies, researchers demonstrated that coating electrode surfaces with porous materials featuring sub-nanochannels can facilitate the pre-desolvation of Li-ions39,40,41. This approach effectively suppresses electrolyte decomposition and prevents the co-intercalation of solvated Li-ions into electrode materials. Yet, these benefits were achieved only on electrode level as the porous materials were directly coated on the electrode surface, which means it cannot achieve real particle-level Li-ions pre-desolvation42. However, attempts to thoroughly mix cathodes with porous materials like metal-organic frameworks (MOFs) face significant challenges. Due to the powdery nature of MOFs, achieving comprehensive and uniform coverage of the cathode particles is difficult. Furthermore, this method only coats the surface of secondary particles, leaving the inner primary particles uncoated, thus failing to ensure complete coverage38. Additionally, the introduction of a coating layer can often impede lithium-ion transport to some extent, which can negatively impact the rate performance of the cathode materials. Complex and time-consuming liquid phase methods, which involve water or other solvents to coat cathodes with MOFs, introduce further complications. These processes can adversely affect the stability of the cathodes due to the presence of additional solvents.
To achieve effective Li-ion pre-desolvation, the pore windows of porous materials should be smaller than the size of solvated Li-ions, which is approximately 7.0 Å42,43. Ideally, the porous coating material should also be capable of transforming into a flowing liquid state. This allows the liquid MOF to diffuse into the inner voids of the secondary cathode particles and infuse into the grain boundaries between the primary cathode particles, thus ensuring 100% MOF coverage. Taking into account the above two prerequisites, a unique MOF namely Zn-P-dmbIm comes into our consideration (Supplementary Fig. 1)44. Besides its narrow 3.4 Å pore window and the one-dimensional (1D) sub-nanochannels, the Zn-P-dmbIm MOF powder can easily transform into liquid state (MOF liquid) upon heating under low temperature of 175 oC (Supplementary Fig. 2). The MOF liquid would transform into glass state (MOF Glass) after a rapid cooling process (Fig. 1c, the top panel). It worth noting that there are several differences between the Zn-P-dmbIm MOF powder and the MOF Glass. Even both of them all possess 1D channels, the MOF Glass exhibits much narrower pore windows of about 2.9 Å (Supplementary Fig. 3 and Supplementary Table 1). On the other hand, unlike crystalline Zn-P-dmbIm MOF powder with long-range ordered channels, MOF Glass possesses only short-range ordered channels, giving it an amorphous characteristic (Fig. 1c, the bottom panel, Supplementary Fig. 4). The MOF liquid demonstrates quite similar amorphous characteristic as that of the MOF Glass. These results inspired us to preparing MOF Glass infused high-voltage cathode using MOF liquid-infusion strategy. By melting the mixture of Zn-P-dmbIm MOF powder and cathode material, the low-viscosity and flowable MOF liquid can easily diffuse into the inner voids of the secondary cathode particles and constantly infuse into the grain boundaries between the primary cathode particles (Supplementary Fig. 5). Before a rapid cooling process, MOF liquid can completely wet the cathode from inside to outside (100% MOF coverage), and finally obtain cathode coated with rigid and thin MOF Glass (Fig. 1d, Supplementary Fig. 6). This unique coating effect cannot be achieved through typical strategies, such as simply mixing MOF particles with cathode materials. For example, when the MOF Glass is physically mixed with the NCM-811 cathode without further low-temperature treatment, it fails to successfully coating the NCM-811 (Supplementary Fig. 7). This highlights the importance of the MOF liquid-infusion strategy with further low-temperature treatment for achieving even coverage. The seamless MOF Glass is expected to facilitate the particle-level Li-ions pre-desolvation (Fig. 1e, f, Supplementary Figs. 8-10). It was found that the thickness of the MOF Glass layer can be controlled by adjusting the amounts of MOF particles added. Obviously, NCM-811 materials with relatively lower MOF additions (0.25, 0.5 and 1 wt%) demonstrate a much thinner and less uniform MOF Glass coating (Supplementary Fig. 11, with parts of the NCM-811 particles left uncovered), while increasing the MOF additions (2, 5 and 10 wt%) leads to a thicker and much uniform MOF Glass layer (Fig. 1g, h, Supplementary Figs. 12-16). We find that a uniformly and thoroughly covered MOF Glass layer would effectively suppress direct contact between the electrolyte and cathode, thus significantly preventing particle cracks (Supplementary Fig. 17) and reducing transition metal dissolution and migration (defined as TM loss) (Supplementary Fig. 18). Notably, batteries utilizing NCM-811 cathodes coated with 5 wt% and 10 wt% MOF glass exhibited nearly the same TM loss as those with a 2 wt% MOF Glass coating. Additionally, NCM-811 cathodes with 2 wt%, 5 wt%, and 10 wt% MOF glass coatings all maintained good stability without significant cracking. These findings highlight the critical importance of a uniform and complete MOF Glass coating. However, increasing the thickness of the MOF Glass coating beyond 2 wt% does not lead to a further reduction in TM loss, likely because the cathodes are already effectively coated. We also find that Glass@NCM-811 cathode material with 2 wt% Glass coating demonstrates almost the same electrochemical performance as Glass@NCM-811 cathode material with 5 wt% and 10 wt% Glass coating (Supplementary Fig. 19). In general, a high mass loading of the MOF Glass coating on the NCM-811 cathode increases fabrication costs and reduces the energy density of batteries utilizing the Glass@NCM-811 cathode material. Therefore, based on these afore-mentioned results, we think that a 2 wt% MOF Glass coating is the optimal thickness, and finally selected 2 wt% MOF Glass coated NCM-811 (shorted as Glass@NCM-811) as the as our main sample. In this article, for consistency, unless specified otherwise, Glass@NCM-811 refers to this configuration thereafter (Fig. 1h, Supplementary Fig. 20 and 21). The MOF Glass coating did not noticeably reduce the electrical conductivity of the cathode material, nor did it contribute additional capacity to the batteries with MOF Glass-coated cathodes. (Supplementary Fig. 22). More importantly, the MOF liquid, heated to 175 °C, can quickly wet the electrodes in just several seconds (Fig. 1i, Supplementary Fig. 23 and 24) benefits from the low viscosity of the MOF liquid44. This rapid wetting highlights the high efficiency of MOF liquid for quick infiltration and complete coating towards cathode primary/secondary particles. The corresponding TEM images and the corresponding Fourier-transform infrared spectroscopy (FTIR)/X-ray photoelectron spectroscopy (XPS) results verify the successful complete MOF Glass coating on NCM-811 (Fig. 1j–m, Supplementary Figs. 25-27). For the pristine bare NCM-811, only apparent layered structure can be clearly observed (Fig. 1j, k). For sharp contrast, a thin 15 nm Glass layer can be clearly found on the Glass@NCM-811 material (Fig. 1l). More importantly, we also find that a special inner region II within the MOF Glass layer (Fig. 1m, highlighted by the green rectangle line) can be clearly observed. Based on the high-resolution TEM (HR-TEM) image, the inner region II that directly attached with the NCM-811 demonstrates slightly narrower layered structure than that of the bare NCM-811. This unique structure is further verified by in-depth etching XPS measurements (Fig. 1n, Supplementary Fig. 28). In the P 2p XPS spectrum collected after the first etching, despite a weak peak related to Zn 3 s, only a single peak corresponding to O-P-O (green curve, situated at about 133.3 eV) is observed45. This peak can be attributed to the P-O interactions within the MOF Glass (Fig. 1n). We also compare the XPS results collected after the first-time etching with that of the pristine MOF Glass and the surface of the Glass@NCM-811 cathode material without etching (Supplementary Fig. 29). The clear overlap of these three curves indicates that the outer layer (surface and after one-time etching) of the Glass@NCM-811 still belongs to the electrically non-conductive MOF Glass. As the etching progresses, two new peaks appear at about 130.9 and 134.7 eV, which can be assigned to Li-P and O-P-Li interactions, respectively (Fig. 1n). It has been reported that Li-P and O-P-Li components can accelerate Li-ion diffusion and enhance the rate performance of batteries46. Therefore, based on these results, we conclude that this MOF liquid-infusion strategy successfully constructs a unique double-layered Glass structure on the surface of the NCM-811, with the outer-layer consisting of electrically non-conductive porous MOF Glass with 2.9 Å pore windows and the inner-layer comprising Li-ion conducting components. It is believed that chemical reactions occurred during the MOF liquid phase at 175 oC and the MOF glass formation process (rapid cooling to 25 oC), leading to the formation of a Li-ion conductive layer containing Li-P and Li-P-O components.
Fast Li-ion desolvation and diffusion of the Glass@NCM-811 cathode
To study the functions of this unique MOF Glass coating, further characterizations are conducted. In the XRD pattern, the intensity ratio between (003) and (104) peaks (I(003)/I(104)) of the Glass@NCM-811 is nearly the same as that of the bare NCM-811, which indicates the Glass coating does not severally damage the layered structure of NCM-811 (Fig. 2a)47. Both the bare NCM-811 and Glass@NCM-811 are coupled with metallic Li to evaluate their rate performances. Compared with the cell based on bare NCM-811 which demonstrates poor rate performance, the Li||Glass@NCM-811 battery exhibits much good rate performance from 0.1 C to as high as 5 C current rates. The biggest difference comes from the capacity obtained under 5 C, in where the Li||Glass@NCM-811 battery delivers about 165 mAh g−1 capacity, while the bare NCM-811//Li battery sustains only 80 mAh g−1 (Fig. 2b). This indicates that the MOF Glass constructs on NCM-811 surface can remarkably enhance the Li-ion transporting rate during battery discharge/charge processes. The state of charge (SoC) vs. time curves of bare Li||NCM-811 and Li||Glass@NCM-811 cells are also tested under constant current-constant voltage (CC-CV) mode (see Experimental Section for detail). As shown in Fig. 2c, the Li||Glass@NCM-811 cell reaches 100% state of charge (SoC) in just 1080 seconds, much faster than the bare Li||NCM-811 cell, which takes 1400 seconds. Additionally, under CC mode, the Li||Glass@NCM-811 cell achieves a much higher SoC of 75%, compared to only 41% for the bare Li||NCM-811 cell. This is benefits from the remarkably low interfacial resistance of Li||Glass@NCM-811 cell38,46. Then, galvanostatic intermittent titration technique (GITT) measurements of two cells are measured after 100 cycles (same cells in Fig. 2b after rate performance test). Obviously, the bare Li||NCM-811 cell exhibits significantly higher battery polarization, with an average voltage loss approximately 3.5 times higher than that of the Li||Glass@NCM-811 cell (0.14 V vs. 0.04 V, Fig. 2d). The corresponding Li+ diffusion coefficient of the battery based on Glass@NCM-811 is one order of magnitude higher than that of the cell assembled with pristine NCM-811 cathode (Supplementary Fig. 30), demonstrates that the MOF Glass coating significantly enhances lithium diffusion kinetics. These results collectively verify the crucial role of the MOF Glass coating, particularly the inner-layer of the Glass, in promoting fast Li-ion diffusion in the NCM-811 cathode. The outer-layer of MOF Glass is also under further investigation by FTIR and Raman spectroscopy. In order to facilitate our data collection, we increase the thickness of the MOF Glass layer (see Experimental Section for detail). Then, by etching away the surface MOF Glass layer (about 2 nm, see Experimental Section for detail), we detect the electrolyte information inside the MOF Glass. Compared with the FTIR result of typical electrolyte (1 M LiPF6-EC/DMC, the bottom panel in Fig. 2e), electrolyte signals detect from inside the surface MOF Glass demonstrate much stronger Li+-solvent interactions (Li-EC and Li-DMC, the top panel in Fig. 2e). This indicates that the electrolyte forms an aggregative configuration inside the MOF Glass41, highlighting the crucial role of MOF Glass in facilitating Li-ion pre-desolvation and accelerating Li-ion migration before reaching the NCM-811 cathode surface46. Raman data of electrolyte inside MOF Glass layer and LiPF6-EC/DMC electrolytes with different concentrations consists well with the FTIR result (Supplementary Fig. 31), which further verifies the effectiveness of MOF Glass in promoting pre-desolvation of Li-ions. Considering the average pore window of the MOF Glass and the sizes of solvent molecules and PF6−, we propose that the pre-desolvated Li-ions confined inside the MOF Glass channel maintain configuration as schematically illustrated in in Supplementary Fig. 32.
a XRD patterns of bare NCM-811 and Glass@NCM-811 cathodes. b Rate performances of batteries based on bare NCM-811 and Glass@NCM-811 cathodes (defined 1 C = 220 mA/g). c The state of charge (SoC) vs. time curves of Li||NCM-811 and Li||Glass@NCM-811 cells. This indicated the much faster Li-ion desolvation and transport of Li||Glass@NCM-811 cell due to the much lower interfacial resistance. d Discharge curves of the GITT measurements conducted after the 100th cycle (same cells used in Fig. 2b). Inset: average voltage loss and its standard deviation over different GITT steps. e FTIR spectra of typical electrolyte (LiPF6-EC/DMC, the bottom panel) and electrolyte formed inside the Glass layer (the top panel). The aggregative electrolyte inside the Glass layer suggested the successfully pre-desolvation enabled by the sub-nanochannels of the Glass layer. f Comparison of activation energies during Li-ion desolvation and its migration across the cathode electrolyte interphase (CEI) for Li||NCM-811 and Li||Glass@NCM-811 cells. g Comparison of the kinetics of desolvation/pre-desolvation and Li-ion transport through the cathode electrolyte interphase (CEI) in Li||NCM-811 cell (the top panel) and Li||Glass@NCM-811 cell (the bottom panel). h Schematic illustration of solvated Li-ions penetrating through the typical CEI formed on cycled bare NCM-811 cathode (top panel) and Glass layer on cycled Glass@NCM-811 cathode (bottom panel).
EIS measurement is used to calculate the activation energy during Li-ion (pre) desolvation and its migration/transport across the cathode electrolyte interphase (CEI) (Supplementary Figs. 33 and 34). To simplify the calculation process, we combine the activation energies from both processes for a unified calculation. Clearly, the activation energy (Ea) of Li||Glass@NCM-811 cell is 45.7 kJ mol−1, which is much lower than 104.3 kJ mol−1 of bare Li||NCM-811 cell (Fig. 2f). The facile Li-ion pre-desolvation and fast Li-ion transport is further verified by data shown in Fig. 2g. Compared with the energy barrier (42.6 kJ mol−1) of Li-ion desolvation within bare NCM-811, the much lower energy barrier (19.3 kJ mol−1) during Li-ions pre-desolvation of the Glass@NCM-811 further verifies the facile Li-ion pre-desolvation inside the 2.9 Å pore windows of MOF Glass outer-layer. Additionally, the significantly lower energy barrier for Li-ion transport in Glass@NCM-811 (26.4 kJ mol−1 compared to 61.7 kJ mol−1 for bare NCM-811) indicates much faster Li-ion diffusion through the dual-layer structure of Glass on NCM-811 surface. We attribute the significantly reduced pre-desolvation and Li-ion transport activation energy of Li||Glass@NCM-811 to the narrow 2.9 Å pore windows of the MOF Glass outer layer, which facilitate facile Li-ion pre-desolvation and create an aggregated electrolyte with a low-solvent-coordination solvation structure46. Additionally, the inner-layer, containing Li-ion transport-accelerating components (Li-P and O-P-Li), promotes fast Li-ion transport. Based on those obtained results, we propose how Li-ions passing through the typical CEI (mainly solvent-derived organics) on bare NCM-811 and Glass layer of Glass@NCM-811 cathode (Fig. 2h).
Glass@NCM-811 cathode suppresses cathode cracks, CEI rapture, cation mixing and side-reactions
The cycled bare NCM-811 and Glass@NCM-811 cathodes underwent further characterization to study the additional positive effects of the MOF Glass coating. After cycling, the bare NCM-811 cathodes show extensive side-reaction byproducts covering their surface (Fig. 3a, Supplementary Fig. 35). In sharp contrast, the cycled Glass@NCM-811 cathodes exhibit much smoother surfaces without various byproducts accumulation (Fig. 3c, Supplementary Fig. 36), indicating significantly suppressed electrolyte decomposition. Elemental mapping results corroborate the SEM images of the two cycled cathodes (Supplementary Figs. 37 and 38). Furthermore, apparent cracks are observed inside the cycled bare NCM-811 cathodes (Fig. 3b, Supplementary Figs. 39 and 40), whereas no obvious cracks were found inside the cycled Glass@NCM-811 cathodes (Fig. 3d, Supplementary Figs. 41 and 42). The TEM elemental mapping of the two cycled cathodes shows substantial differences, with disorganized and uneven distributions of elements, especially P and F, in the cycled bare NCM-811 cathodes, indicating severe byproduct accumulation from electrolyte solvent decompositions (Supplementary Fig. 43). In sharp contrast, the cycled Glass@NCM-811 cathode demonstrates uniform distributions of elements (Fig. 3e–j). The Glass coating remains intact even after various discharge/charge processes, further verifying the MOF Glass’s role in protecting the NCM-811 cathode. Notably, the F element mapping is slightly smaller than the cycled Glass@NCM-811 cathode, and its shape corresponds well with the NCM-811 under the MOF Glass coating (Fig. 3j). Since the only source of F is the LiPF6 salt, we attribute this to the decomposition of the PF6 anion under the electrically non-conductive MOF Glass layer, further confirming the pre-desolvation of Li-ions through the MOF Glass sub-nanochannels. Additionally, the cycled Glass@NCM-811 demonstrates much higher I(003)/I(104) value than the cycled NCM-811 cathode, which suggest it does not experience serious structural degradation even after multiple electrochemical cycles (Fig. 3k). The differential capacity (dQ/dV) curves of two cells based on pristine NCM-811 and Glass@NCM-811 cathode are also recorded (Supplementary Fig. 44). For bare NCM-811 based cell, the H2/H3 (second hexagonal to third hexagonal) gradually disappear and the other redox peaks (hexagonal to monoclinic, H1/M; monoclinic to second hexagonal, M/H2) diminish, which was closely related to cracks apparition48. This indicates that mechanical stress generated cracks in the cathode along the grain boundaries, and the layered structure collapse over 100 cycles. Compared with that of the battery based on bare NCM-811, the cell assembled with Glass@NCM-811 cathode exhibits a reversible redox peak for the H2/H3 transition even after 100 cycles. This indicates the MOF Glass helping in protecting the cathode from structural volume changes, which again underscoring the significant role of MOF Glass in stabilizing the NCM-811 cathode. Most cathode materials are highly susceptible to degradation when exposed to water, with even minimal contact severely impacting their electrochemical performance. Remarkably, the Glass@NCM-811 cathode exhibits good water resistance, maintaining its performance even under such conditions (Supplementary Fig. 45). Specifically, the battery assembled with water-immersed Glass@NCM-811 cathode (after 7 days of immersion) exhibits almost the same electrochemical performance as a battery with a fresh Glass@NCM-811 cathode (Supplementary Fig. 46).
a SEM image and b cross-sectional SEM image of the cycled NCM-811 after 200 cycles. c SEM image and d cross-sectional SEM image of the cycled Glass@NCM-811 after 400 cycles. e–j TEM image and the corresponding elemental mapping images of the cycled Glass@NCM-811. k XRD of cycled Glass@NCM-811 and cycled bare NCM-811. l–n High-resolution transmission electron microscopy (HR-TEM) images of the cycled bare NCM-811. o–q HR-TEM images of the cycled Glass@NCM-811. r In-depth etching FTIR of the cycled bare NCM-811 cathode (the top panel) and the cycled Glass@NCM-811 (the bottom panel).
TEM studies of the two cycled cathodes reveal further insights. For the cycled bare NCM-811 cathode, thick and uneven cathode electrolyte interphase (CEI) layers, varying from 30 to 60 nm and in some cases reaching up to 80 nm, are distinctly observed (yellow arrows highlighted in Fig. 3l and Supplementary Figs. 47, 48). More importantly, various cation-mixed rock-salt phases are observed after cycling (Fig. 3m, n and Supplementary Fig. 47e, 47f, highlighted with yellow circles), indicating serious cation mixing originating from surface oxygen loss followed by surface cation densification. In stark contrast, the layered structure of the cycled Glass@NCM-811 and the Glass layer is well preserved, with no obvious cation-mixed rock-salt phases detected (Fig. 3o–q and Supplementary Figs. 49, 50). This further confirms the MOF Glass coating’s critical role in preventing cation mixing and stabilizing the cathode. In-depth etching FTIR measurement is used to investigate the surface information of cycled two cathodes. For the cycled bare NCM-811 cathode (Fig. 3r, the top panel), tremendous EC electrolyte solvent decomposition induced byproducts (carboxylics (C-O), alkyl carbonates (ROCO2Li) and carbonyls (C = O)) can be clearly found during the whole etching process. Interestingly, during the initial etching of the cycled Glass@NCM-811 cathode, only faint peaks corresponding to the PVDF binder and MOF Glass layer components (P-O, C-N, and N-H) are detected. Notably, there are almost no peaks related to the decomposition products of the EC solvent (Fig. 3r, the bottom panel). Those results together verify the significant role of non-conductive MOF Glass coating in suppressed cathode cracks, CEI growth, cation mixing and side-reactions.
Glass@NCM-811 cathode significantly eliminates gases generation, transition metal dissolution/migration
In-situ differential electrochemical mass spectrometry (In-situ DEMS) measurements are employed to investigate gas generation in cells based on bare NCM-811 and Glass@NCM-811 cathodes. The bare Li||NCM-811 battery produces significant amounts of carbon dioxide and oxygen (CO2/O2) (Fig. 4a, b and Supplementary Fig. 51a, 51b), whereas the Li||Glass@NCM-811 battery generates only negligible amounts of gas during the electrochemical cycling process (Fig. 4c, d). The corresponding 1H Nuclear magnetic resonance (NMR) analysis results of cycled electrolytes collected from the two batteries suggest that the Li||Glass@NCM-811 cell exhibits greatly suppressed electrolyte decomposition than its counterpart (Supplementary Fig. 51c). These results suggest that the MOF liquid infusion coating strategy effectively lowers surface and grain boundary oxygen activity and suppresses electrolyte solvent oxidation. The complete MOF Glass coverage effectively prevents electrolyte penetration into the interior of the NCM-811 cathode, and the aggregative electrolyte formation inside the MOF Glass sub-nanochannels during Li-ion pre-desolvation together contributing to the suppression of electrolyte oxidation and gas generation. We also observe distinct differences in the morphologies of the cycled Li anodes collected from the two batteries. The cycled Li anode from the bare Li||NCM-811cell exhibits a porous layer consisting of mossy Li and unevenly distributed byproducts (Fig. 4e), whereas the Li anode from the Li||Glass@NCM-811 cell maintains a smooth surface without any dendrites and byproducts (Fig. 4f). This smooth, dendrite-free Li anode is likely due to the greatly suppressed transition metal dissolution/migration. Generally, the deposited TM would degrade the performance of the Li anode. To further support this, we assembled Li//Li symmetric cells using electrolytes with/without 400 ppm Ni(TFSI)2 salt (see Experimental Section for detail). As shown in Supplementary Fig. 52, the Li//Li symmetric cell with 400 ppm Ni(TFSI)2 salt added electrolyte short-circuited much faster than the cell without Ni(TFSI)2 salt addition, indicating that the deposited Ni-induced morphological instability would degrade the performance of the lithium-metal anode. Corresponding ICP results of bare Li||NCM-811 and Li||Glass@NCM-811 batteries are also recorded. As shown in Fig. 4g, the bare Li||NCM-811 battery (right panel) experiences more than 20 times higher TM loss than its Li||Glass@NCM-811 counterpart (left panel) after 200 and 500 cycles, respectively. Obviously, the Li anode collected from the bare Li||NCM-811 cell showed significantly higher TM deposits than the Li anode from the Li||Glass@NCM-811 battery (Fig. 4g, the gap between fully-dyed and semi-dyed rectangles). Additionally, we also evaluate the electrochemical performance and transition metal (TM) loss in a bare Li||NCM-811 battery with MOF Glass directly coated on the NCM-811 electrode, termed Glass-NCM-811 (Supplementary Fig. 53a-c). The Li||Glass-NCM-811 cell demonstrates significantly improved cycling stability and considerably reduced TM loss compared to the bare Li||NCM-811 battery (Supplementary Fig. 53d-f). Notably, the cycled Glass-NCM-811 cathode also exhibits much thinner and more uniform CEI layers across various cycles (Supplementary Fig. 54) compared to the bare NCM-811 cathodes (Fig. 3l–n, Supplementary Fig. 48), underscoring the benefits of electrode-level Glass coating. Nevertheless, despite these advantages, the TM loss of Li||Glass-NCM-811 cell is slightly higher and the electrochemical stability lower than that of the Glass@NCM-811//Li battery, emphasizing the superior performance provided by particle-level Glass coating (Supplementary Fig. 53d-f). The XPS of Li anodes from two batteries also demonstrate the same results. By using XPS measurement, we find the SEI of Li anode from Li||Glass@NCM-811 battery exhibits much weaker Ni element signals than the Li anode pairs with bare Li||NCM-811 battery (Fig. 4h). Based on those results we have obtained in Figs. 2–4, we conclude that the MOF liquid infusion strategy effectively suppresses cathode particle cracks, CEI rupture, gas generation, and TM loss. By implementing this perfect particle-level pre-desolvation method, both the stability of cathode and Li anode can be greatly enhanced (Fig. 4i).
a, b Charge curve of bare Li||NCM-811 cell cycled for in-situ Differential Electrochemical Mass Spectrometry (in-situ DEMS) test and the corresponding DEMS data. c, d Charge curve of Li||Glass@NCM-811 cell cycled for in-situ DEMS test and the corresponding DEMS data. SEM of the cycled Li anode harvested from the cycled (e) bare Li||NCM-811 cell and (f) Li||Glass@NCM-811 cell after 400 cycles. g Dissolved transition metals inside the cycled electrolyte and on the cycled Li anode from Li||Glass@NCM-811 (the left panel) and Li||NCM-811 cell (the right panel) by ICP-OES after different cycles. h Ni XPS results collected on the cycled Li anode from Li||Glass@NCM-811 (the top panel) and Li||NCM-811 cell (the bottom panel). i Schematic illustration of the MOF Glass infusion strategy shows stabilization of the NCM-811 cathode by suppressing issues induced by electrolyte penetration and solvated Li-ion/solvent co-insertion, including cathode particle cracks, CEI rupture, oxygen loss, and transition metal migration. By implementing the perfect particle-level pre-desolvation method, both the stability of cathode and Li anode can be greatly enhanced.
Enhanced cycling stability and energy density
The cycling stability and energy density are evaluated in both coin-cell and pouch-cell configurations by pairing MOF Glass-coated high-voltage cathodes with Li anodes. When cycles at a 1 C rate under a 4.4 V cut-off charge voltage, the bare Li||NCM-811 coin-cell demonstrates rapid capacity decay (Fig. 5a, c) after only 400 cycles. In sharp contrast, the Li||Glass@NCM-811 coin-cell exhibits remarkably enhanced cycling performance, retaining 80% of its capacity (calculated from the fourth cycle) after a long 1000 cycles (Fig. 5b, c). This trend is even more apparent when the cells are cycled under an elevated 4.6 V cut-off charge voltage. The bare Li||NCM-811 coin-cell fails quickly after only 300 cycles, while the Li||Glass@NCM-811 coin-cell delivers a high specific capacity of 180 mAh g−1 after 400 cycles (Fig. 5d–f). It worth noting that the MOF Glass layer within Glass@NCM-811 remains intact without any apparent morphological changes, even after cycling at a high voltage of 4.6 V (Supplementary Fig. 55a). The XPS data of the cycled Glass@NCM-811 closely matches that of the uncycled sample, further indicating the good structural stability of the MOF Glass (Supplementary Fig. 55b, 55c). Two other typical high-voltage cathodes, Li-rich manganese (LRMO) and LiCoO2 (LCO), are also coated with MOF Glass and paired with Li anodes to evaluate their cycling stability. Similarly, the Glass@LRMO and Glass@LCO coin-cells demonstrated superior performance compared to cells without MOF Glass coating. It is widely acknowledged that LRMO cathodes tend to suffer from severe voltage attenuation, rapid capacity decay, and poor cycling stability during electrochemical cycling, especially under high cut-off voltage. This is evidenced by the poor cycling performance of the Li||LRMO coin-cell under a high 4.8 V cut-off voltage, which maintains a capacity of 158.3 mAh g−1 after 260 cycles (only 56% capacity retention, calculated from the 3rd cycle). Remarkably, the Li||Glass@LRMO coin-cell demonstrates good cycling performance (Fig. 5g), retaining a high capacity of 237.8 mAh g−1 after 400 cycles (82% capacity retention, calculated from the 3rd cycle). The same trend was observed for the Li||Glass@LCO coin-cell, which also showed much more stable cycling than the Li|| LCO coin-cell (Fig. 5h). The corresponding first galvanostatic curves of different coin cells demonstrated in Fig. 5a-h can also be found as shown in Supplementary Fig. 56. These results further underscore the universality and importance of using MOF Glass coating to improve the stability of high-voltage cathode materials.
Cycling performances and the corresponding discharge/charge curves of Li||Glass@NCM-811 cell (blue/green curves) and Li||NCM-811 cell (light grey curves) cycled under the range of (a–c) 2.7–4.4 V and (d–f) 2.8–4.6 V (defined 1 C = 220 mA/g for NCM-811 based batteries). g Cycling performance of Li||Glass@LRMO cell and Li||LRMO cell under 4.8 V cut-off voltage (defined 1 C = 280 mA/g for LRMO based batteries). h Cycling performance of Li||Glass@LCO cell and Li||LCO cell under 4.6 V cut-off voltage (defined 1 C = 220 mA/g for LCO based batteries). i 385 Wh kg−1-level Li||Glass@NCM-811 pouch-cell and cycling performance of pouch-cell based on bare Li||NCM-811. Inset: the digital photo of the Li||Glass@NCM-811 pouch-cell.
Inspired by these promising results, a 2.0 Ah-level pouch-cell consisting of Glass@NCM-811 and a Li anode was fabricated (Fig. 5i, inset). The Li||Glass@NCM-811 pouch-cell exhibits improved cycling stability, retaining 86.9% of its capacity (calculated from the 2nd cycle) after 300 cycles (Supplementary Fig. 57), compared to 37.6% capacity retention (calculated from the 2nd cycle) after 58 cycles for the bare Li||NCM-811 pouch-cell (grey curve in Fig. 5i, Supplementary Fig. 58). Based on the pouch-cell parameters (Supplementary Table 2), the output energy density of the Li||Glass@NCM-811 pouch-cell is calculated to be as high as 385.5 Wh kg−1 (calculates from the 2nd cycle, 19.579 g for the whole pouch-cell). These electrochemical performances achieved with this MOF liquid infusion strategy rank among the highest compared to other state-of-the-art coating strategies (Supplementary Tables 3-6). Noting that all these performances are achieved on a laboratory scale, we believe that the pouch-cell energy density and cycling stability can be further improved to over 400 Wh kg−1 by optimizing the pouch-cell parameters, especially electrolyte modification. In addition, the MOF liquid infusion strategy used to create MOF glass-coated cathodes in this work is fundamentally different from other studies employing MOF glasses. And the key novelty of our work lies in its unique structure and specialized functions (Supplementary Fig. 59, Supplementary Table 7)49,50,51,52. The MOF liquid infusion strategy for preparing highly stable high-voltage cathodes offers promising prospects for practical industrial battery production, thanks to its good electrochemical performance and its simple, cost-effective, and time-efficient synthesis process.
In summary, we introduce a straightforward and efficient metal-organic framework (MOF, Zn-P-dmbIm) liquid-infusion strategy that fully infuses MOF liquid into the grain boundaries of high-voltage cathodes like NCM-811, LRMO, and LCO, achieving complete MOF Glass coverage (e.g., Glass@NCM-811). The surface electrically non-conductive MOF Glass layer with 2.9 Å pore windows facilitating Li-ion pre-desolvation and enabling highly aggregative electrolyte formation inside the Glass channels, suppressing solvated Li-ion co-insertion and solvent decomposition. While the inner Glass layer attaches with cathode composes of Li-ion conducting components and enhancing fast Li-ion diffusion. This MOF Glass coating prevents cathode particle cracks, CEI rupture, gas generation, and transition metal migration, while promoting rapid Li-ion transport. Consequently, Li||Glass@NCM-811 cells exhibit superior electrochemical performance, with remarkable rate capability and cycling stability, even at high charge rates (5 C) and elevated voltages (4.6 V). Similarly, Li||Glass@LRMO and Li||Glass@LCO batteries demonstrate good cycling stability over 400 cycles at 4.8 V and 4.6 V, respectively. The practical viability of this strategy is underscored by successfully achieving a 385 Wh kg−1-level pouch cell using Glass@NCM-811.