Lithium-ion batteries (LIBs) achieve excellent performances such as high energy density and long cycle lifespan, and have been put into applications in consumer electronics, electric vehicles, and grid-scale energy storage for decades.^[1–3] Notwithstanding the great progresses that have been made in LIBs materials and battery technologies, the common cathode materials used in commercial LIBs still have numerous intractable issues that restrict the further improvement of the battery performances. One of the most arresting problems is the dissolution of transition metal (TM) cations from cathode materials, especially for layered oxide cathodes LiTMO₂ (TM = Ni, Co, Mn, etc.),^[4–7] which significant impacts the battery performances. Explicitly, the dissolution of TM over the battery cycling process leads to deconstruction of the cathode materials and diminish of the valence-variable TM, resulting in capacity fade of the cathodes.^[8,9] Meanwhile, the dissolved TM cations can migrate and deposit on the anode electrode, and subsequently catalyze the decomposition of the electrolyte, leading to uncontrollable growth of solid electrolyte interphase (SEI) and thus deterioration of the interfacial properties,^[10–12] which synergistically contribute to the performance degradation of LIBs.

Numerous previous studies have revealed that the TM dissolution issue is affected by various intertwined factors, such as bulk structure and surface chemistry of cathode, electrolyte, and additives, battery testing protocols and conditions. Accordingly, plenty of strategies and methods have been proposed and developed to address the TM dissolution issue from different aspects. Bulk doping^[13,14] and surface coating^[15,16] on cathode materials are the most widely used methods that are expected to intrinsically improve the structural stability of the material against TM dissolution. Electrolyte additives have been demonstrated to be highly effective to enable the formation of a stable SEI or cathode electrolyte interface (CEI) layer that can physically separate the cathode and electrolyte, and thus prevent the side reaction induced TM dissolution.^[17,18] Employing polycation chelating agents to capture TM ion has also been reported to address the TM dissolution issue. Although this method cannot directly prevent the TM dissolution, the capture of TM and the restrained deposition on the anode can prevent the undesired side reactions between the anode and electrolyte, and thus resulting in enhanced cycle stability.^[19] Recently, coating functional materials onto the polymer separator has been considered as an alternative way to solve the TM dissolution issue.^[6,19] For example, it is reported that the coating of CaCO₃ can consume HF in the electrolyte and thus avoid its reaction with the cathode. All these methods have demonstrated the effectiveness of inhibiting the TM dissolution and thus can greatly improve the battery performances.

In the present work, the oxide solid electrolyte Li_1.5Al_0.5Ti_1.5(PO₄)₃ (LATP), which delivers high ionic conductivity and has a reasonable cost, was employed to be coated on the polyethylene (PE) separator.^[20] The LATP coated PE separator was proposed to inhibit the deterioration of battery performances resulted from TM dissolution. Microscopic and spectroscopic characterizations revealed the significant changes of the LATP surface morphology owing to the reduction of LATP during battery cycling. The as-formed compact surface layer can effectively block the TM ion from getting to the anode electrode for the subsequent deposition, and in the meantime inhibit the TM dissolution from the cathode, leading to the enhanced interfacial stability as well as the cathode structure stability. Accordingly, the LiCoO₂/silicon monoxide (SiO)-graphite and LiMn₂O₄/SiO-graphite cells using the LATP coated separator show superior cycle performances to the cells using regular Al₂O₃ coated separator.

As LATP is regarded as the essential material to optimize the cell performance, XRD was employed to check the phase purity of LATP (Fig. 1(a)). The XRD pattern is overall consistent with the standard pattern of LATP (PDF: 35-0754), with no notable impurities that can be observed. The sharp XRD peaks indicate the good crystallinity of the as-synthesized sample. Further Rietveld refinement confirms the pure phase of LATP with a NASICON structure (S.G.R-3c). The morphology of the LATP was characterized by SEM, showing an irregular particle shape with a size distribution in the range of 500–700 nm (see Fig. 1(b)). Al₂O₃ with similar particle size was deliberately chosen to coat the separator for paralleling comparison. Figures 1(c) and 1(d) show the surface morphologies of Al₂O₃@PE and LATP@PE, respectively. Granularity equivalent Al₂O₃ and LATP particles are coated and uniformly distributed on the surface of the PE separator without any aggregation. The thickness of the LATP and Al₂O₃ coating layer was controlled to be around 2 μm.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (a) The XRD pattern and (b) SEM image of the LATP powder. The surface morphologies of (c) Al2O3@PE and (d) LATP@PE separators.

To evaluate the effectiveness of the LATP coating on the suppression of TM dissolution, charge–discharge cycle tests were performed. Figures 2(a) and 2(b) display the first cycle charge and discharge curves of the LCO-SOC400 cell and the LMO-SOC650 cell using Al₂O₃ coated and LATP coated separators, respectively. The LCO-SOC400 cell with LATP@PE exhibits higher charge capacities (191.8 mAh/g) and discharge capacities (172.4 mAh/g) than the cell with Al₂O₃@PE which shows the first charge capacities of 184.8 mAh/g and discharge capacities of 167.0 mAh/g. The initial coulombic efficiencies (CEs) are comparable between LCO-SOC400 cells using LATP@PE (89.88%) and Al₂O₃@PE (90.36%), but lower than the CE of LiCoO₂ half-cell using Li metal anode. This can be attributed to the formation of a thicker SEI layer on SiO/graphite anode which would consume more lithium ions.^[22,23] The LMO-SOC650 cells show a similar charge–discharge behavior. The first charge capacities (122.1 mAh/g) and discharge capacities (89.4 mAh/g) of the LMO-SOC650 cell with LATP@PE are also larger than those with Al₂O₃@PE (118.8 mAh/g and 84.5 mAh/g). The first cycle CEs of the LMO-SOC650 cells (71.12% for LATP@PE and 73.21% for Al₂O₃@PE) are lower than those of the LCO-SOC400 cell, which is due to the higher content of SiO used in SOC650 than SOC400. The above results indicate that, the LATP@PE separator, no matter utilized in LCO-SOC400 or LMO-SOC650 systems, could enable higher first cycle capacities than the Al₂O₃@PE separator counterparts. And the optimized batteries performance is attributed to the different anode/separator interface properties. The in-depth mechanism will be discussed later combining with the multi-faceted characterizations.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. Charge–discharge curves of (a) LCO-SOC400 cell and (b) LMO-SOC650 cell for the first cycle. The red and blue curves represent the cells using Al2O3 coated and LATP coated separators, respectively.

The cycle performances of the LCO-SOC400 and LMO-SOC650 cells using the Al₂O₃@PE and LATP@PE separators were tested at room temperature. All the cells were charge–discharged at 0.1 C-rate for the first 5 cycles (the shadow in Fig. 3(a) contains more than 5 cycles), and then cycled at 1 C-rate for the subsequent cycles. As shown in Fig. 3(a), the LCO-SOC4000 with LATP@PE shows a much better discharge capacity retention (83.23%/50^th cycle) than the cell with Al₂O₃@PE (59.57%/50^th cycle). Similar behavior can also be observed for LMO-SOC650, where the cell using LATP@PE presents superior capacity retention compared to that using Al₂O₃@PE (73.7% versus 67.45% for 50^th cycle). For comparing the capacity retention ability of these two kinds of separators, the capacity evolution ranges from the 1st to the 50th cycle are put together to obtain a clear view (Figs. 3(c) and 3(d)). In general, the cells using LATP@PE separator display a remarkable advantage in the capacity retention ability over the Al₂O₃@PE cells. Specifically, for LCO-SOC400 cells, the LATP@PE cell has the capacity exacerbation from 172.4 mAh/g to 143.4 mAh/g, which is much better than the Al₂O₃@PE cell. Moreover, the ones using LATP@PE separator manifest a tendency that the capacity fading slows down with a faster speed, which might imply the formation of a more stable SEI layer. These results illustrate that the cycle performances of cells using LATP@PE is better than those using Al₂O₃@PE during the full battery life.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. Discharge capacity retention of (a) LCO-SOC400 cell and (b) LMO-SOC650 cell. Discharge profiles of (c) LCO-SOC400 cells and (d) LMO-SOC650 cells at different cycles.

To understand the impact of the coating layer on the battery performances, EIS measurements were carried out to inspect the interfacial properties of the cells after electrochemical cycling. While all spectra show a similar line shape with two well-separated semi-circles which can be ascribed to the transport resistances from CEI at cathode side and SEI at anode side (denoted as R_a and R_c), much suppressed intermediate-frequency semi-circles can be observed for both cells using the LATP@PE separator. Considering that the difference comes from the coating layer facing to the anode electrode side, the distinct intermediate-frequency resistance behaviors may be due to the different SEI properties of the cells using LATP@PE separator and Al₂O₃@PE separator. Quantitatively analysis was performed by fitting the spectra with an equivalent circuit model as depicted in the insets of Figs. 4(a) and 4(b). The detailed results are listed in Table 1. It can be found that the cells using Al₂O₃@PE separator exhibit slightly larger cell ohmic resistance (R_b) and R_c values, but significant larger R_a values than the cells using LATP@PE separator (for LCO cells, 265.7 Ω versus 174.4 Ω; for LMO cells, 270.7 Ω versus 57.0 Ω), indicating the more severe deterioration of the interfacial property for the Al₂O₃@PE cells over the electrochemical cycling. This is also in good consistent with the more rapid decay of the capacity for the Al₂O₃@PE cells.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. EIS spectra for (a) the LCO-SOC400 cells and (b) the LMO-SOC650 cells after 50 cycles. Rb: ohmic resistance; Ra: resistance of SEI at anode; Rc: resistance of CEI at cathode; CPE: constant phase element.

Table 1.

Table 1.

Table 1. The detailed results of fitting the spectra with an equivalent circuit model. . Sample Rb/Ω Rc/Ω Ra/Ω LCO-SOC400 with LATP@PE 3.6 7.1 174.4 LCO-SOC400 with Al2O3@PE 6.6 15.1 265.7 LMO-SOC650 with LATP@PE 3.4 6.5 57.0 LMO-SOC650 with Al2O3@PE 4.9 10.2 270.7

Table 1. The detailed results of fitting the spectra with an equivalent circuit model. .

Since the main differences between the cells come from the different coating layers on the separators, it is rational to extrapolate that there should be evident difference on the actual working part, that is the separator, after battery cycling. We therefore examined the surface morphology of the separators retrieved from the LCO cells after 50 cycles. Interestingly, the surface morphology of Al₂O₃@PE has barely changed after 50 cycles (Figs. 5(a) and 5(b)) comparing to its original one displayed in Fig. 1(c), whereas significant morphology changes can be seen on LATP@PE as shown in Figs. 5(c) and 5(d) (SEM image of the original LATP@PE is shown in Fig. 1(d)). It can be seen that, a dense interphase layer is formed on the surface of the PE separator, replacing the originate-state LATP particles. Meanwhile, some cracks are also found which may be caused by stress accumulation and volume change. As it is well-known that LATP can be reduced at low voltage (2.5 V vs. Li/Li⁺),^[24,25] the formation of the dense interphase layer should be owing to the reduction of LATP particles during the discharge process. Besides, the volume expansion after the interphase formation was also reported by Lewis et al.,^[26] which is well consistent with our SEM observations.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. SEM images of (a), (b) Al2O3@PE and (c), (d) LATP@PE separators retrieved from the cells after 50 cycles.

The cathode and anode electrodes were harvested from the LCO-SOC400 and LMO-SOC650 cells after 50 cycles. The transition metal element content was detected by SIMS on the anode electrodes after DMC washing and drying. The signal of silicon was chosen as a reference signal because of its consistency during the charge–discharge process. It can be inferred from Figs. 6(a) and 6(b) that the ratio of TM/Si of Al₂O₃@PE cells is much higher than that of LATP@PE, both in the LCO-SOC400 cell and the LMO-SOC650 cell, suggesting the much more suppressed TM deposition on the anode electrodes for the cells with LATP@PE than with Al₂O₃@PE. Raman spectroscopy measurements were performed to inspect the surface structure change of LiCoO₂ after cycling. A relative weak spectrum signal of the Al₂O₃@PE cell indicates the more severe surface structure deconstruction than the LCO electrode in LATP@PE cells. The existence of the characteristic peaks (E_g and A_1g vibration modes in LiCoO₂) suggests the preservation of the main LiCoO₂ layered structure, while the overall suppressed spectrum intensity, for LCO from the Al₂O₃@PE cell, can be attributed to destruction of the LiCoO₂ surface, indicating the more severe Co dissolution.^[27,28] Combining with the aforementioned EIS and SEM results, it can be learned that the reaction of the LATP during battery cycling (very likely the first several cycles) renders the formation of a uniform and dense interphase layer, which can effectively prevent the deposition of TM on the anode electrode. As the catalytic effect of the deposited TM is known to be one of the main causes for the growth of SEI layer, the inhibition of the TM deposition may contribute to the substantially enhanced interfacial stability for cells using Al₂O₃@PE separator over the cycling process. Moreover, since “consumption” of the TM at the anode side is suppressed, the demand for the “supply” from the cathode is also reduced, and therefore, the TM dissolution from the cathode is greatly inhibited. All these benefits synergistically contribute to the much more enhanced cycle stability for the cells using LATP@PE separator.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. SIMS profiles collected on the anode electrodes retrieved from (a) LCO-SOC400 cell and (b) LMO-SOC650 cell after 50 cycles. (c) Raman spectra of LCO electrodes retrieved from LCO-SOC400 with LATP@PE and Al2O3@PE separators. (d) Schematic illustration of the TM dissolution and decomposition process.

In the present work, we demonstrate that a LATP coating layer on the polymer separator can inhibit the undesired transition metal dissolution and deposition process in LIBs. By using various characterizations such as SEM, SIMS, and Raman, it is revealed that the surface morphology of the LATP coating layer has been significantly changed owing to the reduction of LATP at the discharge process. The as-formed dense surface layer can effectively prevent the TM deposition on the anode electrode and inhibit the TM dissolution from the cathode electrode. Besides, the as described TM block function of the LATP@PE separator is mainly attributed to the in situ formed interphase, which is formed by the reduction of LATP under low potential, thus the LATP@PE separator can work with various anodes. As a result, the deconstruction of the cathode material and the deterioration of the electrolyte/anode interface can be largely suppressed. The cells using LATP coated separator show superior cycle performances to the cells using regular Al₂O₃ coated separator. This work demonstrates that the functional coating on the separator, in addition to the traditional strategies through modifications of the cathode materials, can effectively solve the TM dissolution and deposition problem, which is one of the most critical issues that should be addressed to improve the long-term cycle stability of LIBs.