Enhanced Cyclability of Cr8O21 Cathode for PEO-Based All-Solid-State Lithium-Ion Batteries by Atomic Layer Deposition of Al2O3

Cr8O21 can be used as the cathode material in all-solid-state batteries with high energy density due to its high reversible specific capacity and high potential plateau. However, the strong oxidation of Cr8O21 leads to poor compatibility with polymer-based solid electrolytes. Herein, to improve the cycle performance of the battery, Al2O3 atomic layer deposition (ALD) coating is applied on Cr8O21 cathodes to modify the interface between the electrode and the electrolyte. X-ray photoelectron spectroscopy, scanning electron microscope, transmission electron microscope, and Fourier transform infrared spectroscopy, etc., are used to estimate the morphology of the ALD coating and the interface reaction mechanism. The electrochemical properties of the Cr8O21 cathodes are investigated. The results show that the uniform and dense Al2O3 layer not only prevents the polyethylene oxide from oxidization but also enhances the lithium-ion transport. The 12-ALD-cycle-coated electrode with approximately 4 nm Al2O3 layer displays the optimal cycling performance, which delivers a high capacity of 260 mAh g−1 for the 125th cycle at 0.1C with a discharge-specific energy of 630 Wh kg−1.


Introduction
Lithium-ion batteries (LIB) based on inorganic transition-metal oxide cathodes, graphiticcarbon anodes, and organic-liquid electrolytes [1] are widely applied in many fields, including portable electronic devices, power tools, electric vehicles (EV), and large-scale energy storage stations due to their high energy density, long cycle life, and low self-discharge rate [2]. However, the commercialized LIBs still cannot meet the increasing demand for energy storage devices with long cycle life and high energy density for EVs with longer cruising distances.
The energy density of lithium-ion batteries mainly depends on the cathode materials. The state-of-the-art LIB cathodes include Li transition-metal oxides and Li transition-metal phosphates [3], which show a capacity of less than 250 mAh g −1 . However, the energy density above 300 Wh kg −1 is desired to increase the driving range and to decrease the cost [4]. To achieve this target, the specific capacity of the cathode material should reach 500 mAh g −1 . In recent years, Li-free cathodes with high energy density such as S, FeF 2 , FeS 2 , and transition-metal oxides have attracted renewed attention owing to their high theoretical capacity [4][5][6][7][8]. Chromium oxides with high-valence have also been considered as promising cathode materials for lithium-ion batteries due to their ultrahigh theoretical capacity. For example, octachromium henicosaoxide (Cr 8 O 21 ) possesses a theoretical capacity of 642 mAh g −1 when 18 lithium ions insert into the Cr 8 O 21 structure totally. Further researches show that Cr 8 O 21 has an initial capacity of 395 mAh g −1 and a reversible capacity of 260 mAh g −1 [9,10]. Therefore, the use of oxides with high specific capacity as cathode materials for LIBs is a promising choice to improve the energy density and to achieve low cost. Li-free cathode materials show the potential to provide energy densities higher than traditional Li-ion batteries while coupled with Li metal anodes due to the advantages of high specific capacity (3800 mAh g −1 ) and low reduction potential (−3.04 V vs. the standard hydrogen electrode) [11][12][13][14]. While in a conventional Li-ion battery system, the growth of Li dendrites may induce the internal short circuits that may lead to fire, explosion, and uncontrolled electrolyte decomposition. The use of solid-state electrolytes to construct all-solid-state (ASS) batteries is conducive to circumvent these issues [15][16][17]. In addition, the solid electrolyte can be made extremely thin so that it can enhance the energy density of the battery [15,18]. Therefore, the ASS battery consisted of high-capacity metal oxide cathode and lithium metal anode is considered as one of the most promising next-generation battery systems.
Solid electrolytes are usually classified as inorganic solid electrolytes and polymer electrolytes. Polymer electrolytes have attracted much interest owing to the advantages of enhanced resistance to volume changes of the electrodes during the charge/discharge processes, improved safety features, excellent flexibility, and processability [19]. However, the possible chemical incompatibility between oxide cathode materials and polymer electrolytes should be taken into account. Polyethylene oxide (PEO) is the most promising candidate as a solid solvent for lithium salts owing to the flexible ethylene oxide segments and ether oxygen atoms, which have a strong donor character and thus readily solvate Li + cations. More importantly, it is commercially available in a relatively pure state at a very reasonable cost. Therefore, effective measures should be taken to solve the problems of chemical incompatibility between metal oxides and solid electrolytes to improve the specific energy and the cycling performance of the ASS lithium-ion batteries.
Using low-cost, cobalt-free, and high-performance metal oxides as cathode materials is beneficial to the development of solid-state lithium batteries for practical applications [20]. For example, Zhang et al. [21] and Shim et al. [22] investigated Li-V 2 O 5 solid-state lithium batteries with a solid polymer electrolyte (PEO and PEG, respectively). These studies have improved the cycling stability, but the number of cycles was still no greater than 50 cycles. In addition, studies on oxides (e.g., Cr 8 O 21 ) with higher discharge-specific energy and stronger oxidizing used in solid-state batteries have been rarely reported. In general, there is a side reaction between the cathode material and the solid electrolyte in the solid-state battery [17,23,24], which deteriorates the performance of the battery. When oxides with higher oxidation are selected as cathode materials, the interface problems are more prominent (mainly chemical stability) [17,24]. To solve the interface problem and reduce the interface resistance, interfacial modification is an effective approach [17,23,25]. Meanwhile, the coating layer should have suitable Li ionic conductivity, and chemical and electrochemical stability.
Herein, a new all-solid-state battery was prepared by using Cr 8 O 21 and lithium metal as the cathode material and the anode material, respectively, and the PEO-LiTFSI as the solid electrolyte. The performance of Cr 8 O 21 cathode in all-solid-state lithium metal batteries (SSLMBs) was evaluated. The coating of the Al 2 O 3 layer on the Cr 8 O 21 electrode by the ALD technique was introduced to prevent the solid electrolytes from strong oxidization of Cr 8 O 21 , which improves the cycle life of the batteries. The capacity retention rate after 125 cycles is 87.4%, and the energy density is up to 630 Wh kg −1 (based on Cr 8 O 21 materials), indicating a promising safe, high energy density secondary battery system.

Preparation of Cr 8 O 21 Powder
Methods for the preparation of Cr 8 O 21 powder are similar to those reported previously [9]. Typically, CrO 3 flakelets (Alfa Aesar, AR, Haverhill, MA, USA; thin sheets with a diameter of approximately 1 cm) were calcinated under oxygen flow (20 mL min −1 ) at 280 • C for 7 h. After cooling to room temperature, the obtained product was ground with agate mortar and sieved through a 300 mesh screen to obtain the Cr 8 O 21 powder.

Preparation of Cr 8 O 21 Electrode
To prepare the Cr 8 O 21 cathode electrodes, the active material (Cr 8 O 21 ), Ketjen black (KB, Lion Corporation, Tokyo, Japan), and poly(vinylidene fluoride) (PVdF, Arkema, Colombes, Colombes, France) (8:1:1 by weight) were mixed with an appropriate amount of N-methyl-2-pyrrolidinone (Alfa Aesar, AR, Haverhill, MA, USA) to form a slurry. The obtained slurry was then coated on an Al foil using a doctor blade (100 µm thickness). After drying at 120 • C for 12 h in vacuum, the electrode sheet was cut into round sheets with a diameter of 16 mm for both the test and ALD electrodes.

Preparation of PEO-Based Polymer Electrolyte Membrane
PEO (Mw = 100,000, Sigma-Aldrich, St. Louis, MO, USA) and lithium bis (trifluoromethanesulfonyl imide) (LiTFSI, TCI, Tokyo, Japan) were dissolved in tetrahydrofuran (THF, extra dry, TCI, Tokyo, Japan) with a molar ratio of PEO:LiTFSI = 8:1 under continuously stirring at 60 • C until complete dissolution. Then, 50 wt.% Al 2 O 3 nano-powder (Aladdin, Shanghai, China) was added to the above solution until mixing uniform. The mixture was subsequently cast into a thin film using a doctor blade (100 µm thickness) and dried at 80 • C for 12 h. The film was further dried in vacuum at 50 • C for another 12 h.

Preparation of ALD-Coated Cr 8 O 21 Electrode
Al 2 O 3 was deposited on the Cr 8 O 21 electrode or on the Si wafer at 140 • C in a Savannah 100 ALD system (Ultratech/Cambridge Nanotech, USA) using trimethylaluminum (TMA, (CH 3 ) 3 Al, 98%, STREM Chemicals, Newburyport, MA, USA) and distilled water (H 2 O) as precursors. The ALD details are as follows: (i) Put the electrode into the sample chamber, heat to 140 • C; (ii) Perform H 2 O pulse for 0.2 s, and then react for 45 s, followed by purging for 45 s; (iii) Conduct TMA pulse for 0.1 s, react for 45 s, and then purge for 45 s; (iv) Cool down to room temperature, inflate to atmospheric pressure, and then take out the Cr 8 O 21 electrodes. Repeat (ii), (iii) steps to achieve the selected number of cycles. In this study, 8,12,16, and 20 ALD cycles were selected for the Cr 8 O 21 electrode, and 100 ALD cycles were selected for Si wafers.

Preparation of All-Solid-State Cell
The PEO (LiTFSI)/THF solution (molar ratio of PEO:LiTFSI = 8:1) was cast on the Cr 8 O 21 electrodes to obtain composite electrodes with a closely integrated interface, and then the electrodes were dried at 70 • C for 12 h. The obtained Cr 8 O 21 electrode with PEO-based solid polymer electrolyte membrane and Li metal electrode was assembled together under 20 MPa. Finally, all-solid-state cells were assembled to 2430 coin-type cells in an argon-filled glove box. In accordance with the method described above, an all-solid battery prepared by stainless steel sheet instead of Cr 8 O 21 electrode was used for investigating the electrochemical stability of the PEO solid electrolyte. In contrast, the conventional organic electrolyte cells were assembled and tested. The Cr 8 O 21 electrode, Celgard 2320, and lithium metal were used as the cathode, separator, and anode, respectively. A total of 1 M LiPF 6 (Capchem, Shenzhen, China) in ethylene carbonate (EC, TCI, Tokyo, Japan) and dimethyl carbonate (DMC, TCI, Tokyo, Japan) mixture (volume ratio of EC to DMC was 1:1) was used as the electrolyte. The CR2430 coin-type cells were assembled in an Ar-filled glove box.

Characterizations
Powder X-ray diffraction (XRD, Rigaku M2400, Rigaku Co., Tokyo, Japan) was used to characterize the crystal structure of the Cr 8 O 21 powder with Cu Kα radiation. Scanning electron microscope (SEM, S-4800, Hitachi Inc., Tokyo, Japan) and transmission electron microscope (TEM, Philips Tecnai G2 F20, FEI Co., Hillsboro, OR, USA) were used to characterize the morphology and microstructure of both the powder and the electrodes. X-ray photoelectron spectroscopy (XPS, Axis Ultra DLD, Shimadzu-Kratos Co. Ltd., Kanagawa, Japan) was used to derive information on the bonding state of the samples. Fourier transform infrared spectroscopy (FTIR, Nicolet 6700, Thermo Fisher, Waltham, MA, USA) was used to characterize the change of PEO-based membrane. Surface profiler (P-l6+, KLA-Tencor, Milpitas, CA, USA) and atomic force microscope (AFM, Agilent AFM5500, Agilent Technologies, Inc., Santa Clara, CA, USA) in tapping mode at 512 Hz were used to calibrate the thickness of 100-ALD-Al 2 O 3 coated on a silicon plate.

Electrochemical Tests
The electrochemical measurements of all-solid-state cells were performed in a thermostat (Shanghai Bluepard, Shanghai, China) at 60 • C. Before the test, the cells were rest in the thermostat for 30 min. Galvanostatic tests were cycled between 1.5 and 4.2 V by using the LAND-CT2001A battery-testing instrument (Wuhan LAND Electronic Co. Ltd., Wuhan, China). Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) measurements were conducted on a Solartron 1470 + 1400 electrochemical workstation (AMETEK Inc., Berwyn, PA, USA). The scan potential range of CV measurement was 1.5-4.2 V at a scan rate of 0.1 mV s −1 for the Cr 8 O 21 electrode test and 1.0-5.0 V at a scan rate of 0.01 mV s −1 for the electrochemical stability of PEO electrolytes (at 60 • C). The EIS was examined in the frequency range of 10 MHz to 1 Hz with an AC amplitude of 5 mV.

Results and Discussion
The morphology and the crystal structure of the Cr 8 O 21 powder were characterized by SEM, TEM, and XRD. Figure 1a shows the XRD pattern of the Cr 8 O 21 powder, which displays an excellent consistency with the standard card of Cr 8 O 21 (JCPDS card no. 47-1312). Norby et al. determined that the space group of Cr 8 O 21 is P1 by using neutron diffraction techniques and verified the oxide state of chromium atoms in Cr 8 O 21 consisting of two trivalent chromium atoms and six hexavalent chromium atoms [26]. SEM image of the Cr 8 O 21 powder shown in Figure 1b reveals that the average particle size is approximately 10 µm. These particles exhibit angular surfaces and irregular shapes ( Figure 1c). This is mainly due to the fact that the obtained Cr 8 O 21 powder has been subjected to the process of crushing and grinding. The HRTEM image shown in Figure 1d Figure S1a-d present SEM images of the Cr 8 O 21 electrodes with different ALD cycles (0, 8,12,16). The Cr 8 O 21 particles without ALD coating present a rough surface because of pressing and rubbing during the grinding process. After coating with Al 2 O 3 by ALD, the surface of the Cr 8 O 21 particle becomes smooth [27,28], which can be attributed to the high homogeneity of the Al 2 O 3 ALD coating. Meanwhile, the integrity of the Al 2 O 3 layer was improved with the increase in ALD-coated cycles. After 8 ALD cycles, the coating cannot completely cover the surface of the Cr 8 O 21 particles. After 12 ALD cycles and 16 ALD cycles, the Al 2 O 3 coatings become more homogenous and complete. The Al 2 O 3 -coated Cr 8 O 21 electrodes were subjected to surface element analysis by EDS to determine the Al element content. The bar graph of Al element content in the coating with different ALD cycles (0, 8,12,16) is presented in Figure S2, which indicates that the Al element content increased with increasing ALD coating cycles. When the ALD coating cycle is 8, the Al content is 1.67 wt.%. As the ALD coating cycle number increases to 12 and 16, the Al element contents reach 3. 35  Figure S1a-d present SEM images of the Cr8O21 electrodes with different ALD cycles (0, 8,12,16). The Cr8O21 particles without ALD coating present a rough surface because of pressing and rubbing during the grinding process. After coating with Al2O3 by ALD, the surface of the Cr8O21 particle becomes smooth [27,28], which can be attributed to the high homogeneity of the Al2O3 ALD coating. Meanwhile, the integrity of the Al2O3 layer was improved with the increase in ALD-coated cycles. After 8 ALD cycles, the coating cannot completely cover the surface of the Cr8O21 particles. After 12 ALD cycles and 16 ALD cycles, the Al2O3 coatings become more homogenous and complete. The Al2O3-coated Cr8O21 electrodes were subjected to surface element analysis by EDS to determine the Al element content. The bar graph of Al element content in the coating with different ALD cycles (0, 8,12,16) is presented in Figure S2, which indicates that the Al element content increased with increasing ALD coating cycles. When the ALD coating cycle is 8, the Al content is 1.67 wt.%. As the ALD coating cycle number increases to 12 and 16, the Al element contents reach 3.35 and 10.2 wt.%, respectively.
The TEM observation reveals the formation of an amorphous Al2O3 coating layer on the surface of the Cr8O21 particle (Figure 2a, taking the 12 ALD sample as an example) [29][30][31]. The HRTEM shows that the Al2O3 layer has a noncrystalline structure [29], while the Cr8O21 particle shows clear crystalline patterns. The thickness of the Al2O3 layer is approximately 4.13 nm. In the ALD process, the Al2O3 layer thickness can reach 1 nm per 3 cycles (The thickness calibration results by AFM are shown in Figures S3 and S4). Therefore, the thickness of the Al2O3 layer of the 12 ALD cycles sample must be approximately 4 nm, in accordance with the TEM observation. The lattice fringe of the particle core is 0.387 nm,  [29][30][31]. The HRTEM shows that the Al 2 O 3 layer has a noncrystalline structure [29], while the Cr 8 O 21 particle shows clear crystalline patterns. The thickness of the Al 2 O 3 layer is approximately 4.13 nm. In the ALD process, the Al 2 O 3 layer thickness can reach 1 nm per 3 cycles (The thickness calibration results by AFM are shown in Figures S3 and S4). Therefore, the thickness of the Al 2 O 3 layer of the 12 ALD cycles sample must be approximately 4 nm, in accordance with the TEM observation. The lattice fringe of the particle core is 0.387 nm, which corresponds to the (1 1 1)  XPS was used to determine the composition of the ALD layer. Figure S6 shows the XPS spectra of the Cr8O21 electrode with different ALD cycles (0, 8,12,16), and Figure 3a displays the XPS peak of Al2p with different coating cycle numbers (0, 8,12,16). For the electrode without Al2O3 coating, the intensity of the Al2p peak is almost zero. The intensity of Al2p peaks increases with the increase in coating cycle numbers. However, when the coating cycle number reaches 12, the increase in Al2p peak intensity is no longer discernable [30][31][32][33][34]. It is well known that XPS can provide only an analysis depth of a few nanometers. Here, the thickness of Al2O3 ALD film is to be approximately 1 nm per 3 cycles. As 12 ALD coating was performed, the thickness value may achieve the limit of XPS detection depth. Therefore, even if the ALD cycles increase to 16, the intensity of the Al element is no longer significantly increased. The XPS spectra of Cr2p with different coating cycle numbers (0, 8,12,16) are shown in Figure 3b. When there is no ALD coating on the Cr8O21 electrode, the discernable XPS spectrum of Cr2p can be observed. When the Al2O3 ALD layer is coated on the electrode, the signal intensity of chromium decreases obviously (XPS spectra of Cr 2p with different coating cycle numbers (8,12,16) are shown in Figure S7). This is due to the fact that the thick Al2O3 layer on the electrode surface completely blocked the penetration of the X-ray, which corresponds to the variation of the signal intensity of the aluminum element. XPS was used to determine the composition of the ALD layer. Figure S6 shows the XPS spectra of the Cr 8 O 21 electrode with different ALD cycles (0, 8,12,16), and Figure 3a displays the XPS peak of Al2p with different coating cycle numbers (0, 8,12,16). For the electrode without Al 2 O 3 coating, the intensity of the Al2p peak is almost zero. The intensity of Al2p peaks increases with the increase in coating cycle numbers. However, when the coating cycle number reaches 12, the increase in Al2p peak intensity is no longer discernable [30][31][32][33][34]. It is well known that XPS can provide only an analysis depth of a few nanometers. Here, the thickness of Al 2 O 3 ALD film is to be approximately 1 nm per 3 cycles. As 12 ALD coating was performed, the thickness value may achieve the limit of XPS detection depth. Therefore, even if the ALD cycles increase to 16, the intensity of the Al element is no longer significantly increased. The XPS spectra of Cr2p with different coating cycle numbers (0, 8,12,16) are shown in Figure 3b. When there is no ALD coating on the Cr 8 O 21 electrode, the discernable XPS spectrum of Cr2p can be observed. When the Al 2 O 3 ALD layer is coated on the electrode, the signal intensity of chromium decreases obviously (XPS spectra of Cr 2p with different coating cycle numbers (8,12,16) are shown in Figure S7). This is due to the fact that the thick Al 2 O 3 layer on the electrode surface completely blocked the penetration of the X-ray, which corresponds to the variation of the signal intensity of the aluminum element. To evaluate the effects of the ALD coating Al2O3, ASS cells with ALD-Al2O3-coated Cr8O21 cathodes, PEO-based polymer electrolyte membranes, and Li metal anodes were assembled and electrochemically tested. The electrochemical operation window of the PEO-based polymer electrolyte membranes was estimated to be 1.0-4.6 V, as shown in Figure S8. Therefore, the voltage range of the charging/discharging test was set to 1.5-4.2 V for the ASS cells. Considering that Cr8O21 has strong oxidizing property, which may degrade organic substances, the Al2O3 ALD coating on the electrode performed at least eight cycles (labeled 8ALD). Then, the samples with different thicknesses of ALD coating were obtained after 12 cycles (12ALD), 16 cycles (16ALD), and 20 cycles (20ALD). As shown in Figure 4a, the ASS cell with an uncoated ALD electrode displays a similar discharge profile with the conventional organic electrolyte cell ( Figure S9). However, the ASS cell with an uncoated ALD electrode delivers a lower discharge capacity of 405 mAh g −1 compared with that of the organic electrolyte cell (450 mAh g −1 ). The decline of discharge capacity is attributed to the strong oxidizability of Cr8O21 material, which leads to oxidation of PEO electrolyte membrane during the rest at 60 °C for 30 min before testing. As the PEO membrane is oxidized, part chromium atoms in Cr8O21 are reduced to a lower valence state concurrently, which results in a decrease in discharge capacity. When the electrodes are coated by Al2O3 ALD, the discharge voltage decreases with the coating cycle. This phenomenon is ascribed to the Al2O3 ALD that increases the resistance for both electron and ion conduction [35][36][37], which serves as a barrier to ion and electron mobility in the charge/discharge process. For this reason, increasing coating thickness leads to gradually increasing polarization resistance [38][39][40]. The thickness of the ALD coating produces a completely opposite effect on the oxidation resistance and the polarization effect, resulting in a tendency for the discharge capacity to increase first and then decreasing. The discharge capacity of Cr8O21 gradually increases from 470 to 490 mAh g −1 when the coating cycles increase from 8 to 12 ALD. After that, the discharge capacity decreases with the increase in coating cycles. When the electrode was coated by 20 ALD, the discharge capacity was close to zero. The change of capacity can be explained by the synergy of three factors, including temperature, coating cycles, and the reaction of the electrode with the PEO membrane. The discharge capacity of the cell at 60 °C is higher than that at room temperature due to that elevation of temperature not only facilitates the migration of Li ions but also promotes the chemical reaction in the cell. On the one hand, ALD coating can inhibit the reduction in the Cr8O21 electrode and increase the polarization resistance. When the ALD cycles are less than 12, reduction in the Cr8O21 electrode and oxidation of the electrolyte membrane cannot be completely suppressed. The redox reaction between the electrode and the membrane reduces the discharge capacity. When the ALD coating cycle reaches 12, the redox reaction between the electrode and the membrane is completely overwhelmed, which leads to the maximum discharge capacity of 490 mAh g −1 . As the ALD cycle is greater than 12, although the redox reaction can be suppressed, the increase in polarization reduces the discharge capacity. When the coating cycle is 20, To evaluate the effects of the ALD coating Al 2 O 3 , ASS cells with ALD-Al 2 O 3 -coated Cr 8 O 21 cathodes, PEO-based polymer electrolyte membranes, and Li metal anodes were assembled and electrochemically tested. The electrochemical operation window of the PEObased polymer electrolyte membranes was estimated to be 1.0-4.6 V, as shown in Figure  S8. Therefore, the voltage range of the charging/discharging test was set to 1.5-4.2 V for the ASS cells. Considering that Cr 8 O 21 has strong oxidizing property, which may degrade organic substances, the Al 2 O 3 ALD coating on the electrode performed at least eight cycles (labeled 8ALD). Then, the samples with different thicknesses of ALD coating were obtained after 12 cycles (12ALD), 16 cycles (16ALD), and 20 cycles (20ALD). As shown in Figure 4a, the ASS cell with an uncoated ALD electrode displays a similar discharge profile with the conventional organic electrolyte cell ( Figure S9). However, the ASS cell with an uncoated ALD electrode delivers a lower discharge capacity of 405 mAh g −1 compared with that of the organic electrolyte cell (450 mAh g −1 ). The decline of discharge capacity is attributed to the strong oxidizability of Cr 8 O 21 material, which leads to oxidation of PEO electrolyte membrane during the rest at 60 • C for 30 min before testing. As the PEO membrane is oxidized, part chromium atoms in Cr 8 O 21 are reduced to a lower valence state concurrently, which results in a decrease in discharge capacity. When the electrodes are coated by Al 2 O 3 ALD, the discharge voltage decreases with the coating cycle. This phenomenon is ascribed to the Al 2 O 3 ALD that increases the resistance for both electron and ion conduction [35][36][37], which serves as a barrier to ion and electron mobility in the charge/discharge process. For this reason, increasing coating thickness leads to gradually increasing polarization resistance [38][39][40]. The thickness of the ALD coating produces a completely opposite effect on the oxidation resistance and the polarization effect, resulting in a tendency for the discharge capacity to increase first and then decreasing. The discharge capacity of Cr 8 O 21 gradually increases from 470 to 490 mAh g −1 when the coating cycles increase from 8 to 12 ALD. After that, the discharge capacity decreases with the increase in coating cycles. When the electrode was coated by 20 ALD, the discharge capacity was close to zero. The change of capacity can be explained by the synergy of three factors, including temperature, coating cycles, and the reaction of the electrode with the PEO membrane. The discharge capacity of the cell at 60 • C is higher than that at room temperature due to that elevation of temperature not only facilitates the migration of Li ions but also promotes the chemical reaction in the cell. On the one hand, ALD coating can inhibit the reduction in the Cr 8 O 21 electrode and increase the polarization resistance. When the ALD cycles are less than 12, reduction in the Cr 8 O 21 electrode and oxidation of the electrolyte membrane cannot be completely suppressed. The redox reaction between the electrode and the membrane reduces the discharge capacity. When the ALD coating cycle reaches 12, the redox reaction between the electrode and the membrane is completely overwhelmed, which leads to the maximum discharge capacity of 490 mAh g −1 . As the ALD cycle is greater than 12, although the redox reaction can be suppressed, the increase in polarization reduces the discharge capacity. When the coating cycle is 20, the electrode delivers no discharge capacity at all, indicating that the thicker layer of Al 2 O 3 completely impedes the diffusion of lithium ions.
1  When the coating cycle is less than 12, the discharge capacity rapidly declines to approximately 200 mAh g −1 after 10 cycles. Up cycling to 100 cycles, the discharge capacity is 40 mAh g −1 and 26 mAh g −1 for the 0ALD and 8ALD cells, respectively. The 8ALD cell has a worse cycling performance compared with 0ALD, which is mainly attributed to the coincidence of redox reaction and polarization effect. It is worth noting that the discharge capacity becomes stable from the 5th charge/discharge cycle when the ALD coating cycle is 12. It delivers a capacity of 260 mAh g −1 for the 125th cycle at 0.1 C with a discharge-specific energy of 630 Wh kg −1 . However, 16 ALD is found to deteriorate the cycling performance, although it can prevent the redox reaction occurrence completely [27,28,40]. Coulombic efficiency (CE) can be used to estimate the cycling life of LIB. Figure 4e also shows that the CE of the cell with 12 ALD cycles layer is higher than that of the others, which is consistent with the cyclic performance. It was reported that ALD Al 2 O 3 is not thermodynamically stable at low potentials. A stable Li-Al-O phase with high lithium-ion conductivity first forms after lithiation at low potentials, which is beneficial to the rapid transport of lithium ions [41][42][43][44][45][46][47][48]. Then the superfluous lithium ions can flow through the Li-Al-O phase to the adjacent active materials to participate in the charge/discharge reactions. A recent study also suggested that the Li-Al-O phase can be generated directly on the surface of a lithium electrode [49]. This may explain the experimental results when the ALD cycles are less than 12. A thick layer of Al 2 O 3 would inhibit the diffusion of lithium ions, thereby deteriorating the battery performance [43,[50][51][52]. Furthermore, it also increases the distance of ion transport. These negative effects may induce the deterioration of cycling performance with coating cycles of more than 12 ALD.
The effect of resistance on discharge performance was further verified by EIS tests, as shown in Figure 4f and Table 1. The equivalent circuit model of EIS analysis is displayed in Figure 4f. All EIS curves consist of a semicircle in a high-frequency range and an inclined straight line in a low-frequency range. It can be seen that the Al 2 O 3 coating layer does not affect the electrochemical reaction process, which is the same as the previous report [53]. The depressed semicircle in the high-frequency region is assigned to the charge transfer process (Rct), and the inclined line in the low-frequency region is related to the Li-ion diffusion resistance (Zw). Figure 4f and Table 1 indicate that when the ALD coating cycles were less than 20, all the Rs (series ohmic resistances) were approximately 42 Ω. This shows that thin ALD coating has little effect on Rs. The slight change in Rs may be caused by the difference in the preparation process of the coin cells. For the electrodes with ALD coating cycles of more than 16, increasing the thickness of ALD coating obviously leads to the increase in Rs. The Rct increases obviously with the increase in the number of ALD coating cycles. For the electrode without ALD coating, the Rct is 79.2 Ω. When eight cycles of ALD were coated on the electrode, Rct increased to 252.9 Ω. Continually increases the ALD coating cycles to 20, Rct reaches an amazing value of 660.9 Ω. The large resistance of the 20-ALD-coated electrode prevents it from releasing any capacity. The result is consistent with the variation of the first discharge plateau. Table 1 and Figure S10 also show that the 0 ALD electrode has minimum Warburg coefficient (σ) [54], on the other hand, of all ALD-coated electrodes, 12 ALD electrode has a minimum Warburg coefficient (σ), which is consistent with the results of electrochemical performance tests. To further explicate the effect of ALD coating on the cycling performance of ASSB, the PEO membranes at different states were subjected to FTIR analysis, and the associated Cr 8 O 21 electrodes were characterized by XPS analysis. Figure 5a shows the macrographs of the PEO electrolyte membranes directly contacted with the Cr 8 O 21 electrode or the ALD-coated electrode after keeping at 60 • C for a week. The color of the membrane tightly bonded to the electrode with ALD coating remains unchanged, while the color of the membrane tightly bound to the electrode without ALD coating turned yellow. This might be attributed to the degradation of PEO by Cr 8 O 21 oxidation. Figure 5c illustrates the FTIR spectra of the PEO membranes contacting Cr 8 O 21 with and without ALD coating and pure PEO powder. It is obvious that the FTIR patterns are similar. Figures S11 and S12 are partially enlarged spectra in the range of 1560 to 1720 cm −1 and 3200 to 3600 cm −1 , respectively. It is found that there are two weak absorption peaks at 3425 and 1640 cm −1 for the PEO membrane without ALD. The two peaks correspond to the vibration of the hydroxyl group and the carbon double bond, which can be ascribed to oxidative degradation reaction of PEO. This is consistent with the results reported in the literature [55,56]. In contrast, there are no corresponding vibration peaks for the membrane with ALD and pure PEO powder. Therefore, the ALD coating can effectively prevent the PEO from being oxidized. Figure 5b shows the XPS spectrum (Cr 2p) of the pristine Cr 8 O 21 electrodes, in which the intensity ratio of Cr +6 /Cr +3 peaks is higher than that of the Cr 8 O 21 electrode contacting with PEO membranes, as shown in Figure 5d. there are no corresponding vibration peaks for the membrane with ALD and pure PEO powder. Therefore, the ALD coating can effectively prevent the PEO from being oxidized. Figure 5b shows the XPS spectrum (Cr 2p) of the pristine Cr8O21 electrodes, in which the intensity ratio of Cr +6 /Cr +3 peaks is higher than that of the Cr8O21 electrode contacting with PEO membranes, as shown in Figure 5d. On the basis of the above discussion, the roles of the ALD Al2O3 layer in improving the performance of the all-solid-state batteries are summarized as follows: (i) The dense Al2O3 layer can physically insulate the electrode (Cr8O21) from direct contact with the polymer electrolyte membrane (PEO), which avoids side effects between Cr8O21 and PEO. (ii) Although Al2O3 itself is an insulator of electrons and ions, a suitable thickness (it is approximately 4 nm in this study) of the Al2O3 layer could form a stable Li-Al-O phase after lithiation, which is beneficial to the rapid transport of lithium ions. (iii) Li-Al-O has a stable chemical property, which can improve the interfacial stability between the solid electrolyte and the electrode ( Figure S13). (iv) The Al2O3 ALD coating can not only mechanically restrict the pulverization of the active material but also prevent the electrode materials from peeling off the current collector. It eventually improves the cycling stability of the electrode.

Conclusions
In summary, the ALD technique was applied to improve the electrochemical performance of the Cr8O21/Li all-solid-state batteries. The dense Al2O3 layer was evenly coated on the surface of the Cr8O21 electrode. The Al2O3 layer prevents the direct contact between the Cr8O21 electrode and the PEO film; therefore, the oxidation of the PEO electrolyte (ii) Although Al 2 O 3 itself is an insulator of electrons and ions, a suitable thickness (it is approximately 4 nm in this study) of the Al 2 O 3 layer could form a stable Li-Al-O phase after lithiation, which is beneficial to the rapid transport of lithium ions. (iii) Li-Al-O has a stable chemical property, which can improve the interfacial stability between the solid electrolyte and the electrode ( Figure S13). (iv) The Al 2 O 3 ALD coating can not only mechanically restrict the pulverization of the active material but also prevent the electrode materials from peeling off the current collector. It eventually improves the cycling stability of the electrode.

Conclusions
In summary, the ALD technique was applied to improve the electrochemical performance of the Cr 8 O 21 /Li all-solid-state batteries. The dense Al 2 O 3 layer was evenly coated on the surface of the Cr 8 O 21 electrode. The Al 2 O 3 layer prevents the direct contact between the Cr 8 O 21 electrode and the PEO film; therefore, the oxidation of the PEO electrolyte membrane by Cr 8 O 21 is depressed. Meanwhile, the thin Al 2 O 3 layer allows lithium ions to pass through and ensures the electrochemical reactions. The electrochemical tests indicated that the 12-cycle ALD-Al 2 O 3 -coated electrode delivers a high capacity of 260 mAh g −1 after 125 cycles at 0.1 C with a discharge-specific energy of 630 Wh kg −1 . The ALD technique may also be used in other battery systems for stabilizing the interface between the electrode and the solid electrolyte.  Figure S5: The XRD patterns of Cr 8 O 21 electrodes with different ALD cycles. Figure S6: XPS spectra of Cr 8 O 21 electrodes. Figure S7: XPS spectra of Cr 2p with different coating cycle numbers (8,12,16). Figure S8: CV curves of PEO members at a scan rate of 0.1 mV s −1 (60 • C). Figure S9: Galvanostatic charge/discharge profiles of Cr 8 O 21 electrode with liquid electrolyte (1 M LiPF6 EC/DMC). Figure S10: The linear fits of the relationship between Z and ω −0.5 of Cr 8 O 21 electrode with different ALD cycles (0, 8,12,16,20). Figure S11: The partially enlarged spectra in the range of 1560 cm −1 to 1720 cm −1 . Figure S12: The partially enlarged spectra in the range of 3200 cm −1 to 3600 cm −1 . Figure S13