Eliminating graphite exfoliation with an artificial solid electrolyte interphase for stable lithium-ion batteries

www.small-journal.com 2107460 (1 of 9) © 2022 The Authors. Small published by Wiley-VCH GmbH. ReseaRch aRticle Eliminating Graphite Exfoliation with an Artificial Solid Electrolyte Interphase for Stable Lithium-Ion Batteries Junhua Zhou, Keni Ma, Xueyu Lian, Qitao Shi, Jiaqi Wang, Zhujie Chen, Lingli Guo, Yu Liu, Alicja Bachmatiuk, Jingyu Sun, Ruizhi Yang, Jin-Ho Choi, and Mark H. Rümmeli* J. Zhou, K. Ma, X. Lian, Q. Shi, J. Wang, Z. Chen, L. Guo, Y. Liu, J. Sun, R. Yang, J.-H. Choi, M. H. Rümmeli College of Energy Soochow Institute for Energy and Materials InnovationS (SIEMIS) Key Laboratory of Advanced Carbon Materials and Wearable Energy Technologies of Jiangsu Province Soochow University Suzhou 215006, P. R. China E-mail: [email protected] A. Bachmatiuk LUKASIEWICZ Research Network PORT Polish Center for Technology Development Stablowicka 147, Wroclaw 54-066, Poland J. Sun Beijing Graphene Institute (BGI) Beijing 100095, P. R. China M. H. Rümmeli Leibniz Institute for Solid State and Materials Research Dresden P.O. Box 270116, Dresden D-01171, Germany M. H. Rümmeli Centre of Polymer and Carbon Materials Polish Academy of Sciences M. Curie-Sklodowskiej 34, Zabrze 41-819, Poland M. H. Rümmeli Institute of Environmental Technology VSB-Technical University of Ostrava 17. Listopadu 15, Ostrava 708 33, Czech Republic The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/smll.202107460. DOI: 10.1002/smll.202107460 1. Introduction Graphite with a gentle theoretical specific capacity (372 mAh g−1, LiC6), a very flat and low potential (≈0.2V vs Li/Li+), and a low cost, has been the major market share (>90%) of anode materials since lithiumion batteries (LIBs) were commercialized in 1991.[1,2] Although well-known next-generation anode materials are Si/graphite or SiO/graphite composites, the weight ratio of graphite in these composites exceeds 50%, because both Si and SiO demonstrate huge volume expansion (≈370% for Si; ≈200% for SiO) upon cycling.[3–7] Therefore, graphite anodes cannot be fully replaced in the near future. However, severe capacity loss and lithium plating during fast charging degrade the properties and increase safety risks of graphite-based LIBs.[8,9] These challenges are closely related to the sluggish kinetics and the low redox potential (≈0 V) of graphite. Specifically, large polarization along with high C-rate charging yields an electrochemical plateau loss and reduces the potential sufficiently to enable Li metal deposition. Interface modification, including direct coating of the graphite surface or introduction of novel electrolyte additives for robust SEI, is among the most popular methods for enhancing the fast-charging ability and cyclability of graphite anodes. For instance, graphite cannot cycle normally in an early PC-based electrolyte owing to serious graphite exfoliation, while the following EC-based electrolyte can solve this problem efficiently by forming a stable SEI.[10] Moreover, coating has also been demonstrated to effectively improve the lithium diffusion rate, by offering 3D diffusion, rather than 2D diffusion in bare graphite anodes.[11] Still, the thickness of the SEI layer and the uniformity of the coating layer cannot be controlled easily by simply adjusting the electrolyte compositions and coating agents.[12–15] Here, we report an ultrathin and uniform coating layer on graphite surfaces as an artificial SEI, which can effectively solve severe graphite exfoliation and continuously increasing SEI during prolonged usage, enabling extremely stable operation of graphite anodes. Specifically, potato-like natural graphite (G), (Figure S1, Supporting Information) was first coated with a polydopamine (PDA) layer (the optimized mass ratio is 14%, Figure S4, Supporting Information) with a self-polymerization reaction, which finally transformed into amorphous carbon (G@C) by calcination at 1000°C under an Ar atmosphere (Figure1a, see the Experimental Section for Although graphite materials with desirable comprehensive properties dominate the anode market of commercial lithium-ion batteries (LIBs), their low capacity during fast charging precludes further commercialization. In the present work, natural graphite (G) is reported not only to suffer from low capacity during fast charging, but also from charge failure after many charging cycles. Using different characterization techniques, severe graphite exfoliation, and continuously increasing solid electrolyte interphase (SEI) are demonstrated as reasons for the failure of G samples. An ultrathin artificial SEI is proposed, addressing these problems effectively and ensuring extremely stable operation of the graphite anode, with a capacity retention of ≈97.5% after 400 cycles at 1 C. Such an artificial SEI modification strategy provides a universal approach to tailoring and designing better anode materials for next-generation LIBs with high energy densities. Small 2022, 18, 2107460 © 2022 The Authors. Small published by Wiley-VCH GmbH. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. www.advancedsciencenews.com www.small-journal.com 2107460 (2 of 9) © 2022 The Authors. Small published by Wiley-VCH GmbH. details). The modified G@C anode exhibited excellent operation, with a capacity retention of ≈97.5% after 400 cycles at 1 C, whereas the G compounds simply cycled stably for 250cycles, owing to a strange charge failure. Further failure analysis using X-ray diffraction (XRD), Raman spectroscopy, scanning electron microscopy (SEM), and transmission electron microscopy (TEM) revealed that the bare G sample experienced severe graphite exfoliation, which was fully inhibited by our interface-modification strategy (Figure1b). Such an artificial SEI strategy provides a universal approach to tailoring and designing better anode materials for next-generation LIBs with high energy densities. 2. Results and Discussion 2.1. PhysicalCharacterization of the Artificial SEI The diffraction peaks of both G and G@C in the XRD patterns (Figure2a) suggest hexagonal graphite with the P63/mmc space group (PDF card: 08-0415), despite the notably reduced (002) peak intensity of G@C compared with that of G, resulting from the poorly crystallized external coating layer. Based on the Raman profiles in Figure 2b, the G sample exhibits a typical graphite spectrum, with a very strong G peak (at ≈1580 cm−1) and a weak D peak (at ≈1350 cm−1),[16] and with a low ID/IG intensity ratio of ≈0.12. As for the G@C anode, both the G and D peaks are very strong, with a relatively high ID/IG value of ≈0.76, further demonstrating the highly noncrystallized nature of the artificial SEI. Low-magnification SEM and TEM images in Figures S1 and S2 (Supporting Information) show that both G and G@C exhibit a potato-like morphology, which characterizes natural graphite, with a spheroidization process of raw graphite flakes for reducing their specific surface area and increasing their Coulombic efficiency. High-magnification SEM images in Figure2c,d reveal clear graphite flakes for G but not for G@C, owing to the complete amorphous carbon coating layer on the surface of G. High-resolution TEM (HRTEM) images of G and G@C, taken from the [010] zone axis and shown in Figure2e,f, reveal that the interlayer spacing along the c-axis is ≈0.35nm in both cases, corresponding to the (002) lattice plane of hexagonal graphite. Moreover, a less-crystalized coating layer with a higher contrast than the bulk graphite phase is evident for G@C (thickness, ≈4nm). Such a coating layer is even thinner than many SEIs driven by electrolytes after cycling,[17] which cannot inhibit the diffusion of Li+. The composition of the G and G@C samples was further characterized by surface-sensitive X-ray photoelectron spectroscopy (XPS), and the results are shown in Figure 2g–i. The G compound exhibits the typical CC (≈283.8eV), CO (≈284.8eV), CO (≈289.6eV)[18,19] species in the C 1s spectra (Figure2g), and CO (≈532eV) in the O 1s spectra (Figure2h), where the content of C is as high as 97.58%, and the tiny oxygen signal contributes to contamination. Such a high ratio of C is advantageous for improving the Coulombic efficiency of graphite, because some other elements such as O, N, and S will store Li+ irreversibly.[17] For the G@C sample, clear peaks arising from the CC, CO, and CO species in the C 1s spectra, as well as CO and CO (≈533.1eV) species in the O 1s spectra are also present. Furthermore, the newly present CO signatures in the O 1s spectra, and smaller CC, higher CO, and CO species ratio in Figure2i reveal that the oxygen content of G@C is obviously higher than that of G, owing to the elemental O in the coating layer. Overall, the artificial SEI for G@C is an amorphous, thin, and relatively oxygen-rich (compared with the bulk region) carbon coating layer. 2.2. Electrochemical Characteristics of the G and G@C Anodes 2025-type LIBs were assembled with 1 m LiPF6 in EC-DMC (1:1 by volume) as the electrolyte. The charge–discharge curves at 0.1 C (1 C = 350 mAh g−1, Figure3a) show that the initial reversible specific capacities of G and G@C are 349 and 355 mAh g−1, with the respective Coulombic efficiencies of Small 2022, 18, 2107460 Figure 1. Synthesis a) and functional mechanism b) of artificial SEI. www.advancedsciencenews.com www.small-journal.com 2107460 (3 of 9) © 2022 The Authors. Small published by Wiley-VCH GmbH. 88.1% and 86.4%. The slightly higher irreversible capacity of G@C compared with that of G contributes to its lower C content and higher O content (Figure 2i). Differential curves (V≈ dQ/dV, Figure3b) and in situ XRD tests (Figure S3, Supporting Information) reveal that G and G@C experience multistep lithiation processes, yielding a series of lithium-graphite intercalation compounds with a special stage structure, that is, graphite→dilute stage-1→stage-4→stage-3→“liquid like” stage2→stage-2(LiC12)→stage-1 (LiC6).[20,21] From the amplified plot in Figure3b, the peak intensity (especially for peak I) of G@C is higher than that of G, but its voltage polarization between the oxidation and reduction peaks is smaller, suggesting better kinetics. The rate performance in Figure3c reveals that G@C exhibits a larger capacity than G samples in the 0.1–3 C range, despite of both showing low capacity at high C-rates, owing to the disappearance of the low potential plateau in Figure S5 (Supporting Information). Carbon coating layer modification for improving the rate performance has been reported elsewhere, owing to the 3D Li+ diffusion rather than the 2D diffusion in bare graphite anodes.[11] In view of the relatively low capacity of the G and G@C anodes at high C-rates, we applied a staircase discharge mode (SD mode, see the Experimental Section for details), for characterizing the behavior of these devices with respect to prolonged cycling. The SD mode refers to discharging cells first at a high rate and then at a low rate, to take full advantage of the electrode’s capacity, which is similar to the commonly used constant-voltage discharging in commercial tests (Figure 3e). As shown in Figure3d, the G@C anode is extremely stable, with a capacity retention of ≈97.5% after 400 cycles at 1 C in the SD mode, whereas the G anode cycles normally only for 250cycles, owing to a strange charge failure (Figure 3f). Figure S6 Small 2022, 18, 2107460 Figure 2. Physicalcharacterization of G and G@C. a,b) XRD (a) and Raman (b) patterns. c–f) SEM and TEM images of G (c–e) and G@C (d–f). The amplified view in (e,f) shows the (002) lattice plane of graphite. The yellow dashed lines in (f) indicate that the typical thickness of artificial SEI is ≈4nm. g–i) C 1s (g), O 1s (h) XPS spectra, and corresponding species ratios (i). The scale bars in panels (c,d) and (e,f) are 1µm and 10nm, respectively. www.advancedsciencenews.com www.small-journal.com 2107460 (4 of 9) © 2022 The Authors. Small published by Wiley-VCH GmbH. (Supporting Information) also reveals the excellent stability of the G@C anode, with an average Coulombic efficiency of ≈99.97% over 400 cycles. Our strategy ensures an attractive cycling performance from a graphite anode without at the expense of initial Coulombic efficiency as compared with the latest works reported in the literature (Table S1, Supporting Information). To illustrate the universality of our strategy, 2025-type potassium-ion batteries (PIBs) were also assembled with a K metal disk and KPF6 (0.8 m) in an EC-DMC mixture (1:1 by volume) acting as the counter electrode and electrolyte, respectively. In the PIB case, the G anode also exhibited a worse rate performance than the G@C anode, and its charge failure was more severe than that for LIBs (Figure S8, Supporting Information). 2.3. Surface Structure Evolution During Initial Cycling A special cell setup with a glass window enabling the transmission of a laser light was used for synchronously recording the samples’ Raman signals during the battery operation. From Figure4a,b, both the G and G@C anodes exhibit variational D (≈1350 cm−1, nongraphitized peak, A2g breathing modes of in-plane graphene rings, Figure 4e) and G (≈1580 cm−1, graphitized peak, high-frequency E2g phonon at Γ, Figure 4e) peaks during the charge-discharge process.[22,23] The extremely strong background signal marked with a dashed elliptical ring in Figure4b is attributed to the fluorescence effect associated with a short excitation wavelength (532nm in this work). To better demonstrate the variations, including the position and intensity of the Raman peaks, especially for the G anode, amplified views are shown in Figure 4c,d, for the Raman shift from 1540 to 1620 cm−1. For the G anode (Figure4c; and Figure S9, Supporting Information), the G peak gradually shifts toward higher wavenumbers, from 1580 to 1591 cm−1, during the discharging process, increasing the concentration of dopants (including Li+) and increasing the risk of other faults owing to the lithium intercalation.[24–26] Moreover, this trend is irreversible and specific, and the G peak can be used to reliably infer the wavenumber during the entire charging process. The fact that the irreversible shift of the G peak contradicts the corresponding reversible electrochemical curves (Figure 4a) and the ex situ XRD results (Figure S10, Supporting Information) reveals that the Raman signature is relatively surface-sensitive, compared with the large particle sizes of the G and G@C samples (≈15µm). In other words, such irreversibility only applies to the surface, rather than to the bulk region. However, the G peak position for the G@C anode is relatively constant, compared with that for the G anode, demonstrating that our artificial SEI layer improved the reversibility of the surface structure. From Figure 4f,g, the irreversibly decreased G peak intensity and the increased ID/IG ratio for the G anode deliver its continuously enhanced defect concentration, whereas the G@C anode exhibits much better reversibility, in agreement with the G peak position. 2.4. Evidence of Graphite Exfoliation after Long Cycling The full XRD spectra (10°–80°) in Figure S11 (Supporting Information) reveal that both G and G@C exhibit decreased (002) diffraction intensity peaks, corresponding to the stress accumulation Small 2022, 18, 2107460 Figure 3. Electrochemical performance of G and G@C in LIBs. a,b) Charge–discharge curves (a) and corresponding differential plots (b) at 0.1 C (1 C = 350 mAh g−1), for the 0.005–2V range. c,d) C-rate (c) and cycling (d) performance at 1 C. e,f) Charge–discharge curves of G (f) and G@C (e) at the 250th cycle. The stair discharge (SD) mode was used in this work for the cycling test (see the Experimental Section for details). www.advancedsciencenews.com www.small-journal.com 2107460 (5 of 9) © 2022 The Authors. Small published by Wiley-VCH GmbH. with cycling. Further amplified views of the XRD patterns (26°–28°) in Figure5a show that the (002) peak of the G sample shifts toward lower diffraction angles after 100 cycles, compared with that without cycling, yielding a wider layer space along the c-axis according to the well-known Bragg equation (2d·sinθ= n·λ). In Figure5b, both the ID/IG ratio and the full width at half maximum for the G anode increase notably after cycling, whereas in the G@C case (Figure5c), the ID/IG ratio almost does not change, demonstrating a super-stable surface structure even after 100 cycles. Graphene layer exfoliation for the G anode after cycling was directly observed using SEM and TEM. In Figure5d, the rough surface of the G anode contributes to the formation of the SEI layer, owing to the decomposition of the electrolyte at a low potential (≈0.7V vs Li/Li+),[17] resulting in a low initial Coulombic efficiency. More importantly, internal graphene layers coated by the SEI film are also evident, implying the exfoliation of the G anode owing to the intercalation of Li+ and electrolyte molecules upon cycling, consistent with the increased layer spacing inferred from the XRD results. The exfoliated graphene layer is also observed in the TEM images (Figure5e; and Figure S13, Supporting Information), with a typical thickness of ≈5.7 nm (17 graphene layers, Figure 5f). As for the G@C sample, the SEM image in Figure 5g reveals that it is also coated by a rough SEI layer, but no graphene layers are observed, suggesting that the artificial SEI effectively inhibits graphite exfoliation. The HRTEM images in Figure5h,i further reveal that all graphene layers are tightly organized, consistent with the local structure of natural graphite. Of note, batteries, including the LIBs and PIBs presented in this work, operate using a common electrolyte without any additives such as fluoroethylene carbonate (FEC) and vinylene carbonate (VC), which are very important for forming SEIs.[27,28] The G samples in such electrolytes cannot yield stable SEI Small 2022, 18, 2107460 Figure 4. In situ Raman test results for G and G@C. a–d) In situ Raman results (a,b), 1000–2000 cm−1, and corresponding amplified views (c,d) 1540–1620 cm−1. e) Carbon motions in the G and D modes. f,g) The intensity of the G peak and ID/IG value statistics for the special state of charge for data in panels (a,b). www.advancedsciencenews.com www.small-journal.com 2107460 (6 of 9) © 2022 The Authors. Small published by Wiley-VCH GmbH. layers, giving rise to the intercalation of both Li+ and electrolyte molecules, resulting in severe graphite exfoliation and unwanted charging failures. Interestingly, our artificial SEI significantly alleviates this problem, yielding normal operation for more than 400 cycles, for the G@C sample. 2.5. Interphase Composition and Kinetics The interphase compositions of the G and G@C anodes after cycling were characterized by surface-sensitive XPS. The deconvoluted C 1s spectra (Figure6a) for both G and G@C contain three peaks, belonging to CC, CO, and CO species with various electronegativities. Nevertheless, for each species, the ratios with respect to the peak areas are very different; specifically, CO is the highest compound for the G sample, but in the G@C case, the highest compound is CC. This means that the G anode has a higher O but a lower C content than the G@C anode (Figure6c), indicating that its thicker SEI is driven by the decomposition of the electrolyte, because the O compounds in the interphase mainly originate from electrolyte solvents. This is reasonable because the G anode without an artificial SEI is likely to experience severe graphite exfoliation, which is likely to increase the specific surface area, thus causing electrolyte depletion and battery death after many cycles. From the O 1s, P 2p, F 1s, and Li 1s spectra (Figure S14, Supporting Information), the SEIs of these samples also contain the PFx−, LixPOyFz, and LiF species, and the extra N 1s patterns observed in G@C (Figure6b) are attributed to the carbon coating layer decomposition from PDA. The interfacial impedance values for G and G@C can be inferred from Nyquist plots. All of the analyzed samples (Figure6d,e) exhibit an intercept with the x-axis, a small semicircle, a large semicircle, and a slope from high to low frequencies, corresponding to the resistances attributed to the electrolyte (Re), SEI (RSEI), charge transfer (Rct), and solid electrode diffusion process, respectively.[29] From Figure6f; and Figure S17 (Supporting Information), the impedances for both G and G@C are notably reduced, especially for RSEI and Rct, contributing to the typical activation process. Meanwhile, G@C always exhibits a smaller resistance regardless of the number of cycles, consistent with the above-mentioned electrochemical behavior. Small 2022, 18, 2107460 Figure 5. Bulk structure evolution of G and G@C, after 100 cycles. a–c) XRD (a) and Raman (b,c) patterns. d–i) SEM (d,g), TEM (e,h), and HR-TEM (f,i) images. Panels (f) and (i) show amplified views of areas marked with yellow dashed rectangles in panels (e) and (h), respectively. The scale bars in (d,g), (e,h), and (f,i) are 1µm, 20nm, and 2nm, respectively. www.advancedsciencenews.com www.small-journal.com 2107460 (7 of 9) © 2022 The Authors. Small published by Wiley-VCH GmbH. Small 2022, 18, 2107460 Li+ diffusion in solid electrodes is the rate-determining step in electrochemistry, which can be explored using the galvanostatic intermittent titration technique (GITT). Each step in the GITT test includes a discharge or charge at a small constant current and a long-time rest process, providing information about Δt (voltage change during discharge/charge at time t), Δs (change in the steady-state voltage at this step), Ohmic polarization, and voltage hysteresis (Figure 6f,g).[30] From Figure 6h,i; and Figure S18 (Supporting Information), the G@C anode exhibits smaller Ohmic polarization and voltage hysteresis, but a higher Li+ diffusion coefficient at almost all states of charge (SOC) than the G anode, suggesting that the artificial SEI can also improve kinetics. 3. Conclusions In this study, the interphase property was demonstrated to be crucial for improving both the long cycling and rate performance of graphite anodes. The modified G@C anode operated stably, with a capacity retention of ≈97.5% after 400 cycles at 1 C, whereas the bare G compounds were stable only for 250 cycles, owing to strange charge failures. Using different characterization techniques, including in situ Raman spectroscopy, XRD, TEM, and XPS, severe graphite exfoliation and continuously increasing SEI were demonstrated as causative factors underlying the failure of G anodes after many operation cycles. new A novel ultrathin (≈4nm) artificial SEI was designed, which effectively inhibited graphite exfoliation and increased the interphase stability, thus ensuring long-term operational stability of G@C anodes. The artificial SEI layer can also improve the battery’s C-rate performance, reducing the interface impedance, but increasing the bulk Li+ diffusion coefficient. Such an artificial SEI modification strategy provides a universal approach to tailoring and designing better anode materials for next-generation LIBs with high energy densities. Figure 6. Interphase composition and kinetics of the G and G@C anodes. a–c) C 1s (a), N 1s (b) XPS spectra, and corresponding species ratios (c). d,e) Nyquist plots (d), the corresponding electrolyte resistance (Re), SEI (RSEI), and charge transfer (Rct) (e). f–i) GITT curves (f,g), the corresponding Ohmic polarization (h), and Li+ diffusion coefficient (i). www.advancedsciencenews.com www.small-journal.com 2107460 (8 of 9) © 2022 The Authors. Small published by Wiley-VCH GmbH. Small 2022, 18, 2107460 4. Experimental Section Material Synthesis: Natural graphite (3 g, Iopsilion Corporation) and dopamine hydrochloride (0.5g) were dispersed in the Tris buffer (500mL, 20 × 10−3 m, Sigma-Aldrich) at a pH of ≈8.5. The mixture was continuously stirred for 24 h at room temperature. After the self-polymerization of dopamine on natural graphite was completed, the solid product was collected, washed with deionized water several times, lyophilized, and finally annealed under Ar at 1000°C for 2 h, at a heating rate of 5 °C min−1. Structure Characterization: All XRD patterns in this study were obtained using a Bruker D8 Advance diffractometer with a Cu Kα radiation source operating at 40kV and 40mA. For ex situ XRD tests, coin cells after cycling were first disassembled in a glove box to obtain a working electrode, which was then washed with the DMC solvent to remove the remaining electrolyte, and dried at room temperature. The dried electrode was then sealed on an XRD specimen stage using a polyimide tape, to avoid contact with air. For in situ XRD tests, a stainless-steel Swagelok-type battery with a Be window was assembled and tested at 0.1 C, in the 0.005–2 V range, with >100 XRD scans recorded synchronously. In situ Raman spectra were measured using an in situ Raman battery setup with a micro-Raman spectrometer (Horiba Jobin Yvon, HR Evolution), using a neon laser with a wavelength of 532nm. SEM and corresponding EDS mapping images were obtained using a Hitachi SU8010 SEM operating at 10kV and 10mA, and 15kV and 15 mA, respectively. XPS patterns were obtained using a Thermo Fisher Escalab 250Xi XPS with an Al Kα radiation source in a vacuum transfer chamber, to avoid contact with air. All binding energies in this work were calibrated based on C contamination using the C1s peak at 284.8eV. TEM images were taken using a spherical aberration-corrected FEI Titan Themis Cubed G2 300 TEM operating at 300kV. Electrochemical Measurements: The G and G@C electrodes were prepared by first mixing G or G@C active materials, Super P (TIMCAL) conductive additives, and a carboxymethyl cellulose (CMC, DAICEL) binder, at a mass ratio of 90:5:5, in deionized water solvent, using magnetic stirring for 12 h. The obtained slurry was then cast on a Cu current collector using a scraper and dried at 80°C for 2 h. Subsequently, the electrode was cut into disks (diameter, 14cm) with an active material loading of ≈3mg cm−2 (thickness, ≈40µm). Subsequently, it was further dried at 120°C for 6 h in a vacuum oven prior to being transferred to the glove box. 2025-type LIBs were assembled in the glove box with the G or G@C work electrodes, a Li metal disk (diameter, 1.5cm; thickness, 0.5mm; Alfa Aesar; 99.9%) as the counter electrode, a PP-PE-PP separator (Celgard 2325), and 1 m LiPF6 in an EC-DMC (1:1 by volume) electrolyte (the amount of electrolyte in each battery was 80µL). 2025-type PIBs were assembled using the same procedure as for LIBs, except that the replacement of KPF6 (0.8 m) in an EC-DMC mixture (1:1 by volume) acted as the electrolyte, and a K metal disk (diameter, 1.5cm; thickness, 1mm; Alfa Aesar; 99.9%) acted as the counter electrode. The cycling and C-rate performances at 0.5 C, 1 C, 2 C, 5 C, and 10 C (1 C = 350mA g−1) in the 0.005–2V range (vs Li/Li+) with the initial three formation cycles at 0.1 C were tested at 25°C using a CT3001A system (Wuhan LAND). Electrochemical impedance spectroscopy (EIS) was performed in the 0.01–1000000Hz, with a bias voltage of 0.005V, using the CHI800D system (CH Instruments). The “stair” discharge mode (SD mode) was: rest 3min → constant current discharge at 1 C to 0.005V, constant current discharge at 0.5 C to 0.005V, constant current discharge at 0.1 C to 0.005V. Supporting Information Supporting Information is available from the Wiley Online Library or from the author. Acknowledgements J.Z., K.M., and X.L. contributed equally to this work. This work was supported by the National Natural Science Foundation of China (Grant No. 52071225), the Czech Republic from ERDF “Institute of Environmental Technology – Excellent Research” (No. CZ.02.1.01/0.0/0 .0/16_019/0000853). M.H.R. thanks the Sino-German Research Institute for support (Project GZ 1400). J.H.C. thanks the National Natural Science Foundation of China (Grant No. 11874044). J.Y.S. thanks the National Natural Science Foundation of China (No. 51702225), Beijing Municipal Science and Technology Commission (No. Z161100002116020), and the Natural Science Foundation of Jiangsu Province (No. BK20170336). R.Z.Y. thanks the National Natural Science Foundation of China (Grant Nos. 51972220 and 51572181), the National Key Research and Development Program of China (No. 2016YFB0100200), and the Key University Science Research Project of Jiangsu Province (No. 20KJA480003). Open access funding enabled and organized by Projekt DEAL. Conflict of Interest The authors declare no conflict of interest. Data Availability Statement Data sharing is not applicable to this article as no new data were created or analyzed in this study. Keywords artificial solid electrolyte interphase (SEI), graphite anodes, graphite exfoliation Received: December 2, 2021 Revised: December 28, 2021 Published online: February 27, 2022 [1] Y. Li, Y. Lu, P. Adelhelm, M. M.Titirici, Y. S. Hu, Chem. Soc. Rev. 2019, 48, 4655. [2] J. B.Goodenough, K. S.Park, J. Am. Chem. Soc. 2013, 135, 1167. [3] S.Chae, M.Ko, K.Kim, K.Ahn, J.Cho, Joule 2017, 1, 47. [4] M.Ko, S.Chae, J.Ma, N.Kim, H.-W.Lee, Y.Cui, J.Cho, Nat. Energy 2016, 1, 16113. [5] N.Kim, S.Chae, J.Ma, M.Ko, J.Cho, Nat. Commun. 2017, 8, 812. [6] J. W.Choi, D.Aurbach, Nat. Rev. Mater. 2016, 1, 16013. [7] S.Zhu, J.Zhou, Y.Guan, W.Cai, Y.Zhao, Y. Zhu, L.Zhu, Y.Zhu, Y.Qian, Small 2018, 14, 1802457. [8] D.Aurbach, E.Zinigrad, Y. Cohen, H.Teller, Solid State Ion 2002, 148, 405. [9] H.Zhu, J. A.Russell, Z.Fang, P.Barnes, L.Li, C.Efaw, A.Muenzer, J.May, K.Hamal, I. F.Cheng, P. H.Davis, E.Dufek, H.Xiong, Small 2021, 17, 2105292. [10] J.Zhang, Z.Cao, L.Zhou, G.Liu, G.-T.Park, L.Cavallo, L.Wang, H. N.Alshareef, Y.-K.Sun, J.Ming, ACS Energy Lett. 2020, 5, 2651. [11] W. Cai, C. Yan, Y.-X. Yao, L. Xu, R. Xu, L.-L. Jiang, J.-Q. Huang, Q.Zhang, Small Struct. 2020, 1, 2000010. [12] H. Zeng, J. He, D. Fang, Y. Liang, R. Zhao, Y. Cai, D. Lu, Energy Technol. 2019, 7, 1801078. [13] H. Lu, X. Chen, Y. Jia, H. Chen, Y. Wang, X. Ai, H. Yang, Y. Cao, Nano Energy 2019, 64, 103903. [14] F. S.Li, Y. S.Wu, J.Chou, M.Winter, N. L.Wu, Adv. Mater. 2015, 27, 130. [15] M.Yoshio, H.Wang, K.Fukud, Y.Har, Y.Adachi, J. Electrochem. Soc. 2000, 147, 1245. [16] R.Baddour-Hadjean, J.-P.Pereira-Ramos, Chem. Rev. 2010, 110, 1278. [17] S. J. An, J. Li, C. Daniel, D. Mohanty, S. Nagpure, D. L. Wood, Carbon 2016, 105, 52. www.advancedsciencenews.com www.small-journal.com 2107460 (9 of 9) © 2022 The Authors. Small published by Wiley-VCH GmbH. Small 2022, 18, 2107460 [18] S. Jiao, X. Ren, R. Cao, M. H. Engelhard, Y. Liu, D. Hu, D. Mei, J. Zheng, W. Zhao, Q. Li, N. Liu, B. D. Adams, C. Ma, J. Liu, J.-G.Zhang, W.Xu, Nat. Energy 2018, 3, 739. [19] J.Zheng, M. H.Engelhard, D.Mei, S.Jiao, B. J.Polzin, J.-G.Zhang, W.Xu, Nat. Energy 2017, 2, 17012. [20] J. R.Dahn, Phys. Rev. B: Condens. Matter 1991, 44, 9170. [21] J.Park, S. S.Park, Y. S.Won, Electrochim. Acta 2013, 107, 467. [22] A. C.Ferrari, J.Robertson, Phys. Rev. B 2000, 61, 14095. [23] X.-q. Cheng, H.-j. Li, Z.-x. Zhao, Y.-z. Wang, X.-m. Wang, New Carbon Mater. 2021, 36, 93. [24] K.Ni, X.Wang, Z.Tao, J.Yang, N.Shu, J.Ye, F.Pan, J.Xie, Z.Tan, X.Sun, J.Liu, Z.Qi, Y.Chen, X.Wu, Y.Zhu, Adv. Mater. 2019, 31, 1808091. [25] E.Pollak, B.Geng, K. J.Jeon, I. T.Lucas, T. J.Richardson, F.Wang, R.Kostecki, Nano Lett. 2010, 10, 3386. [26] A. Das, S. Pisana, B. Chakraborty, S. Piscanec, S. K. Saha, U. V.Waghmare, K. S.Novoselov, H. R.Krishnamurthy, A. K.Geim, A. C.Ferrari, A. K.Sood, Nat. Nanotechnol. 2008, 3, 210. [27] E. Markevich, G. Salitra, D. Aurbach, ACS Energy Lett. 2017, 2, 1337. [28] A. L. Michan, B. S. Parimalam, M.Leskes, R. N. Kerber, T. Yoon, C. P.Grey, B. L.Lucht, Chem. Mater. 2016, 28, 8149. [29] A. Funabiki, M. lnaba, Z.Ogumi, S.-i. Yuasa, J. Otsuli, A. Tasaka, J. Electrochem. Soc. 1998, 145, 172. [30] W.Weppner, R. A.Huggins, J. Electrochem. Soc. 1977, 124, 1569.