ZnO‐Embedded Expanded Graphite Composite Anodes with Controlled Charge Storage Mechanism Enabling Operation of Lithium‐Ion Batteries at Ultra‐Low Temperatures

As lithium (Li)‐ion batteries expand their applications, operating over a wide temperature range becomes increasingly important. However, the low‐temperature performance of conventional graphite anodes is severely hampered by the poor diffusion kinetics of Li ions (Li+). Here, zinc oxide (ZnO) nanoparticles are incorporated into the expanded graphite to improve Li+ diffusion kinetics, resulting in a significant improvement in low‐temperature performance. The ZnO–embedded expanded graphite anodes are investigated with different amounts of ZnO to establish the structure‐charge storage mechanism‐performance relationship with a focus on low‐temperature applications. Electrochemical analysis reveals that the ZnO–embedded expanded graphite anode with nano‐sized ZnO maintains a large portion of the diffusion‐controlled charge storage mechanism at an ultra‐low temperature of −50 °C. Due to this significantly enhanced Li+ diffusion rate, a full cell with the ZnO–embedded expanded graphite anode and a LiNi0.88Co0.09Al0.03O2 cathode delivers high capacities of 176 mAh g−1 at 20 °C and 86 mAh g−1 at −50 °C at a high rate of 1 C. The outstanding low‐temperature performance of the composite anode by improving the Li+ diffusion kinetics provides important scientific insights into the fundamental design principles of anodes for low‐temperature Li‐ion battery operation.


Introduction
Lithium (Li)-ion batteries (LIBs) have been established as indispensable energy storage devices due to their high energy and power densities and stable cycle life. [1][2][3] LIBs have traditionally been used as power sources for portable electronic devices and have recently expanded to applications in electric vehicles. [4] In addition, LIBs are expanding their application range to various fields in extreme environments that require stable operation in a wide range of temperature range (À60 to 100°C), such as Earth's polar regions, high-altitude, deep sea exploration, and space missions. [5,6] However, conventional LIBs have limited performance at sub-zero temperatures. [7,8] For example, a conventional graphite anode retains about 12% of its room temperature capacity at À20°C. [9] The poor low-temperature performance of the graphite anode is attributed to the relatively slow Li-ion (Li + ) transfer rate at the electrode and electrode/electrolyte interface during intercalation. [10,11] Such sluggish electrochemical kinetics often led to unstable solid electrolyte interphase (SEI) and metallic Li plating issues. [12] As a result, graphite anodes with poor low-temperature performance limit the stable operation of LIBs under extremely cold conditions. [13][14][15] To improve the low-temperature performance of anodes, structural engineering of carbon materials has been considered a promising strategy to either promote diffusion-controlled charge storage mechanisms or utilize surfacecontrolled charge storage mechanisms. [11] For example, multilayer crystalline graphene (GRAL) anodes were fabricated by enlarging the d-spacing of graphite to facilitate Li + diffusion kinetics under low-temperature conditions. [16] The GRAL anode maintained a capacity of 210 mAh g À1 at À20°C, demonstrating improved performance compared to a conventional graphite anode. In another way, the capacitive charge storage mechanism was exploited through stacking engineering of graphene layers to improve low-temperature performance. Sun et al. [17] achieved a high capacity of 155 mAh g À1 at 0°C for 300 cycles using electrochemically exfoliated graphene as the anode. Lee et al. [18] synthesized crumpled graphene as an anode for low-temperature operation, where the crumpled morphology prevents restacking of graphene layers. The crumpled graphene anode exhibited a high capacity of 154 mAh g À1 at À40°C due to the effective utilization of the surface-controlled charge storage mechanism. Structural control of carbon therefore modifies the charge storage mechanisms, offering opportunities for effectively exploiting the diffusive and capacitive charge storage mechanisms at low temperatures.
Zinc oxide (ZnO) has been extensively studied because of its high theoretical capacity (987 mAh g À1 ), high Li + diffusion coefficient, environmental abundance, safety, and low cost. [19][20][21] However, ZnO is vulnerable to polarization and cracking, resulting in significant capacity drop and poor cycle life. [22,23] The poor electrical conductivity and drastic volume changes of ZnO during cycling hinder its practical application in LIBs. [24,25] Researchers have mainly focused on overcoming these challenges by incorporating ZnO into the carbon supports. [26,27] Li et al. [26] synthesized ZnO nanocrystals confined in nitrogen (N)- As lithium (Li)-ion batteries expand their applications, operating over a wide temperature range becomes increasingly important. However, the lowtemperature performance of conventional graphite anodes is severely hampered by the poor diffusion kinetics of Li ions (Li + ). Here, zinc oxide (ZnO) nanoparticles are incorporated into the expanded graphite to improve Li + diffusion kinetics, resulting in a significant improvement in lowtemperature performance. The ZnO-embedded expanded graphite anodes are investigated with different amounts of ZnO to establish the structurecharge storage mechanism-performance relationship with a focus on lowtemperature applications. Electrochemical analysis reveals that the ZnOembedded expanded graphite anode with nano-sized ZnO maintains a large portion of the diffusion-controlled charge storage mechanism at an ultra-low temperature of À50°C. Due to this significantly enhanced Li + diffusion rate, a full cell with the ZnO-embedded expanded graphite anode and a LiNi 0.88 Co 0.09 Al 0.03 O 2 cathode delivers high capacities of 176 mAh g À1 at 20°C and 86 mAh g À1 at À50°C at a high rate of 1 C. The outstanding low-temperature performance of the composite anode by improving the Li + diffusion kinetics provides important scientific insights into the fundamental design principles of anodes for low-temperature Li-ion battery operation.
doped carbon, where the N-doped carbon improved the conductivity of ZnO while providing additional active sites for Li + storage. The composite delivered a high capacity of 687 mAh g À1 compared to pristine ZnO (180 mAh g À1 ) after 500 cycles. Fan et al. [27] developed mulberry-like ZnO particles immobilized in graphene aerogels that exhibited a specific capacity of 445 mAh g À1 after 500 cycles. In this composite, the graphene aerogel increased the electrical conductivity of ZnO through the C-O-Zn linkage and buffered the volume expansion for enhanced reversibility. [27] These previous studies on the ZnO-based composite anodes focused on achieving high capacity by using high contents of ZnO (76.3-86.3 wt.%), but at the expense of rate capability. Thereby, the structures of these composite anodes have not been optimized for low-temperature operation, which requires fast Li + diffusion kinetics, and furthermore, the structure-charge storage mechanism-performance relationship for ZnO-carbon composite anodes under low-temperature conditions has not been established.
In this study, we investigate the ZnO-embedded expanded graphite (ZnO-EG) composite anodes to establish the structure-charge storage mechanism-performance relationship for targeting stable operation of LIBs at sub-zero temperatures. The packing structure of EG and the size of ZnO were controlled by changing the amount of ZnO within the EG matrix. We reveal that a small amount of ZnO nanoparticles inserted into the EG matrix can significantly alter the ratio between the diffusive and capacitive charge storage mechanisms of the composite anodes under high-rate or low-temperature conditions. The optimized ZnO-EG composite anodes show significantly improved rate and cycling performance at both room and low temperatures. Specifically, the ZnO-EG anode maintains a high capacity of 580 mAh g À1 at 0.5 A g À1 after 400 cycles at room temperature. At sub-zero temperatures, the ZnO-EG anode retains the capacities of 360, 238, 122, and 55 mAh g À1 at 0, À20, À40, and À50°C, respectively. Determining the contributions of the diffusive and capacitive currents at different rates and temperatures reveals that the superior performance of the ZnO-EG anodes is due to the effective retention of diffusion-controlled charge storage mechanism under high rate and low-temperature conditions. Furthermore, the full cells comprising of the ZnO-EG composite anode and a high nickel-LiNi 0.88 Co 0.09 Al 0.03 O 2 (NCA-88) cathode deliver high capacities of 176 and 86 mAh g À1 under a high 1 C rate at 20 and À50°C, respectively. The results provide significant insights into the design of composite anodes for use under low-temperature conditions by establishing the structure-mechanism-performance relationships.

Design of the Zinc Oxide-Embedded Expanded Graphite (ZnO-EG) Composites
ZnO-EG composites were synthesized by controlling the amount of the Zn precursor relative to the EG matrix ( Figure S1, Supporting Information). Briefly, graphite oxide dispersed in deionized water was stirred with the ZnO precursor (ZnNO 3 ) at room temperature for 12 h. The mixture was then freeze dried and subsequently reduced at a temperature of 800°C for 6 h to obtain ZnO-EG composites. The different compositions of ZnO-EG composites were denoted by ZnO-EG-X, where X (X = 1 and 2) represents the mass ratio between the ZnO precursor and graphite oxide. ZnO-free EG was also synthesized to study the role of ZnO in the composite on the charge storage mechanisms and electrochemical performances. The morphology of EG and ZnO-EG composites was investigated using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Comparing the SEM images of graphite and EG, it was found that in the case of EG, the graphite was significantly expanded and wrinkled ( Figure 1a and Figure S2, Supporting Information). In the SEM images of ZnO-EG composites (ZnO-EG-1 and ZnO-EG-2), white crystals representing ZnO nanoparticles were observed around and within the EG layers (Figure 1b,c). These white crystals in ZnO-EG composites were also verified using energy-dispersive spectroscopy (EDS) mapping images (Figure 1d).
High-resolution TEM (HR-TEM) was employed to determine the d-spacing of expanded graphite and ZnO to verify the successful synthesis of ZnO-EG composites (Figure 1e,f). The expanded graphite showed a d-spacing of 0.352 nm for the d 002 layers (Figure 1f), while the ZnO nanoparticles, with a hexagonal close-packed crystal structure, exhibited a d 0002 spacing of 0.52 nm (Figure 1g). The HR-TEM images confirm the presence of expanded graphite wrapping around the ZnO nanoparticles, further supporting the formation of the ZnO-EG composites ( Figure S3). TEM analysis showed that the amount and size of ZnO nanoparticles in ZnO-EG-2 (∼ 150-250 nm) were larger than those in ZnO-EG-1 (∼ 40-70 nm) ( Figure S4, Supporting Information). The larger size of ZnO particles abundantly present in ZnO-EG-2 can be attributed to the increased amount of ZnO precursor that causes agglomeration of ZnO upon thermal reduction.

Physical and Chemical Structures of the ZnO-EG Composites
The formation of ZnO nanoparticles and the stacking structure of EG and ZnO-EG were investigated using X-ray diffraction (XRD) (Figure 2a).  . [28,29] The XRD spectrum of graphite showed a sharp (002) peak with a d-spacing of 0.332 nm, reflecting graphitic carbon ( Figure S5, Supporting Information). In contrast, the XRD spectra of EG, ZnO-EG-1, and ZnO-EG-2 exhibited a broadened (002) Figure 2. a) X-ray diffraction (XRD), b) Raman, and c) X-ray photoelectron spectroscopy (XPS) survey scan, d) high-resolution Zn 2p spectra, and e) highresolution O 1s spectra of the EG and ZnO-EG composites.
Energy Environ. Mater. 2023, 6, e12662 3 of 9 diffraction peak at 24.9°, 24.7°, and 24.4°with a d-spacing of 0.354, 0.357, and 0.362 nm, respectively. The d-spacing obtained using the XRD profiles are well correlated with the values obtained from the HR-TEM images. The larger d-spacing of EG, ZnO-EG-1, and ZnO-EG-2 than graphite indicates successful synthesis of the EG structures. In addition, the increased d-spacing with more ZnO precursor (ZnO-EG-2) can be attributed to the formation of more ZnO nanoparticles within the EG layers, further expanding the EG layers in the ZnO-EG composite. The defect structure of carbon on the samples was investigated using Raman spectroscopy (Figure 2b). All samples showed two distinct peaks corresponding to defective (D) and graphitic (G) peaks at ∼ 1360 and ∼ 1580 cm À1 , respectively. The D band originates from the disorder in the carbon lattice, while the G band is assigned to the symmetric stretching of sp 2 C-C bonds. [30] The integrated ratios of G to D peak (A G /A D ) for the samples were calculated to compare the relative amount of defects. The A G /A D values for EG, ZnO-EG-1, and ZnO-EG-2 were 0.6, 0.56, and 0.54, respectively, indicating that the incorporation of ZnO nanoparticles into EG increased the carbon disorder.
Thermogravimetric analysis (TGA) was performed to quantify the ZnO content in the ZnO-EG composites ( Figure S6, Supporting Information). The weight loss observed corresponds to be 95.0% and 89.9% for ZnO-EG-1 and ZnO-EG-2, respectively, which is attributed to the carbon content within the composites. Accordingly, the weight percentage of ZnO is determined as 5.0% for ZnO-EG-1 and 10.1% for ZnO-EG-2. Notably, these values are significantly lower than those reported in previous studies that focused on using carbon supports to enhance ZnO performance (Table S1, Supporting Information). [24,[31][32][33][34][35][36][37][38] The surface chemistry of the EG and ZnO-EG composites was studied using X-ray photoelectron spectroscopy (XPS). The XPS survey scans showed a gradual increase in the atomic oxygen to carbon (O/C) ratio with increasing amounts of ZnO nanoparticles, from 0.05 for EG to 0.07 for ZnO-EG-1 to 0.1 for ZnO-EG-2 ( Figure 2c). The atomic zinc to C (Zn/C) ratio increased from 0.01 for ZnO-EG-1 to 0.03 for ZnO-EG-2. High-resolution XPS Zn 2p spectra showed the Zn 2p 3/2 and Zn 2p 1/2 peaks at 1021.8 and 1044.9 eV, respectively (Figure 2d). The energy difference of 23.1 eV between the two Zn 2p peaks indicates the successful formation of ZnO. [26,39] The binding energies of the ZnO-EG composites are higher than the typical binding energies of ZnO particles (1021 and 1044 eV), which can be attributed to the strong interaction between ZnO and C from expanded graphite. [40,41] High-resolution C 1s spectra were fitted by the four characteristic peaks at 284.5 AE 0.1 for C (sp 2 ), 285.  (Figure 2e). [43,44] The C-O-Zn peak increased with higher ZnO content, indicating that more ZnO nanoparticles were successfully bound to the EG matrix ( Figure S7c, Supporting Information).

Electrochemical Characterizations of the ZnO-EG Composites
To investigate the redox reactions of ZnO, cyclic voltammetry (CV) experiments were conducted for the ZnO-EG composites in the potential range of 0.01-3 V versus Li/Li + at a scan rate of 0.1 mV s À1 ( Figure S8, Supporting Information). ZnO reversibly interacts with Li through the conversion reaction (ZnO + 2Li + + 2e À ↔ Zn 0 + Li 2 O) and the (de)alloying reaction (Zn 0 + xLi + + xe À ↔ Li x Zn). [45] Two cathodic peaks at 0.3 and 0.8 V were commonly assigned to the Zn-Li alloying and conversion reactions, respectively. [27] The anodic CV scan showed the multi-step dealloying reaction between 0.4 and 0.7 V and the conversion reaction at 1.35 V. [29] In addition, the redox peaks at low potential, below 0.3 V, can be ascribed to the Li + intercalation reactions to EG. The CV scans of the ZnO-EG composites showed a peak voltage shift and a decrease in anodic peak intensity after the first cycle. This can be attributed to the polarization and pulverization of ZnO nanoparticles and irreversible side reactions, including SEI formation. [27,46] However, the relatively small peak voltage shift and peak intensity of ZnO-EG-1 compared to ZnO-EG-2 indicates that ZnO-EG-1 has higher interfacial stability against the electrolyte than that of ZnO-EG-2. [47,48] Galvanostatic rate-capability tests of the samples were performed at room temperature with current densities ranging from 0.05 to 10 A g À1 (Figure 3a,b, and Figure S9, Supporting Information). The initial Coulombic efficiencies (CEs) were calculated to be 57%, 59%, and 51% for EG, ZnO-EG-1, and ZnO-EG-2, respectively ( Figure S10, Supporting Information). The higher CE of ZnO-EG-1 compared to ZnO-EG-2 is consistent with the CV results. The galvanostatic charge/ discharge (GCD) curves of EG showed a sloped profile typically observed in EG or graphene electrodes (Figure 3a). [18,49] The GCD profiles of the ZnO-EG composites exhibited multiple plateaus on the sloped profile, implying the presence of the conversion and (de)alloying reactions. At a slow rate of 0.05 A g À1 , EG, ZnO-EG-1, and ZnO-EG-2 delivered capacities of 507, 526, and 534 mAh g À1 , respectively (Figure 3b). The high capacity of ZnO-EG-2 is due to the higher amount of ZnO nanoparticles within EG. At a high rate of 10 A g À1 , ZnO-EG-1 showed superior capacity retention (340 mAh g À1 , 65% of its capacity at 0.05 A g À1 ) compared to those of EG (285 mAh g À1 ; 56%) and ZnO-EG-2 (245 mAh g À1 ; 46%) ( Figure S11, Supporting Information). The activation overpotential calculated from the GCD profiles at 10 A g À1 were 281, 223, and 299 mV for EG, ZnO-EG-1, and ZnO-EG-2, respectively. After a rate resumed to a moderate rate of 0.5 A g À1 , ZnO-EG-1 exhibited a high capacity of 437 mAh g À1 . Such excellent rate-performance of ZnO-EG-1 can be attributed to the enhanced Li + diffusion kinetics through 3D EG structure, whereas the excessive amount of ZnO in ZnO-EG-2 impedes charge storage kinetics. The cycling stability tests of the samples were conducted up to 400 cycles at room temperature (Figure 3c and Figure S12, Supporting Information). The specific capacity of EG and ZnO-EG-2 gradually decreased after the first 100 cycles. On the other hand, ZnO-EG-1 exhibited remarkable cycling stability with an average CE of 99.7% (1st-400th) and maintained a high capacity of 576 mAh g À1 after 400 cycles with negligible capacity decay (Figure 3d). Interestingly, a gradual increase in the capacity was observed upon cycling of ZnO-EG-1, which indicates the activation process upon cycling. [18,50] Temperature-dependent GCD tests of the EG and ZnO-EG composites were further performed at a current density of 0.05 A g À1 from 20 to À40°C (Figure 4a and Figure S13, Supporting Information). At a low temperature of À40°C, ZnO-EG-1 delivered the highest capacity of 119 mAh g À1 compared to those of EG (61 mAh g À1 ) and ZnO-EG-2 (100 mAh g À1 ) (Figure 4b). These capacity values correspond to 19.5% of the room-temperature capacities for ZnO-EG-1, 14.7% for EG, and 17.8% for ZnO-EG-2. In addition, the rate performance of ZnO-EG-1 at À50°C outperformed EG and ZnO-EG-2 (Figure 4c,d and Figure S14, Supporting Information). At a low rate of 0.05 A g À1 , EG, ZnO-EG-1, and ZnO-EG-2 delivered capacities of 44, 62, and 69 mAh g À1 , respectively. At a high rate of 0.5 A g À1 , the activation overpotentials for EG, ZnO-EG-1, and ZnO-EG-2 were 1.01, 0.88, and 0.93 V versus Li/Li + , respectively (Figure S15, Supporting Information). When the current density returned to 0.05 A g À1 after 40 cycles at various current densities, ZnO-EG-1 Figure 4. a) Temperature-dependent GCD of the EG and ZnO-EG composites at 0.05 A g À1 . b) Corresponding GCD profiles at À40°C. c) Rate and cycling performances of the EG and ZnO-EG composites at À50°C. d) Corresponding GCD profiles at 120 th cycle.
Energy Environ. Mater. 2023, 6, e12662 5 of 9 retained 95% of its initial capacity, which is much higher than those of EG (84%) and ZnO-EG-2 (61%) (Figure 4c and Figure S16, Supporting Information). When the cells were further cycled up to 120 cycles at 0.05 A g À1 , ZnO-EG-1 maintained 96% of the capacity with a very high average CE of 99.9 (41st-120th), superior to EG (77% with a CE of 99.2%) and ZnO-EG-2 (81% with a CE of 99.4%) (Figure 4d). It should be noted that the ZnO-EG-1 anode showed the significantly improved low-temperature performance compared to previously reported carbon-based anodes (Table  S2, Supporting Information). [10,[16][17][18]51] The results show that incorporating a small amount of ZnO into the EG matrix can efficiently improve the lowtemperature performance of the composite electrode.

Charge Storage Mechanisms of the ZnO-EG Composites
To elucidate the charge storage mechanisms of the EG and ZnO-EG composites, rate-dependent CV scans were measured at varied scan rates from 0.1 to 1.0 mV s À1 in the potential range of 0.01-3 V versus Li/ Li + at 20 and À50°C (Figures S17 and S18, Supporting Information). The dependence of the current response on the scan rate can be used to quantify the surface-and diffusion-controlled current contributions using Equation (1). [52,53] where i is the normalized current, k 1 and k 2 are constraints, and ν is the scan rate. The terms k 1 ν and k 2 ν 1/2 refer to the current contributions to the surface-and diffusion-controlled charge storage mechanisms, respectively. At 20°C, the diffusive contributions from the CV scans at 0.1 mV s À1 were calculated to be 76% for EG, 72% for ZnO-EG-1, and 62% for ZnO-EG-2 (Figure 5a-c).
The CV scans of the samples revealed that the diffusive contribution is dominant in the low potential region below 0.1 V due to the intercalation, (de)alloying, and conversion reactions. As the scan rate increased to 1 mV s À1 , the diffusive contributions decreased to 43% for EG, 46% for ZnO-EG-1 and 37% for ZnO-EG-2 ( Figures S19-S21, Supporting Information). This trend indicates that the capacitive charge storage capability becomes dominant as the scan rate increases. As the temperature decreased to À50°C, diffusive contributions at 0.1 mV s À1 were calculated to be 38% for EG, 63% for ZnO-EG-1, and 46% for ZnO-EG-2 (Figure 5d-f). At this low-temperature condition, EG and ZnO-EG-2 lost a large portion of their diffusive contribution due to a significant decrease in current in the low potential region, whereas ZnO-EG-1 still maintained a high diffusive contribution (63%). Under low-temperature and high-rate conditions (À50°C and 1 mV s À1 ), EG and ZnO-EG-2 showed further reduced diffusive contributions of 17 and 22%, respectively. In contrast, ZnO-EG-1 exhibited a significantly higher diffusive contribution of 40%, indicating a balanced diffusive and capacitive charge storage mechanism (Figure 5g and Figures S22-S24, Supporting Information).
The relationship between the current and scan rate can be empirically expressed using the following Equation (2): where i represents the current density, v represents the scan rate, and a and b are numerical constants. In this equation, the i diffusive represents the diffusion-controlled process following the Fick's law and has a well-defined b-value of 0.5, whereas the i capacitive represents the surface-controlled current required to charge the double layer with a b-value of 1.0. There exists a transitional region between b-values of 0.5 and 1.0, where b-values close to 0.5 indicate a diffusion-controlled charge storage mechanism, and b-values close to 1.0 indicate a surface-controlled charge storage mechanism. Therefore, the b-values provide insights for interpreting diffusion-and surface-controlled charge storage mechanisms.
There exists a transitional area between b-value of 0.5 and 1.0; in other words, b-values close to 0.5 indicates diffusive-controlled charge storage mechanism and b-values close to 1.0 indicates surface-controlled charge storage mechanism. The b-values provide guidance for the interpretation of diffusion-and surface-controlled charge storage mechanisms.
The b-values for EG, ZnO-EG-1, and ZnO-EG-2 were calculated to be 0.73, 0.70, and 0.78, respectively ( Figure S25a). ZnO-EG-1, with the lowest b-value of 0.7, exhibits enhanced charge storage through diffusion compared to EG or ZnO-EG-2. These results indicate that the incorporation of an appropriate amount of ZnO nanoparticles into the EG structure effectively enhances the diffusive charge storage mechanism. To quantitatively assess the diffusive and capacitive contributions to the charge storage mechanism at low temperatures, the b-values of EG, ZnO-EG-1, and ZnO-EG-2 were estimated to be 0.88, 0.75, and 0.82 at À50°C, respectively ( Figure S25b). Even at a low temperature of À50°C, the ZnO-EG composites demonstrate a smaller change in bvalue (0.7-0.75) compared to EG, which shows the largest deviation in b-value between 20 and À50°C, increasing from 0.73 to 0.88. These results suggest that the addition of ZnO nanoparticles facilitates Li + diffusion, enabling effective utilization of the diffusion-controlled charge storage mechanism under low-temperature conditions. Therefore, the excellent rate and low-temperature performances of ZnO-EG-1 can be attributed to the effective retention of the diffusion-controlled charge storage mechanism by introducing small-sized ZnO nanoparticles into the EG matrix.

Full Cell Demonstration Using the ZnO-EG Composites
To demonstrate the stable operation of the ZnO-EG composite anode in LIBs, the full cells were assembled with ZnO-EG-1 and a high-Ni NCA-88 cathode with a negative/positive (N/P) ratio of 1.0. The full cells were cycled at a high rate of 1 C at different temperatures. At 20°C, the full cell exhibited an average capacity of 176 mAh g À1 with an average CE of 99.7% for 100 cycles (Figure  6a,b). When the temperature decreased to À50°C, the full cell maintained a high capacity of 86 mAh g À1 with an average CE of 95% for 60 cycles (Figure 6c,d). It should be emphasized that this is the best cycling performance for carbon-based anode research under severe conditions, including a low N/P ratio of 1.0, a high C rate of 1 C, and an ultra-low temperature of À50°C (Table S3). [17,44] The results thus suggest that engineered ZnO-EG composites can be utilized as an anode combined with state-of-theart cathodes for stable low-temperature operation of LIBs.

Conclusion
The ZnO-embedded expanded graphite (ZnO-EG) composite anodes were assembled by adjusting the ZnO content in the ZnO-EG composites. The composite anode containing small-sized and low content of ZnO (ZnO-EG-1) enabled faster Li + diffusion kinetics, demonstrating outstanding rate and cycling performances at both room and low temperatures. At room temperature, ZnO-EG-1 exhibited a high capacity of 340 mAh g À1 at a high current density of 10 A g À1 and excellent cycling performance up to 400 cycles. At an ultra-low temperature of À50°C, ZnO-EG-1 still maintained a capacity of 55 mAh g À1 at 0.05 A g À1 due to the effective utilization of the diffusion-controlled charge storage mechanism. The full cell consisting of the ZnO-EG-1 anode and the NCA-88 cathode with an N/P ratio of 1.0 delivered a high capacity of 87 mAh g À1 at 1 C for 60 cycles, demonstrating superior low-temperature performance. The established structure-mechanism-performance relationship of composite anodes by tuning the charge storage mechanisms provides useful insights into the stable operation of LIBs under extreme low-temperature conditions.

Experimental Section
Detailed information related to the synthesis of active electrodes, physicochemical characterization, and electrochemical evaluation of bifunctional electrodes towards UOR and supercapacitor application is provided in Supporting Information.