Graphene nanotube array assists all‐wood supercapacitors to access high energy density and stability

Porous carbons with advanced nanostructures and volumetric performance are particularly attractive and essential for miniature supercapacitors to access high energy densities and capacitances, both for portable electronics and massive electrical equipments. However, the electrochemical performances and the pore structure are closely bound up, both restricted by pore volume and pore density. Herein, the wood slice (~0.7 mm) with the periodic porous structure is chosen as the basic framework with rich macropores and the graphene nanotube array (GNTA) with mesopores is used as an intermediate structure in situ synthesized to form the substructure in macropores; therefore, the biomass and nanotube array together construct a porous carbon with hierarchical pores and large surface area. On this basis, Cu‐Co oxides are coated on the surface of the pores, to increase the capacitance of electrodes for supercapacitor applications. Because of the GNTA, the specific surface area increases from 38.2 to 1086.0 m2 g−1, which is quite helpful for the deposition of Cu‐Co oxide nanosheets and effectively alleviates their typical self‐stacking phenomenon. Meanwhile, the GNTA creates multiscale pores that served as channels for the rapid electron transfer and ion shuttling; as a result, the resistance obviously induces and capacitance increased by 131% (from 323.4 to 747.5 mF cm−2). For the assembled all‐wood asymmetric supercapacitor, the specific capacitance is 151.2 F g−1 (1 A g−1), the energy density is 53.8 Wh kg−1 with a power density of 900 W kg−1, and the specific capacitance remains extremely stable during the cycling. Our work provides a practical structure–design strategy for high‐performance supercapacitors.


| INTRODUCTION
To realize carbon neutrality and green energy transition, efforts from all over the world have been made. As one of the most efficient and typical energy storage and conversion technologies, the supercapacitor (SC) plays a crucial role in the fields of electric vehicles and auxiliary power due to its high reliability, fast charge/discharge rate, high power density, and environmental friendliness. [1][2][3] With the growing demand for miniaturized electronic devices and superior human-machine interaction experiences, the demand for the high energy density of SCs is rising rapidly. Therefore, it is essential to simultaneously regulate the pore architecture of porous carbons and introduce pseudocapacitance.
To meet this requirement, the high porosity and large surface area of porous carbon are important, 4,5 but also the inner space of porous carbon is nonnegligible and needed to be further explored for higher energy density. Introducing hollow nanostructures into porous carbon frameworks could be considered a promising approach. In general, the porous carbon framework is prepared by template methods, because its morphology and pore structure could be adjusted by the templates. 6 However, the ideal carbon structure is difficult to hold and the physiochemical damages to that are hardly avoided during the removal process of the template. Compared with artificial porous carbons, natural biomass is a cost-efficient choice to prepare porous carbons because of the advantages of eco-friendly resources and simple processing. 7 Above all, some biomasses have naturally unique and expectant pore structures; 8,9 thus, they are suitable to serve as porous carbon frameworks and accommodate the introduction of varied substructures. Besides, biomassderived porous carbons generally contain abundant heteroatoms such as O, N, S, and P, and so on, 7,8 which could produce a synergistic effect to modify the electron localization of the carbon skeleton, raise extra electrochemically active sites, and broaden the potential window of the carbon materials. 5,10,11 In terms of enhancing capacitance performance, apart from the heteroatom doping in porous carbons, transition metal oxides, especially bimetal oxides, 12 are one of the most efficient choices to introduce the pseudocapacitance, due to their rapid surface redox chemical process. 12,13 In recent years, researchers have made great efforts on transition metal (e.g., Co, 14,15 Ni, 16,17 Fe, 18,19 and so on) oxides and achieved superior results. However, the aggregation problem of nanomaterials makes their performance quite inferior to what has been expected; thus, the distribution of metal oxide nanostructures is essential for their practical performances and needs to be fine-tuned. 12 As to the structural problem, one representative example is the Cu-Co oxides. Among transition metal nanostructures, cobalt oxides appeal to great interest because of their high theoretical specific capacitance, 20,21 but the expensive exploitation of raw materials calls for substitutes with high energy density and low cost. Copper has outstanding conductivity and a narrow band gap, 22 and more importantly, possesses large natural reserves. Thus, replacing partial Co atoms with Cu atoms to form Cu-Co oxides is a promising substitute for energy storage devices. 23 However, the mismatching of electron clouds between Co and Cu atoms leads to the poor structural stability of Cu-Co nanomaterials. Because of the problem mentioned above, Cu-Co oxides are seldom reported as electrode materials in the field of electrochemical energy storage. The synthesis of Cu-Co oxides with stable nanostructure is a challenge, but on the contrary, the structural defects caused by mismatching are a chance to obtain a high electron activity. In addition, the structural problem of Cu-Co oxides could also be alleviated by carbon nanomaterials. Besides, the Fe-based oxides are studied a lot in the reported research and they are good candidates for negative electrodes. On the basis of the above, it is practical and valuable to choose Fe-based oxides to further verify the structural reliability and portability.
On the foundation of the above concept, an SC electrode with high energy density could be prepared by constructing and fine-tuning the graded pores in carbon materials, uniformly depositing pseudocapacitive materials, and adjusting the interrelationship between those. Herein, a practical electrode material was synthesized, in which the wood slice with thorough-holes was chosen as the raw material to construct the biomassderived carbon framework, the graphene nanotube array (GNTA) was in-situ synthesized innerin the thorough holes inside the carbon framework, and nanosheets of transition metal oxides was deposited on the surface of the carbon structure to act as the pseudocapacitive part. Because of the synergetic contribution of the porous carbon, the hollow substructure, and the pseudocapacitive substance, the specific surface area of the wood-based material was raised to 1086.0 m 2 g −1 from 38.2 m 2 g −1 .
The specific capacitance of the all-wood asymmetric SC (ASC) is 151.2 F g −1 (1 A g −1 ) and the energy density is 53.8 Wh kg −1 with a power density of 900 W kg −1 . The specific capacitance remains extremely stable during the cycling. The design and architecture of the electrode material and all-wood SC are promising for eco-friendly, low-cost, high-security, high-energy/power-density energy storage applications.

| Pretreatment of the wood carbon slices
First, the Balsa wood was cut perpendicular to the growth direction into slices with a thickness of 0.7 mm and thermally treated in an Ar atmosphere. The pyrolyzing temperature was raised to a heating rate of 3°C/min and held at 500°C for an hour, then held at 900°C for 2 h at a heating rate of 5°C/min. After cooling down, the pyrolyzed biomass was obtained and the sample was marked as W. For comparison, the wood slice was heated in 450°C, 600°C, and 750°C, separately, which was named as W-450, W-600, and W-750, respectively.

| Synthesis of GNTAs
The synthesizing process contains two steps, that is, the loading of the catalyzer and the in situ synthesis of GNTAs by a CVD method. The electro-deposition method was adopted to synthesize NiCo-LDH nanosheets on the W, catalyzing the synthesis of nanotubes. After drying at 60°C for 6 h, the as-prepared slice and 2 g dicyandiamide (C 2 H 4 N 4 ) were separately placed in the tube furnace and C 2 H 4 N 4 is put between the gas inlet and the slice. Being kept in an Ar atmosphere at 900°C for 25 min, the GNTAs were synthesized. The Sample W deposited with in situ GNTAs was marked as W/GNTAs.

| Synthesis of pseudocapacitive substance (Cu-Co oxides)
The Cu-Co oxides were synthesized via CV. One hundred milliliters of water containing 1 mmol Cu(NO 3 ) 2 and 2 mmol Co(NO 3 ) 2 is served as the electrolyte. The asprepared W/GNTAs is served as the working electrode, the Pt electrode as the counter electrode, and the Ag/AgCl (saturated with KCl) as the reference electrode. The potential range of CV is −1.1~−0.5 V. After the 15-time cycles, the slice was washed with water and ethanol and dried at 60°C. Then the slice was thermally treated at 300°C for 2 h with a heating rate of 3°C/min in air. The obtained sample was marked as W/GNTAs@Cu-Co-O. For comparison, Cu-Co oxides are also deposited on the W and the final product was marked as W@Cu-Co-O. The mass of the Cu-Co oxide is 0.74 mg cm −2 .

| Synthesis of Wood/ GNTAs@FeOOH (W/GNTAs@Fe)
The Fe-based compound was obtained via a chronoamperometry process. 100 mL deionized water with 1 mmol FeSO 4 was heated to 70°C to serve as the electrolyte. The W/GNTAs served as the working electrode, the Pt electrode as the counter electrode, and the Ag/AgCl (saturated with KCl) as the reference electrode. The depositing potential is +1.5 V. After 12-min electrodeposition, the slice was dried at 60°C in air, then thermally treated at 300°C for 2 h with a heating rate of 3°C/min in air. The obtained sample was marked as W/ GNTAs@Fe. The active mass of the W/GNTAs@Fe is 1.08 mg cm −2 .

| Electrochemistry measurements
Electrochemical performances of samples were analyzed by CV, GCD, and EIS methods on an electrochemical workstation (CS310, Corrtest) in a three-electrode cell with 2 mol L −1 KOH aqueous solution. The as-prepared samples directly served as the working electrode (1 × 1 cm 2 ) without any conductive additives and polymer binders. The Ag/AgCl (saturated with KCl) electrode and Pt electrode served as the reference electrode and the counter electrode, respectively. The specific capacitance (C sp , F g −1 ) and specific capacity (C s , mAh g −1 ) in a three-electrode system can be calculated by the equations: 24 is the current density based on mass, t (s) the discharging time, and v (V) the applied potential with the initial v i and the final v f .
To assemble the all-wood ASC, the W/GNTAs@Cu-Co-O and W/GNTAs@Fe served as the cathode and anode, respectively. The mass loadings of the anode and cathode are 3.21 and 0.74 g, respectively. The mass loadings of both electrodes were determined according to the equation: 24 The energy density (E, Wh kg −1 ) and power density (P, W kg −1 ) of the ASC are calculated by the following equations: 24 YUAN ET AL.
where v(V) represents the potential window, t(s) refers to the discharging time, I(A) is the current, and m(g) is the mass of the total mass of active materials on both cathode and anode.

| Structural characterization
The principle of the experiment scheme is illustrated in Figure 1. First of all, the Balsa wood slices were carbonized and activated. During the carbonization, the lignin in wood slices was thermally decomposed and the cellulose-derived carbon skeleton was left, exposing enough regular macro thorough-holes for the construction of advanced micro-/nanostructures. Then, the GNTAs were introduced by an in-situ chemical vapor depoitin (CVD) process into the porous carbon as a hollow substructure to enrich pores and broaden specific surface area. Finally, the pseudocapacitive layer was uniformly and thinly electrodeposited on the carbon frameworks to further improve the energy density. The Cu-Co bimetal oxides and FeOOH nanostructures were separately electro-deposited on the GNTAs, serving as the cathode and anode to assemble all-wood ASCs. The microstructure of the cellulose-derived carbon is shown in Figure 2A,B. The open channels in the skeleton are regular and steady and the length of the long edge is about 50 μm. Seen as the scanning electron microscopic (SEM) image of the wood slice without carbonization shown in Supporting Information: Figure S1, the holes in the wood slices are blocked. During the thermal treatment, the lignin was decomposed and thorough holes were exposed (Supporting Information: Figure S1c). With the increase of the thermal-treatment temperature, the graphitization degree and electroconductivity raise (Supporting Information: Figure S2). Finally, the cellulose-derived carbon framework shown in Figure 2A,B was formed. In addition, the woodderived carbon has reserved rich surface functional groups (Supporting Information: Figure S2c), which provides active sites for the following synthesis process and electrochemical applications. For the purpose of enriching the pore structure, the GNTAs were in situ synthesized in wood slices, constructing the sample W deposited with in situ GNTAs (W/GNTAs) material. As the SEM images are shown in Figure 2C-E, the nanotubes regularly and evenly array on all walls of the thorough holes, pointing to the circle centers. The nanotubes are longer than 10 μm, whereas the diameter is about 120 nm ( Figure 2F), several times that of the current commercial carbon nanotubes (10~30 nm). As shown in Figure 2F, vivid wrinkles are spread on the surface of nanotubes. The high-resolution transmission electron microscopic (HRTEM) image in Figure 2G,H illustrates the thickness of the wrinkle is about 5 nm and the spacing between atom layers is about 0.357 nm, which together indicate that these nanotubes are made of graphene, and the nanotube array is GNTA. Besides, at the beginning of the growth, the nanotubes are growing straight, but with changeable diameters as the time prolonged, the natubes are longer and the diameters are more uniform (Supporting Information: Figure S3). With the GNTAs, the W/GNTAs carbon substrate was prepared and the pseudocapacitive layer would be following introduced. The microstructure of the W/ GNTAs@Cu-Co-O is illustrated in Figure 3. As shown in Figure 3A,B, the metal oxides wrap up every GNTA and extend into the center with the guidance of the GNTAs. The TEM pattern in Figure 3C displays that Cu-Co oxide nanosheets is soft and flimsy, and uniformly covered over the GNTA. The enlarge pattern in Figure 3D indicates that the nanosheet is just about 5 nm. The HRTEM image F I G U R E 1 Illustration of the experimental scheme.  in Figure 3E indicates that the lattice spacing is 0.231 nm, corresponding to the inter-planner spacing of (222) plane for cubic CuCo 2 O 4 (PDF: Copper Cobalt Oxide, 00-001-1155). 24 The ultrathin nanostructure is favorable for ions shuttling forth and back rapidly. The element mappings in Figure 3G illustrate that the Cu, Co, and O atoms distribute evenly in the nanostructure and the atom ratio of Cu and Co atoms, calculated from Figure 3H, is about 1:1.
Because of the soft and ultrathin features of the pseudocapacitive layer and the array structure of GNTAs, the surface area of the composites is expanded and the surface area of the W/GNTAs@Cu-Co-O is 1086.0 m 2 g −1 , which is 38 times as that of the W/Cu-Co-O (Supporting Information: Figures S4 and S5). As seen in Supporting Information: Figure S4, the Cu-Co-O nanosheets spread uniformly on the surface of the wood-derived carbon, but without the assistance from GNTAs with the electron transform, the nanosheets are a little different from those in W/GNTAs@Cu-Co-O. As seen in Supporting Information: Figure S5, the proportion of macropores decreased and that of micropores and mesopores raised, which indicates that the introduction of the array substructure leads to full exploitation of the inner space of the carbon skeleton. Thus, the accessible surface area and active sites significantly increased; meanwhile, the ion exchange and electron transfer processes would be faster.
The phase structure of the W/GNTAs@Cu-Co-O is first characterized by using a Raman spectrometer, and its Raman spectrum is shown in Figure 4. Peaks at about 1368 and 1582 cm −1 , respectively, refer to the D and G bands generated from the wood carbon and GNTAs. Peaks at 187, 461, 505, and 663 cm −1 respond to the F g , E g , and A 1g vibrational mode of CuCo 2 O 4 , 25 confirming the formation of CuCo 2 O 4 . Besides, the peak at 275 cm −1 is corresponding to the A g vibrational mode of CuO. 25 To further analyze the surface chemical states of the sample, X-ray photoelectron spectroscopy (XPS) was carried out ( Figure 5). In the survey spectrum in Figure 5A, signals of C 1s, N 1s, O 1s, Co 2p, and Cu 2p were detected. N atoms mainly exist in the wood carbon. In the high-resolution region of C 1s ( Figure 5B), chemical bonds and functional groups such as C=O, C-OR were detected, indicating plenty of oxygen-containing functional groups existing in the carbon framework, which are effective to improve the surface activity and thus boost ion adsorption. As shown in Figure 5C,D, Cu 2p and Co 2p spectra can be fitted into two spin-obit doublets and two shake-up satellites (defined as 'Sat.') by the Gaussian Fitting method. 24 In the Cu 2p spectrum ( Figure 5C), Sat. peaks at 961.8 and 941.7 eV demonstrate the existence of Cu 2+ . 26,27 In the Co 2p spectrum ( Figure 5D), peaks at 779.8 and 796.2 eV correspond to the Co 2p 3/2 and Co 2p 1/2 orbitals of Co 3+ , whereas those at 780.9 and 794.5 eV accord with the Co 2+ . The XPS results indicate two chemical states of Co ions existed in the sample and, generally, metal oxides with variable valances would present superb electrochemical activity. 26,27 As the effect of the GNTA remarkably enlarges the specific surface area, this structure is also adopted to fabricate anode electrode, loading Fe-based nanostructures as the pseudocapacitive layer. The morphology of the nanostructure is shown in Figure 6. As seen in the SEM images ( Figure 6A-C), same as the W/GNTAs@Cu-Co-O, the Fe-based nanoneedles also uniformly adhere to the surface of GNTAs and the structure seemed like spiked clubs formed. Because of the nanotubes in the center of the nanostructure, the spiked clubs are hollow and advantageous to shorten paths of ion transfers and improve conductivity. The TEM images in Figure 6D,E also evidence the nanoneedle structure and spiked type of the nanomaterial. The element mappings in Figure 6F illustrate that the sample mainly contains elements of Fe, O, and C.
The XPS technique was used to confirm the surface chemical states of the sample. The survey spectrum in Figure 7 indicates the presence of C, O, and Fe elements, corresponding to the energy dispersive X-ray spectroscopy results. The high resolution of the Fe 2p spectrum is shown in Figure 7B, and the Fe 2p spectrum was fitted into two spin-obit doublets by the Gaussian Fitting method, suggesting two kinds of chemical states of Fe in the sample. Besides, the difference in binding energy between the two peaks is 13.6 eV, declaring the existence of FeOOH. 28 The X-ray diffraction (XRD) pattern and Raman spectra in Supporting Information: Figures S6 and S7 also evidenced the above results. The XRD pattern shows that apart from the diffraction peaks of carbon  Figure S7) also support this result.

| Electrochemical behaviors
To illustrate the effectiveness of the GNTA, a comparison of electrochemical behaviors between W@Cu-Co-O  and W/GNTAs@Cu-Co-O was first carried out in a three-electrode system. Figure 8A shows the cyclic voltammetry (CV) curves of different samples at a scan rate of 20 mV s −1 in a potential window of −0.2~0.7 V. The overall shapes of these curves are the same, suggesting similar reversible Faradaic redox reactions. The CV curve of the W/GNTAs@Cu-Co-O displays a larger area and indicates a better electrochemical performance. Same to the CV results, as shown in Figure 8B, the W/GNTAs@Cu-Co-O illustrates a longer discharging time in the galvanostatic charge-discharge (GCD) test, demonstrating a much-improved specific capacitance. The specific capacitances of W@Cu-Co-O and W/GNTAs@Cu-Co-O are 323.4 and 747.5 mF cm −2 at 1 mA cm −2 , respectively.
To kinetically analyze the samples, the electrochemical impedance spectroscopy (EIS) technique was utilized and the Nyquist plots were shown in Figure 8C. In the high-frequency region, the first intersection of the semicircle and the horizontal axis refers to the equivalent seclusion resistance (R s ), corresponding to the intrinsic resistance of the electrode and the interfacial resistance between the electrolyte and electrode, while the radius of the semicircle denotes the charge-transfer resistance (R ct ). 29 In the low-frequency region, the slope of the line corresponds to the Warburg impedance (R w ) 29 and is bound up with the diffusion resistance of ions shuttling between the electrolyte and the electrode. The larger the slope, the smaller the R w . 29 As seen in the EIS patterns, the difference of the slopes in the low-frequency region is less, which is due to the same chemical composition, microstructure, and synthesizing method of the two electrodes. While the differences in the high frequency are distinct. The W/GNTAs@Cu-Co-O reveals a lower R s (2.441 Ω) and R ct , suggesting the improved electronic conductivity, thus the array structure of GNTAs indeed acts as the bridge for electron transfer between the substrate and metal oxides and effectively accelerates the electron transport.
Meanwhile, the comparison among W, W/GNTAs, and W/GNTAs@Cu-Co-O was also carried out. As shown in Supporting Information: Figure S8, the comparison of W/GNTAs with W in CV and GCD curves illustrate that the GNTA is useful in improving the electrochemical F I G U R E 7 X-ray photoelectron spectroscopy (XPS) spectra of sample W deposited with in situ graphene nanotube arrays (W/GNTAs)@Fe. performance. The EIS pattern in Supporting Information: Figure S8c indicates that the introduction of GNTAs decreases the internal resistance of the carbon substrate. More importantly, comparing the W/GNTAs with the W/ GNTAs@Cu-Co-O in CV and GCD curves (Supporting Information: Figure S8a,b), the electrochemical performance of the electrode material has been further improved, suggesting the Cu-Co-O is significant for electrochemical energy storage. The analysis above evidence that the GNTA substructure and the carbon framework constructed in this paper are practically credible, as well as the Cu-Co-O nanostructure is a good energy storage candidate.
The electrochemical performance of W/GNTAs@Cu-Co-O is detailed and displayed in Figure 9. Figure 9A shows the CV curves of the sample at scan rates from 2 to 50 mV s −1 in a potential window of 0~0.45 V and the shape of CV curves is well maintained with the increase of scan rates. For Cu-Co oxides, a pair of typical redox peaks usually can be seen in CV curves. 24,30,31 However, the peaks in Figure 9A are not quite clear, which may come down to the involvement of the biomass carbon and GNTAs in the double-layer energy storage process. The reaction kinetics and relative contributions of the capacitive and diffusion-controlled parts are shown in Supporting Information: Figure S9. Besides, Figure 9B presents GCD curves of W/GNTAs@Cu-Co-O with different current densities in a potential window of 0~0.45 V and the specific capacitance is calculated at 932.4 F g −1 (116.5 mAh g −1 ) at 1 A g −1 . The specific capacitance of the improved sample is much better than that of the reported Cu-Co-O-based materials, such as the CuO/Co 2 O 4 @N-CNT, 32 Co/Cu-MOF/Cu 2+1 O, 33 Cu-Co 2 O 4 /CuO, 29 CuCoO Nanowires, 34 and so on. The detailed comparison is shown in Supporting Information: Table S1. Finally, its cycling stability was tested at a current density of 5 A g −1 . As shown in Figure 9C, the specific capacitance retains 80.4% after 6000 cycles, and the coulombic efficiency is about 100%. The after-cycling morphology of the sample is shown in Supporting Information: Figure S10, which has little structural change compared to the original structure. The nanostructure is kept in good shape and covers the nanotubes uniformly during the cycling test.
As for the negative electrode, GNTAs also perform well. To verify the advance of the structure, the Fe-based components were chosen for negative electrodes, which have been widely explored in the formerly related research. The electrochemical performance of W/ GNTAs@Fe is demonstrated and the result displays in Figure 10. As shown in Figure 10A, a pair of redox peaks illustrates in the potential range of −1.2~0 V, whereas at high scan rates, the shape of CV curves is expanded and the redox peaks are undercover, suggesting a significant double-layer effect generated from biomass substrate and GNTAs at high scan rates. The reaction kinetics and relative contributions of the capacitive and diffusioncontrolled parts are shown in Supporting Information: Figure S11. Figure 10B,C presents the GCD curves of W/ GNTAs@Fe at current densities from 1 to 50 mA cm −2 in a potential window of −1.2~0 V. In these GCD curves, the charging and discharging parts are symmetrical, and curves are kept in good shape with the current density increasing, indicating a high Coulombic efficiency and stability. Besides, specific capacitances of W/GNTAs@Fe at different current densities have been calculated and the results are displayed in the inset of Figure 10D. The W/GNTAs@Fe exhibits a high specific capacitance of 1208.5 F g −1 (402.8 mAh g −1 ) at 1 A g −1 . Although the specific capacitance slightly decreases with the rising of the current density, that still retains 454.3 F g −1 (151.4 mAh g −1 ) at 20 A g −1 . The specific capacitance of W/GNTAs@Fe is significantly improved by the unique structure, which is much larger than those reported before, such as CNTs-Fe 3 O 4 (330 F g −1 , 0.2 A g −1 ), 35 G@Fe 3 O 4 (732 F g −1 , 2 A g −1 ), 36 Carbon Cloth/Fe 3 O 4 @C (463 F g −1 , 1 A g −1 ), 37 FeOOH (1066 F g −1 , 1 A g −1 ), 38 FeOOH/G (365 F g −1 , 1Ag −1 ). 39 The stability of W/GNTAs@Fe is shown in Figure 10D. After 5500 cycles F I G U R E 9 Electrochemical performance of sample W deposited with in situ graphene nanotube arrays (W/GNTAs)@Cu-Co-O: (A) cyclic voltammetry (CV) curves at different scan rates; (B) galvanostatic charge-discharge (GCD) curves at current densities; (C) cycling stability and coulombic efficiency at 5 A g −1 . Inset: specific capacitances at different current densities. of the charging-discharging process, the specific capacitance maintains great stability without obvious loss. The after-cycling morphology of the sample is shown in Supporting Information: Figure S12. The nanotube-active substance structure has little structural change compared to the original structure. The active substance wraps the nanotubes up uniformly during the cycling test. Thus, the electron transfer and mass transform are a little distorted during the cycling, which are essential for a long-life device.
To further analyze the practical applications of both samples, an all-wood ASC was assembled as illustrated in Figure 11A. In the ASC, W/GNTAs@Cu-Co-O served as the cathode, W/GNTAs@Fe as the anode, and an original wood carbon slice as the separator. The electrochemical performance of the ASC is illustrated in Figure 11B-I. The CV curves of both electrodes at a scan rate of 20 mV s −1 are displayed in Figure 11B and the potential windows of W/GNTAs@Cu-Co-O and W/GNTAs@Fe are well-matched. The CV and GCD curves of the ASC device with various potential windows are respectively illustrated in Figure 11C,D. Obviously, the shapes of CV curves are kept stable till the potential is up to 1.6 V, and in this potential range, the discharging and charging parts in GCD curves are all quasi-symmetric. When the potential is over 1.6 V, the CV curve becomes a little deformed, and so do the GCD curves ( Figure 11D), which may be due to the oxygen evolution reactions in the high potential range. 29 Thus, the potential of 1.6 V was chosen as the operating potential for the ASC to analyze electrochemical performance. Figure 11E presents the CV curves of the device at scan rates ranging from 2 to~50 mV s −1 . Their shapes are quasi-rectangle and well retained with the scan rate increases, so the anode and cathode are in good electrostatic compatibility. In addition, it can be concluded from Figure 11E that the double-layer charging storage originating from the wood carbon and the GNTAs is considerable in the entire energy storage process, which is advantageous to the rapid electron transport and redox reactions, as well as the good stability for the device.
The energy-storage property of the ASC was tested and Figure 11F displays the GCD curves of the device at different current densities. The GCD curves at every current density are symmetric, indicating a high Coulombic efficiency. The device delivered a high specific capacitance of 151.2 F g −1 at 1 A g −1 . The cycling stability of the device was demonstrated by GCD at a current density of 7 A g −1 ( Figure 11G) and the device has no significant capacitance deterioration and coulumbic efficiency after 5000 charging-discharging cycles. Besides, the energy and power densities have been calculated and exhibited in Figure 11H. Impressively, the energy density of the W/GNTAs@Cu-Co-O//W/GNTAs@Fe all-wood ASC is up to 53.8 Wh kg −1 and the power density is over 7.9 kW kg −1 , which are competitive compared with many reported works, such as CuCo 2 O 4 //Carbon Nanofiber (25.1 Wh kg −1 at 400 W kg −1 ), 34  Finally, as exhibited in Figure 11I, SC packs are constructed and series connected with a blue lightemitting diode (LED), and the LED glared for over 50 s before the reduction in brightness. Besides, with half of the size, the device can also power a blue light. Hence, F I G U R E 10 Electrochemical performance of sample W deposited with in situ graphene nanotube arrays (W/GNTAs)@Fe: (A) cyclic voltammetry (CV) curves at different scan rates; (B) galvanostatic charge-discharge (GCD) curves at low current densities; (C) GCD curves at larger current densities; (D) cycling stability at 20 mA cm −2 . Inset: specific capacitances at different current densities. the practical value of the electrode and the device has been evaluated, illustrating excellent reliability in practical manufacturing applications. The W/GNTAs@Cu-Co-O//W/GNTAs@Fe all-wood ASC is an environmentfriendly, nonflammable, reliable, and practical candidate in the field of electronics and energy storage with highsecurity requirements and high power/energy densities.
According to these CV curves in Figure 11E, the detailed energy-storage mechanisms and reaction kinetics could be further analyzed. In the charge-discharge process, the total charged energy could be divided into two parts, that is, the capacitive-controlled and diffusion-controlled parts. The contribution of the two parts can be qualified according to the equation: 26  to the capacitive and diffusion-controlled processes, respectively. Herein, a typical CV curve at 2 mV s −1 is shown in Figure 12A, the capacitive controlled part is 68.6%, and the proportion decreased with the scan rate rising ( Figure 12B), which means that at a lower scan rate, the metal oxides play a leading role in energy storage, while porous carbon framework takes a major role at a higher scan rate. Besides, the EIS pattern was tested and shown in Figure 12C. Compared with the EIS patterns of the single electrode, the EIS pattern of the ASC displays a similar tendency, but a larger internal resistance, which is perhaps due to the double-electrode structure of the ASC and the resultant more solid-liquid layers. As schematically illustrated in Figure 13, the structural advantage of W/GNTAs-based materials was discussed in detail, which are the main reasons for the enhanced electrochemical performances. First, wood-derived carbon with regular thorough holes serves a porous carbon framework with high specific surface area and rich functional groups, providing a conductive substrate and amounts of active sites to construct substructures. Second, the GNTA is synthesized to optimize the pore structure and together with wood-derived carbon, play as a substrate. The GNTA, on the one hand, boosts the directional electron transfer and thus improves the electroconductivity of the electrode; on the other hand, leads metal oxides to go deep into the center of the thorough holes to fully use the inner space, assisting to enhance the volumetric energy density. More importantly, the GNTA acts as a bridge connecting the metal oxide nanostructure and the electrolyte, which enlarges the electrode/electrolyte interface and shortens the diffusion path of ions and electrons, resulting in an accelerated electrochemical reaction process and improved conductivity. Third, the introduction of the pseudocapacitive nanostructure compositionally enhances the capacitive property, and structurally provides extra active sites for electrochemical reactions. Finally, due to the in situ synthesized products and hollow structure, the shape and volume changes during the electrochemical process could be alleviated, and the product illustrates good stability. Therefore, benefiting from the structural superiorities, the W/GNTAs-related electrode exhibits outstanding and comprehensive performances.

| CONCLUSION
To sum up, we have successfully established an advanced design and constructed the concept of a hierarchical allwood ASC with a fully exploited pore structure, fabricated by a facile, low-cost, and sustainable process. Wood carbon has unique merits, such as the natural thorough holes in the wood-growth direction, high porosity, ion/electron conductivity, and structural stability. With the assistance of the GNTAs, the pore structure of the electrode is optimized. The cooperation of the macro-and mesopores is essential for the expanded specific surface area and accommodation of the electrolyte. Besides, the metal oxides access hole centers along with the nanotube array, and thus the inner space has been fully explored, the electrode/electrolyte interface is enlarged so that the ion exchange and chemical reactions are accelerated. Meanwhile, the hollow structure absorbs the shape exchange during the electrochemical process to maintain the structural and functional stability. As a result, the all-wood ASC displays a high specific capacitance (151.2 F g −1 at 1 A g −1 ), an outstanding energy density (53.8 Wh kg −1 at 900 W kg −1 ), and a great cycling stability, representing the top-class values among Cu-Co-O-based SCs. This structural design of electrodes and SCs takes full advantage of the structural features of biomass and the electrochemical value of metal oxides, and the products are to be promising candidates for high-performance devices in energy storage, catalysis, and electronics fields.