Preparation of Tough, Binder‐Free, and Self‐Supporting LiFePO4 Cathode by Using Mono‐Dispersed Ultra‐Long Single‐Walled Carbon Nanotubes for High‐Rate Performance Li‐Ion Battery

Keywords: binder‐free, lithium‐ion batteries, lithium iron phosphate, mono‐dispersed single‐walled carbon nanotubes, self‐supporting

2.1. Preparation and Molecular Dynamics Simulation of SWCNT Mono‐Dispersion

Battery performance is affected not only by the intrinsic properties of the active materials but also by the electrode structure.^{[ 14 ]} Well‐dispersed conductive networks in electrodes have been proven important for fast electron transport and charge transfer.^{[ 15 ]} In this work, ultra‐long SWCNTs were synergistically dispersed by dipole/dipole electrostatic interactions and steric hindrance, and the mono‐dispersion in NMP solution was prepared by a two‐step milling method, named as SWCNT‐NMP (Figure 1a). The resulting dispersion in NMP was suction‐filtered into a film and dried, and the dispersion state of SWCNTs was observed by field emission scanning electron microscopy (SEM). The SWCNTs are staggered with each other, even under the action of force during the vacuum filtration process, they can still maintain an individual dispersion state with no obviously observed aggregates or tube bundles (Figure 1b). The transmission electron microscope (TEM) reveals an individual distribution of SWCNTs in the solvent, and the dispersant molecules are partially helically adsorbed on the surface of SWCNTs (Figure 1c). There is a very low‐intensity D peak and high‐intensity G peak in the bare SWCNT powders, indicating a very small amount of amorphous carbon or defects.^{[ 16 ]} After dispersion, the D peak disappears and the intensity of the G peak remains almost unchanged, which may be ascribed to the reason that mono‐dispersed SWCNTs are more inclined to a single distribution state and with an improved degree of order (Figure 1d).^{[ 5 ]} Meanwhile, as‐prepared mono‐dispersed SWCNTs in NMP solution appear as colloidal dispersion with obvious Tyndall phenomenon (Figure 1e). The particle size distribution was confirmed by nanoparticle size and zeta potential analyzer, as shown in Figure 1f, in which the particle size is normally distributed in a narrow range (D90 < 220 nm). The excellent stability of the dispersion was tested after the aqueous solution was left to stand for 1 week and 6 months, respectively, showing a negligible increase in the average particle size (Figure 1g). Figure 1h shows that the addition of dispersants prevents SWCNTs from re‐aggregating, while bare SWCNTs obviously subsided and agglomerated after being placed for a week. The UV–vis absorption spectrum was used to estimate the dispersion state of SWCNTs (Figure 1i). When the dispersion is diluted to 0.005% (mass fraction), the transmittance (at 500 nm of visible light) does not exceed 20%, showing homogenous dispersion. The dispersion efficiency was calculated after centrifugation at 10 000 rpm, and it was found that the polyvinylpyrrolidone (PVP) + sodium cholate (SC) mixed dispersant had the highest dispersion efficiency, reaching 96.4% (Figure 1j). It is attributed to the synergistic effect of the two dispersants. PVP increases the compatibility of SWCNTs in NMP and the dissociation of SO3−Na ⁺ electron pairs in SC provide electrostatic repulsion, preventing SWCNTs from approaching each other, thus achieving stable mono‐dispersion.

The mono‐dispersion fabrication process for ultra‐long SWCNTs is shown in Figure 2a. First, the dispersant molecules gradually approach the SWCNTs in the solvents and adsorb on their surfaces in the pre‐dispersion stage. With the input of external energy (such as grinding), the distance between the SWCNTs becomes larger, and the dispersant molecules gradually cover the entire surface of the individual SWCNTs, which promotes their further disassembling from the bundles and finally forms a uniform mono‐dispersed SWCNT dispersion.^{[ 17 ]} MD simulations were used to investigate the effect of dispersant molecules on the dispersion behavior of SWCNTs in the NMP solvent (Figure 2b). Under vacuum conditions, the interaction energy between the individual tubes is −97.1 kcal mol⁻¹, showing the mutual attraction between nanotubes (Table S2, Supporting Information). To maintain the lowest energy in the system, the SWCNTs tend to bundle together. In the NMP‐filled conditions, the addition of dispersants decreases the interaction energy of SWCNTs and NMP, indicating that dispersant molecules can expel part of the NMP molecules around SWCNTs (Table S4, Supporting Information). The radial distribution function (RDF) also confirms that the C and H atoms of the dispersants are closer to the surface of SWCNTs (Figure 2f,g).^{[ 18 ]} We measured the energy changes of the system when two individual SWCNTs were brought into close proximity in NMP solvent (Figure S9, Supporting Information). The energy increases sharply when the distance between individual SWCNTs is reduced to 14.5 Å, indicating that the spacing cannot be further reduced, while the distance between bare SWCNTs can reach 7.5 Å (Figure 2e). MD simulations have shown that SWCNTs can be efficiently disassembled into individual SWCNTs in NMP under dipole/dipole electrostatic interactions of both SC and PVP, which are used as the driving force to disassemble the bundles of SWCNTs.^{[ 19 ]}

2.2. Conductive Network Construction and Low‐Addition, High‐Rate Performance of LFP Cathode

In order to explore the effectiveness of constructing conductive networks of mono‐dispersed SWCNTs and their effects on the electrochemical performance of electrodes, LFP cathodes composed of bare SWCNTs powders and dispersions with various ratios of SWCNTs as conductive additives were prepared, named LFP‐SWCNT _x and LFP‐SWCNT _x ‐NMP, respectively (x = 0.5, 1, 2, 3, 5, x represents the mass ratio of SWCNTs). As shown in Figure 3a,b, large agglomerates and bundles are generated when using bare SWCNTs powders as conductive additives. It can be observed that there is a smooth and bare surface of LFP particles without any SWCNTs wrapping, indicating the inhomogeneous distribution of SWCNTs in the electrode. In contrast, for the LFP‐SWCNT‐NMP cathode, no obvious agglomerates or bundles can be observed (Figure 3e,f). The as‐prepared SWCNT dispersion produced by our method constructs an unbundled network in the whole regions of the electrode, even in the electrode regions with a higher density of SWCNTs. To evaluate the performance of conductive networks and gain the high‐performance electrode at low contents as conductive additives, a systematic study of LFP electrodes with various contents from 0.5 to 5 wt% of SWCNTs was performed. In LFP‐SWCNT _x , the rate performance deteriorates significantly with decreasing conductive additive contents, especially at higher rates when electrons and ions need to be transported more efficiently (Figure 3c). However, through the construction of the mono‐dispersed SWCNTs conductive network, the resultant deterioration is not obvious as LFP‐SWCNT _x . The rate performance of LFP‐SWCNT _x ‐NMP increases first and then decreases with the decrease of the addition amount, with an optimal addition amount of 2 wt% (Figure 3g). Moreover, the capacity gap of electrodes with different addition amounts was smaller, that is to say, the conductive network constructed by well‐dispersed SWCNTs in LFP electrodes was less affected by the addition amount.^{[ 20 ]} At low contents, the rate performance of LFP‐SWCNT_0.5‐NMP is 72.9 and 59.7 mAh g⁻¹ at high charge/discharge rates of 15 and 20 C, respectively, significantly more superior to that of LFP‐SWCNT_0.5 (16.9 and 8.1 mAh g⁻¹ at 15 and 20 C). It can be ascribed to the complete and well‐established stable conductive network by mono‐dispersed ultra‐long SWCNTs at low additions. The LFP‐SWCNT_0.5‐NMP electrode shows a higher specific capacity retention of 81.4% after 200 cycles at a high charge/discharge rate of 2 C, compared with 22.2% retention for the LFP‐SWCNT_0.5 electrode (Figure 3j). The charge transfer resistance (R _ct) of the LFP‐SWCNT_0.5‐NMP electrode was calculated as 62.8 Ω, which was lower than that of the other composites (87.4, 86.7, 64.8, and 71.8 Ω for LFP‐SWCNT₅‐NMP, LFP‐SWCNT₃‐NMP, LFP‐SWCNT₂‐NMP, and LFP‐SWCNT₁‐NMP, respectively) and much lower than that of all the LFP‐SWCNT electrodes (Figures S10d and S11d, Supporting Information).

The galvanostatic intermittent titration technique (GITT) was further used to investigate the fast electron/ion transporting behaviors of the LFP electrode. The kinetic behavior of LFP‐SWCNT₅ and LFP‐SWCNT₅‐NMP were investigated by GITT and the Li⁺ diffusion coefficient (D _GITT) was calculated based on the following equation^{[ 21 ]}

where L is the Li⁺ diffusion length (cm), τ is the relaxation time (s), ΔE _s is the steady‐state voltage change (V) of the current pulse, and ΔE _t is the potential change (V) during the cross‐current pulse after removing the iR drop.

In the GITT process, the quiescent voltage of LFP‐SWCNT₅ and LFP‐SWCNT₅‐NMP was 3.43 V for all quiescent steps during charging and discharging, showing the occurrence of Li⁺ intercalation/deintercalation at this potential. The average diffusion coefficients of Li⁺ in LFP‐SWCNT₅ and LFP‐SWCNT₅‐NMP were calculated to be 1.29 × 10⁻¹⁰ and 1.61 × 10⁻¹⁰ cm² s⁻¹, respectively, indicating that the LFP‐SWCNT₅‐NMP has better ion transport behavior and smaller electrode polarization.^{[ 15c ]}

2.3. Preparation and Performance of Binder‐Free Electrode (LFP‐SWCNT _x ‐BF)

In the conventional preparation of electrodes, an electrochemically inert polyvinylidene difluoride (PVDF) binder needs to be added to connect the active material, conductive agent, and current collector.^{[ 11} , ^{22 ]} In addition to decreasing the electron transfer rate, they also increase the weight of the electrode and thus reduce the overall energy density.^{[ 10a ]} Herein, a binder‐free electrode was prepared, named LFP‐SWCNT_x‐BF (Figure 4a), where the LFP particles were wrapped and fixed by the interconnected ultra‐long SWCNT networks, providing physical bonding and stable support during the charging and discharging.^{[ 4 ]} As shown in Figure 4b,c, the distribution of ultra‐long SWCNTs in the cathode was relatively homogeneous with no observed SWCNT bundles. The individual ultra‐long SWCNTs were interconnected to form a 3D network, and thus, physical binding was achieved by wrapping and pinning LFP active particles. Moreover, the excellent mechanical strength of SWCNTs provides a certain strengthening for the cathode structure, which is beneficial for maintaining the stability of the electrode structure. As shown in Figure 4d, the cathode mixed slurry was coated on the insulated PET film for conductivity measurement. Under the same amount of addition, the electrode without PVDF binder had significantly higher conductivity than the electrode with PVDF binder. As confirmed by X‐ray diffraction (XRD), the introduction of dispersants during disassembling of the bundles into individual tubes had no adverse effect on the crystal structure of the LFP, and would not affect the redox reaction during the charging and discharging process (Figure 4e). To characterize the chemical composition and state of the LFP‐SWCNT _x ‐BF cathodes, X‐ray photoelectron spectroscopy (XPS) was used for testing. A wide spectrum shows the peaks of Li 1s, P 2p, O 1s, and Fe 2p elemental compositions, without typical peaks of F 1s (Figure S13, Supporting Information). The C 1s peaks at 284.6, 285.8, 288, and 289.5 eV can be ascribed to the C=C, C—O/C—O—C, C=O, and O—C=O bonds, respectively (Figure 4f).^{[ 15} , ^{22 ]} In the high‐resolution XPS spectrum of Fe 2p, the diffraction peaks of Fe 2p_3/2 and Fe 2p_1/2 are located at 710 and 724.1 eV, respectively, with a peak spacing of 13.9 eV, consistent with the LFP feature (Figure 4g).^{[ 15} , ^{21 ]}

Due to the absence of PVDF binder, all the LFP‐SWCNT _x ‐BF cathodes have a higher specific capacity than the results mentioned above and LFP‐SWCNT₅‐BF has a closer capacity to the theoretical value of LFP, reaching 169.2 mAh g⁻¹ at 0.2 C. With the decrease of the SWCNTs contents from 5 to 0.5 wt%, there is no obvious deterioration in the rate performance. LFP‐SWCNT₃‐BF (mass loading ≈1.5 mg cm⁻²) exhibited the best rate capacities of 161.5, 156.1, 146.8, 130.2, 111.9, 99.9, and 90.7 mAh g⁻¹, at 0.5, 1, 2, 5, 10, 15, and 20 C, respectively (Figure 4h). Moreover, the voltage profiles for typical charge/discharge cycles at 0.2, 0.5, 1, 2, 5, 15, and 20 C are shown in Figure 4i. However, the binder‐free LFP‐SWCNT₃‐BF electrodes do not contain PVDF, so only 3 wt% of the non‐conducting dispersant molecules remain. As a result, the direct contact area between the active material and the SWCNTs exposure is larger, and thus the active material shows the best performance at higher addition amount (3 wt%). The potential intervals between the two redox peaks of the LFP‐SWCNT _x ‐BF electrodes with contents of 0.5, 1, 2, 3, and 5 wt% were 0.464, 0.305, 0.280, 0.257, and 0.304 V, respectively, showing that the LFP‐SWCNT₃‐BF electrode has the lowest internal potential, which can be ascribed to the increasing Li⁺ diffusion depth and thus reduced polarization (Figure 4j). ​At the addition of 3 wt%, a sufficiently continuous and efficient 3D conductive network with enough pores can be formed in the positive electrode, which provides faster Li⁺ diffusion behavior during electrochemical reactions, and thus the shortening of the diffusion time to the surface of the active material. The charge transfer resistance (R _ct) of the LFP‐SWCNT₃‐BF electrode was calculated as 83.88 Ω, which is lower than that of the other composites (Figure 4k). According to GITT, the average diffusion coefficient of LFP‐SWCNT₃‐BF was calculated to be 8.41 × 10⁻¹⁰ cm² s⁻¹ (Figure S10, Supporting Information). After the removal of the binders, the coverage area of the electrical insulating substance inside the electrode is reduced, which can greatly improve the electron transport efficiency and the electrochemical reaction rate. The LFP‐SWCNT₃‐BF electrode shows a higher specific capacity retention of 87.4% after 200 cycles at 2 C and 71.3% after 500 cycles at 5 C, respectively, compared with other cathodes (Figure 4l,m). These results demonstrated that it can achieve comparable or higher values than most references listed in Table S5, Supporting Information, with a lower content as a conductive additive and provide a practical strategy for the application of conductive additives in high‐rate charge/discharge LIBs.

2.4. Performance of Tough, Binder‐Free, and Self‐Supporting Electrode (LFP‐SWCNT‐SS)

The slurry with lower viscosity is hard to cast onto the collector with a blade owing to the absence of binder, thus causing a low LFP mass loading in the cathodes (<3 mg cm⁻²).^{[ 15} , ^{20 ]} In order to prepare the thick electrodes with a density over 3 mg cm⁻², vacuum filtration was adopted to get the self‐supporting cathodes with higher LFP mass loadings (>8 mg cm⁻²), named LFP‐SWCNT_x‐SS. The synthesis method is illustrated in Figure 5a. Different from the traditional coating process, the organic solvent can be collected and reused after filtration, so it is more environmentally friendly and economical.^{[ 13b ]} Reducing the amount of SWCNTs resulted in brittle membrane electrodes due to the absence of a metal collector. Therefore, 5 wt% was chosen as the addition amount of SWCNTs in the self‐supporting electrodes. Various high mass loadings of self‐standing, binder‐free LFP‐SWCNT₅‐SS cathodes were obtained (mass loadings ≈8.1, 18.7, 26.5, and 39.1 mg cm⁻², respectively). With the increase in electrode thickness, there was negligible shrinkage in the radial direction (Figure S17, Supporting Information). The surface (Figure 5c–e) and cross‐sectional (Figure 5f–h) microstructures of the LFP‐SWCNT₅‐SS cathode were observed, and as prepared mono‐dispersed ultra‐long SWCNTs can form a long‐range film‐like network by tightly wrapping adjacent LFP active particles or by establishing connections between the active particles as conductive bridges.^{[ 10} , ^{23 ]} Due to the force of vacuum filtration, the individual SWCNTs or film‐forming bundles were roughly distributed parallel to the surface (Figure 5f–h).

The elemental distribution of the self‐standing, binder‐free LFP‐SWCNT₅‐SS electrode was measured by mapping C, O, Fe, and P elements, revealing a uniform distribution in the cross‐sectional electrode (Figure S16, Supporting Information). N₂ adsorption/desorption isotherms are shown as a shape of V with distinct H₃‐typed hysteresis loops, and a high specific surface area of 18.7 m² g⁻¹ and a pore volume of 0.045 cm³ g⁻¹, providing rapid and uniform diffusion of ions in the porous LFP cathode and accelerating electrochemical reactions at the electrode/electrolyte interface (Figure 5l,m).^{[ 3c ]} XRD patterns of LFP and LFP‐SWCNT₅‐SS electrodes were matched with PDF Card No. 40–1499, showing the absence of an impurity phase (Figure S21, Supporting Information). The self‐supporting cathode exhibits strong bending resistance and durability, which can be curled and folded into any shape such as “airplane” without any damage (Figure 5k).

Tensile experiments also show that the self‐supporting electrode LFP‐SWCNT₅‐SS has excellent mechanical properties, which can withstand at least 7.2 MPa stress and 5% strain (Figure 5i). The electro‐rheological properties of the LFP‐SWCNT₅‐SS cathodes at both dry and wet states were investigated, which are important parameters to evaluate whether the electrodes can maintain good structures under pressure in the electrolyte.^{[ 24 ]} The compression and relaxation curves of LFP‐SWCNT₅‐SS in the wet state are very similar to those in the dry state (Figure 5j). Electrochemical tests were performed in half‐cells without collectors. Rate performance was obtained for the LFP‐SWCNT₅‐SS electrode, showing high capacitance of 162.9, 158.8, 150.6, 139.1, and 114.7 mAh g⁻¹, at 0.2, 0.5, 1, 2, and 5 C, respectively, and much lower R _ct of 40.53 Ω compared with that of LFP‐SWCNT₅‐BF (107.3 Ω) (Figure 5n). The half‐cell was charged and discharged at 1 C for 200 cycles to test the stability of the cathode, and the cycling performance at 1 C shows that the LFP‐SWCNT₅‐SS electrode has a retention of 85.2% with a coulombic efficiency of ≈99% (Figure 5O).

Preparation of Dispersion

SWCNTs (Tuball, OCSiAl, diameter <2 nm) were prepared as mono‐dispersion via the following steps: i) 3.3 g of the as‐purchased SWCNT powders, 1.65 g of SC (Acros), and 1.65 g of PVP (TCI Chemicals) were pre‐dispersed into 1000 mL NMP using a continuously operating bead mill system for 24 h; ii) the pre‐dispersed slurry was further grinded using a bead mill (Multi Lab DYNO‐Mill, 0.6 mm zirconium beads) for 16 h to ensure the SWCNTs were dispersed individually. ​The resulting homogeneous monodispersed SWCNT slurry was named SWCNT‐NMP.

Preparation of Cathodes

The components of the cathodes with a total solid content of 100 mg included 5 wt% PVDF as the binder, the different SWCNT powder mass ratios ranging from 5, 3, 2, 1, and 0.5 wt%, and the corresponding LiFePO₄ mass ratios 90, 92, 93, 94, and 94.5wt %, named as LFP‐SWCNT _x , in which x (mg) represented the mass of SWCNTs per 100 mg of the cathode. In a similar way, the cathode that SWCNT powder was replaced with dispersion named LFP‐SWCNT _x ‐NMP. Furthermore, in order to reduce the content of inactive substances, the addition of PVDF was removed, and the corresponding binder‐free cathode was named LFP‐SWCNT _x ‐BF. The details are shown in Table S1, Supporting Information. Meanwhile, a self‐supported and flexible LFP cathode with high mass loading without collector and binder was prepared with SWCNT dispersion, named LFP‐SWCNT₅‐SS (x = 5). The cathode slurry of LFP‐SWCNT _x , LFP‐SWCNT _x ‐NMP, and LFP‐SWCNT _x ‐BF was obtained by grinding the mixture of LFP powder and conductive additive for 30 min, and then adding PVDF solution of NMP (solid content was 20 mg mL⁻¹), after that, fully mixed for 30 min. The cathodes were prepared by depositing them on a carbon‐coated aluminum foil and dried in a vacuum at 80 °C for 12 h. Self‐supported cathode slurry was obtained by stirring magnetic stirring for 30 min, the mixture of LFP powder and SWCNT dispersion, which was diluted to 50 mL, and then stirred for 30 min after ultrasonic processing (on/off = 2s/2s at 5% power) in the Ultrasonic Homogenizer (JY92‐IIDN, 900 W) for 10 min. The self‐supported cathodes were fabricated by vacuum filtration and dried to completely remove the solvent. For the specific mass additions of each component of preparation, refer to Table S1, Supporting Information.

Materials Characterizations

The morphology and the distribution of the components in the dispersion and the electrode were obtained with field emission SEM (FEI Nova Nano 450) and energy dispersive spectrometer (EDS). Individual SWCNTs in the dispersion were observed by TEM observations (FEI Talos F200S Super‐X) at an accelerating voltage of 200 kV. Degrees of dispersion was confirmed via size distribution analysis using a nanoparticle size and zeta potential analyzer (BeNano 90 Zeta, Bettersize Instruments Ltd.). Elemental composition was examined by XPS (Thermo Scientific K‐Alpha) with an Al Kα X‐ray source. XRD was acquired on a Smart Lab. The specific surface area and pore structures of electrodes were analyzed using Nitrogen adsorption/desorption by the Brunauer–Emmett–Teller (BET, Micromeritics ASAP 2460) method at 77 K.

Electrochemical Measurements

All the cathodes used for electrochemical measurements were punched as 12 mm diameter discs. The as‐prepared cathodes were assembled into Coin‐type (CR2032) half‐cells, which used lithium foil as the counter electrode, a glass fiber membrane (Whatman, GF/D) as the separator, and 1 m LiPF₆ in 1 m LiPF₆ in a mixture of ethyl carbonate/dimethyl carbonate (EC/DMC, 1:1, in volume). The electrochemical performances of obtained batteries were tested on the Neware battery tester in the voltage range of 2–3.8 V (vs Li/Li⁺) for LFP electrodes at different current densities (the current density of 1 C was ≈0.17 A g⁻¹) on the Neware battery tester at room temperature (25 ± 2 °C). All the batteries were activated for five cycles at 0.2 C before the rate and cycling performance were tested. Cyclic voltammetry (CV) measurements were performed on an electrochemical workstation (CHI660E) at a scan rate of 0.1 mV s⁻¹. The electrochemical impedance spectrum (EIS) tests were also conducted over a frequency range from 100 kHz to 0.01 Hz.

Molecular Dynamics Simulation

MD simulations were performed using the LAMMPS package. The potential function polymer consistent force field (PCFF) was used to describe the various bonding and non‐bonding interactions involving dispersants and SWCNTs.^{[ 17 ]} The simulation process was carried out in a box of size 6 × 6 × 5 nm. The box space was filled with NMP molecules, and the SWCNTs were placed in the center. Two systems were considered, one contained dispersants and the other one was without dispersants. For the specific parameters of the simulation, refer to Tables S2–S4, Supporting Information.

Supporting Information

• 1.a) Xu T., Du H., Liu H., Liu W., Zhang X., Si C., Liu P., Zhang K., Adv. Mater. 2021, 33, 2101368; [DOI] [PMC free article] [PubMed] [Google Scholar]; b) Si J., Xu L., Zhu M., Zhang Z., Peng L.‐M., Adv. Electron. Mater. 2019, 5, 1900122. [Google Scholar]
• 2. Dunn B., Kamath H., Tarascon J.‐M., Science 2011, 334, 928. [DOI] [PubMed] [Google Scholar]
• 3.a) Zhang L., Qin X., Zhao S., Wang A., Luo J., Wang Z. L., Kang F., Lin Z., Li B., Adv. Mater. 2020, 32, 1908445; [DOI] [PubMed] [Google Scholar]; b) Hwang C., Song W.‐J., Han J.‐G., Bae S., Song G., Choi N.‐S., Park S., Song H.‐K., Adv. Mater. 2018, 30, 1705445; [DOI] [PubMed] [Google Scholar]; c) Chao D., Xia X., Liu J., Fan Z., Ng C. F., Lin J., Zhang H., Shen Z. X., Fan H. J., Adv. Mater. 2014, 26, 5794. [DOI] [PubMed] [Google Scholar]
• 4. Park S.‐H., King P. J., Tian R., Boland C. S., Coelho J., Zhang C., McBean P., McEvoy N., Kremer M. P., Daly D., Coleman J. N., Nicolosi V., Nat. Energy 2019, 4, 560. [Google Scholar]
• 5. He Y., Lin X., Feng Y., Luo B., Liu M., Adv. Sci. 2022, 9, 2204675. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 6.a) Tang R., Yun Q., Lv W., He Y.‐B., You C., Su F., Ke L., Li B., Kang F., Yang Q.‐H., Carbon 2016, 103, 356; [Google Scholar]; b) Qi X., Blizanac B., DuPasquier A., Oljaca M., Li J., Winter M., Carbon 2013, 64, 334; [Google Scholar]; c) Shi Y., Wen L., Pei S., Wu M., Li F., J. Energy Chem. 2019, 30, 19; [Google Scholar]; d) Liu X.‐M., Huang Z. d., Oh S. w., Zhang B., Ma P.‐C., Yuen M. M. F., Kim J.‐K., Compos. Sci. Technol. 2012, 72, 121; [Google Scholar]; e) Liu Y., Zhang H., Huang Z., Wang Q., Guo M., Zhao M., Zhang D., Wang J., He P., Liu X., Terrones M., Wang Y., J. Alloys Compd. 2022, 905, 164205. [Google Scholar]
• 7. Wang Y., Fugetsu B., Carbon 2015, 82, 152. [Google Scholar]
• 8. Lin L., Peng H., Liu Z., Nat. Mater. 2019, 18, 520. [DOI] [PubMed] [Google Scholar]
• 9. Zhang Q., Huang J.‐Q., Qian W.‐Z., Zhang Y.‐Y., Wei F., Small 2013, 9, 1237. [DOI] [PubMed] [Google Scholar]
• 10.a) Wang H., Fu J., Wang C., Wang J., Yang A., Li C., Sun Q., Cui Y., Li H., Energy Environ. Sci. 2020, 13, 848; [Google Scholar]; b) Brahim S., Maat S., ECS Meet. Abstr. 2022, MA2022‐01, 379; [Google Scholar]; c) Jin T., Han Q., Jiao L., Adv. Mater. 2020, 32, 1806304; [DOI] [PubMed] [Google Scholar]; d) Shen K., Zhai S., Wang S., Ru Q., Hou X., Hui K. S., Hui K. N., Chen F., Batteries Supercaps 2021, 4, 860. [Google Scholar]
• 11.a) Yang B. Z., Xie P., Luo Q., Li Z. W., Zhou C. Y., Yin Y. H., Liu X. B., Li Y. S., Wu Z. P., Adv. Funct. Mater. 2022, 32, 2110665; [Google Scholar]; b) Han J. H., Shin K. H., Lee Y. J., ACS Appl. Mater. Interfaces 2021, 13, 6375. [DOI] [PubMed] [Google Scholar]
• 12.a) Joshi B., Samuel E., Kim Y.‐i., Yarin A. L., Swihart M. T., Yoon S. S., Adv. Funct. Mater. 2021, 31, 2008181; [Google Scholar]; b) Deng W., Wang Y. L., Guo W., Zhao C. F., Liu X. B., Hu Y. Y., Li Y. S., Xia B. Y., Wu Z. P., Carbon 2021, 178, 573; [Google Scholar]; c) Chandrasekaran S., Khandelwal M., Dayong F., Sui L., Chung J. S., Misra R. D. K., Yin P., Kim E. J., Kim W., Vanchiappan A., Liu Y., Hur S. H., Zhang H., Bowen C., Adv. Energy Mater. 2022, 12, 2200409; [Google Scholar]; d) Wang J., Wang B., Liu X., Bai J., Wang H., Wang G., Chem. Eng. J. 2020, 382, 123050. [Google Scholar]
• 13.a) Zhu B., Liu G., Lv G., Mu Y., Zhao Y., Wang Y., Li X., Yao P., Deng Y., Cui Y., Zhu J., Sci. Adv. 2019, 5, eaax0651; [DOI] [PMC free article] [PubMed] [Google Scholar]; b) Luais E., Mery A., Abou‐Rjeily J., Sakai J., Tran‐Van F., Ghamouss F., J. Electrochem. Sci. Technol. 2019, 10, 373; [Google Scholar]; c) Wang L., Bock D. C., Li J., Stach E. A., Marschilok A. C., Takeuchi K. J., Takeuchi E. S., ACS Appl. Mater. Interfaces 2018, 10, 8770. [DOI] [PubMed] [Google Scholar]
• 14.a) Chabi S., Peng C., Hu D., Zhu Y., Adv. Mater. 2014, 26, 2440; [DOI] [PubMed] [Google Scholar]; b) Meng J., Guo H., Niu C., Zhao Y., Xu L., Li Q., Mai L., Joule 2017, 1, 522. [Google Scholar]
• 15.a) Qin J., Zhang Q., Cao Z., Li X., Hu C., Wei B., Nano Energy 2013, 2, 733; [Google Scholar]; b) Wang Z., Lu Y., ACS Appl. Mater. Interfaces 2019, 11, 13225; [DOI] [PubMed] [Google Scholar]; c) Liu Y., Zhang H., Huang Z., Wang Q., Guo M., Zhao M., Zhang D., Wang J., He P., Liu X., Terrones M., Wang Y., J. Alloys Compd. 2022, 905, 164205; [Google Scholar]; d) Zhang M., Ning G., Xiao Z., Energy Technol. 2020, 8, 2000589. [Google Scholar]
• 16. Dresselhaus M. S., Jorio A., Hofmann M., Dresselhaus G., Saito R., Nano Lett. 2010, 10, 751. [DOI] [PubMed] [Google Scholar]
• 17. Pramanik C., Gissinger J. R., Kumar S., Heinz H., ACS Nano 2017, 11, 12805. [DOI] [PubMed] [Google Scholar]
• 18. Tummala N. R., Morrow B. H., Resasco D. E., Striolo A., ACS Nano 2010, 4, 7193. [DOI] [PubMed] [Google Scholar]
• 19. Giordani S., Bergin S., Nicolosi V., Lebedkin S., Blau W. J., Coleman J. N., Phys. Status Solidi B 2006, 243, 3058. [Google Scholar]
• 20. Yang S., Zhou C., Wang Q., Chen B., Zhao Y., Guo B., Zhang Z., Gao X., Chowdhury R., Wang H., Lai C., Brandon N. P., Wu B., Liu X., Energy Environ. Mater. 2022, 5, 1332. [Google Scholar]
• 21. Li H., Peng L., Wu D., Wu J., Zhu Y.‐J., Hu X., Adv. Energy Mater. 2019, 9, 1802930. [Google Scholar]
• 22. Yin B., Zhang S., Ke K., Xiong T., Wang Y., Lim B. K. D., Lee W. S. V., Wang Z., Xue J., Nanoscale 2019, 11, 19723. [DOI] [PubMed] [Google Scholar]
• 23. Zhu K., Luo Y., Zhao F., Hou J., Wang X., Ma H., Wu H., Zhang Y., Jiang K., Fan S., Wang J., Liu K., ACS Sustainable Chem. Eng. 2018, 6, 3426,. [Google Scholar]
• 24. Jing L., Ji Y., Feng L., Fu X., He X., He Y., Zhu Z., Sun X., Liu Z., Yang M., Yang W., Wang Y., Energy Storage Mater. 2022, 45, 828. [Google Scholar]

Supplementary Materials

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.