A two-dimensional assembly of ultrafine cobalt oxide nanocrystallites anchored on single-layer Ti₃C₂T_x nanosheets with enhanced lithium storage for Li-ion batteries

Abstract

Recently, Ti-based MXenes were expected to compete with graphene and other carbonaceous materials towards Li-ion batteries (LIBs) due to their two-dimensional (2D) open structure, cost efficiency, superior conductivity and low Li⁺ diffusion barrier. However, the relatively moderate capacity and aggregation tendency hamper their practical applications in next-generation LIBs. Herein, we explore for the first time a scalable bottom-up approach to fabricate a series of Co₃O₄@single-layer Ti₃C₂T_x (s-Ti₃C₂T_x) hybrids, where numerous homogeneous Co₃O₄ nanocrystallites (NCs), serving both as a spacer and electroactive phase, are anchored uniformly on the surface of s-Ti₃C₂T_x nanosheets (NSs) through the Co–O–Ti interfacial bonds. Furthermore, detailed experimental analyses clearly shed light upon the formation mechanism of the hybrid Co₃O₄@s-Ti₃C₂T_x NSs. Thanks to the structural and compositional merits, the 2D Co₃O₄@s-Ti₃C₂T_x NSs even exhibit a remarkable high-rate capacity of ∼223 mA h g⁻¹ at an ultra-high current density of 10 A g⁻¹, and a long-span cycle life with a high reversible capacity of 550 mA h g⁻¹ at 1 A g⁻¹ after 700 consecutive cycles. Corresponding density functional theory calculation further confirms that the Co–O–Ti interfacial function leads to an even higher pseudocapacitive contribution and faster lithium storage behavior due to the enhanced interfacial electron transfer.

---

1. Introduction

The emerging market of large-scale energy storage applications, such as electric vehicles and electric grids, encourages a great deal of interest in research on versatile energy storage systems (ESSs).¹ Li-ion batteries (LIBs), as a promising platform for ESSs, have drawn significant attention in terms of energy/power density and lifetime merits. Unfortunately, the low-capacity graphite-based anodes (∼372 mA h g⁻¹) seriously limit the applications of commercial LIBs. Many researchers are making enormous efforts to explore eligible anodes for advanced LIBs to meet rigorous future demands.^2–4

Among the available anode candidates, transition-metal-oxides (TMOs) have been extensively investigated due to their large specific theoretical capacities,⁵ in which Co₃O₄ is deemed as an anode for LIBs owing to its large capacity and facile synthesis.^6,7 However, the direct utilization of Co₃O₄, like all other TMOs, still results in several ineluctable disadvantages, including sluggish Li⁺ kinetics,⁸ severe volumetric expansion/shrinkage and an unstable solid electrolyte interphase (SEI) layer over successive lithiation/delithiation cycles.^9,10 To date, two general efficient strategies have been developed well to address these issues. One is reducing the particle size of Co₃O₄ from micro- to nano-dimensions, effectively shortening the Li⁺ diffusion path and facilitating the accommodation of volumetric changes, and the other is the hybridization of Co₃O₄ with other conductive and flexible matrices, which can improve the conductivity of the Co₃O₄-based anodes and buffer the internal stress caused by the volume change meanwhile. Although both exhibit positive functions on enhancing the rate and cycle behaviors of LIBs,¹¹ however, the synthesis of well-dispersed nanoscale Co₃O₄ in appealing matrices still remains a great challenge towards enhancing its lithium storage via smart integration of the two methods mentioned above.

Recently, MXenes, a representative type of two-dimensional (2D) transition metal carbide/nitride, have gained widespread attention in energy-related applications.^12–16 Typically, Ti₃C₂T_x (T for F, OH and O) is one of the most investigated MXenes due to its mature preparation process and low cost. Besides, Ti₃C₂T_x intrinsically possesses metallic conductivity, continuous 2D charge transfer channels,^17,18 a robust structure¹⁹ and a low ionic diffusion barrier,^20,21 which make it especially serve as a matrix material to support other electroactive materials, such as TMOs, for elegantly engineering robust LIB electrodes. Commonly, multi-layer Ti₃C₂T_x (m-Ti₃C₂T_x) is applied for LIBs. However, the huge interlayer space in the m-Ti₃C₂T_x is just a “dead surface”, which cannot be fully exploited for loading other electroactive phases, or even utilized as electroactive sites for lithium storage. By contrast, single-layer Ti₃C₂T_x (s-Ti₃C₂T_x) nanosheets (NSs) emerge with an even larger naked surface. Unfortunately, numerous hydrogen bonds attached on their surface Ti atoms make the s-Ti₃C₂T_x NSs prone to restacking, resulting in a substantial loss of active area and inaccessibility of electrolyte ions.^22–24 Therefore, it is of huge significance to devise several simple yet efficient strategies to construct nanoarchitectured hybrids consisting of s-Ti₃C₂T_x NSs and nanosized Co₃O₄.

Inspired by the comprehensive considerations above, in this work, we developed a facile freeze-drying strategy along with the subsequent calcination for large-scale fabrication of 2D hybrid Co₃O₄@s-Ti₃C₂T_x NSs, where the s-Ti₃C₂T_x NSs were uniformly decorated with size-controlled Co₃O₄ nanocrystallites (NCs). The negatively charged functional groups on the surface of s-Ti₃C₂T_x NSs anchored and restrained the growth and amalgamation of the Co₃O₄ NCs; meanwhile the in situ growth of Co₃O₄ NCs minimized the strong tendency to restacking and/or agglomeration of the s-Ti₃C₂T_x itself. Owing to the synergetic contributions from their compositional and structural advantages, the as-obtained Co₃O₄@s-Ti₃C₂T_x hybrid NSs exhibited a superb rate capability of ∼223 mA h g⁻¹ at 10 A g⁻¹, high reversible capacity and long-term cycling stability (up to ∼550 mA h g⁻¹ after 700 cycles at 1 A g⁻¹ and ∼160 mA h g⁻¹ at 5 A g⁻¹ after 1000 cycles) as a competitive anode material for advanced LIBs.

2. Experimental section

2.1 Material preparation

2.1.1 Preparation of 2D s-Ti₃C₂T_x NSs. Typically, m-Ti₃C₂T_x was synthesized by etching 3 g of Ti₃AlC₂ (>98 wt% purity, Shanghai Yuehuan New Materials Technology Co., Ltd) powder in 50 mL of hydrofluoric acid (HF, 40 wt%) for 72 h at 50 °C. The obtained sediment was washed with de-ionized (DI) water and then centrifuged until the pH of the supernatant became 6–7, followed by drying at 60 °C in a vacuum. To obtain s-Ti₃C₂T_x, 0.2 g of m-Ti₃C₂T_x was further added into 10 mL of aqueous tetramethylammonium hydroxide (TMAOH) (25 wt%) and stirred for 24 h at 30 °C. The mixture was washed twice via centrifugation (10 000 rpm, 5 min per cycle) to ensure that the pH reaches ∼7.0. A stable s-Ti₃C₂T_x colloidal solution was collected by centrifugation (10 000 rpm, 5 min per cycle) after sonication (1 h) of the sedimentation. Although the texture of s-Ti₃C₂T_x could be preserved largely in the initial 5 days in aqueous solution at room temperature (RT), the colloidal solution was stored in a refrigeration environment (at ∼4 °C) to improve the stability and lifetime of s-Ti₃C₂T_x. To obtain freeze-dried Ti₃C₂T_x (f-Ti₃C₂T_x), the resultant colloidal solution (2.5 mg mL⁻¹) was rapidly frozen using liquid nitrogen, and further freeze-dried at −90 °C.

2.1.2 Preparation of Co₃O₄@s-Ti₃C₂T_x nanohybrids. A certain amount of Co(NO₃)₂·6H₂O was added to the s-Ti₃C₂T_x colloidal solution (50 mg) under stirring at RT. Before the freeze drying process, the resulting homogeneous mixture was first subjected to rapid freezing using liquid nitrogen. After being freeze-dried for 48 h, Ti₃C₂T_x/Co salt composites were obtained. Subsequently, the composites were thermally annealed at 270 °C with a ramp rate of 2° min⁻¹ for 3 h under a N₂ atmosphere to finally obtain the Co₃O₄@s-Ti₃C₂T_x nanohybrids. The Co₃O₄ contents in the Co₃O₄@s-Ti₃C₂T_x nanohybrids were changed by using 100, 200, 400 and 600 mg of the Co(NO₃)₂·6H₂O, respectively, and the corresponding Co₃O₄@s-Ti₃C₂T_x nanohybrids were labeled as H-2, H-4, H-8 and H-12, respectively. The specific contents of Co₃O₄ in H-2, H-4, H-8 and H-12 are about 35.5, 52.5, 68.8 and 76.8 wt%, respectively, which are calculated based on their theoretical stoichiometry. Pure Co₃O₄ was also prepared with the similar procedure mentioned above for the Co₃O₄@s-Ti₃C₂T_x nanohybrids just without the s-Ti₃C₂T_x.

2.2 Material characterization

The zeta potentials of the s-Ti₃C₂T_x colloidal solution and the Ti₃C₂T_x/Co salt mixture were measured using a Zetasizer Nano-ZS90 (Malvern Instruments). The thickness of the Ti₃C₂T_x NSs was measured using an atomic force microscope (AFM, Bruker Dimension Icon). The crystalline phases of all samples were characterized by powder X-ray diffraction (XRD, Cu Kα radiation, Rigaku Ultima IV) at a step scan of 10° min⁻¹ ranging from 5 to 80°. The morphologies of the samples were investigated by field-emission scanning electron microscopy (FESEM, JEOL-6300F, 15 kV), transmission electron microscopy (TEM), scanning TEM (STEM), and high-resolution TEM (HRTEM) (JEOL JEM 2100 system) with energy dispersive X-ray spectroscopy (EDS). The chemical composition and surface elemental electronic states were characterized by EDS and X-ray photoelectron spectroscopy (XPS, Thermo, an Escalab 250xi spectrometer) with an Al Kα monochromatic X-ray source (1486.6 eV). Raman spectroscopy was performed using a Raman spectroscope (Lab RAM HR, excited by a 514.5 nm laser).

2.3 Electrochemical evaluation

Coin-type cells (CR2032) were fabricated in an argon-filled glove box (MBRAUN, Germany) with oxygen and moisture contents being under 0.5 ppm to evaluate the electrochemical performance of the electrodes. The working electrodes were obtained by mixing the active materials with acetylene black and carboxymethyl cellulose (CMC, average M_w: ∼250 000) (80 : 10 : 10 in weight) in DI water. Then the homogeneous slurry was coated on copper foil and dried in a vacuum oven at 110 °C. The typical loading mass of the electroactive materials was ∼1.5 mg cm⁻². In a half-cell configuration, metallic lithium foil was used as the counter/reference electrode and a microporous membrane (Celgard 2302) as the separator. The electrolyte was a solution of 1.0 M LiPF₆ in ethylene carbonate and diethylene carbonate with a volumetric ratio of 1 : 1. Galvanostatic charge–discharge cycling tests were carried out using the LAND CT2001A battery testing system (Wuhan, China) with a voltage window range of 0.01–3.00 V (vs. Li/Li⁺). Cyclic voltammetry (CV) tests were also performed in a voltage range between 0.01 and 3.00 V (vs. Li/Li⁺). The electrical impedance spectroscopy (EIS) measurements were performed between 0.01 and 100 kHz with a potential amplitude of 5.0 mV. Both CV and EIS were conducted on an electrochemical workstation (IviumStat.h, The Netherlands). All electrochemical measurements were conducted at RT.

2.4 Computational method

All the spin-polarized density functional theory (DFT) calculations were performed with the Perdew–Burke–Ernzerhof (PBE)²⁵ function within the generalized gradient approximation (GGA) using the Vienna ab initio simulation package (VASP).²⁶ The projector augmented wave (PAW) method was used to describe the electron–ion interaction.²⁷ The van der Waals (vdW) interaction was described using the opt-B86b approach.²⁸ The Hubbard-U (GGA+U, U = 3 eV)²⁹ correction was applied for a better description of the localized d-electrons of the Co element. A hydroxylation structure consisting of Ti₃C₂ and the (222) surface of Co₃O₄ was used to model the Co₃O₄/Ti₃C₂ periodic interface. A relatively small supercell, i.e., 1 × 1 of Co₃O₄ (222) on 4 × 4 of Ti₃C₂, is large enough to maintain the lattice mismatch between Ti₃C₂ and Co₃O₄ (222) within 2.5%. What's more, among all the low index facets, the (222) surface of Co₃O₄ has the lowest surface energy.³⁰ A plane wave basis set with a cut-off energy of 500 eV was used. The Brillouin zone was sampled by using a 3 × 3 × 1 Monkhorst–Pack k-point mesh. All structures were optimized based on the electronic and ionic degrees of freedom using thresholds for the total energy of 10⁻⁵ eV and a force of 0.02 eV Å⁻¹. A vacuum region of about 15 Å was adopted to avoid unnecessary interactions.

3. Results and discussion

3.1. Physicochemical and structural characterization

The overall synthetic procedure for the assembly of uniform Co₃O₄ NCs anchored on 2D s-Ti₃C₂T_x NSs is schematically illustrated in Scheme 1. Firstly, m-Ti₃C₂T_x is etched from the bulk Ti₃AlC₂ with the HF solution. After being delaminated by TMAOH under sonication via an ion exchange between protons and TMA⁺ in the interlayer of m-Ti₃C₂T_x, swelling and exfoliation take place, and the s-Ti₃C₂T_x colloidal solution is obtained with an average zeta potential of −3.34 μm cm V s⁻¹, owing to the negatively charged functional groups on the surface of s-Ti₃C₂T_x. With the subsequent addition of cobalt ions into the above solution, the Co²⁺ ions were spontaneously attached on the s-Ti₃C₂T_x NSs through the electrostatic effect, rendering the mixture with a decreased zeta potential of −0.539 μm cm V s⁻¹. The hybrid solution was rapidly frozen by using liquid nitrogen. After being dried via a freeze-drying process, the sample is then annealed in N₂, which induces the in situ decomposition of the cobalt salt into the Co₃O₄ NCs on the s-Ti₃C₂T_x surface, and the 2D Co₃O₄@s-Ti₃C₂T_x hybrid NSs are produced accordingly.

Scheme 1 Schematic illustration for the fabrication process of the Co₃O₄@s-Ti₃C₂T_x hybrid material.

Fig. 1 shows the wide-angle X-ray diffraction (XRD) patterns of Co₃O₄@s-Ti₃C₂T_x with various Co₃O₄ loadings (i.e., H-2, H-4 and H-8), along with those for the pure Co₃O₄ and f-Ti₃C₂T_x. All these discernable peaks for H-2, H-4 and H-8 are indexed and assigned to the reflections of the cubic Co₃O₄ phase (JCPDS file No. 76-1802) with a space group of Fd3m. Besides, typical diffraction signals from Co₃O₄ become more and more intense as the Co₃O₄ content increases in the hybrids. It is worth noting that the (002) peak from Ti₃C₂T_x in the XRD pattern of Co₃O₄@s-Ti₃C₂T_x hybrids obviously disappears. This unique phenomenon may be ascribed to the fact that the in situ growth of Co₃O₄ NCs, as a smart spacer, efficiently restrains the restacking of the s-Ti₃C₂T_x NSs in the 2D hybrids.³¹ In addition, the defect introduction in s-Ti₃C₂T_x components due to the formation of Co₃O₄ NCs would weaken the (002) peaks from the Ti₃C₂T_x.¹ It is noteworthy that no signals for TiO_x phases can be detected in the reflections for the hybrids, which confirms that the oxidation of s-Ti₃C₂T_x can be prevented by low-temperature annealing under a N₂ atmosphere.

Fig. 1 XRD patterns of the f-Ti₃C₂T_x, H-2, H-4, H-8 and single-phase Co₃O₄ samples.

Raman spectroscopy was conducted to further figure out the surface structure of the products, as comparably shown in Fig. 2. The characteristic peak at 205 cm⁻¹ results from the out-of-plane A_1g vibration of the Ti atoms in the f-Ti₃C₂T_x, and those peaks at 370 and 571 cm⁻¹ correspond to the C vibrations_.^32,33 As noted, the peaks detected at around 193, 481, 523, and 670 cm⁻¹ can be attributed to the F_2g, E_g, F¹_2g, and A_1g modes of the Co₃O₄, respectively.³⁴ Strikingly, the H-4 specimen combines the prominent Raman features of the Co₃O₄ and f-Ti₃C₂T_x, while the peak intensities of the Co₃O₄ in H-4 become even weaker, compared to that for the pure Co₃O₄, for which the Co₃O₄ NCs separated well by s-Ti₃C₂T_x NSs should be well responsible.³⁵ Notably, in the H-4 specimen, the characteristic peaks for the Co₃O₄ exhibit an obvious shift to the lower wavenumber while those for the f-Ti₃C₂T_x shift to the higher wavenumber, indicating the strong chemical interactions between the s-Ti₃C₂T_x and Co₃O₄.

Fig. 2 Raman spectra of the f-Ti₃C₂T_x Co₃O₄ and H-4 samples.

X-ray photoelectron spectroscopy (XPS) was further performed to explore the chemical composition and valence information of the obtained products. Compared with m-Ti₃C₂T_x, the increase in the intensity of O 1s is obviously observed for the f-Ti₃C₂T_x and H-4, as visualized in the survey spectra (Fig. 3a), which is related to more oxygen functional groups (such as –OH and –O) terminating the larger naked surface of the s-Ti₃C₂T_x and Co₃O₄ NCs in the H-4 sample, respectively. The high-resolution XPS spectra and the corresponding fitted peaks of Ti 2p for the three samples are shown in Fig. 3b, and the corresponding quantitative analysis is shown in Fig. 3c. In contrast to m-Ti₃C₂T_x, the peaks attributed to Ti³⁺ and Ti–O from the f-Ti₃C₂T_x and H-4 samples become even stronger, which increase from 13.3 at% (m-Ti₃C₂T_x) to 23.9 at% (f-Ti₃C₂T_x)/28.7 at% (H-4) for Ti³⁺, and from 9.2 at% (m-Ti₃C₂T_x) to 35 at% (f-Ti₃C₂T_x)/71.3 at% (H-4) for Ti–O, respectively. In contrast, the peaks corresponding to Ti and Ti²⁺ become significantly weakened in the f-Ti₃C₂T_x, and even disappear in H-4, revealing the increase in the valence of Ti atoms in the hybrid. In particular, the shift in the binding energy for the H-4 sample confirms the existence of chemical bonding between the s-Ti₃C₂T_x and Co₃O₄ NCs, which will result in the spontaneous charge transfer from the adsorbates to the substrate via the Co–O–Ti bonds at the interface between the two components.^36,37 To straightforwardly elucidate the impact of –OH terminations and Co–O–Ti bonds on the valence state of Ti atoms, atomic charge calculation was conducted using the Bader charge analysis.³⁸ The analysis for the relaxed configuration about the two-layer bare Ti₃C₂ and the hydroxylation structure of the hybrid is conducted. As for the bare Ti₃C₂ (Fig. 3d), the Bader charge of Ti atoms ranges from 1.03 to 1.57, in which the Ti₃ surface from the interlayer possesses the lowest valence state due to the vdW interaction between the two layers. While for the hydroxylated Ti₅ surface, the Bader charge increases from 1.13 to 1.50 since the O draws electrons from the atoms of Ti₅, which is consistent with the report by Wang's group.³⁹ As shown in Fig. 3e, the Co atoms on the Co₃O₄ (222) surface are anchored on the surface of s-Ti₃C₂T_x, forming Co–O–Ti bonds at the interface between the two components. As a result, with the above calculation results in mind, it is easy to conclude that the charges of Co atoms would transfer to Ti₄ through O ions via the interfacial bonds, and that the surface Ti₄ in the hybrid has the highest valence state of 1.59, which provides further evidence for the strong interaction between the s-Ti₃C₂T_x NSs and Co₃O₄ NCs. Accordingly, Ti elements in the hybrids display a higher binding energy in average, which is in good agreement with the XPS spectra (Fig. 3b).

Fig. 3 (a) XPS survey spectra, (b) high-resolution Ti 2p spectra and (c) percentages of various atomic Ti species (Ti, Ti²⁺, Ti³⁺ and Ti–O) of the m-Ti₃C₂T_x, f-Ti₃C₂T_x and H-4. Bader charge results of (d) two-layer bare Ti₃C₂ and (e) the hydroxylation structure of the hybrid.

The C 1s region of the m-Ti₃C₂T_x, f-Ti₃C₂T_x and H-4 samples (Fig. S1a and Table S1, ESI†) displays four components corresponding to the C–Ti bonds from Ti₃C₂, C–C bonds from the testing type (284.6 eV),^40,41 and C–O and O–C=O bonds from oxygen terminations, respectively. The component assigned to C–Ti shows an obvious decline in f-Ti₃C₂T_x and H-4, which can possibly be attributed to the defect introduction in the Ti–C layers due to the exfoliation and annealing processes.³⁶ The O 1s XPS spectra (Fig. S1b and Table S1, ESI†) can be fitted with four kinds of peaks, corresponding to the Co–O and/or Ti–O bonds (M–O), oxygen termination bonds (–O_x, –(OH)_x), adsorbed H₂O (H₂O_ads) and organic compounds (OR) due to atmospheric contaminations, which overlap with, and obscure, many other peaks.^36,37 Notably, the peak intensity of the M–O bonds obviously increases in H-4, indicating the formation of Co–O bonds. Quantitative analysis (Fig. S1c, ESI†) implies that the specific content of Co₃O₄ in the H-4 sample is ∼58.95 wt% based on its stoichiometry, which is in good agreement with the theoretical data.

Fig. 4a shows the field-emission scanning electron microscopy (FESEM) image of the s-Ti₃C₂T_x. Micro-sized NSs are apparently observed, which is wholly distinct from the accordion-like m-Ti₃C₂T_x (Fig. S2, ESI†). The typical Tyndall effect, also known as the Willis–Tyndall scattering phenomenon (Fig. 4b), visually evidences the good dispersion of the s-Ti₃C₂T_x in aqueous solution. The atomic force microscopy (AFM) measurement was further performed to detect the topography of the resulting s-Ti₃C₂T_x NSs. As shown in Fig. 4c, the thickness of the s-Ti₃C₂T_x NSs with a lateral dimension of approximately 500 nm is approximately 1.5 nm, somewhat larger than that of unilamellar Ti₃C₂T_x NSs (∼0.98 nm).^42,43 The increase in the thickness of the s-Ti₃C₂T_x observed here should mainly result from the presence of surface adsorbates, such as water molecules, similar to the observations on other 2D materials as well.⁴⁴Fig. 4d shows the FESEM image of the f-Ti₃C₂T_x specimen. Loose and microscale NSs are still evident. The TEM images (Fig. 4e and f) illustrate the thin and electron-beam transparent but slightly restacked morphology of f-Ti₃C₂T_x, which proves the vdW interaction between the NSs. In addition, the flakes of f-Ti₃C₂T_x, as marked in Fig. 4f, are composed of only about 2–3 layered NSs, which can be verified from the HRTEM image (Fig. 4g), which authenticates the restacking few-layer structure of the f-Ti₃C₂T_x itself.

Fig. 4 (a) FESEM image, (b) digital photograph for the Tyndall effect of the s-Ti₃C₂T_x colloidal solution, (c) AFM image and line-scan height profiles for the s-Ti₃C₂T_x deposited on the Si/SiO₂ substrate, (d) FESEM, (e, f) TEM and (g) HRTEM images of the f-Ti₃C₂T_x. The images in panels (f) and (g) are taken from the orange rectangle in panels (e) and (f) as indicated.

Fig. 5a and b show the FESEM images with various magnifications for the H-4 sample. Delightfully, the H-4 sample exhibits a kinetically favorable 2D structure with a crumpled surface (Fig. 5a and b). Obviously, it is wholly different from the frozen product obtained using the refrigerator pre-freezing method (Fig. S3, ESI†), which shows a simple agglomerated structure. These distinct observations fully corroborate that the rapid pre-freezing process is critical in the efficient formation of 2D Co₃O₄@s-Ti₃C₂T_x hybrids. As is well known, the liquid nitrogen pre-freezing process shows a higher freezing rate than the common refrigerator pre-freezing avenue, and the slow freezing rate would allow water molecules to exclude foreign particles and eventually lead them to aggregation.⁴⁵ Therefore, if the dispersion was frozen at a slower freezing rate, such as in a refrigerator, the hybrids may have enough time to form strong aggregates. Delightfully, the lateral size of H-4 can reach over 50 μm, which is much larger than that of f-Ti₃C₂T_x flakes (Fig. 4c). This result indicates that the negatively charged s-Ti₃C₂T_x NSs are assembled with each other due to the induction of positively charged metal ions, and accordingly the 2D assembly of the Co₃O₄@s-Ti₃C₂T_x NSs is obtained. Notably, the extended 2D structure of hybrids is vitally important to provide sufficient and effective electron conduction channels for Co₃O₄ NCs during cycling. Interestingly, after a careful inspection into the TEM images (Fig. 5c and d) with various magnifications, no discernable particles could be found on the surface of the s-Ti₃C₂T_x NSs. However, the scanning TEM (STEM) image (Fig. 5e) and the corresponding EDS elemental (Ti and Co) mappings (Fig. 5f) evidently authenticate the uniform distribution of the element Co over the s-Ti₃C₂T_x substrate, and the atomic Co/Ti ratio is found to be ∼3.9 in the H-4 sample (Fig. S4b, ESI†). Further HRTEM observations (Fig. 5g and h) suggest that the high-quality Co₃O₄ NCs with an ultra-small size less than 5 nm (the inset of Fig. 5g) are uniformly dispersed on the flexible s-Ti₃C₂T_x matrix without obvious agglomeration. With the strong Co–O–Ti bonds, numerous separated Co₃O₄ NCs are strongly anchored on the surface of s-Ti₃C₂T_x, which in turn self-limits the growth of Co₃O₄ NCs. The lattice distance of 0.20 and 0.23 nm, respectively, corresponds to the (400) and (222) crystalline plane distance of Co₃O₄, as well demonstrated in Fig. 5e. In sharp contrast, only the aggregates with large particles are obtained in the absence of the s-Ti₃C₂T_x matrix (Fig. S5, ESI†).

Fig. 5 (a, b) FESEM, (c, d) TEM, (e) STEM, (f) the corresponding elemental (Ti Co) mapping images and (g, h) HRTEM of the H-4 specimen. The image in panel (d) is taken from the green rectangle in panel (c). The image in panel (h) is taken from the yellow rectangle in panel (g). The inset in panel (g) for the lateral size distribution of the Co₃O₄ NCs in the H-4 sample.

For comparison, the micro-morphologies of the H-2 and H-8 samples are also characterized, and shown in Fig. 6. As presented in Fig. 6a, similar to that of the H-4 sample, the H-2 sample exhibits a 2D open NS structure with an atomic Co/Ti ratio of ∼2.3 (Fig. S4a, ESI†), which is lower than H-4. With regard to the H-8 sample (Fig. 6b), the Co₃O₄ NCs seriously aggregate and overlap on the surface of the s-Ti₃C₂T_x NSs, which can be attributed to the limited surface of the s-Ti₃C₂T_x and the high loading (atomic Co/Ti ratio = ∼7.8, Fig. S4c, ESI†) of the Co₃O₄ in the hybrid. Furthermore, the TEM (Fig. 6c and d) and HRTEM (Fig. 6e and f) images clearly reveal that the size of the Co₃O₄ NCs becomes even larger and denser as the concentration of the cobalt salt gradually increases. Specifically, for the H-2 sample, the Co₃O₄ NCs are ultrafine with an average size less than 3 nm (the inset in Fig. 6c), isolated from each other (Fig. 6c and e), and uniformly dispersed on the s-Ti₃C₂T_x matrix (Fig. 6g). While for the case of H-8 (Fig. 6d and f), the average size of the Co₃O₄ NCs is even up to ∼9 nm (the inset in Fig. 6d), but were still well dispersed throughout the s-Ti₃C₂T_x NSs, as supported by the STEM and corresponding EDS elemental (Ti and Co) mapping images (Fig. 6h). In contrast, for the H-12 sample (Fig. S6, ESI†), namely the specimen with the highest loading of Co₃O₄, the Co₃O₄ NCs agglomerated severely and separated with the Co₃O₄@s-Ti₃C₂T_x hybrids due to the excessive introduction of Co₃O₄, which produces huge volume expansion in the charging and discharging processes.

Fig. 6 FESEM images of the (a) H-2 and (b) H-8 products. (c–f) HRTEM images of the (c, e) H-2 and (d, f) H-8. The images in panels (e, f) are taken from the yellow rectangles in panel (c, d) as indicated. STEM and the corresponding element (Ti and Co) mapping images of the (g) H-2 and (h) H-8 products. The insets in panels (c, d) for the size distribution of the Co₃O₄ NCs in the (c) H-2 and (d) H-8 as indicated.

3.2. Electrochemical evaluation

Detailed electrochemical evaluations were conducted at RT to investigate the suitability of Co₃O₄@s-Ti₃C₂T_x hybrids as an anode material for LIBs. Fig. 7a shows the initial three CV profiles (0.01 mV s⁻¹) of the H-4 electrode within the potential range from 0.01 to 3.0 V (vs. Li/Li⁺). The representative appearance of CV profiles generally corresponds to the co-contributions from the cobalt oxide and Ti₃C₂T_x itself.^5,10,46–52 Specifically, the distinct reduction peak at around 0.50 V in the first cathodic scan corresponds to the reduction of Co₃O₄ to metallic cobalt (see eqn (1)) and formation of a solid electrolyte interphase (SEI) film due to the electrolyte decomposition. The small tail located at 0.07 V is ascribed to the reaction of Li⁺ ions with Ti₃C₂T_x (see eqn (2)). There are two distinct anodic peaks in the oxidation process. The one, appearing between 1.07 and 1.80 V, is assigned to the reactions of Co and Li₂O, and the delithiation process of the s-Ti₃C₂T_x, and the other, situated near 2.11 V, is attributed to the transformation of Co into Co₃O₄, indicating the multi-step electrochemical reaction between Li₂O and Co. The cyclic voltammetry (CV) curves of the 3^rd and 5^th cycles are almost the same and well overlapped, confirming the desirable cycling stability and electrochemical reversibility of the H-4 anode. For the f-Ti₃C₂T_x (Fig S7a, ESI†), only a couple small peaks located at 0.07 V and 0.12 V, corresponding to the lithiation and delithiation reaction between Li⁺ and the surface active sites of the s-Ti₃C₂T_x,^53,54 are detected along with the embodied capacitive contribution. Furthermore, the low current responses in the CV curves of the f-Ti₃C₂T_x verify its modest Li-storage behavior. In contrast, the phase-pure Co₃O₄ sample (Fig S7b, ESI†) exhibits a series of strong and sharp peaks, corresponding to the typical redox process. Obviously, the CV profiles of the hybrid materials combine all the characteristic anodic/cathodic peaks belonging to the Ti₃C₂T_x and active Co₃O₄, indicating the combination of Co₃O₄ and Ti₃C₂T_x phases. The CV profiles of the H-2 (Fig. S7c, ESI†) and H-8 samples (Fig. S7d, ESI†) comparatively seem to be closely related to the Co₃O₄/s-Ti₃C₂T_x ratios in the hybrids. Typically, the H-8 sample shows similar CV curves to the Co₃O₄, owing to the high loading of electroactive Co₃O₄ NCs upon the surface of s-Ti₃C₂T_x NSs.

Co₃O₄ + 8Li⁺ + 8e⁻ ↔ 3Co + 4Li₂O (1)

Ti₃C₂T_x + xLi⁺ + xe⁻ ↔ Li_xTi₃C₂T_x (2)

Fig. 7 (a) The 1^st, 3^rd and 5^th CV curves (0.1 mV s⁻¹) and (b) galvanostatic charge–discharge plots (0.1 A g⁻¹) for the H-4 sample. (c) Rate performance of the H-2 H-4 H-8 and f-Ti₃C₂T_x anodes within the current range from 0.1 to 10 A g⁻¹. (d) Comparison of the H-4 and H-8 anodes in terms of rate properties with other recently reported cobalt-based/MXene-based anodes. (e) Long-term cycling performance for all specimens and CE data of the H-4 anode at 1 A g⁻¹ and (f) electrochemical stability of the H-4 anode at a high rate of 5 A g⁻¹.

Fig. 7b displays the typical galvanostatic charge–discharge curves of the H-4 anode at a current density of 100 mA g⁻¹. The H-4 anode exhibits a discharge capacity of ∼1168 mA h g⁻¹ and a charge capacity of ∼747 mA h g⁻¹ indicating an initial coulombic efficiency (CE) of ∼64%. As observed in Fig. 7b the discharge plateaus appearing at ∼0.9 V in the first cycle move up to ∼1.4 V in the subsequent cycles and the charge plateaus are always located at ∼1.3 and ∼2.0 V in all cycles, which is consistent with the starting position of redox peaks in the CV results. In comparison with the hybrid anode, the f-Ti₃C₂T_x and bare Co₃O₄ (Fig. S8 ESI†) show only the first discharge/charge capacities of ∼167/∼71.0 mA h g⁻¹ and ∼1267/∼924 mA h g⁻¹ corresponding to the initial CE values of ∼43% and ∼73% respectively. The larger irreversibility of f-Ti₃C₂T_x should arise from numerous functional groups such as –OH and –F on its surface. Compared to the pure Co₃O₄ in micron dimension, the lower initial CE of the hybrids can be reasonably attributed to two main reasons. One is the low CE of the Ti₃C₂T_x NSs and the other is the enhanced electrode/electrolyte surfaces/interfaces resulting from ultrafine Co₃O₄ NCs, which leads to the serious consumption of Li⁺ due to the formation of a SEI film.

The rate capabilities within a wide current range from 0.1 to 10 A g⁻¹ for the H-2, H-4, H-8 and f-Ti₃C₂T_x samples are also compared and collected in Fig. 7c. It should be noted that the f-Ti₃C₂T_x electrode delivers decreasing discharge specific capacities (DSCs) from ∼70 to ∼21 mA h g⁻¹ with the current density up to 10 from 0.1 A g⁻¹. Through a preliminary inspection it is easy to find that the DSCs increase with the increasing content of Co₃O₄ gradually at the initial current rates due to the high loading of the large-capacity Co₃O₄. Strikingly, the H-8 electrode achieves the highest capacity of ∼772 mA h g⁻¹ at 100 mA g⁻¹ but it declines quickly down to ∼235 mA h g⁻¹ equaling to that of the H-4 electrode at a high rate of 10 A g⁻¹. This should be related to the failure of efficient electronic conductive channels due to the coverage of Co₃O₄ with a larger size on the surface of the s-Ti₃C₂T_x NSs in the H-8 electrode. Because of the low weight ratio of Co₃O₄ to s-Ti₃C₂T_x the H-2 electrode exhibits minimum DSCs (i.e. ∼772 mA h g⁻¹ at 100 mA g⁻¹ and ∼87 mA h g⁻¹ at 10 A g⁻¹) in all hybrids. Encouragingly the specific capacities are estimated as ∼672, ∼629, ∼559, ∼486, ∼408, ∼297 and ∼223 mA h g⁻¹ for the H-4 anode at current rates of 0.1, 0.2, 0.5, 1.0, 2.0, 5.0 and 10.0 A g⁻¹, respectively, which are much larger than those of the f-Ti₃C₂T_x and H-2. Furthermore with the current density of 0.1 A g⁻¹, the capacity of the H-4 anode could recover to ∼682.2 mA h g⁻¹ indicating its superb rate capability. As summarized in Fig. 7d benefiting from the fast charge storage and electron transfer properties, the H-4 and H-8 electrodes exhibit appealing capacity retention with the current rate increasing, which is superior to many other cobalt oxide-based/MXene-based anodes reported previously especially at the high rates.^23,46–52 Therefore the s-Ti₃C₂T_x NSs can be highly expected as an ideal matrix for fabricating 2D hybrid anodes with superb electrochemical behavior for LIBs.

Fig. 7e shows the cycling performance of the H-2, H-4, H-8, pure Co₃O₄ and f-Ti₃C₂T_x samples over 700 consecutive cycles which were all conducted at a high current density of 1 A g⁻¹ after 4-cycle activation at 100 mA g⁻¹. The pure Co₃O₄ and H-8 anodes present much higher capacities than other anodes in the initial several cycles but suffer a dramatic decline as the charge/discharge cycles continue. This significant capacity degradation may result from the destruction of the structure caused by the strain during the electrochemical cycling. As noted the capacities of ∼420 and ∼550 mA h g⁻¹ can be achieved by the H-2 and H-4 electrodes after 700 cycles respectively which are much higher than that of f-Ti₃C₂T_x (∼44.9 mA h g⁻¹). It is worth mentioning that the cycling plots of the H-2 and H-4 anodes are almost as stable as the f-Ti₃C₂T_x sample and the CE value of the H-4 anode was close to 100% after the first few activation cycles until the 700^th cycle, indicating their ultrafast lithium storage and excellent long-term cycling stability. Remarkably the H-4 anode holds excellent capability to balance the high reversible capacity and stability exhibiting its superb lithium storage properties for practical application. What's more, even after evaluating under 5 A g⁻¹, the H-4 anode can still deliver a discharge capacity as large as ∼160 mA h g⁻¹ after the 1000^th cycle (Fig. 7f). The obviously improved rate performance and the cycling stability of H-4 could be ascribed to its desirable kinetic properties as evidenced by the EIS study (Fig. S9, ESI†). For each spectrum there is a depressed semicircle at a high frequency and a linear Warburg factor at a low frequency (Fig. S9a, ESI†). In the circuit model, R_s corresponds to the intercept with the x-axis which can reflect the conductivity of electrodes under the same test conditions. The size of the semicircle is attributed to the charge transfer resistance (R_ct) at the interface between the reaction sites and electrodes. The fitted results (Table S2, ESI†) indicate that the f-Ti₃C₂T_x electrode shows the lowest R_s (1.16 Ω) and R_ct (99.3 Ω) due to the excellent conductivity. Note that the much lower R_ct (130.0 Ω) and R_s (1.30 Ω) of the H-4 anode than the pure Co₃O₄ (R_ct of 238.2 Ω and R_s of 1.60 Ω) is beneficial for the enhancement of the charge transfer kinetics. Moreover the R_ct value is only 175.4 Ω even after 700 cycles at a current density of 1 A g⁻¹ which can be well responsible for the large reversible capacity, superb rate behavior and better stability of the H-4 anode. On the other hand the lithium diffusion kinetics can be investigated by plotting the real part of the complex impedance (Z′) vs. angular frequency (ω^−1/2) (Fig. S9b, ESI†). The slope of the fitted line represents the Warburg coefficient (δ_ω). The H-4 electrode shows a lower Warburg factor (δ = 80.6) than the Co₃O₄ electrode (δ = 114.4). It corresponds to a faster lithium diffusion rate in H-4 with respect to Co₃O₄ since the lithium diffusion coefficient (D_Li⁺) is inversely proportional to δ² with a relationship of D_Li⁺ = R²T²/2A²n⁴F⁴C²δ², where C (concentration of lithium ions), A (area of the electrode), F (Faraday constant), T (absolute temperature) and R (universal gas constant) are all constant under the same conditions and n (number of electrons per molecule in the reaction) has little effect on the overall value.¹ The higher D_Li⁺ implies that the layered structure of the H-4 electrode ensures fast transport of Li⁺ during cycling.

To further understand the charge-storage properties of the optimized H-4 anode, the corresponding kinetics analyses based on CV curves were carried out. Fig. 8a shows the CV curves at various sweep rates ranging from 0.4 to 2.4 mV s⁻¹. The contributions of the surface-induced pseudo-capacitive process (k₁v) and the diffusion-controlled insertion process (k₂v^0.5) of the H-4 electrode can be distinguished quantitatively by using the following formula: i (V) = k₁v + k₂v^0.5 = av^b (ref. 55 and 56), where i (V) and v are the response current at a fixed potential V and the specific scan rate respectively. k₁ and k₂ are constants and a and b are adjustable values. Generally the case of b = 0.5 indicates an ideal diffusion controlled intercalation process while the b value approaching 1.0 implies the surface-induced pseudo-capacitance controlled process. The b value can be obtained by the slope of the fitted line of log(i) and log(v). As illustrated in Fig. 8b, the b values for typical anodic and cathodic peaks are 0.67 and 0.74 respectively. The result verifies that the dual model charge storage mechanism is involved for the H-4 electrode. As calculated from the CV profile (Fig. 8c), the pseudo-capacitive contribution ratio is estimated as ∼68.1% at a scanning rate of 1.8 mV s⁻¹, indicative of a favored surface/subsurface-induced charge storage process of the hybrid.⁵⁷ Notably upon increasing the sweep rate the H-4 anode displays enlarged capacitive contributions of ∼44.3%, ∼56.7% and ∼79.4% at the scan rates of 0.4, 1.0 and 2.4 mV s⁻¹ respectively as plotted in Fig. 8d. As a result, the high-ratio surface induced pseudo-capacitive process would highly promote the rate performance of the H-4 electrode.

Fig. 8 Kinetic analysis of the electrochemical behavior of the H-4 electrode. (a) CV curves of H-4 at various scan rates. (b) Log(i) vs. log(v) plots for cathodic and anodic peaks and the slopes of the fitted lines. (c) CV profile for the H-4 electrode at a scan rate of 1.8 mV s⁻¹ and the capacitive contribution is highlighted by the green region. (d) Contribution ratio of the capacitive and diffusion-controlled capacities of the H-4 electrode at various scan rates.

All in all the enhanced electrochemical behaviors of the H-4 anode, especially high-rate behavior and long-span cycling stability, can be rationally ascribed to the following aspects. First the intrinsic Co–O–Ti interaction at the s-Ti₃C₂T_x/Co₃O₄ interfaces effectively limits the growth of the Co₃O₄ NCs and further prevents these electroactive NCs from shedding from the s-Ti₃C₂T_x surface during cycling, which in turn avoids the aggregation of ultrathin s-Ti₃C₂T_x NSs. Meanwhile the unique Co–O–Ti interfacial bonds allow convenient interfacial electron transfer between the Co₃O₄ and s-Ti₃C₂T_x. Second the 2D opened s-Ti₃C₂T_x as a matrix of hybrid materials not only promotes the electronic conductivity of the full electrode but also allows accessibility of the electroactive sites by Li⁺ ions in the electrolytes. Finally the Co₃O₄ NCs with a tiny size of ∼5 nm provide an extremely shortened diffusion pathway over charge storage, and the internal strain of the hybrid electrode over a continuous lithiation/de-lithiation process can be efficiently buffered with the synergistic contribution of the flexible s-Ti₃C₂T_x NSs. These results highlight the promising application of the Co₃O₄@s-Ti₃C₂T_x NS hybrids as high-performance anodes for next-generation LIBs.

4. Conclusions

In summary we developed for the first time an extremely facile method for the large-scale fabrication of 2D Co₃O₄@s-Ti₃C₂T_x hybrid NSs where ultrafine Co₃O₄ NCs are dispersed and anchored evenly upon the s-Ti₃C₂T_x surface through the Co–O–Ti interfacial bonds. The formation process of the Co₃O₄@s-Ti₃C₂T_x hybrids has been rationally clarified here. Benefiting from the appealing structural and compositional merits, the resultant H-4 anode exhibited remarkable Li-storage behavior even superior to many other MXene-based/cobalt oxide-based anodes in terms of rate capability and cycling stability. Impressively the H-4 electrode delivered a superb rate capability of ∼223 mA h g⁻¹ at a rate of 10 A g⁻¹ and maintained a reversible capacity of ∼550 mA h g⁻¹ after 700 cycles at 1 A g⁻¹. These results convincingly highlighted the promising application of the Co₃O₄@s-Ti₃C₂T_x NSs as advanced anodes for LIBs. More significantly the methodologies we devised here hold enormous promise in generalizing to other 2D metal oxides/s-Ti₃C₂T_x hybrids for versatile energy-related applications.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 51572005, 51772127), Major Program of Shandong Province Natural Science Foundation (No. ZR2018ZB0317), Taishan Scholars (No. ts201712050), Natural Science Doctoral Foundation of Shandong Province (ZR2019BEM038), Natural Science Doctoral Foundation of the University of Jinan (XBS1830), and Collaborative Innovation Center of Technology and Equipment for Biological Diagnosis and Therapy in Universities of Shandong.

References

1. S. Niu, Z. Wang, M. Yu, M. Yu, L. Xiu, S. Wang, X. Wu and J. Qiu, ACS Nano, 2018, 12, 3928–3937.
2. Y. Wu, P. Nie, L. Wu, H. Dou and X. G. Zhang, Chem. Eng. J., 2018, 334, 932–938.
3. L. Liang, W. Chen, X. Sun, S. Xuan, L. Hou, J. Sun and C. Yuan, Adv. Mater. Interfaces, 2017, 4, 1700382.
4. J. Meng, Z. Liu, C. Niu, L. Xu, X. Wang, Q. Li, X. Wei, W. Yang, L. Huang and L. Mai, Mater. Horiz., 2018, 5, 78–85.
5. X. Xu, S. Chen, C. Xiao, K. Xi, C. Guo, S. Guo, S. Ding, D. Yu and R. V. Kumar, ACS Appl. Mater. Interfaces, 2016, 8, 6004–6010.
6. H. Long, M. Zhang, Q. Wang, L. Xing, S. Wang and X. Xue, J. Alloys Compd., 2017, 701, 200–207.
7. L. Sun, Q. Deng, Y. Li, H. Mi, S. Wang, L. Deng, X. Ren and P. Zhang, Electrochim. Acta, 2017, 241, 252–260.
8. G. Liu and J. Shao, J. Mater. Chem. A, 2017, 5, 9801–9806.
9. K. Xie, P. Wu, Y. Zhou, Y. Ye, H. Wang, Y. Tang, Y. Zhou and T. Lu, ACS Appl. Mater. Interfaces, 2014, 6, 10602–10607.
10. W. Kang, Y. Zhang, L. Fan, L. Zhang, F. Dai, R. Wang and D. Sun, ACS Appl. Mater. Interfaces, 2017, 9, 10602–10609.
11. M. Xu, Q. Xia, J. Yue, X. Zhu, Q. Guo, J. Zhu and H. Xia, Adv. Funct. Mater., 2018, 29, 1807377.
12. M. Xu, S. Lei, J. Qi, Q. Dou, L. Liu, Y. Lu, Q. Huang, S. Shi and X. Yan, ACS Nano, 2018, 12, 3733–3740.
13. B. Ding, J. Wang, Y. Wang, Z. Chang, G. Pang, H. Dou and X. Zhang, Nanoscale, 2016, 8, 11136–11142.
14. W. Bao, L. Liu, C. Wang, S. Choi, D. Wang and G. Wang, Adv. Energy Mater., 2018, 8, 1702485.
15. A. Lipatov, H. Lu, M. Alhabeb, B. Anasori, A. Gruverman, Y. Gogotsi and A. Sinitskii, Sci. Adv., 2018, 4, eaat0491.
16. X. Sun, Y. Liu, J. Y. Zhang, L. R. Hou, J. F. Sun and C. Z. Yuan, Electrochim. Acta, 2019, 295, 237–245.
17. B. Anasori, M. R. Lukatskaya and Y. Gogotsi, Nat. Rev. Mater., 2017, 2, 16098.
18. D. Xiong, X. Li, Z. Bai and S. Lu, Small, 2018, 14, e1703419.
19. H. Wang, Y. Wu, X. Yuan, G. Zeng, J. Zhou, X. Wang and J. W. Chew, Adv. Mater., 2018, 30, e1704561.
20. Q. Tang, Z. Zhou and P. Shen, J. Am. Chem. Soc., 2012, 134, 16909–16916.
21. S. Sun, C. Liao, A. M. Hafez, H. Zhu and S. Wu, Chem. Eng. J., 2018, 338, 27–45.
22. X. T. Gao, Y. Xie, X. D. Zhu, K. N. Sun, X. M. Xie, Y. T. Liu, J. Y. Yu and B. Ding, Small, 2018, 14, e1802443.
23. M. Q. Zhao, M. Torelli, C. E. Ren, M. Ghidiu, Z. Ling, B. Anasori, M. W. Barsoum and Y. Gogotsi, Nano Energy, 2016, 30, 603–613.
24. C. J. Zhang, S. Pinilla, N. McEvoy, C. P. Cullen, B. Anasori, E. Long, S.-H. Park, A. Seral-Ascaso, A. Shmeliov, D. Krishnan, C. Morant, X. Liu, G. S. Duesberg, Y. Gogotsi and V. Nicolosi, Chem. Mater., 2017, 29, 4848–4856.
25. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
26. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
27. P. E. Blochl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 50, 17953–17979.
28. J. Klimeš, D. R. Bowler and A. Michaelides, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 83, 195131.
29. T. Sharifi, Y. Xie, X. Zhang, H. R. Barzegar, J. Lei, G. Coulter, S. Sun, C. Tiwary, A. Zettl, B. Yakobson and P. M. Ajayan, Carbon, 2019, 141, 266–273.
30. S. L. Dudarev, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 57, 1505–1509.
31. X. Wu, Z. Wang, M. Yu, L. Xiu and J. Qiu, Adv. Mater., 2017, 29, 1607017.
32. C. Chen, X. Xie, B. Anasori, A. Sarycheva, T. Makaryan, M. Zhao, P. Urbankowski, L. Miao and J. Jiang, Angew. Chem., Int. Ed., 2018, 57, 1846–1850.
33. X.-D. Zhu, Y. Xie and Y.-T. Liu, J. Mater. Chem. A, 2018, 6, 21255–21260.
34. Y. Jiang, X. Yan, W. Xiao, M. Tian, L. Gao, D. Qu and H. Tang, J. Alloys Compd., 2017, 710, 114–120.
35. X. Guo, X. Xie, S. Choi, Y. Zhao, H. Liu, C. Wang, S. Chang and G. Wang, J. Mater. Chem. A, 2017, 5, 12445–12452.
36. J. Halim, K. M. Cook, M. Naguib, P. Eklund, Y. Gogotsi, J. Rosen and M. W. Barsoum, Appl. Surf. Sci., 2016, 362, 406–417.
37. R. Meng, J. Huang, Y. Feng, L. Zu, C. Peng, L. Zheng, L. Zheng, Z. Chen, G. Liu, B. Chen, Y. Mi and J. Yang, Adv. Energy Mater., 2018, 8, 1801514.
38. G. Henkelman, A. Arnaldsson and H. Jónsson, Comput. Mater. Sci., 2006, 36, 354–360.
39. T. Hu, Z. Li, M. Hu, J. Wang, Q. Hu, Q. Li and X. Wang, J. Phys. Chem. C, 2017, 121, 19254–19261.
40. J. Sun, G. Zheng, H. W. Lee, N. Liu, H. Wang, H. Yao, W. Yang and Y. Cui, Nano Lett., 2014, 14, 4573–4580.
41. J. Liu, H. B. Zhang, R. Sun, Y. Liu, Z. Liu, A. Zhou and Z. Z. Yu, Adv. Mater., 2017, 29, 1702367.
42. M. Ghidiu, M. R. Lukatskaya, M. Q. Zhao, Y. Gogotsi and M. W. Barsoum, Nature, 2014, 516, 78–81.
43. X. Wang, X. Shen, Y. Gao, Z. Wang, R. Yu and L. Chen, J. Am. Chem. Soc., 2015, 137, 2715.
44. A. Lipatov, M. Alhabeb, M. R. Lukatskaya, A. Boson, Y. Gogotsi and A. Sinitskii, Adv. Electron. Mater., 2016, 2, 1600255.
45. J. Lee and Y. Cheng, J. Controlled Release, 2006, 111, 185–192.
46. S. Chen, Y. Zhao, B. Sun, Z. Ao, X. Xie, Y. Wei and G. Wang, ACS Appl. Mater. Interfaces, 2015, 7, 3306–3313.
47. D. Wang, Y. Yu, H. He, J. Wang, W. Zhou and H. D. Abruna, ACS Nano, 2015, 9, 1775–1781.
48. X. Yang, K. Fan, Y. Zhu, J. Shen, X. Jiang, P. Zhao, S. Luan and C. Li, ACS Appl. Mater. Interfaces, 2013, 5, 997–1002.
49. M. Liu, X. Deng, Y. Ma, W. Xie, X. Hou, Y. Fu and D. He, Adv. Mater. Interfaces, 2017, 4, 1700553.
50. M. Jing, M. Zhou, G. Li, Z. Chen, W. Xu, X. Chen and Z. Hou, ACS Appl. Mater. Interfaces, 2017, 9, 9662–9668.
51. H. Zhang, P. Zhang, W. Zheng, W. Tian, J. Chen, Y. Zhang and Z. Sun, Electrochim. Acta, 2018, 285, 94–102.
52. J. Xiong, L. Pan, H. Wang, F. Du, Y. Chen, J. Yang and C. Zhang, Electrochim. Acta, 2018, 268, 503–511.
53. B. Ahmed, D. H. Anjum, Y. Gogotsi and H. N. Alshareef, Nano Energy, 2017, 34, 249–256.
54. C. Yang, Y. Liu, X. Sun, Y. R. Zhang, L. R. Hou, Q. A. Zhang and C. Z. Yuan, Electrochim. Acta, 2018, 271, 165–172.
55. V. Augustyn, J. Come, M. A. Lowe, J. W. Kim, P. L. Taberna, S. H. Tolbert, H. D. Abruna, P. Simon and B. Dunn, Nat. Mater., 2013, 12, 518–522.
56. T. Brezesinski, J. Wang, J. Polleux, B. Dunn and S. H. Tolbert, J. Am. Chem. Soc., 2009, 131, 1802–1809.
57. D. Chao, C. Zhu, P. Yang, X. Xia, J. Liu, J. Wang, X. Fan, S. V. Savilov, J. Lin, H. J. Fan and Z. X. Shen, Nat. Commun., 2016, 7, 12122.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9nr04377b

---