Qifei
Li‡
a,
Xiangxiang
Ye‡
a,
Siling
Cheng
a,
Hong
Yu
b,
Weiling
Liu
c,
Cheng-Feng
Du
*b and
Xianhong
Rui
*a
aSchool of Materials and Energy, Guangdong University of Technology, Guangzhou 510006, China. E-mail: xhrui@gdut.edu.cn
bCenter of Advanced Lubrication and Seal Materials, Northwestern Polytechnical University, 127 West Youyi Road, Xi’an 710072, P. R. China. E-mail: cfdu@nwpu.edu.cn
cSchool of Materials Science and Engineering, Nanyang Technological University, Singapore, 639798, Singapore
First published on 12th November 2021
Given the abundant natural reserves of potassium and its similar properties to lithium, potassium-ion batteries (KIBs) are expected to be an emerging energy storage system to replace lithium-ion batteries (LIBs). Here, to overcome the sluggish diffusion kinetics and structure collapse of the anode in current KIBs, copper tetrathiovanadate (Cu3VS4) nanoparticles uniformly loaded in carbon nanofibers (CVS/CNF) are designed as a new anode for KIBs through electrospinning and subsequent carbonization/sulfidation treatment. Benefiting from the highly conductive Cu species and CNF network, an outstanding charge transfer kinetics is achieved in CVS/CNF composites. Meanwhile, the nanosized CVS particles and robust CNF networks are believed to be beneficial for enhanced structural stability during potassium insertion, and thus endow the composite anode with an excellent long-term capability. The CVS/CNF composite anode displays a high capacity (486 mA h g−1 at 0.1 A g−1), great rate feature (162 mA h g−1 at 6 A g−1) and outstanding cycling performance (193 mA h g−1 after 2000 cycles at 1 A g−1), which are superior to those of most of the reported transition metal sulfide anodes.
Recently, a variety of anode materials have been reported in KIBs, mainly including carbon-based materials (e.g., graphene, hard carbon, porous carbon, etc.),13–16 metallic materials (e.g., Sb, Sn, Bi, etc.),17–19 transition metal compounds (e.g., oxides, phosphides, chalcogenides, etc.),20–25 MXene derivatives26–28 and so on. For example, Lian et al. firstly fabricated alkalized Ti3C2 MXene nanoribbons possessing expanded interlayer spacing, with three-dimensional porous frameworks, which obtained the capacities of 136 and 78 mA h g−1 at 20 and 200 mA g−1 as anode materials for KIBs, respectively.27 Dong et al. reported the transformation of accordion-like Ti3C2 MXene into ultrathin nanoribbons of K2Ti4O9 as a high-performance anode material for KIBs, which achieved capacities of 151 and 88 mA h g−1 at 50 and 300 mA g−1, respectively.28 Compared with them, transition metal chalcogenides (TMCs) possess a large theoretical specific capacity and moderate volumetric change upon the charge/discharge process. Vanadium sulfides (such as V5S8 and VS2), as a member of the TMC family, not only carry the common character, but also have been widely explored as anode materials for KIBs because of their multiple valence states (+2 to +5) and rich spatial structures.22–24 For instance, Li et al. reported a honeycomb-like V5S8@C as the anode for KIBs, which achieved an acceptable K+ ion storage property of 351.9 mA h g−1 at 0.1 A g−1.22 Chen et al. reported a nanoflower-like VS2, showing a decent capacity of 383 mA h g−1 at 0.1 A g−1.24 Xie et al. synthesized ultra-enlarged interlayer carbon-containing VS2, which exhibited an enhanced electrochemical performance (439.2 mA h g−1 at 0.1 A g−1) for KIBs.23 Although these achievements have been made, there is still a long way to go before reaching a satisfactory rate performance because of the poor intrinsic conductivity and volume expansion during charge/discharge. Up to now, the combination of vanadium sulfides with a highly electronic conductive carbonaceous substrate, such as carbon nanofibers (CNFs), has been proved to be an effective route to enhance the electrochemical properties of vanadium sulfide materials.29 Meanwhile, atomic doping has also shown great potential towards enhanced conductivity and reaction kinetics for energy storage devices in recent years.30–32 Remarkably, with an ordered atom distribution, the bimetallic compounds provide a buffering substrate to ease volume expansion as well. Due to the introduction of Cu, the electrical conductivity and structural stability of some other metal sulfides can be significantly improved, such as Cu2SnS3 and Cu3SnS4.33,34
Inspired by the above views, using electrospinning and subsequent one-step carbonization/sulfidation treatment, a novel composite of bimetallic sulfide and carbon nanofibers, i.e., copper tetrathiovanadate nanoparticles (Cu3VS4) uniformly dispersed in carbon nanofibers (denoted as CVS/CNF), was designed and synthesized as the anode for KIBs. Derived from the good electrical conductivity of Cu and the CNF network, the CVS/CNF with high Cu content has the characteristic of high electrical conductivity and fast charge transfer. Additionally, the Cu atoms accompanied by CNFs can act as buffer matrixes for easing volume expansion during cycling, which further contributes to the excellent stability. Supported by all these merits, the CVS/CNF composite anode shows an outstanding K+ ion storage performance in terms of a large K+ ion storage capacity (486 mA h g−1 at 0.1 A g−1) and impressive rate property (162 mA h g−1 at 6 A g−1). Remarkably, the CVS/CNF composite anode delivered an ultra-stable long-term cycling capability (193 mA h g−1 after 2000 cycles at a relatively high rate of 1 A g−1), which is better than those of most of the reported transition metal sulfide KIB anodes.
Fig. 1 (a) XRD pattern of CVS/CNF. High resolution XPS spectra of (b) Cu 2p, (c) V 2p, (d) S 2p, (e) C 1s and (f) N 1s. |
Meanwhile, to examine the chemical compositions and valence state of the elements in all samples, X-ray photoelectron spectroscopy (XPS) techniques were performed (Fig. 1b–f and Fig. S2–S4, ESI†). From the full XPS spectrum (Fig. S3, ESI†), Cu, V, S, C, N and O elements are present in the CVS/CNF sample, in which the O element comes from the surface oxidation and adsorption. As for the other elements, the high-resolution spectrum of Cu 2p can be decomposed into two sets of peaks, as given in Fig. 1b. The two peaks located at 932.7 and 952.4 eV belong to the characteristic peaks of 2p3/2 and 2p1/2 for Cu+, respectively, while the other two peaks at 933.5 and 953.7 eV are appointed to Cu2+.36 The V 2p core-level spectrum reveals three states of V species, as shown in Fig. 1c. The two peaks at 517.3 and 516.4 eV correspond to the 2p3/2 of V4+ and V3+, respectively. Moreover, the two obvious peaks at around 514.2 and 521.7 eV are identified as the V–C bond, which is associated with bonding between Cu3VS4 and carbon nanofibers.37 The S 2p high-resolution spectrum is divided into three sets of characteristic peaks, representing three types of S elements of S2− (161.7 and 163.0 eV), C–S (163.8 and 165.0 eV), and SOx (168.4 and 169.5 eV), as shown in Fig. 1d, which are derived from the bonds between S and metal, S doped carbon nanofibers and surface oxidation, respectively.38 As for the C 1s spectrum (Fig. 1e), it consists of three peaks, i.e., 284.6 eV for C–C, 285.6 eV for C–N and C–S, and 288.4 eV for CO, where the formation of the bonds of C–N and C–S may result from that N element in PVP and excess S element in CS2 gas are doped into CNFs during the high-temperature carbonization/sulfidation process.39 The high-resolution spectrum of the N 1s region was also analysed (Fig. 1f), and it can be divided into three fitted peaks at 398.4, 400.5 and 401.5 eV, which correspond to three types of N elements, i.e., pyridinic, pyrrolic, and graphitic, respectively. It has been proved that pyridinic N and pyrrolic N can deliver more abundant vacancies, resulting in more diffusion channels and defects for K+ ions.40,41 In contrast, the XPS spectra of CS/CNF are given in Fig. S4 (ESI†), which show similar characteristic spectra of Cu 2p, S 2p and C 1s with CVS/CNF. Moreover, the XPS spectra of VS/CNF are also shown in Fig. S5 (ESI†), in which only V, S, C and N elements are detected, and the V, S, and C elements show a similar state to CVS/CNF. The S2− and C–S peaks are detected at 161.1/162.1 and 163.8/165.1 eV (Fig. S3c, ESI†), respectively, demonstrating the successful synthesis of VS species on CNFs.
The scanning electron microscope (SEM) and transmission electron microscope (TEM) technologies were used to investigate the morphologies and microstructures of the as-prepared CVS/CNF sample (Fig. 2). As exhibited in the low-magnification SEM images (Fig. 2a and b), the CVS/CNF exhibits a uniform 1D nanofiber structure with lengths of about 10–20 μm. Magnified SEM (Fig. 2c) and TEM (Fig. 2d) images reveal that the diameter of the nanofibers is about 250 nm. CVS nanoparticles with a size of dozens of nanometers are uniformly distributed in the CNFs. Meanwhile, the morphologies of both VS/CNF and CS/CNF composites also show a similar 1D nanofiber appearance (Fig. S6, ESI†). As for the high-resolution TEM (HRTEM) images (Fig. 2e and f), two groups of lattice fringes with an interplanar distance of 0.56 nm and 0.32 nm are observed from the nanoparticles, which are coincident with the (100) and (111) planes of Cu3VS4, respectively. In the selected area electron diffraction (SAED) image, as displayed in Fig. 2g, the obvious diffraction rings of the (111) and (220) faces can be observed, proving the polycrystalline nature of Cu3VS4. Moreover, the scanning TEM (STEM) image and the corresponding elemental mapping of a nanofiber are depicted in Fig. 2h. In the regions of the nanoparticles, the dominant elements are Cu, V, and S. As for the other regions of nanofibers, S, C and N dominate, further clarifying that the CNFs are successfully doped with S and N elements.
Fig. 2 (a–c) SEM, (d) TEM, (e and f) HRTEM pictures, (g) SAED pattern and (h) STEM image and corresponding elemental mapping images of CVS/CNF. |
To evaluate the K storage properties of the CVS/CNF anode, the coin-type batteries were packed using K foil and 3 M KFSI in DME as the counter electrode and electrolyte, respectively. Fig. 3a shows the cyclic voltammetry (CV) curves under a relatively low sweep rate of 0.1 mV s−1. During the first reduction process, a broad reduction peak located at 1.1 V is detected and vanishes in the subsequent cycles, which can be described as the generation of a solid electrolyte interphase (SEI) membrane and irreversible K+ ion insertion.42,43 In the next three cycles, the CV curves well overlap, indicating the strong reversibility and cycling stability of the CVS/CNF anode. Fig. 3b exhibits the galvanostatic charging-discharging profiles of the CVS/CNF anode under 0.1 A g−1 for different cycles. During the first cycle, the discharge/charge capacities of 906/459 mA h g−1 are delivered, corresponding to a relatively low coulomb efficiency (CE) of 50.7%, which can be explained by the large consumption of K+ ions by SEI membrane generation and irreversible reactions.43 In the subsequent several cycles, owing to the stepwise stabilization of the SEI membrane and the gradual activation of the CVS/CNF anode, the charge/discharge profiles change slightly and stabilize gradually, and basically follow the same trail after the first 15 cycles, indicating the excellent reversibility and structural stability of the CVS/CNF anode. As presented in Fig. 3c, the CVS/CNF anode delivers a high K+ storage capacity of 430 mA h g−1 after 120 cycles, outdistancing 364 mA h g−1 for VS/CNF and 275 mA h g−1 for CS/CNF under the same conditions. Moreover, the specific capacity delivered in the CVS/CNF anode is also higher than that of recently reported bimetallic sulfides, such as FeCoS2@rGO (372 mA h g−1 at 0.1 A g−1),44 NiCo2S4 (264 mA h g g−1 at 0.1 A g−1),30 and Ni–Fe sulfide (297 mA h g−1 at 0.1 A g−1),45 showing great potential for K+ ion storage.
The rate performance of the CVS/CNF composite was also investigated, as displayed in Fig. 3d. The reversible capacities of 435, 427, 390, 368, 353, 307, 275, 241 and 179 mA h g−1 are delivered at various rates of 0.1, 0.2, 0.5, 0.8, 1, 2, 3, 4 and 5 A g−1, respectively. It is noteworthy that a large K+ ion storage capacity of 162 mA h g−1 is maintained even at a high charging-discharging rate of 6 A g−1, and the capacity can recover to 463 mA h g−1 when the rate decreases to 0.1 A g−1, reflecting the excellent rate adaptability and structural stability of CVS/CNF. As a comparison, VS/CNF shows only a tiny capacity of 13 mA h g−1 left at 6 A g−1. Even worse, the CS/CNF anode is no longer able to charge/discharge properly when the rate exceeds 3 A g−1. The long-term cycling performance of CVS/CNF electrodes was also assessed at a relatively large current density of 1 A g−1. As displayed in Fig. 3e, the rising reversible capacities in the initial several laps can be ascribed to the hysteretic activation behaviours under high current density.46,47 After 2000 cycles, the CVS/CNF anode can still maintain a large specific capacity of 193 mA h g−1 with a stable CE of nearly 100%. In contrast, VS/CNF and CS/CNF anodes exhibit poor cycling stability, in which the former can only retain a low capacity of 142 mA h g−1 upon 500 cycles, and CS/CNF was even damaged in 300 cycles.
The high capacity and superior rate performance of CVS/CNF anodes arouse our interest. To investigate the electrochemical reaction kinetics of K+ ions in a CVS/CNF anode, the CV technique was performed during the sweep rates of 0.2–1.0 mV s−1 (Fig. 4a). As the scanning rate increased, the redox peaks shifted slightly, but the general shapes of the CV curves are maintained. Theoretically, the peak current (i) follows a power-law relationship with a sweep rate (ν), as follows: i = aνb.48 The b-value is related to the K+ ion storage behaviour, where b = 1.0 represents the capacitive process, and b = 0.5 stands for the diffusion behavior. As shown in Fig. 4b, the b values, calculated from the marked peaks in Fig. 4a, are 0.88 and 0.94, respectively, which illustrates that the main capacity is contributed by the capacitive process. The capacitive contribution of the electrode material during the quantified chemical reaction is then calculated based on the relationship: i = k1ν + k2ν1/2, in which the K+ ion storage behaviour can be divided into two types, i.e., capacitive control behaviour (k1ν) and diffusion control behavior (k2ν1/2).49 At a sweeping rate of 0.2 mV s−1, 65.5% of the capacity is contributed by the capacitive behaviour (Fig. 4c). Furthermore, the capacitive contributions grew from 65.5% to 80.8% as the sweep speed increased (Fig. 4d). Such a large proportion of capacitive-controlled capacity might be in charge with the fast reaction kinetics and excellent high-rate charging/discharging property.
Moreover, electrochemical impedance spectroscopy (EIS) technologies were also used to further explore the charge transfer and reaction kinetics of K+ ions in the CVS/CNF anode. In the Nyquist diagram, the charge transfer resistance (Rct) at the interface between the electrode and electrolyte is represented by a semicircle formed in the high-frequency range, and the K+ ion diffusion resistance (Warburg resistance, W) is represented by a sloping line formed in the low-frequency range. As exhibited in Fig. 4e, the value of Rct for the fresh KIBs assembled with CVS/CNF is relatively large (up to 1411 Ω), and surprisingly drops to 639 Ω after the initial fully charging step, indicating the rapid generation of the fresh SEI membrane on the surface of the negative electrode.35 Furthermore, the cell presents a reduced Rct value of 523 Ω after 10 cycles, indicating the gradual stabilization of the SEI film and CVS/CNF anode.44Fig. 4f shows the Nyquist diagrams of CVS/CNF, VS/CNF, and CS/CNF anodes after 10 cycles under the same conditions. The Rct value of CVS/CNF (523 Ω) is much smaller than that of VS/CNF (1069 Ω) and CS/CNF (650 Ω), indicating the faster charge transfer of CVS/CNF than that of VS/CNF and CS/CNF. In addition, the apparent diffusion coefficients of the K+ ions (DK+) of all samples were analysed and determined using the following formula: DK+ = 0.5 R2T2/S2n4F4C2σ2.36 Here, R, T, F, S, n, C, and σ stand for the gas constant, the absolute temperature, the Faraday constant, the electrochemical reaction area, the number of charges transferred, the K+ ion concentration and the Warburg coefficient, respectively. Moreover, the σ is determined by the formula: Z′ = R + σω−1/2,50 where ω is associated with the low-frequency region, R is a kinetic parameter independent of frequency, which is numerically equal to Rct in the Nyquist plot. Based on the Z′–ω−1/2 curves in Fig. S7 (ESI†), the K+ ion diffusion coefficients are calculated to be 1.82 × 10−19, 7.68 × 10−21, and 4.61 × 10−20 cm2 s−1 for the CVS/CNF, VS/CNF and CS/CNF anodes, respectively, which further confirms the rapid diffusion kinetics of the K+ ions in the CVS/CNF anode.
In addition, to explore the K-storage mechanism of the CVS/CNF anode, the ex situ XRD and XPS under full-discharge/charge states were carried out. As shown in the results of ex situ XRD (Fig. S8, ESI†), only one diffraction peak is observed at the full discharge state, corresponding to the VSx phase. During the subsequent charging process, instead of returning to its original state, the electrode material forms a mixture of CuSy and VSx. The XRD patterns of the electrode during the 5th cycle also show the same transformation, implying the high reversibility of the CVS/CNF anode. Therefore, during the first K+ ion insertion process, we believe that the CVS/CNF anode converts to the composite of VSx and CuSy (i.e., Cu3VS4 → VSx + CuSy). After that, the VSx phase undergoes an intercalation reaction (i.e., VSx + me− + mK+ ↔ KmVSx) and the CuSy phase undergoes a conversion reaction (i.e., CuSy + nye− + nyK+ ↔ Cu + yKnS), which are also confirmed by ex situ XPS. As shown in Fig. S9 (ESI†), the strong intensity of the K element at full discharge demonstrates the insertion of large amounts of K+ ions. Subsequently, the weak signal of the K element at full charge mainly comes from the SEI film and implies the reversible extraction of K+ ions. For Cu spectra (Fig. S10, ESI†), the copper metal is detected at the full discharge state, demonstrating the conversion reaction from CuSy to Cu and KnS. Additionally, the Cu2+ peaks are mainly due to the incomplete conversion reaction. And at the subsequent full charge state, the copper metal returns to the Cu2+ state, implying the reversible conversion reaction. For V spectra (Fig. S11, ESI†), V3+ and V4+ are dominant at full discharge and charge states, respectively, and no vanadium metal is detected, denying the conversion reaction and indicating the insertion reaction of K+ in the VSx phase.
Footnotes |
† Electronic supplementary information (ESI) available: Schematic crystal structures, XRD patterns, XPS spectra, SEM images, TEM images, Z′–ω−1/2 curve, in situ XRD and in situ XPS. See DOI: 10.1039/d1qm01096d |
‡ Q. Li and X. Ye contributed equally to this work. |
This journal is © the Partner Organisations 2022 |