Shuhao
Xiao‡
ab,
Ying
Zhu‡
b,
Ximeng
Liu‡
c,
Ruichuan
Zhang
b,
Jiaqian
Qin
d,
Haiyuan
Chen
b,
Xiaobin
Niu
b,
John
Wang
c,
Jinxia
Jiang
*a and
Jun Song
Chen
bef
aChongqing Medical and Pharmaceutical College, Chongqing 401331, China. E-mail: 2020364@cqmpc.edu.cn
bSchool of Materials and Energy, University of Electronic Science and Technology of China, Chengdu, 610054, China
cDepartment of Materials Science and Engineering, National University of Singapore, 117574, Singapore
dCenter of Excellence on Advanced Materials for Energy Storage, Metallurgy and Materials Science Research Institute, Chulalongkorn University, Bangkok 10330, Thailand
eInterdisciplinary Materials Research Center, Institute for Advanced Study, Chengdu University, Chengdu 610106, China
fShenzhen Institute for Advanced Study, University of Electronic Science and Technology of China, Shenzhen, China
First published on 8th October 2024
Constructing heterostructures has been considered a promising strategy to enhance the electrochemical performance of active materials for energy storage. However, the mechanism for boosting K storage by the heterointerfaces between two phases with different crystal structures is still unclear. Herein, we designed a heterostructure consisting of non-layered CoSe2 and layered VSe2 decorated on the surface of N-doped carbon nanofibers. Interestingly, the lattice mismatch between these two phases creates a large amount of lattice distortion at phase boundaries, which gives rise to significantly enhanced electrochemical performance in potassium-ion batteries with a high reversible capacity of ∼400 mA h g−1 for 200 cycles at 1 A g−1. Density functional theory calculations and molecular dynamics simulations suggest that K diffusion is facilitated at the interface with a more reduced energy barrier, thus achieving high reversible capacity and stable cycling performance.
To date, numerous materials have been explored as anode materials for PIBs, such as hard carbon,13–16 metal alloys,10,17 phosphorus,18,19 transition metal oxides and dichalcogenides.12,20–23 Among them, transition metal selenides are considered suitable anode materials for PIBs because of their high theoretical capacity, low cost and environmental friendliness.24–26 Compared with metal sulfides, most of the metal selenides show narrow band gaps. Furthermore, some of them (such as CoSe2) possess metal-like electron structures,27 offering enhanced electronic conductivity. Moreover, metal–Se bonds are weaker than metal–S bonds since the Se atom is much larger than the S atom, which would facilitate conversion reactions.28 However, the generally poor ionic diffusion during charge and discharge originating from the large atomic radius of K has always been a major obstacle to realizing superior electrochemical performance. In addition, low initial coulombic efficiency and inferior cycling stability are issues in the application of metal selenides.
To address these drawbacks, various strategies have been attempted, such as carbon coating,29,30 structure designing,31–33 and electrolyte engineering.34 Very recently, constructing heterostructures has attracted much attention owing to the resultant favorable properties, such as high electron mobility and charge spatial separation ability at the interface between the participating components.35–38 Indeed, it has been demonstrated that the electrochemical performance could be greatly improved compared to those of the single-phase counterparts.39 Inspired by this idea, different heterostructures such as SnS/SnO2,40 MnS/MoS2 (ref. 41) and Sb2S3/FeS2 (ref. 42) have been constructed. However, most of these previous studies were limited to components with different chemical compositions. Meanwhile, very few studies have focused on the heterointerface between two phases with different crystal structures. Generally, transition metal selenides can be categorized into layered type (such as VSe2) and non-layered type (such as CoSe2). Compared with the non-layered type, the layered transition metal selenides could provide K diffusion channels along their (001) planes, resulting in relatively low energy barriers for K transport.43 Therefore, an optimal combination between a non-layered type selenide with high electrical conductivity and a layered counterpart able to facilitate K diffusion kinetics would offer new opportunities at the interface.
In this work, we purposely fabricated a CoSe2/VSe2@N-doped carbon nanofiber (CoSe2/VSe2@NCNF) via electrospinning method. The N-doped carbon nanofibers (NCNF) constituted an electron-conducting network, which was decorated with VSe2 nanoplates attached to CoSe2 nanoparticles, giving rise to a heterointerface between the non-layered CoSe2 and layered VSe2. The lattice distortions between these two phases are shown to generate many defects at the interface, which would be beneficial for K storage. Such a hetero-nanocomposite structure displayed a superior rate capability of 325 mA h g−1 at 5 A g−1, and a high reversible capacity of 400 mA h g−1 at 1 A g−1 with 97.9% capacity retention even after 200 cycles when employed as the anode in PIBs, suggesting its overall excellent performance and cyclic stability. Moreover, we have established the mechanism for boosting K storage by the layered/non-layered heterostructure, by first-principles calculations and ab initio molecular dynamics simulations, which confirmed the reduced energy barrier for K diffusion at the interface and the enhanced electronic conductivity via electron transfer between the two phases.
From the high-resolution TEM (HRTEM) studies (Fig. 1e), the nanoparticles possessed lattice fringes with the interplanar spacing of 0.26 nm, corresponding to the (210) plane of orthorhombic CoSe2 (JCPDS: 53-0449). Meanwhile, the nanoflake showed an interplanar spacing of 0.20 nm, which could be assigned to the (003) plane of hexagonal VSe2 (JCPDS: 89-1641). The interface between these two phases was marked by the dashed lines with different colors in the inverse fast Fourier transformations (IFFTs; Fig. 1f). The lattice mismatch between the layered and non-layered phases leads to a significant lattice distortion at the boundary region, which could be beneficial for K storage.45 As shown in the element mapping (Fig. 1g), the C and N elements are distributed uniformly in the carbon nanofiber. Moreover, the Co and V elements are mainly located in the nanoparticles and nanoflakes, respectively, further suggesting the formation of the respective selenide phases that constitute the heterostructure.
The crystallographic information of the as-prepared samples was gathered by X-ray diffraction (XRD). As depicted in Fig. S2,† both control samples containing only Co or V were completely transformed into the corresponding selenides after selenization without any other impurity phases. Similarly, all of the diffraction peaks of CoSe2/VSe2@NCNF could be assigned to orthorhombic CoSe2 (JCPDS: 53-0449) and hexagonal VSe2 (JCPDS: 89-1641), revealing the coexistence of the CoSe2 and VSe2 phases (Fig. 2a). The metal content in the CoSe2/VSe2@NCNF composites was investigated by inductively coupled plasma-atomic emission spectrometry (ICP-AES), confirming that the atomic ratio of Co to V was about 1:
1. Subsequently, thermogravimetric (TG) analysis was conducted from room temperature to 600 °C at a heating rate of 10 °C min−1 in air (Fig. S3†). The sample displayed a small decrease in weight up to 150 °C, which could be attributed to the evaporation of moisture. After that, the metal selenides started to decompose and SeO2 began to form, leading to a ∼4% increase in weight. The rapid weight loss from 310 to 430 °C can be ascribed to the combustion of carbon and the volatilization of SeO2. Finally, the residue with a mass content of 27 wt% could be due to the oxide phases of Co3O4 and V2O5. In tandem with the ICP-AES results, the mass content of CoSe2 and VSe2 were calculated to be 34.6 wt% and 33.4 wt%, respectively, and the content of carbon in CoSe2/VSe2@NCNF was determined to be 32 wt%.
The specific surface area and the pore structure of the as-prepared samples were analyzed by Brunauer–Emmert–Teller (BET) method. As shown in Fig. S4a,† CoSe2@NCNF exhibits a type III adsorption–desorption isotherm, and possesses the lowest specific surface area and total pore volume of 20.6 m2 g−1 and 0.08 cm3 g−1, respectively. Meanwhile, both CoSe2/VSe2@NCNF and VSe2@NCNF display a typical IV adsorption isotherm, indicating the presence of micro- and mesopores in their structures. CoSe2/VSe2@NCNF exhibits a specific surface area of 32.0 m2 g−1 and a total pore volume of 0.15 cm3 g−1, which are comparable to that of VSe2@NCNF (36.4 m2 g−1 and 0.16 cm3 g−1). Additionally, the pore size distributions in the three samples (Fig. S4b†) reveal that CoSe2/VSe2@NCNF shows a relatively narrow pore size distribution at about 2.5–5 nm.
Furthermore, X-ray photoelectron spectroscopy (XPS) was utilized to determine the valence and chemical states of the samples. Two peaks at 284.8 and 286.1 eV can be identified in the high-resolution spectrum of C 1s, which can be attributed to the C–C and C–N bonding, respectively (Fig. 2b). The high-resolution spectrum of N 1s in CoSe2/VSe2@NCNF can be split into three peaks at the binding energies of 397.5, 399.2 and 401.3 eV, which correspond to pyridinic-N, pyrrolic-N and graphite-N (Fig. 2c),46 respectively. A comparison of the Co 2p and V 2p spectra among CoSe2/VSe2@NCNF, CoSe2/NCNF and VSe2/NCNF is displayed in Fig. 2d and e, respectively. The deconvoluted peaks at the binding energies of 797.2 and 780.3 eV in the Co 2p spectrum can be attributed to the Co 2p1/2 and 2p3/2 spin orbits of Co2+, respectively. Meanwhile, the peaks centered at 740.0 and 778.8 eV are ascribed to Co3+.47 The Co2+ state belongs to the CoSe2 phase, while the Co3+ state corresponds to the surface oxidation of CoSe2 in air.44 Moreover, the V 2p spectrum shows two peaks at 524.4 and 517.1 eV, corresponding to V 2p1/2 and 2p3/2, respectively.48 Notably, compared to CoSe2/NCNF and VSe2/NCNF, the peaks of the Co 2p spectra had shifted to lower binding energies in CoSe2/VSe2@NCNF, while the peaks in the V 2p spectra showed a shift in the opposite direction. Such shifts in binding energies indicated that the electronic density was lowered in Co2+, while it increased in V4+,49 revealing the spontaneous transfer of electrons from CoSe2 to VSe2. Such a phenomenon could be mainly attributed to the redistribution of charges at the interface that would optimize the electronic state of VSe2, giving rise to an improvement of the electrical conductivity, and thus accelerated reaction process.45
The local structures of CoSe2/VSe2@NCNF, CoSe2@NCNF and VSe2@NCNF were further investigated by X-ray absorption fine structure (XAFS) measurements. Compared to CoSe2@NCNF, the Co K-edge X-ray absorption near-edge structure (XANES) spectra of CoSe2/VSe2@NCNF exhibited a shift to higher energy (Fig. 2f). Meanwhile, a shift of the absorption edge to lower energy can be observed in the V K-edge XANES (Fig. 2h), indicating a transfer of electrons from CoSe2 to VSe2 in CoSe2/VSe2@NCNF,45 which was consistent with the XPS results. In addition, the coordination environment of the Co and V atoms in the three samples was investigated by Fourier-transformed extended X-ray absorption fine structure (EXAFS) spectra. The peak intensity of the Co–Se bonds in CoSe2/VSe2@NCNF was higher than that of CoSe2@NCNF (Fig. 2g), revealing that the Co species in CoSe2/VSe2@NCNF had higher coordination numbers.50 However, compared to VSe2@NCNF, a lower peak intensity of the V–Se bonds in CoSe2/VSe2@NCNF was identified, suggesting a lower coordination environment, which might be ascribed to the induced defects.51 The XAFS results further revealed that the charge transfer and lattice distortions were generated in the CoSe2/VSe2 interface, which were expected to improve the K storage capability.
To investigate the potassium storage properties of the three samples, coin-type half-cells with K metal as the counter and reference electrodes were first assembled and tested between 0.01 to 3 V (vs. K/K+). Fig. 3a shows the cyclic voltammetry (CV) curves of CoSe2/VSe2@NCNF with the first three cycles. In the first cathodic scan, the peak at 1.65 V, which showed an upshift to 1.91 V in the following scans, could be attributed to the intercalation of K+ ions into the interlayer spacing of VSe2 and the formation of KxVSe2 (eqn (1)). Meanwhile, the two peaks located at 1.01 and 0.42 V are correlated to multistep electrochemical reactions, including the conversion reaction of CoSe2 and VSe2 (eqn (3) and (4)), and the formation of the solid electrolyte interphase (SEI).52 In the second scan, these two anodic peaks split into three peaks, revealing an activation-stabilization process during the first cycle.53 Among them, the peaks located at 1.23 and 1.45 V are correlated to the intercalation of CoSe2, and the peak at 0.63 V corresponds to the conversion of VSe2 and CoSe2 to K2Se, V and Co (eqn (2), (5) and (6)). In the anodic scan, the three peaks at 1.60, 1.91 and 2.1 V are related to the de-potassiation process of CoSe2 and VSe2 (eqn (7)–(10)).47,48 Moreover, the CV curves almost overlapped in the following cycles, suggesting a good stability of CoSe2/VSe2@NCNF. Therefore, the overall reactions of K+ storage in CoSe2/VSe2@NCNF during the charge–discharge cycles can be expressed as follows:
Discharge:
VSe2 + xK + xe− → KxVSe2 | (1) |
CoSe2 + xK + xe− → KxCoSe2 | (2) |
KxCoSe2 + (2 − x)K+ + (2 − x)e− → CoSe + K2Se | (3) |
KxVSe2 + (2 − x)K+ + (2 − x)e− → VSe + K2Se | (4) |
CoSe + 2K+ + 2e− → Co + K2Se | (5) |
VSe + 2K+ + 2e− → V + K2Se | (6) |
Charge:
V + K2Se → VSe + 2K+ + 2e− | (7) |
Co + 2K2Se → KxCoSe2 + (4 − x)K+ + (4 − x)e− | (8) |
KxCoSe2 → CoSe2 + xK+ + xe− | (9) |
VSe + K2Se → VSe2 + 2K+ + 2e− | (10) |
The CV curves of CoSe2/NCNF and VSe2/NCNF (Fig. S5†) display similar patterns as those of CoSe2/VSe2@NCNF.
The galvanostatic charge–discharge profiles of CoSe2/VSe2@NCNF (Fig. 3b) show voltage plateaus that are consistent with the CV results. The initial discharge and charge capacities are 550 and 483 mA h g−1, respectively, corresponding to a high initial coulombic efficiency (ICE) of 87.8%. Remarkably, in the following cycles, no observable difference could be identified and a stable capacity of about 425 mA h g−1 was observed, indicating the high reversibility of the electrochemical reactions. The rate performance of these three samples was studied at various current densities, ranging from 0.1 to 5 A g−1 (Fig. 3c). CoSe2/VSe2@NCNF showed the highest capacities of 583, 434, 412, 373, 352 and 325 mA h g−1 at 0.5, 1, 2, 3, 4 and 5 A g−1, respectively. Even at a high current density of 5 A g−1, it could still deliver a high reversible capacity of 325 mA h g−1. In contrast, CoSe2/NCNF and VSe2/NCNF exhibited significantly lower capacities at these current densities. Notably, CoSe2/NCNF displayed a rapid capacity fading in the first few cycles in these tests, which could well be due to the low ionic conductivity and poor structural stability of CoSe2. Such a deterioration in capacity was greatly mitigated in CoSe2/VSe2@NCNF, revealing that the heterostructure between CoSe2 and layered VSe2 with lattice distortions facilitated K insertion/deinsertion and improved the overall structural stability, thus leading to a more stable cycling performance. The long-term cycling tests of CoSe2/VSe2@NCNF at the current densities of 0.2 and 1 A g−1 are shown in Fig. 3d and e, respectively. It delivered a high reversible capacity of 489 mA h g−1 at 0.2 A g−1 after 50 cycles, 404 mA h g−1 at 1 A g−1 after 200 cycles, and a high reversible capacity of 325 mA h g−1 at 5 A g−1 during the rate performance test, indicating the outstanding cycling stability. The comparison of the electrochemical performance between CoSe2/VSe2@NCNF and other reported anode materials of PIBs, especially metal selenides and heterostructures, is presented in Table S2,† illustrating the superiority of our sample.
The dynamic processes in the three samples were further investigated by impedance measurement. The Nyquist plots of the three samples display comparable profiles, which consist of a depressed semi-circle and an up-sloping line (Fig. S6a†), corresponding to the charge transfer resistance (Rct) and the Warburg diffusion process (Zω),54 respectively. After fitting, the Rct of CoSe2/VSe2@NCNF is 4.91 Ω (Table S1†), which is much lower than those of CoSe2@NCNF (6.77 Ω) and VSe2@NCNF (9.52 Ω). Additionally, the linear fits of the real part of the impedance and low frequencies are shown in Fig. S6b.† The slopes of CoSe2/VSe2@NCNF, CoSe2@NCNF and VSe2@NCNF are 2.8, 16.16 and 23.49 Ω s−1/2, respectively, which could be used to calculate the apparent K+ diffusion coefficient based on the following equation:42,55
![]() | (11) |
To further understand the effects of such a non-layered/layered heterostructure on the potassium storage, density functional theory (DFT) calculations were performed. The (010) plane of CoSe2 and (001) plane of VSe2 were chosen to construct the CoSe2–VSe2 heterostructure. The diffusion barriers of the K ions were first investigated by the NEB method, and the migration path of K at the CoSe2–VSe2 interface is displayed in Fig. 4a. The K diffusion showed an energy barrier of 0.02 eV at the CoSe2–VSe2 interface (Fig. 4b), which is much lower than that at the (001) plane of VSe2 (0.17 eV) and (010) plane of CoSe2 (0.16 eV), suggesting that K diffuses faster at the CoSe2–VSe2 interface. Meanwhile, the calculated density of states (DOS) of the CoSe2–VSe2 heterostructure, CoSe2, and VSe2 are shown in Fig. 4c. Apparently, VSe2 possessed a significant band gap (0.3 eV) near the Fermi level, indicating a low intrinsic electrical conductivity. After forming the heterostructure with CoSe2, the band gap disappeared and the electron density near the Fermi level was greatly enhanced, suggesting that the electrical conductivity had been improved in the heterostructure.56–58 From the partial density of states (PDOS), the V d orbital showed a downshift, while the Co d orbital displayed an upshift. This could be attributed to the electron transfer between CoSe2 and VSe2, which was further confirmed by the charge density differences of the CoSe2–VSe2 heterostructure. The DFT results revealed that constructing heterostructures with layered VSe2 and non-layered CoSe2 could not only enhance the electronic conductivity, but also accelerate the diffusion of K+, contributing to the optimized electrochemical properties.
![]() | ||
Fig. 4 (a) K migration path along the CoSe2–VSe2 interface. (b) Migration energy barriers in K migrations, and (c) calculated DOS of different materials. |
To further study the diffusion of the K+ ions, we conducted ab initio molecular dynamics (AIMD) simulations on three samples within the temperature range of 1000 K to 2400 K, as illustrated in Fig. 5a–c. At these temperatures, a distinct K+ ion migration phenomenon was observed by measuring the mean square displacement (MSD) of the K+ ion relative to their initial positions over time. As the temperature increased, the diffusion tendency of the K+ ion gradually intensified. To determine the diffusion coefficient (D), each curve was fit over the time range from 0 to 8 ps, yielding the corresponding slope with the following equation:59,60
MSD(t) = 〈r2(t)〉 = 〈|ri(t) − ri(0)|2〉 | (13) |
![]() | (14) |
![]() | ||
Fig. 5 MSD of FPMD simulations for (a) CoSe2–VSe2, (b) CoSe2, and (c) VSe2 at 1000, 1200, 1800, and 2400 K. (d) Arrhenius plot of diffusion coefficients. |
Furthermore, an investigation was conducted on the radial distribution functions (g(r)) of K–Se, K–V, and K–Co atomic pairs at 1000 K, building a quantitative assessment of the microscale distribution of K+. As illustrated in Fig. S7a,† K ions were coordinated with Se atoms in CoSe2–VSe2 at an average distance of 3.6 Å, exceeding the coordination distances observed in VSe2 (3.0 Å) and CoSe2 (3.2 Å). It was observed that the binding energy of K–Se in the heterostructure decreases, facilitating K+ ion transport and thereby enhancing the dynamic performance of CoSe2–VSe2.62 As shown in Fig. S7b,† the g(r) of the K–V atomic pairs in both CoSe2–VSe2 and VSe2 were presented. In comparison to VSe2, K+ ions in CoSe2–VSe2 may have been distributed more uniformly, displaying reduced aggregation.63 Simultaneously, in both the heterostructure of CoSe2–VSe2 and the pure phase of CoSe2, the distribution trend of the K–Co atomic pairs was observed to be similar to that of K–V (Fig. S7c†). This microscopic distribution difference provided evidence that the construction of heterostructures contributed to the enhancement of the electrochemical performance.
Footnotes |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ta06049k |
‡ These authors contributed equally to this work. |
This journal is © The Royal Society of Chemistry 2024 |