High-Performance Cathode Material of FeF3·0.33H2O Modified with Carbon Nanotubes and Graphene for Lithium-Ion Batteries

The FeF3·0.33H2O cathode material can exhibit a high capacity and high energy density through transfer of multiple electrons in the conversion reaction and has attracted great attention from researchers. However, the low conductivity of FeF3·0.33H2O greatly restricts its application. Generally, carbon nanotubes (CNTs) and graphene can be used as conductive networks to improve the conductivities of active materials. In this work, the FeF3·0.33H2O cathode material was synthesized via a liquid-phase method, and the FeF3·0.33H2O/CNT + graphene nanocomposite was successfully fabricated by introduction of CNTs and graphene conductive networks. The electrochemical results illustrate that FeF3·0.33H2O/CNT + graphene nanocomposite delivers a high discharge capacity of 234.2 mAh g−1 in the voltage range of 1.8–4.5 V (vs. Li+/Li) at 0.1 C rate, exhibits a prominent cycling performance (193.1 mAh g−1 after 50 cycles at 0.2 C rate), and rate capability (140.4 mAh g−1 at 5 C rate). Therefore, the electronic conductivity and electrochemical performance of the FeF3·0.33H2O cathode material modified with CNTs and graphene composite conductive network can be effectively improved.


Introduction
Rechargeable lithium-ion batteries (LIBs) are the most effective power storage systems for portable electronic devices and considered as promising candidates for electric vehicles (EVs) and hybrid electric vehicles (HEVs) [1]. Compared with traditional fossil energy, LIBs are renewable and clean energy and friendly to the environment. Recently, with the rapid development of LIBs technology, demands for both energy and power density have continuously increased. One key challenge is developing high-performance electrode active materials, and the cathode material is a vital factor for improving the electrochemical properties of LIBs, including the specific capacity, cycling capability, rate capability, etc. [2,3]. Commercialized cathode materials, such as LiCoO 2 [4], LiMn 2 O 4 [5], and LiFePO 4 [6], suffer from low theoretical capacities due to the intercalation reaction involving only a single electron reaction, which cannot satisfy the demands of EVs. In the past several years, multi-electron materials have attracted substantial interest because they can realize the transfer of more than one electron through the conversion reaction [7]. Metal fluorides are ideal cathode materials with high theoretical capacities, energy densities, and operating voltages. Among them, FeF 3 has been regarded as the most suitable cathode material due to its high theoretical specific capacity of 712 mAh g −1 (3e − transfer) and 237 mAh g −1 (1e − transfer), high discharge voltage plateau at approximately 2.7 V, and superb thermal stability [8][9][10].
Despite these remarkable merits, FeF 3 as a cathode material still has several shortcomings, which have restricted its practical application. The main drawback of FeF 3 is its electronic insulating behavior caused by a high ionicity, which induces a large band gap of the Fe-F bond and eventually leads to a low actual specific capacity, an inferior rate capability, and poor energy efficiency [11][12][13]. In order to resolve these issues, various strategies have been adopted to overcome the poor electronic and ionic conductivities. Generally, the methods to improve conductivity can be summarized in three aspects as follows: (1) element doping. Element doping can effectively decrease the band gap and actively effect microcrystal growth [14,15]. Rahman et al. prepared Co-doped iron fluoride (Fe 0.9 Co 0.1 F 3 ·0.5H 2 O) by a non-aqueous precipitation method, resulting in a high discharge capacity of 227 mAh g −1 at 0.1 C between 1.8 and 4.5 V [14]. (2) Surface coating. Modification by introducing a coating layer can significantly shorten the Li + transport length and alleviate volume changes [16]. Ma et al. successfully fabricated FeF 3 coated with poly (3,4-ethylenedioxythiophene) (PEDOT) via a novel in situ polymerization method, and the sample exhibited a high power capability of 120 mAh g −1 at 1 A g −1 at room temperature due to the improved ionic and electronic transport in the electrode [17]. (3) Fabricating composite with conductive additives. It can substantially enhance the cycling and rate performance of the FeF 3 cathode material [18][19][20][21]. Jung et al. obtained FeF 3 /ordered mesoporous carbon (OMC) nanocomposite that showed a high reversible specific capacity (178 mAh g −1 at 0.1 C during the second cycle in the voltage range of 2.0-4.5 V) and better cycling stabilities (capacity fading of 8.8%) than bulk FeF 3 (capacity fading of approximately 42%) at 30 cycles [22]. Noticeably, the fabrication of composite with conductive network is the most beneficial approach to improve both the ionic and electronic conductivities to eventually enhance the electrochemical performance of the cathode material.  [26], have been extensively reported. Among them, hexagonal tungsten bronze-type FeF 3 ·0.33H 2 O demonstrated that with the best electrochemical property, its characteristic one-dimensional hexagonal cavity is convenient for efficient Li + transport and can facilitate electrolyte penetration [27]. In addition, the unique structure can effectively limit the movement of water and stabilize the crystal structure. Different functionalized carbon matrixes have been used as conductive networks, but overall, carbon nanotubes (CNTs) and graphene exhibit significant potential as conductive medium due to their distinguished electronic conductivities and excellent stabilities [28][29][30][31]. Graphene, with its large specific area, can promote sufficient contact at the electrode and electrolyte interface, and the graphene network plays an important role in electron transfer and ion migration. Furthermore, graphene provides excellent mechanical stability, which contributes to the bend and stretch of electrode [32,33].
In this study, nanostructured FeF 3 ·0.33H 2 O cathode material was synthesized via a liquid-phase method, and then, the precursor was milled with CNTs followed by sintering to obtain FeF 3 ·0.33H 2 O/CNT composite that was further mixed with graphene conducting paste without a binder. Finally, the CNTs and graphene co-modified FeF 3 ·0.33H 2 O nanocomposite was successfully prepared. The CNTs with intrinsic flexibility and large specific surface area can greatly facilitate the electron transport, and graphene with high mechanical strength and high chemical stability can effectively buffer the volume change and provides a support for electrochemical reaction [31,34,35]. Moreover, the interconnecting of CNTs and graphene sheets can construct an integrated three-dimensional conductive framework, which tremendously promotes Li + diffusion and simultaneously increases the structure stability. Therefore, compared to the FeF 3 ·0.33H 2 O composite with a single conductive network of CNTs and pure FeF 3 ·0.33H 2 O, FeF 3 ·0.33H 2 O nanocomposite with CNTs and graphene networks exhibits superior electrochemical properties. The morphologies, crystal structures, and electrochemical performances of all the samples were systematically investigated.

Results and Discussion
Structural and Morphology Analysis Thermogravimetric-differential scanning calorimetry (TG-DSC) measurement was carried out to confirm the dehydration temperature of the FeF 3 ·3H 2 O precursor and the result is shown in Fig. 1a. Four stages of the weight loss curve are found in the regions of 30-110°C, 110-250°C, 250-450°C, and 450-700°C. In the first stage of 30-110°C, a slight weight loss of approximately 3% can be attributed to phase transformation of crystal. In the second stage of 110-250°C, the TG curve has a rapid weight loss of about 15% and the DSC curve shows an evident endothermic peak around 170°C; the corresponding reaction process is the removal of hydration water (2.67 H 2 O) from FeF 3 ·3H 2 O. In the third stage of 250-450°C, the weight loss is about 6% which may be due to the removal of hydration water for FeF 3 ·0.33H 2 O transforming to FeF 3 and a weak exothermic peak is observed from the DSC curve. In the last stage of 450-700°C, a little weight loss of about 4% is probably due to the decomposition of FeF 3 . According to these results, the precursor was dried at 80°C in a vacuum oven to remove the absorbed water and calcinated at 240°C to obtain FeF 3 ·0.33H 2 O.
X-ray diffraction (XRD) measurements were conducted to investigate the crystal structure of the synthesized samples. The XRD patterns of FeF 3 ·0.  [36]. No evident characteristic peak of CNTs and graphene are observed in the XRD pattern of the FeF 3 ·0.33H 2 O/CNT and FeF 3 ·0.33H 2 O/C + G samples, which is mainly due to the low contents of CNTs and graphene.
The SEM and EDS measurements were performed to analyze the microstructure of the composites. The morphologies and particle sizes of FeF 3 ·0.33H 2 O, FeF 3 ·0.33H 2 O/CNT and FeF 3 ·0.33H 2 O/C + G nanocomposites are shown in Fig. 2. As distinctly seen from Fig. 2a, the particle size of pure FeF 3 ·0.33H 2 O is around 100 nm, and the particles are uniform in size and well distributed, slight aggregation is observed, and the particle size of FeF 3 ·0.33H 2 O can be further confirmed by particle size distribution diagram shown in Fig. 2e. Figure 2b presents the morphology of FeF 3 ·0.33H 2 O/CNT nanocomposite. Clearly, the conductive network of CNTs is intimately intertwined on the surface of the FeF 3 ·0.33H 2 O particles. For the FeF 3 ·0.33H 2 O/C + G nanocomposite, the surface of the FeF 3 ·0.33H 2 O particles is wrapped by CNTs and graphene sheets; as shown in Fig. 2c, the FeF 3 ·0.33H 2 O particles and CNTs are well covered by graphene sheets. In addition, the graphene sheets are preserved well-layered structure in the FeF 3 ·0.33H 2 O/C + G nanocomposite, which can provide a fast channel for Li + transport. The conductive contact between the FeF 3 ·0.33H 2 O material and current collector can be significantly improved by CNTs and graphene due to their outstanding electronic conductivity. Especially graphene with a large surface area can provide an additional transport channel for Li-ion diffusion, which makes the FeF 3 ·0.33H 2 O/C + G nanocomposite with superior electrochemical performance. EDS test was carried out to further investigate the elemental composition of the FeF 3 ·0.33H 2 O/C + G nanocomposite.
The elements of Fe, F, O, and C can be observed from the EDS image in Fig. 2d.
The morphology and detailed microstructure of the FeF 3 ·0.33H 2 O/C + G nanocomposite were further studied by TEM, and the TEM images are displayed in Fig. 3. As shown in Fig. 3a, b, the FeF 3 ·0.33H 2 O particles and CNTs and graphene sheets are closely interconnected with each other, which is consistent with the result of the SEM images. The HRTEM image shown in Fig. 3c offers no evident delineation between the bulk and wrapping layer; the lattice fringe spacing of 0.64 nm coincides with the (110) facet of FeF 3 ·0.33H 2 O. The SAED pattern of the FeF 3 ·0.33H 2 O/C + G nanoparticle is shown in Fig. 3d; the planes of (110), (002), (220), (132), and (004) correspond to the XRD results, which indexed to the hexagonal tungsten bronze structure FeF 3 ·0.33H 2 O. The FeF 3 ·0.33H 2 O/C + G nanocomposite with small particle size and superb conductive network structure, favors sufficient contact between the electrode material and electrolyte and facilitates Li-ion transport; therefore, better electrochemical performance can be achieved.

Electrochemical Characterization
To investigate the electrochemical properties of all samples, galvanostatic charge/discharge tests were implemented in the voltage range of 1.8-4.5 V (vs. Li + /Li), and this voltage range allows only one electron reaction to occur. The charge-discharge profiles of all samples are shown in Fig. 4. The initial charge-discharge curves of the three electrodes at 0.1 C (1 C = 237 mAh g −1 ) rate are shown in Fig. 4a; the pristine FeF 3 ·0.33H 2 O electrode exhibits the lowest initial discharge capacity of 217.5 mAh g −1 , which may be due to the poor electronic conductivity of FeF 3 ·0.33H 2 O. While the FeF 3 ·0.33H 2 O/CNT and FeF 3 ·0.33H 2 O/C + G electrodes deliver higher initial discharge capacities of approximately 225.1 mAh g −1 and 234.2 mAh g −1 , respectively. In our test, the initial discharge capacity of the FeF 3 ·0.33H 2 O/C + G electrode is only 16.7 mAh g −1 higher than that of the pristine FeF 3 ·0.33H 2 O electrode, illustrating CNTs and graphene almost deliver no capacity in the FeF 3 ·0.33H 2 O/ C + G sample. The slightly increased capacity can be attributed to the CNTs and graphene incorporation enhanced the electron transport and reduced electrochemical polarization. From the initial charge-discharge curves of all electrodes, all the curves have an evident discharge plateau at 2.7 V due to the insertion The charge-discharge profiles of different cycles at 0.2 C rate are presented in Fig. 4b Fig. 4c. It is worth noting that the FeF 3 ·0.33H 2 O/C + G electrode still retain a capacity of 193.1 mAh g −1 even after 50 cycles shown in Fig. 4d. In addition, the FeF 3 ·0.33H 2 O/C + G electrode presents the lowest charge voltage plateau and the highest discharge voltage plateau, demonstrating that it has the smallest electrochemical polarization and excellent reversibility, thus mitigate the voltage hysteresis. The better performance of FeF 3 ·0.33H 2 O/CNT and FeF 3 ·0.33H 2 O/C + G electrodes demonstrate that adding CNTs and graphene can effectively improve the conductivity of the FeF 3 ·0.33H 2 O cathode material. Particularly, the FeF 3 ·0.33H 2 O/C + G electrode exhibits the best electrochemical performance due to the interlacing of CNTs and graphene forms a three-dimensional conductive structure, which tremendously facilitate the transport of Li-ion, and thus resulting in promoting the intercalation process of Li-ions [37,38].
To further demonstrate the excellent cycling stability of the FeF 3 ·0. Notably, the FeF 3 ·0.33H 2 O/C + G electrode achieve the highest capacity retention of 85.48% (only 0.29% fading per cycle) after 50 cycles. Moreover, the coulombic efficiency of the FeF 3 ·0.33H 2 O/C + G electrode can reach up to over 99% during the Li + insertion and extraction processes. The above results demonstrate that CNTs and graphene can improve the electronic conductivity and enhance the discharge capacities of FeF 3 ·0.33H 2 O. Particularly, the FeF 3 ·0.33H 2 O/C + G electrode exhibits the best cycling performance, illustrating higher electrical conductivity, better reversibility, and lower polarization after adding of CNTs and graphene. CNTs with high surface area supply sufficient pathway for electron transfer and the graphene works as an excellent conductive network for enabling fast Li + transport between the electrolyte and electrode [28,35]. Moreover, the FeF 3 ·0.33H 2 O particles and CNTs can work as spacers to impede the stacking of graphene sheets and The rate capabilities of the FeF 3 ·0.33H 2 O, FeF 3 ·0.33H 2 O/ CNT, and FeF 3 ·0.33H 2 O/C + G electrodes were evaluated at 0.1 C, 0.5 C,1 C, 3 C, and 5 C rates and then again at 0.1 C rate and results are displayed in Fig. 5b. The discharge capacities of all samples are decreased with increased current density. As expected, the FeF 3 ·0.33H 2 O/C + G electrode presents a superior rate performance among the three electrodes and delivers average discharge capacities of 228 mAh g −1 , 210.7 mAh g −1 , 194.4 mAh g −1 , 170.5 mAh g −1 , and 140.4 mAh g −1 at 0.1 C, 0.5 C, 1 C, 3 C, and 5 C rates. When the rate is returned to 0.1 C, the electrode can still deliver a discharge capacity of 226.7 mAh g −1 . For comparison, FeF 3 ·0.33H 2 O and FeF 3 ·0.33H 2 O/CNT electrodes show inferior rate performance; they deliver poor discharge capacities of 81.7 mAh g −1 and 115.7 mAh g −1 at 5 C rate, which are remarkably lower than that of FeF 3 ·0.33H 2 O/C + G electrode. As a result, the rate capability of the FeF 3 ·0.33H 2 O/C + G electrode is significantly improved compared to those of FeF 3 ·0.33H 2 O without or with a single CNT conductive network. Therefore, the good rate capability of the FeF 3 ·0.33H 2 O/C + G electrode result from the CNTs and graphene conductive networks, which enhanced the electronic conductivity, and above all, the constructed three-dimensional conductive network is beneficial for Li-ion insertion and extraction between electrodes.
Cyclic voltammogram (CV) measurements were carried out to further examine the electrochemical properties of the FeF 3 ·0.33H 2 O, FeF 3 ·0.33H 2 O/CNT, and FeF 3 ·0.33H 2 O/C + G electrodes at a scan rate of 1 mV s −1 between 1.8 V and 4.5 V (vs. Li + /Li) which are shown in Fig. 6. The three curves display similar shapes with apparent oxidation/reduction peaks corresponding to delithiation/lithiation processes. The oxidation and reduction peaks of the FeF 3 ·0.33H 2 O/C + G electrode are detected at 3.32 V and 2.78 V, and the potential interval (ΔE p ) is 0.54 V. While the ΔE p values of the FeF 3 ·0.33H 2 O and FeF 3 ·0.33H 2 O/CNT electrodes are 0.59 V and 0.62 V, respectively. Smaller ΔE p value indicates a smaller electrochemical polarization and a better reversibility of the electrode. In addition, the FeF 3 ·0.33H 2 O/ C + G electrode exhibits a higher current and a larger area than those of the FeF 3 ·0.33H 2 O and FeF 3 ·0.33H 2 O/CNT electrodes. The area surrounded by the CV curve represents the capacity of the material; the larger area is related to the higher capacity, and the change rate of area represents the decay rate of capacity. The results reveal that the FeF 3 ·0.33H 2 O/C + G electrode has a higher capacity and better reversibility, which is well consistent with the galvanostatic charge/discharge tests.
Electrochemical impedance spectroscopy (EIS) measurements were performed to explore the electrochemical reaction kinetics behavior of the FeF 3 ·0.33H 2 O, FeF 3 ·0.33H 2 O/CNT, and FeF 3 ·0.33H 2 O/C + G electrodes after the 3rd cycle and 50th cycle, and results are shown in Fig. 7a, b. All the Nyquist plots of the electrodes after activation are consisted of a semicircle and a sloping line. The semicircle in the high frequency is related to the charge transfer resistance (R ct ), which represents the reaction kinetics of the electrode. The smaller radius of semicircle demonstrates the easier transport of Li + and the electron transfer between the electrolyte and electrode interface, and the sloping line in the low frequency is associated with the Warburg resistance (Z w ) of Li + diffusion in the bulk of cathode material [39]. The corresponding equivalent circuit model was constructed to illustrate the impedance spectra shown in Fig. 7e; the uncompensated ohmic resistance (R s ) represents the resistance of the electrolyte and electrode material, and the constant phase-angle element (CPE) represents the double-layer capacitance and passive film capacitance [40]. The impedance values of R s and R ct for the three electrodes after the 3rd and 50th cycle are listed in Table 1. No significant difference of R s values for the three electrodes after the 3rd cycle is noted. However, the R ct value (50.9 Ω) of the FeF 3 ·0.33H 2 O/C + G electrode is evidently lower than those of the FeF 3 ·0.33H 2 O (115.7 Ω) and FeF 3 ·0.33H 2 O/CNT (68.2 Ω) electrodes, which indicated less polarization of the FeF 3 ·0.33H 2 O/C + G electrode. Moreover, the R ct value of the FeF 3 ·0.33H 2 O/C + G electrode is 86.5 Ω after the 50th cycle, which is also the smallest among the three electrodes. The lower R ct value of the electrode after activation suggested better charge transfer kinetics behavior. The lithium ion diffusion coefficients (D Li+ ) of the FeF 3 ·0.33H 2 O, FeF 3 ·0.33H 2 O/CNT, and FeF 3 ·0.33H 2 O/C + G electrodes are calculated from the following equation [41], In Eq. (1), R is the gas constant, T is the absolute temperature, A is the surface area of electrode, n is the number of electrons involved in the redox reaction, F is the Faraday constant, C is the molar concentration of Li + , and σ ω is the Warburg coefficient which can be obtained from the following relationship, where Z' is the real part of impedance and ω is the angular frequency in the low-frequency region. The linearity of Z' and ω − 1/2 after the 3rd cycle and 50th cycle are shown in Fig. 7c, d. The Li + diffusion coefficients of the three electrodes are listed in Table 1. The D Li+ value (1.67 × 10 −12 cm 2 s −1 ) of the FeF 3 ·0.33H 2 O/C + G electrode after the 3rd cycle is higher than those of the FeF 3 ·0.33H 2 O/CNT (1.19 × 10 −12 cm 2 s −1 ) and FeF 3 ·0.33H 2 O (7.63 × 10 −13 cm 2 s −1 ). In addition, the D Li+ values of the FeF 3 ·0.33H 2 O, FeF 3 ·0.33H 2 O/CNT, and FeF 3 ·0.33H 2 O/C + G electrodes after the 50th

Conclusions
In summary, the FeF 3 ·0.33H 2 O cathode material was successfully synthesized by a liquid-phase method, and the FeF 3 ·3H 2 O precursor was milled with CNTs conductive network followed by sintering to obtain FeF 3 ·0.33H 2 O/ CNT nanocomposite, and then mixed with graphene conducting paste without a binder to obtain the FeF 3 ·0.33H 2 O/ C + G nanocomposite. The functional network consisted of CNTs and graphene provides an effective strategy to improve the electronic conductivity of FeF 3 ·0.33H 2 O cathode material. The FeF 3 ·0.33H 2 O/C + G nanocomposite exhibits better electrochemical performances with increased specific capacity, extended cyclic lifespan, and enhanced rate capability than that of pure FeF 3 ·0.33H 2 O. The EIS results also indicate that the FeF 3 ·0.33H 2 O/C + G electrode has the best electrochemical reaction kinetics behavior. The outstanding electrochemical performances of FeF 3 ·0.33H 2 O/C + G

Characterization
Thermogravimetric-differential scanning calorimetry (TG-DSC) measurement of the precursor was carried out in the temperature range from 30 to 700°C at a heating rate of 10°C min −1 under an argon atmosphere. The crystal structures of all the samples were characterized by X-ray diffraction (XRD, Bruker AXS D8, Germany) with Cu Kα radiation in the 2θ range of 10°-80°at a scan rate of 8°min −1 . The morphologies and particle sizes of the materials were observed by scanning electron microscopy (SEM, JEOL JSM-6610 LV) and energy-dispersive spectroscopy (EDS, JEOL JSM-6610 LV). Transmission electron microscopy (TEM) and selected area electron diffraction (SAED) were carried out to further investigate the microstructure of materials by using a transmission electron microscope (JEOL JSM-2100F).

Electrochemical Measurement
The electrochemical performances of the prepared cathode materials were characterized by CR2032 coin-type half-cells. The working electrodes were made by mixing the cathode materials (FeF 3 ·0.33H 2 O or FeF 3 ·0.33H 2 O/ CNT), carbon black (Super P Li carbon), and polyvinylidene fluoride (PVDF) at a weight ratio of 90:5:5 in N-methyl pyrrolidinone (NMP). When the slurry was stirred uniform, it was pasted on an Al foil and dried at 85°C overnight. The FeF 3 ·0.33H 2 O/C + G combination electrode was fabricated as mentioned above. The cathode electrodes were pressed and cut into several disks and weighted, and then they were dried at 85°C for 4 h in a vacuum oven. The coin-type cells were assembled in an argon-filled glove box, where the oxygen and water contents were controlled to less than 0.1 ppm, metal Li foils as anodes and Celgard 2400 membrane as separator; 1.0 M LiPF 6 in ethylene carbonate (EC), propylene carbonate (PC), and diethyl carbonate (DEC) with a volume ratio of 1:1:1 were used as electrolyte. All the coin cells were aged for 4 h before testing. Galvanostatic charge/discharge tests were performed in the voltage range of 1.8-4.5 V (vs. Li + /Li) on a Land battery test system (LAND CT-2001A, Wuhan, China) at room temperature. The specific capacities of the working electrodes were calculated based on the mass of the active cathode materials. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were measured by an electrochemical workstation (CorrTest CS310). The scanning rate of the CV tests was 1 mV s −1 in the voltage range of 1.8-4.5 V (vs. Li + /Li). The frequency range of EIS was from 100 kHz to 0.01 Hz at potentiostatic signal amplitudes of 5 mV.