Graphitization of carbon-based catalyst for improving the performance of negative electrode in vanadium redox flow battery

Carbon-based materials were prepared to catalyze the V3+/V2+ couple of vanadium redox flow battery using chitosan as the preliminary material and FeCl3 as activating agent. Graphite microcrystals were the main structures of the obtained catalyst (CTS-Fe-900) activated by FeCl3, and they contained a large number of curled and overlapped carbon nanosheets. Compared with CTS-D-900 (without FeCl3 as an activating agent), CTS-Fe-900 exhibits a better structure, higher graphitization degree, stronger current response, and smaller charge transfer resistance. The charge/discharge measurements indicates that the performances of the cell are improved by using the CTS-Fe-900-modified negative electrode in terms of increased discharge capacity and energy efficiency. Compared with pristine cell, the capacity retention for the CTS-Fe-900 modified cell maintains 82.2% at 50 mA cm−2 after 50 cycles, increased by 5.2%, and the corresponding energy efficiency reaches 81.3%, enhanced by 5.1%. This work reveals that CTS-Fe-900 catalyst can improve the comprehensive energy storage performance. The excellent electrocatalytic properties are mainly attribute to the effect of the FeCl3 template and the increase in the degree of graphitization.


Introduction
With the impoverishment of resources and the extensive utilization of new energy (wind energy, solar energy, etc), the technology of energy storage has been extensively developed to resolve the discontinuity and geographical distribution of new energy [1][2][3][4][5]. Vanadium redox flow battery (VRFB), which was first proposed by Skyllas-Kazacos et al in 1985 [6], has been investigated extensively because of its many advantages such as long life, high efficiency and environmental friendliness [7][8][9]. VRFB uses V 3+ /V 2+ and VO 2+ /VO 2 + redox couples in sulfuric acid as the negative and positive half-cell electrolytes, respectively, and it exhibits an open circuit voltage of approximately 1.26 V produced from the potential separation of two couples [10][11][12]. Two reactions take place the on electrode surface; however, the electrode does not participate the in reaction [13]. Hence, the electrochemical activity of the electrode is crucial for the performance of the cell. VRFB mainly employs carbonbased fiber such as graphite, graphite felt, carbon cloth, and carbon paper, etc as electrode [14]. Nevertheless, the electrochemical activity of these raw electrodes is not perfect enough. Thus, various methods for enhancing the electrochemical properties of the carbon electrodes, such as acid treatment, electrochemical oxidation, and heating treatment, have been reported [15][16][17]. Moreover, the catalyst introduced onto the surface of the electrode is a research hotspot for enhancing the electrochemical activity of the electrode. The catalysts used in VRFB include carbon-based materials (graphene [18], carbon nanotubes (CNTs) [19], and carbon fibers [20] etc), metals (Pt [21], Ir [22], and Sb [23], etc), and metal oxides (ZrO 2 [24], expensive. Consequently, more researchers have focused on designing many low-price metal oxides and nitrides catalysts to replace the noble metal catalysts. However, some metal oxides and nitrides demonstrate relatively poor conductivity and stability on the electrode surface. Carbon-based catalysts demonstrates excellent electrochemical catalytic activity for V 3+ /V 2+ redox couples because of their high conductivity, high specific surface areas, and excellent corrosion resistance. Nevertheless, graphene and carbon nanotubes are limited in commercial applications due to their tedious preparation and high cost. Recent research has mainly focused on improving the electrochemical performance of positive electrodes [22,25,27,28]. However, improving the performance of the negative electrode is crucial for improving the performance of the battery. This is because the negative reaction takes larger proportion of the voltage loss of the cell [29]. In this work, chitosan was used as the preliminary material and FeCl 3 as the activating agent to obtain carbon-based catalysts for VRFB. The carbon-based catalyst was acted the as electrocatalyst for the V 3+ /V 2+ couple of VRFB. Notably, the obtained carbon-based catalyst shows excellent electrocatalytic performances for V 3+ /V 2+ couple of VRFB.

Preparation of materials
Chitosan (1.5 g) and FeCl3 (3 g) were soaked in 10 ml of deionized water and sufficiently stirred, the resulting mixture was then dried at 80°C. The solid materials were calcined at 900°C (5°C min −1 ) for 2 h in argon. Thereafter, the resultant product was soak in 1 mol l −1 HCL solution for 24 h, washed with deionized water to pH 7, and dried at 80°C before use (named as CTS-Fe-900). The chitosan (without FeCl 3 ) was calcined directly at 900°C (5°C min −1 ) for 2 h in argon for comparison (named as CTS-D-900).

Characterizations of materials
The morphology of the samples was studied through scanning electron microscopy (SEM, S-4800, Hitachi, Japan). The degree of surface defects in the samples was investigated by Raman spectroscopy (DXR, Thermo Fisher Scientific). The phase of the as-prepared sample was verified via x-ray diffraction (XRD, DX-2700, Dandong Haoyuan, China). Hydrophilicity performance was evaluated using water contact angle measurement (JC2001, Dong guan saite testing equipment co. LTD).

Electrochemical measurements
The as-prepared samples (each 10 mg) was dissolved into the 5 ml N,N-dimethylformamide (DMF) via ultrasonication for 3 h to ensure that the samples was well dispersed in the solution, and 20 μl of dispersed solution was dropped onto the surface of glass carbon electrode and dried at 80°C for 4 h.
Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) experiments were performed on CHI660E electrochemical workstation (Shanghai Chenhua, China) in 1.6 mol l −1 V 3+ +3.0 mol l −1 H 2 SO 4 electrolyte. Electrochemical measurements were conducted using typical three-electrode system, GCE coated with samples as the working electrodes, Pt sheet (1×1 cm 2 ) as the counter electrodes, and a saturated calomel electrode (SCE) as the reference electrodes.

Single cell tests
CTS-Fe-900 (6 mg) was ultrasonically dispersed in 15 ml DMF, and the pristine graphite felt (4 cm×4 cm) was immersed in the solution until it was completely drained by the graphite felt (GF). The modified GF was dried in an oven at 80°C for 8 h. Using the CTS-Fe-900 modified GF as the negative electrode, the performance of the VRFB was evaluated in a static cell. The pristine GF and CTS-Fe-900 modified GF were employed as positive and negative electrodes, respectively. Two half cells were separated by ion-exchange membrane. 1.6 mol l −1 V 1 O 2+ +3 mol l −1 H 2 SO 4 and 1.6 mol l −1 V 3+ +3 mol l −1 H 2 SO 4 were stored in corresponding positive and negative electrode. Galvanostatic charge/discharge measurement was conducted on CT2001A testing system (Land, Wuhan). The upper and lower limits of the charge and discharge voltage was 1.7 and 0.7 V. The charge/ discharge performance was evaluated at a series of current density of 50-90 mA cm −2 and 50 mA cm −2 for 50 cycles. Additionally, the pristine GFs were also used as positive and negative electrodes for comparison.

Results and discussion
The XRD patterns of CTS-D-900, CTS-Fe-900 and CTS-Fe-900 precursor (the sample was not washed with deionized water) are shown in figure 1. There are two diffraction peaks at 23°and 44°of 2θ for CTS-D-900, and the diffraction peaks at 23°are formed by the superposition of the amorphous carbon diffraction peak at 22°and  the crystallo-graphitc planes diffraction peak at 26°. Compared with CTS-D-900, CTS-Fe-900 has a sharp and intense diffraction peak at 26°and no diffraction peaks of other substances, which indicates a higher organized crystal than CTS-D-900. The diffraction peaks at 2θ=26°for CTS-Fe-900 precursor is also attributed to crystallo-graphitc planes, and other peaks for CTS-Fe-900 precursor are attributed to Fe and oxides ( FeO, Fe 2 O 3 , and Fe 3 O 4 ). Fe is a good graphitization-reinforcing catalyst that converts the amorphous carbon into graphite carbon; therefore, the electronic conductivity of carbon materials can be improved by improving the graphitization degree of carbon materials. The increase in electron conductivity can accelerate the electron transfer during electrochemical reaction, and thus accelerate electrode reaction process.
SEM images of CTS-D-900 and CTS-Fe-900 samples at different magnifications are shown in figure 2. The CTS-D-900 sample is an irregular carbon block, some of which are up to 10 μm in size (figures 2(a), (b)). The CTS-Fe-900 sample is composed of a large number of curled and overlapped carbon nanosheets (figures 2(c), (d)). As can be seen, the size of the carbon sheet of the CTS-Fe-900 sample is significantly smaller than that of the CTS-D-900 sample. Moreover, the CTS-Fe-900 sample exhibits a better porous structure. Evidently, the CTS-Fe-900 sample has a larger specific surface area than the CTS-D-900 sample. The carbon nanosheets are formed because FeCl 3 not only acts as an oxidant at high temperatures and a template, but also facilitates redox reactions with biomass materials. FeCl 3 is embedded into the chitosan during the high temperature reaction, however, FeCl 3 and its intermediate products can be removed by pickling after the high-temperature reaction, leaving behind a porous lamellar structure.
The graphitization of the carbon materials of CTS-D-900 and CTS-Fe-900 was studied through Raman spectroscopy, and the results are shown in figure 3. Figure 3 shows that the Raman spectra of the two samples are similar, with two strong peaks at 1350 and 1580 cm-1 representing disordered amorphous carbon (D band) and crystalline graphitic carbon (G band), respectively [30]. In general, a smaller ratio of the D to G peak intensity (I D /I G ) can reveal more crystalline graphitic carbon [31]. Notably, I D /I G for CTS-Fe-900 is 0.62, which is smaller than that for CTS-D-900 with I D /I G of 0.97. This indicates that the CTS-Fe-900 sample has a higher degree of graphitization than the CTS-D-900 sample. This is consistent with the XRD results.
A test of water contact angle was conducted to examine the wetting performance of the two samples, and the results are shown in figure 4. A larger water contact angle indicates worse hydrophilicity; the worse hydrophilicity reflects a higher degree of graphitization. As seen, the water contact angle of CTS-Fe-900 is 139.4°, which is larger than that for CTS-D-900 ( 99.1°). It indicates that CTS-D-900 has a higher degree of  graphitization. This is because the FeCl 3 generates Fe at high temperatures. Fe can improve the graphitization degree of the material, and the increase in the degree of graphitization causes a decline in hydrophilicity.
The electrocatalytic activity of the two samples ware compared via CV measurements with scanning rate of 5 mV s −1 in 1.6 mol l −1 V 3+ +3.0 mol l −1 H 2 SO 4 . The cyclic voltammograms of the two samples are presented in figure 5. As seen, CTS-D-900 displays a pair of relatively poor redox peaks for the V 3+ /V 2+ couple. Compared with CTS-D-900, CTS-Fe-900 presents a clear peak shape, showing a higher peak current density and better symmetry for the redox peak of V 3+ /V 2+ couple. The reduction peak currents of the CTS-D-900 and CTS-Fe-900 are 1.235 and 3.165 mA, respectively. The oxidation peak currents of the CTS-D-900 and CTS-Fe-900 are 0.377 and 1.532 mA, respectively. Evidently, CTS-Fe-900 exhibits better electrocatalytic performance than CTS-D-900. The peak potential gap is a curial parameter for evaluating the electrochemical reversibility of the V 3+ /V 2+ reaction [18]. In general, a smaller peak potential gap corresponds to better electrochemical reversibility. The peak potential gap for CTS-Fe-900 is 265 mV, which is smaller than that of CTS-D-900, implying that the electrochemical reversibility of the V 3+ /V 2+ reaction on CTS-Fe-900 is better than that on CTS-D-900. The superior electrocatalytic performance of CTS-Fe-900 can be attributed to two reasons. First, CTS-Fe-900 has a larger specific surface area, which can provide a larger reaction site for the electrode reaction. Second, improving the graphitization degree of carbon materials improves the electronic conductivity of the materials.
To investigate the electrochemical kinetics of the V 3+ /V 2+ couple, EIS measurements were carried out for the CTS-D-900 and CTS-Fe-900 electrodes. Nyquist plots of the two samples and the corresponding equivalent electrical circuit are shown in figure 6. Nyquist plots are made of two parts for each sample, including a semicircle ascribed to the charge transfer process in the high frequency range and a sloped line reflecting the diffusion process of vanadium ions in the low frequency range, respectively [32]. It implies that the V 3+ /V 2+ couple is conjointly dominated by the diffusion process of active species and charge transfer processes. In the electric equivalent circuit, R s is the ohmic resistance composed of the solution, electrode, and contact resistance,   R ct is the charge transfer resistance, Q t and Q m are constant-phase elements, where Q t is the diffusion capacitance of vanadium ions, and Q m is the electric double-layer capacitance between the electrode/electrolyte surface between electrode/electrolyte surface. The electrochemical parameters are listed in table 1. It can be seen that R s of CTS-Fe-900 is 7.84 Ω, which is smaller than that for CTS-D-900 (8.19 Ω). This is because the higher graphitization degree of CTS-Fe-900 reduces the ohmic resistance of the GCE/CTS-Fe-900 electrode. Meanwhile, R ct for CTS-Fe-900 (7.57 Ω) is significantly smaller than that for CTS-D-900 (199.30 Ω), demonstrating quicker charge transfer for redox reaction on CTS-Fe-900, which is because of larger surface area and higher graphitization. In addition, Q m and Q t for CTS-Fe-900 increased significantly compared with those of CTS-D-900, which indicates that the increase in the specific surface area and graphitization degree of the material could improve the diffusion process of vanadium ions and the electrochemical reaction of the V 3+ /V 2+ redox couple.
The adhesion of CTS-Fe-900 to graphite felt was studied by SEM. Figure 7 shows the blank graphite felt and CTS-Fe-900 modified graphite felt. The surface of the blank graphite felt is relatively clean, and there is no obvious impurity ( figure 7(a)). As displayed in figure 7(b), a large number of nanoscale sheets of CTS-Fe-900 electrocatalyst attached to the surface of blank graphite felt evenly, which can greatly improve the surface area of the graphite felt. The thickness of the electricity catalyst is nanosized, therefore, it did not affect the flow of the electrolyte on the electrode surface.   Charge-discharge tests were carried out to study the effect of CTS-Fe-900 on the rate performance of the cell. Figure 8 displays the charge-discharge voltage profiles of the pristine cell and CTS-Fe-900 modified cells at different current densities. As can be seen, the related voltage separation of the two cells increases with increasing current density, but the cell using CTS-Fe-900 exhibits a lower charge plateau and higher discharge plateau at 50-90 mA cm −2 due to lower overpotential. The mean discharge voltages at different current densities are listed in table 2. The mean discharge voltage of the two cells decreased with increasing current density because of the electrochemical polarization. The mean discharge voltage for the CTS-Fe-900-modified cell was higher than that of the pristine cell at 50-90 mA cm −2 . In particular, the mean discharge voltage for the CTS-Fe-900-modified cell was 1.162 V at 90 mA cm −2 , with an increase of 0.035 V compared with that of the pristine cell (1.127 V). Evidently, CTS-Fe-900 reduced the overpotential and enhanced the mean discharge voltage owing to its larger surface area and higher graphitization, and it can weaken the electrochemical polarization and improve the energy density of the cell.
To study the cycling performance of the pristine-and CTS-Fe-900-modified cells, charge-discharge tests were performed at 50 mA cm −2 for 50 cycles, as shown in figure 9. As shown in figure 9(a), the discharge capacity of both cells gradually decreased with an increasing number of cycles. This is because the charging and discharging processes weaken the electrode activity, and some vanadium ions in the electrolyte penetrate the membrane with an increasing number of cycles, resulting in a decline in the discharge capacity. However, the CTS-Fe-900-modified cell demonstrated a higher discharge capacity than the pristine cell. The original discharge capacity of the CTS-Fe-900-modified cell is 220.0 mA h, which is 15.5 mA h higher than that of the pristine cell. Moreover, the capacity retention of the CTS-Fe-900 modified cell maintains 82.2% after 50 cycles, 5.2% larger than that of the pristine cell. The energy efficiency of two cells are compared in figure 9(b). Energy efficiency of the CTS-Fe-900 modified cell is 81.3% at 50 mA cm −2 , which is 5.1% higher than that of the pristine cell (76.2%). The above results demonstrate that the CTS-Fe-900 modified cell has more excellent cycling stability than the pristine cell during the charge and discharge processes.

Conclusions
In summary, carbon-based materials were prepared to catalyze V 3+ /V 2+ using chitosan as the preliminary material and FeCl 3 as the activating agent. The main structure of CTS-D-900 without FeCl 3 as an activating agent was amorphous carbon with a lower degree of graphitization and irregular micron carbon particles. The main structure of the obtained catalyst (CTS-Fe-900) activated by FeCl 3 was graphite microcrystals and it contained a large number of curled and overlapped carbon nanosheet materials, there, it had a better structure and higher graphitization degree. Compared with those of CTS-D-900, the current response of CTS-Fe-900 was stronger and the charge transfer resistance was smaller, therefore it had stronger electrocatalytic activity than CTS-D-900. As a result, the CTS-Fe-900 modified cell demonstrates higher discharge capacity and energy efficiency than the pristine cell at 50 mA cm −2 for 50 cycles. The capacity retention of the CTS-Fe-900 modified cell maintains 82.2% after 50 cycles, 5.2% larger than that of the pristine cell. The energy efficiency of the CTS-Fe-900 modified cell is 81.3% at 50 mA cm −2 , which 5.1% higher than that of the pristine cell. This indicates that using CTS-Fe-900 catalyst can improve the performance of comprehensive energy storage. The excellent electrocatalytic properties mainly come from the effect of the FeCl 3 template and the increase in the degree of graphitization.