coatings Discrete Coating of CNT on Carbon Fiber Surfaces and the Effect on Improving the Electrochemical Performance of VRFB Systems

: Carbon ﬁber, as an electrode material, has been widely used in all-vanadium liquid ﬂow batteries. In order to further reduce the size of the all-vanadium storage system, it is imperative to increase the current density of the battery and to achieve high conductivity and large electrostatic capacitance. The graphitization of the electrode material and the improvement in the speciﬁc surface area of the electrode surface also greatly affect the performance of all-vanadium redox liquid ﬂow batteries. Therefore, in this paper, carbon nanotubes (CNTs) with a small diameter and a large speciﬁc surface area were coated on the electrode surface of the VRFB system by the dispersion method to improve the cell performance. The performance of the surface-modiﬁed electrode was also veriﬁed by Raman spectroscopy, XRD and SEM surface observations and charge/discharge experiments.


Introduction
The liquid flow battery is an electrochemical energy storage technology proposed by Thaller (NASA Lewis Research Center, Cleveland, OH, USA) in 1974 [1,2]. The liquid flow battery is also known as a battery active material renewable fuel cell [3]. When vanadium ions are used as the active material in the battery, called a vanadium liquid flow battery (VRFB), the vanadium electrolyte in the liquid storage tank is pressed into the battery stack through an external pump during system operation to complete the electrochemical reaction. V 4+ in the positive electrode is oxidized to V 5+ during charging, and V 3+ in the negative electrode is reduced to V 2+ . The reverse reaction occurs during the discharge process. The reaction is as follows.
Positive electrode reaction: Negative electrode reaction: As an energy storage battery, the all-vanadium liquid flow battery has been widely used in various energy storage fields, including large power stations, photovoltaic power generation and wind power generation [4][5][6][7]. However, the disadvantages of a low energy density and a large size need to be further addressed [8][9][10]. Therefore, reducing the size of all-vanadium liquid flow batteries and increasing the energy density and current density of

Experimental Methods
In this study, the dispersion of multi-walled carbon nanotubes was treated with different dispersions and the dispersions were analyzed using a ZC-3000 ZETA potentiostat from Kyowa Interface Science Co., Ltd. (Niiza, Japan). To ensure the uniformity of the coating, we applied the dispersed CNT solution in batches on both the front and back sides of the electrodes, which were heated and dried. The VRFB used for charge/discharge experiments was a homemade spiral pump-driven cell with an electrode area of 5 × 10 cm. For the data of the charge/discharge experiments, SEM-EDS(JEOL Ltd. JSM-5510 Tokyo, Japan) observation analysis, Raman spectroscopy analysis(JASCO Corporation. NRS-4100, Tokyo, Japan) and XRD (Rigaku Corporation, RINT2500HF, Tokyo, Japan) surface crystallography analysis were performed on the electrode surface. The obtained analytical data were calculated, and the calculated results were compared in order to summarize the experimental results.

Experimental Conditions
The CNTs used in this experiment were multi-walled carbon nanotubes produced by Wako Pure Chemical Industries, Ltd (Tokyo, Japan), commercially available multi-walled CNTs with an average diameter of 10 nm and an average length of 1 µm. The dispersion medium mainly contained deionized water and ethanol solution, and the surfactant was SDBS (sodium dodecylbenzene sulfonate). Here, 0.006 g of CNTs was placed in three different solvents, stirred for 15 min, sonicated for 20 min and left for 1 h. The stability of the dispersed solution was determined by measuring the ZETA potential. The ultrasonic treatment instrument was model W-113A manufactured by Honda Electronics, Ltd. (Tokyo, Japan) and the ultrasonic treatment conditions were 60 • C, 45 KHZ and 20 min, as shown in Figure 1.
experiments was a homemade spiral pump-driven cell with an electrode area of 5 × 10 cm. For the data of the charge/discharge experiments, SEM-EDS(JEOL Ltd. JSM-5510 Tokyo, Japan) observation analysis, Raman spectroscopy analysis(JASCO Corporation. NRS-4100, Tokyo, Japan) and XRD (Rigaku Corporation, RINT2500HF, Tokyo, Japan) surface crystallography analysis were performed on the electrode surface. The obtained analytical data were calculated, and the calculated results were compared in order to summarize the experimental results.

Experimental Conditions
The CNTs used in this experiment were multi-walled carbon nanotubes produced by Wako Pure Chemical Industries, Ltd (Tokyo, Japan), commercially available multi-walled CNTs with an average diameter of 10 nm and an average length of 1 μm. The dispersion medium mainly contained deionized water and ethanol solution, and the surfactant was SDBS (sodium dodecylbenzene sulfonate). Here, 0.006 g of CNTs was placed in three different solvents, stirred for 15 min, sonicated for 20 min and left for 1 h. The stability of the dispersed solution was determined by measuring the ZETA potential. The ultrasonic treatment instrument was model W-113A manufactured by Honda Electronics, Ltd. (Tokyo, Japan) and the ultrasonic treatment conditions were 60 °C, 45 KHZ and 20 min, as shown in Figure 1.

Calculation Method
The charge/discharge experiment was performed with a constant current charge/discharge, an electrolyte flow rate of 2.4 mL min −1 cm −2 and a carbon felt compression ratio of 30%. The charge/discharge experiments were carried out on untreated carbon felt and treated carbon felt, and the charge/discharge data results were calculated to compare the experimental results. The charge and discharge internal resistance (IR), voltage efficiency (EV), coulombic efficiency (EC), energy efficiency (EE), energy density (ED) and output density (OD) of the battery were calculated based on the constant current charge and discharge curves. The calculation equations are shown in Equations (3)-(8).

Calculation Method
The charge/discharge experiment was performed with a constant current charge/ discharge, an electrolyte flow rate of 2.4 mL min −1 cm −2 and a carbon felt compression ratio of 30%. The charge/discharge experiments were carried out on untreated carbon felt and treated carbon felt, and the charge/discharge data results were calculated to compare the experimental results. The charge and discharge internal resistance (I R ), voltage efficiency (E V ), coulombic efficiency (E C ), energy efficiency (E E ), energy density (E D ) and output density (O D ) of the battery were calculated based on the constant current charge and discharge curves. The calculation equations are shown in Equations (3)- (8).

Dispersion Experiments of CNTs and SEM-EDS Surface Observation and Elemental Analysis
The dispersions of the three dispersions at different concentrations were compared by measuring the ZETA potential. The results show that the 80% ethanol solution (a), 0.1 mol/L SDBS ethanol solution (b) had the best dispersion. The dispersion potential is negative and arranged in a gradient. The measurement results are shown in Figure 2. Since the SDBS ethanol solution produced a large number of bubbles after sonication, the solution started to solidify after 2 h of standing, and the precipitation was obvious after 6 h of standing. Therefore, the 80% ethanol solution was used in this experiment.
The 100 mL of dispersed solution was heated and dried after being dropped into the carbon felt electrode several times. The surface state and elements of the carbon fiber electrodes were observed and analyzed by SEM-EDS. The SEM images of the untreated electrode, Figure 3(a1-a3), and the treated electrode, Figure 3(b1-b3), are shown in Figure 3. The SEM comparison shows that the dispersed CNTs completely adhered to the surface of the carbon felt fibers.
The results of the EDS elemental analysis are shown in Table 1 and Figure 4. It can be seen from the elemental analysis results that the carbon element percentage of the treated carbon felt increased by 4.25%.    The results of the EDS elemental analysis are shown in Table 1 and Figure 4. It can be seen from the elemental analysis results that the carbon element percentage of the treated carbon felt increased by 4.25%.

Charge and Discharge Experiments and Impedance Measurement Evaluation
The experimental results are shown in Figure 5 ((a) untreated carbon felt, (b) treated carbon felt) after conducting a constant current charge/discharge experiment on a single cell with a current density of 100-500 mA/cm 2 . The maximum power density of the untreated carbon felt electrode is 300 mA/cm 2 , and the charging voltage is 1.8 V. The treated carbon felt electrode can reach 400 mA/cm 2 , and the charging voltage is 1.6-1.7 V. The charging and discharging data were calculated according to Equations (3)-(8), as shown in Table 2. It can be seen from Table 2 that the IR of the treated carbon felt is two times lower than that of the treated carbon felt, and the EV and EE are increased by about 15%. This is because the surface coating of the carbon felt with CNTs improves the conductivity of the electrode and increases the specific surface area, thus reducing the charge/discharge internal resistance.

Charge and Discharge Experiments and Impedance Measurement Evaluation
The experimental results are shown in Figure 5 ((a) untreated carbon felt, (b) treated carbon felt) after conducting a constant current charge/discharge experiment on a single cell with a current density of 100-500 mA/cm 2 . The maximum power density of the untreated carbon felt electrode is 300 mA/cm 2 , and the charging voltage is 1.8 V. The treated carbon felt electrode can reach 400 mA/cm 2 , and the charging voltage is 1.6-1.7 V. The charging and discharging data were calculated according to Equations (3)-(8), as shown in Table 2. It can be seen from Table 2 that the I R of the treated carbon felt is two times lower than that of the treated carbon felt, and the E V and E E are increased by about 15%. This is because the surface coating of the carbon felt with CNTs improves the conductivity of the electrode and increases the specific surface area, thus reducing the charge/discharge internal resistance.

Charge and Discharge Experiments and Impedance Measurement Evaluation
The experimental results are shown in Figure 5 ((a) untreated carbon felt, (b) tr carbon felt) after conducting a constant current charge/discharge experiment on a cell with a current density of 100-500 mA/cm 2 . The maximum power density of th treated carbon felt electrode is 300 mA/cm 2 , and the charging voltage is 1.8 V. The tr carbon felt electrode can reach 400 mA/cm 2 , and the charging voltage is 1.6-1.7 V charging and discharging data were calculated according to Equations (3)-(8), as s in Table 2. It can be seen from Table 2 that the IR of the treated carbon felt is two lower than that of the treated carbon felt, and the EV and EE are increased by about This is because the surface coating of the carbon felt with CNTs improves the conduc of the electrode and increases the specific surface area, thus reducing the charge/disc internal resistance.   For the charge/discharge experimental data, the AC impedance of the battery was measured, and the results (Nyquist plot) are shown in Figure 6 for the test range of 10 5 → 0.1 Hz. The intersection of the starting position of the curve with the x-axis is the on-state resistance (R L ) of the battery itself. In the figure, the half-circle part is the charge movement resistance (R P ) in the high-frequency region, and the diagonal part is the material movement resistance (R D ) in the low-frequency region. σ is the Warburg coefficient, ω is the frequency and C d is the double-layer capacitance. The calculation equation is shown in (9) and (10).
The main purpose of this experiment is to review the charge movement resistance R P in the high-frequency region and the matter movement resistance R D in the low-frequency region. From the measured Nyquist diagram, it can be calculated that R P = 0.7 Ω and R D = 1 Ω for the untreated carbon felt and R P = 0.025 Ω and R D = 0.035 Ω for the treated carbon felt. It can be seen that the charge and discharge impedance of the coated carbon felt is reduced by about five times. For the charge/discharge experimental data, the AC impedance of the battery was measured, and the results (Nyquist plot) are shown in Figure 6 for the test range of 10 5 → 0.1 Hz. The intersection of the starting position of the curve with the x-axis is the on-state resistance (RL) of the battery itself. In the figure, the half-circle part is the charge movement resistance (RP) in the high-frequency region, and the diagonal part is the material movement resistance (RD) in the low-frequency region. σ is the Warburg coefficient, ω is the frequency and Cd is the double-layer capacitance. The calculation equation is shown in (9) and (10).
The main purpose of this experiment is to review the charge movement resistance RP in the high-frequency region and the matter movement resistance RD in the low-frequency region. From the measured Nyquist diagram, it can be calculated that RP = 0.7 Ω and RD = 1 Ω for the untreated carbon felt and RP = 0.025 Ω and RD = 0.035 Ω for the treated carbon felt. It can be seen that the charge and discharge impedance of the coated carbon felt is reduced by about five times.

Raman Spectroscopic Detection and XRD Surface Crystallization Analysis
For the results of constant current charging and discharging, we performed Raman spectroscopy and XRD surface crystallographic analysis on two carbon felts.
In this paper, we use the common method of Raman spectroscopy to analyze carbon materials. There are two characteristic peaks in the Raman spectrum of carbon materials: the G peak near 1580 cm −1 and the D peak near 1360 cm −1 [35]. Researchers usually go through the D and G peaks. The integrated area ratio ID/IG is used to determine the integrity of the carbon material. The larger the ID/IG ratio, the lower the structural integrity of the carbon material. The smaller the ID/IG ratio, the higher the structural integrity of the carbon material [36,37]. In this study, the treated carbon fiber electrodes were tested by Raman spectroscopy using the NRS-4100 from JASCO (JASCO Corporation. Tokyo, Japan). The measured Raman spectra are shown in Figure 7, and the measured data are

Raman Spectroscopic Detection and XRD Surface Crystallization Analysis
For the results of constant current charging and discharging, we performed Raman spectroscopy and XRD surface crystallographic analysis on two carbon felts.
In this paper, we use the common method of Raman spectroscopy to analyze carbon materials. There are two characteristic peaks in the Raman spectrum of carbon materials: the G peak near 1580 cm −1 and the D peak near 1360 cm −1 [35]. Researchers usually go through the D and G peaks. The integrated area ratio I D /I G is used to determine the integrity of the carbon material. The larger the I D /I G ratio, the lower the structural integrity of the carbon material. The smaller the I D /I G ratio, the higher the structural integrity of the carbon material [36,37]. In this study, the treated carbon fiber electrodes were tested by Raman spectroscopy using the NRS-4100 from JASCO (JASCO Corporation. Tokyo, Japan). The measured Raman spectra are shown in Figure 7, and the measured data are shown in Table 3. Combining the Raman spectra and the measured data, it can be seen that the peak shapes of the D and G peaks of the coated treated carbon felt became relatively sharp and obvious, the integrated area ratio and half-height width of the peaks also became smaller, the characteristic peak G' appeared near 2700 cm −1 of the carbon felt and the ratio of I D /I G decreased by 0.068. shown in Table 3. Combining the Raman spectra and the measured data, it can b that the peak shapes of the D and G peaks of the coated treated carbon felt becam tively sharp and obvious, the integrated area ratio and half-height width of the peak became smaller, the characteristic peak G' appeared near 2700 cm −1 of the carbon fe the ratio of ID/IG decreased by 0.068.  For the surface crystal structure, untreated and treated carbon fibers were ana using the RINT 2500 VHF from RIGAKU Corporation, Japan, under the followin conditions: voltage: 40 kV: 30 mA (Cu). The measurement results are shown in Fig  It can be seen that the shape of the (002) diffraction peak for 2θ = 25° is basically the However, the intensity of the peak increases after the coating treatment. There is n nificant difference in the (100) diffraction peak at 2θ = 43°.
According to Equations (11) and (12), the layer spacing d002 of the carbon fibber g ite crystallites and the thickness of the crystallite stack Lc can be calculated [38]. Th peak in the XRD spectrum can be used to calculate the direction of the graphite crys along the axis, the crystal plane width La and the diffraction angle of the θ crystal diffraction peak; λ is the wavelength (λ = 0.1541 nm); K is the shape factor, K = 0.94 Lc is calculated, and K = 1.84 when La is calculated; and β is the measured full wi half maximum. The calculation results are shown in Table 4.  For the surface crystal structure, untreated and treated carbon fibers were analyzed using the RINT 2500 VHF from RIGAKU Corporation, Japan, under the following test conditions: voltage: 40 kV: 30 mA (Cu). The measurement results are shown in Figure 8. It can be seen that the shape of the (002) diffraction peak for 2θ = 25 • is basically the same. However, the intensity of the peak increases after the coating treatment. There is no significant difference in the (100) diffraction peak at 2θ = 43 • .
Coatings 2021, 11, 736 From the comparison of the calculated results in Table 4, it can be found that th no significant difference between the crystal spacing d002 and the crystal spacing d100 untreated and treated electrodes; the radial size Lc and the axial size La of the crys the treated carbon fiber electrodes become larger. The larger crystallite size indic more complete development and a higher degree of graphitization [39]. Therefore, theoretically demonstrated that the performance of the treated carbon cellulose elect was higher than that of the untreated electrodes.  According to Equations (11) and (12), the layer spacing d 002 of the carbon fibber graphite crystallites and the thickness of the crystallite stack Lc can be calculated [38]. The (100) peak in the XRD spectrum can be used to calculate the direction of the graphite crystallite along the axis, the crystal plane width La and the diffraction angle of the θ crystal plane diffraction peak; λ is the wavelength (λ = 0.1541 nm); K is the shape factor, K = 0.94 when Lc is calculated, and K = 1.84 when La is calculated; and β is the measured full width at half maximum. The calculation results are shown in Table 4. From the comparison of the calculated results in Table 4, it can be found that there is no significant difference between the crystal spacing d 002 and the crystal spacing d 100 of the untreated and treated electrodes; the radial size Lc and the axial size La of the crystals of the treated carbon fiber electrodes become larger. The larger crystallite size indicates a more complete development and a higher degree of graphitization [39]. Therefore, it was theoretically demonstrated that the performance of the treated carbon cellulose electrodes was higher than that of the untreated electrodes. frequency Ω ohm Superscripts and Subscripts avg average charge process discharge discharge process