Improving Electrochemical Activity in a Semi-V-I Redox Flow Battery by Using a C–TiO2–Pd Composite Electrode

This study developed composite electrodes used in a semi-vanadium/iodine redox flow battery (semi-V-I RFB) system and designed semi-V-I RFB stacks to provide performance comparable to that of an all-vanadium redox flow battery (all-VRFB) system. These electrodes were modified using the electroless plating method and sol-gel process. The basic characteristics of the composited electrodes, such as the surface structural morphology, metal crystal phases, and electrochemical properties, were verified through cyclic voltammetry, field emission-scanning electron microscopy, energy-dispersive X-ray spectrometry, and X-ray diffraction. The results show that the sintering C–TiO2–Pd electrode improved the electrocatalytic activity of the semi-V-I RFB system, thereby effectively increasing the energy storage ability of the system. The C–TiO2–Pd electrode was used as a negative electrode in a single semi-V-I RFB and exhibited excellent cyclic performance in a charge-discharge test of 50 cycles. The average values for coulomb efficiency, voltage efficiency, and energy efficiency were approximately 96.56%, 84.12%, and 81.23%, respectively. Moreover, the semi-V-I RFB stacks were designed using series or parallel combination methods that can effectively provide the desired operating voltage and linearly increase the power capacity. The amount of vanadium salt required to fabricate the semi-V-I RFB system can be reduced by combining large stack modules of the system. Therefore, this system not only reduced costs but also exhibited potential for applications in energy storage systems.


Introduction
The energy of a redox flow battery (RFB) is stored in separated positive and negative electrolytes, which provide the driving force that initiates the oxidation-reduction reaction [1]. The use of energy stored in the form of chemical energy, a mode of storage that does not entail any geographical restrictions, is expected to facilitate large-scale energy storage and has significant environmental and socioeconomic advantages. RFBs have been utilized as large-scale energy storage systems in the last decade [2][3][4][5]. The all vanadium RFB (all-VRFB) is one of the most promising technologies for mid-to-large-scale (kilowatt-megawatt) energy storage and was first proposed by Skyllas-Kazacos in 1985 [6,7]. The power output of a vanadium RFB increases with the size of the cell stack and the active electrode surface area, whereas its energy storage capacity increases with the volume of the storage tanks and the electrolyte concentrations [8][9][10]. The all-VRFB can be modified for use in specific large-scale utility applications; however, these batteries are more expensive than secondary batteries. Consequently, the technologies of the all-VRFB system should be further advanced to realize low-cost fabrication and to improve performance and durability, thus enabling their widespread application. The key materials of the all-VRFB system include an ion exchange membrane, electrodes, electrolyte, and a flow channel. The battery's energy density is determined not only by the characteristics of these key materials but also by their intimate interrelationships. An ideal electrode is one that exhibits low resistance, high conductivity, and acid-resisting property. Many studies have focused on modifying electrode materials, such as increasing their electrocatalytic activity or developing a new, low-cost electrode with more robust mechanical strength in concentrated acid solutions [11][12][13][14][15]. In our previous research, we developed a novel semi-vanadium-iodine RFB (semi-V-I RFB) [16]. This battery has a high coulomb efficiency and significantly reduces the fabrication cost; however, its voltage efficiency is low (68%). The overall reactions of vanadium-iodine, iodine-vitamin C (C 6 H 8 O 6 ), and vanadium-vitamin C oxide (C 6 H 6 O 6 ) can be expressed as shown in equations (1)-(6) on the charging mode.

Vanadium-Iodine.
Oxidation reaction 2VO 2+ + 2H 2 O → 2VO 2 A study demonstrated that the reversibility of the oxidation-reduction reaction can be effectively increased if vitamin C is added to the iodine solution [17]. Therefore, the standard electromotive forces of the semi-V-I RFB system are between 0.46 to 0.94 V on the discharging mode.
This study focused on developing composite electrodes that can be used as the negative electrode of the semi-V-I RFB system to effectively improve the voltage efficiency of the system. Furthermore, we designed semi-V-I RFB stacks and compared the charge-discharge performance of the stacks with that of the all-VRFB system.

Experimental
2.1. Fabrication of Composite Electrodes. Graphite carbon paper electrodes (C-electrodes; Shenhe Carbon Fiber Materials Co. Ltd., Liaoning, China) were modified by depositing palladium (Pd) through electroless plating. The effective exposed surface area of the electrodes was 5 × 5 cm 2 . SnCl 2 and PdCl 2 were used as sensitized and active reagents in the deposition of the C-Pd electrode. The composition of the plating solution was 19 g/L Na 2 C 4 H 4 O 4 ·6H 2 O, 10 g/L PdCl 2 , 8.5 g/L HCl, and 25.6 g/L C 2 H 4 (NH 2 ) 2 . The C-Pd, C-Pd-TiO 2 , and C-TiO 2 -Pd electrodes were prepared through electroless metal deposition by using the composition of the plating solution [18][19][20][21]. Among them, the C-TiO 2 electrode was fabricated using the following steps: tetrabutyric acid was mixed with EtOH/HCl (pH = 1) aqueous solution at a tetrabutyric acid : EtOH : HCl molar ratio of 1 : 8 : 4 in a flask and mechanically stirred to carry out the hydrolysis reaction at room temperature. Then, carbon paper was added to the premixed tetrabutyric acid solution under vigorous stirring for 1 h. Finally, it was completely dried in an oven at 60°C, and a C-TiO 2 composite electrode was obtained. The C-Pd-TiO 2 and C-TiO 2 -Pd electrodes were fabricated using a two-step deposition method. In the first step, the sol-gel route was employed to prepare the C-TiO 2 -electrode, following which the Pd was redeposited on the C-TiO 2 electrode through the electroless technique to form the C-TiO 2 -Pd modified electrode. The C-Pd-TiO 2 electrode was fabricated under the same conditions, but with the process reversed. The plating solution had a pH in the range of 4.8-5.1 at 80°C. Finally, the modified electrodes were heat treated at 400°C in an oven for 1 h. All chemicals used were of analytical reagent grade.

Morphology and Composition of Composite Electrodes.
The surface morphologies and components of the composite electrodes were characterized through scanning electron microscopy coupled with energy dispersive X-ray analysis (JSM-6700F, JEOL, Tokyo, Japan). The composition of the electrode was identified through X-ray diffraction (XRD) with a filtered Cu-Ka radiation source (X'Pert PRO, PANalytical, Almelo, the Netherlands), and diffraction patterns were collected from 20°to 90°.

Measurement of the Electrochemical Properties.
Cyclic voltammetry (CV) experiments were conducted in a threeelectrode cell and by using an electrochemical analyzer (CHI 6273 C, CH Instruments Inc., US). The composition of the cell was as follows: an Ag-AgCl electrode was used as the quasireference electrode, the reference electrode and the platinum gauze were used as the counter electrode, and the C-Pd or C-TiO 2 -Pd modified electrode with a surface area of 1 × 1 cm 2 was used as the working electrode. The positive electrolyte was a solution of 0.01 M I 2 with a variable amount of vitamin C. The scan rate was varied from 1 to 10 mV·s −1 . Electrochemical properties, such as the potential interval (ΔE p ), ratio of the anodic-cathodic current (I pa -I pc ), diffusion coefficient (D o ), and double-layer charged capacity (C d ), were obtained from the CV experimental data.

Charge-Discharge Test.
We used the iodine-vitamin C solution as a cathode electrolyte and vanadium ion (V 4+ -V 5+ ) solution as an anodic electrolyte to set up a single cell. The (V 4+ -V 5+ ) vanadium solutions were prepared by dissolving 1 M VOSO 4 in 2 M H 2 SO 4 aqueous solution. Multiple electrodes, that is, C and C-TiO 2 -Pd electrode, were used. The electrodes were separated by a cationic exchange Nafion 117 membrane (DuPont, USA). In addition, the chargedischarge test was conducted on two cell stacks-series and parallel.

Crystal Structure Identification of Composite Electrodes.
The XRD diffraction patterns of all the composite electrodes, including C, C-TiO 2 , C-Pd, C-Pd-TiO 2 , and C-TiO 2 -Pd with and without 400°C sintering, are shown in Figure 1. The diffraction pattern in Figure 1 reveals that (1) before sintering, C (002) possessed two sharp diffraction peaks at 27°-29°and two weak peaks, C (100)/C (004), at 43°and 54°, respectively. TiO 2 in the C-TiO 2 region was amorphous; thus, the diffraction patterns were similar to those of C. The (111), (200), and (220) facets of Pd with a face-centered cubic lattice structure were assigned to the diffraction peaks at 40°, 47°, and 68°, respectively. Three regions (i.e., C-Pd, C-Pd-TiO 2 , and C-TiO 2 -Pd) can verify a part of the Pd crystal structure (111, 200); nevertheless, the intensity of the peaks was weak in the C-TiO 2 -Pd regions, as depicted in Figure 1(a). (2) After sintering, the C electrode's diffraction patterns remained unchanged. The C-TiO 2 electrode exhibited a strong diffraction peak at 25°, which was the anatase lattice structure of TiO 2 (A101), and a relatively weak diffraction peak at 38°-39°for TiO 2 (A103) and TiO 2 (A004). In the C-Pd-TiO 2 region, fingerprint peaks similar to those of TiO 2 were observed; however, these were not observed in the C-TiO 2 -Pd region. In particular, in the C-Pd and C-TiO 2 -Pd regions, a strong new diffraction peak appeared at 34°, which is the diffraction characteristic peak of PdO (101), implying that Pd oxidized to form PdO, as presented in Figure 1(b). (3) A significant difference was observed between the diffraction peaks of C-TiO 2 -Pd crystal structures with and without sintering. The intensity of the diffraction peaks of Pd (111) and Pd (200) for the sintering C-TiO 2 -Pd electrode was stronger and more stable, and the diffraction pattern of Pd (220) was more apparent than without sintering. Moreover, a strong new PdO (101) diffraction peak was observed at 34°in the C-TiO 2 -Pd regions but not in the C-Pd-TiO 2 region, as shown in Figure 1(c). Thus, the experimental results demonstrate that the stable Pd metals and PdO crystal structures for the C-TiO 2 -Pd electrode could be formed through high-temperature sintering. These may affect electrocatalytic activity and generate different electrochemical properties [22][23][24].

Surface Morphology Analysis of Composite Electrodes.
The scanning electron microscopy (SEM) images of the composite electrodes are presented in Figure 2. The Pd particles with white spots were deposited onto the carbon   Figure 2(c), implying that the inner Pd particles are covered by the outer TiO 2 layers. Additionally, we observed that the growth in TiO 2 was derived from the inner layer and layer-by-layer stacking of TiO 2 , and a significant number of TiO 2 grains were uniformly embedded into the carbon paper network matrix through the sol-gel method, as presented in Figure 2(b). Furthermore, to identify the component of the C-TiO 2 -Pd electrode, we performed energy-dispersive spectrometry (EDS) analysis of the C-TiO 2 -Pd electrode with sintering, as depicted in Figure 3. The spectrum and element analysis presented in Figure 3(a) reveals that the elements of the sintering C-TiO 2 -Pd electrode included C (47.58 wt%), Ti (17.64 wt%), Pd (2.71 wt%), and O (32.07 wt%). O elements exhibited a high wt% under a low wt% Ti element, implying that Pd was oxidized to form PdO. These results were confirmed because the PdO (101) facet was assigned to the diffraction peak at 34°, as presented in Figure 1(c). In addition, the Pd particles were deposited uniformly on the TiO 2 layers, as illustrated in the mapping analysis of Figures 3(b)-3(f). The deposition of Pd and PdO particles on the TiO 2 layers reduced the aggregation of Pd particles and increased the active area of the C-TiO 2 -Pd composite electrode.

Electrochemical Characteristics of Various Electrodes.
The CV curves of various electrodes in 10 mL of 0.01 M iodine with 1 wt% vitamin C solution at a 10 mV/s scanning rate in the range of 0-1 V are shown in Figure 4, and the data are summarized in Table 1. The oxidation pattern of electrolyte solutions is referred to as an oxidized or anodic current peak (I pa ) on a tested electrode, corresponding to the anodic voltage (E pa ) when CV is scanned from a low to high potential. Conversely, if it is scanned from a high to low potential, then the electrolyte is reduced to obtain a reduced or cathodic current peak (I pc ) and a cathodic voltage (E pc ). The I pa /I pc ratio tended to unity, revealing that the quasireversible redox reactions occurred because of the effects of the iodine electrolyte solutions on the tested electrodes. The values of I pa and I pc were low, meaning that the redox efficiency of the composite electrodes in the iodine solutions was poor. The I pa values were in the following order: C-TiO 2 -Pd > C-Pd > C-Pd-TiO 2 > C > C-TiO 2 . Moreover, the potential interval (ΔE p ) was in the following order: C-TiO 2 > C-Pd-TiO 2 > C-Pd > C > C-TiO 2 -Pd, as listed in Table 1.
The low ΔE p value demonstrated that the energy barriers of redox reactions were lower because of the effects of the iodine electrolyte solutions on the tested electrodes. These results imply that the C-TiO 2 -Pd electrode has higher redox currents and lower ΔE p than other electrodes. Moreover, the C-TiO 2 -Pd electrode had a symmetric I pa /I pc value, which was close to unity, meaning that this is a reversible redox reaction for the iodine solutions. In addition, the outer TiO 2 layer of the C-Pd-TiO 2 electrodes could not improve the redox kinetic reaction of the iodine solution; thus, the redox current of the C-Pd-TiO 2 electrode was lower 15    Journal of Nanomaterials than that of the C-TiO 2 -Pd electrode. Therefore, TiO 2 particles between a carbon matrix and Pd metal layers increase the active surface area of an electrode and can enhance the electrocatalytic effect of the electrode [19,22]. Consequently, the C-TiO 2 -Pd electrode performed better in the redox kinetic reaction than the others.   where i p , n, A, C, D o , and υ are the peak current, electron mobility number, reaction area (cm 2 ), concentration of reactants (mole/cm 3 ), diffusion coefficient (cm 2 /s), and scan rate (V/s), respectively. The kinetic reactions of iodine with vitamin C electrolyte solutions on various electrodes were studied through a series of CV experiments at different scan rates. The relationship between I pa and υ 1/2 was plotted in Figure 5, and the value of D o was obtained using the slope of I pa and υ 1/2 and Randles-Sevcik equations [16]. As shown in Table 2, the modified C-TiO 2 -Pd electrode increased the overall mass transfer and iodide ion electrochemical activity. The D o value of iodine-vitamin C species was enhanced by approximately 60% compared with that of the C-electrode, from 12.8 × 10 −6 to 20.4 × 10 −6 cm 2 /s, on the C-TiO 2 -Pd electrode.   Journal of Nanomaterials An electric double layer exists between the electrode and electrolyte. The electric double layer capacitor can be represented by equation (8).
where i cd , A, C d , and υ are the electric double layer charging current, reaction area (cm 2 ), electric double layer capacitor value (F/cm 2 ), and scan rate (V/s), respectively. On the basis of our previous research [16], the relationship between i cd and υ can be derived by forming a straight line with the slope proportional to C d in Figure 6 for the iodine-vitamin C electrolyte for the different electrodes. Table 2 summarizes the C d data for comparison. C d was the highest for the C-TiO 2 -Pd electrode in the iodine-vitamin C electrolyte. Moreover, C d increased from 0.132 to 0.249 F/cm 2 , 89% higher than the C d for the C-electrode.
3.5. Charge-Discharge Performance. Battery energy storage systems entail numerous design requirements, such as high voltage, current, power, and capacity. The capacity units of a large-scale energy storage battery system should meet the commercial operation requirements of the unit cells. Ordinarily, these batteries are fabricated in series or parallel to increase the storage capacity of the battery module and supply voltage group for achieving commercial requirements.
In this section, we fabricated the C-TiO 2 -Pd composite electrode as a negative electrode in the semi-V-I RFB system to design a single cell and series and parallel cell stacks and to evaluate the charge-discharge performance of the cell and stacks.    Figure 6: Plots of the anodic peak current (I pa ) versus the scan rate (υ) for the iodine-vitamin C electrolyte for various electrodes. 7 Journal of Nanomaterials 3.5.1. Single Cell. The RFB energy storage capacity is proportional to the volume and concentration of the electrolyte. The C-TiO 2 -Pd composite electrode is employed as a negative electrode in the semi-V-I RFB system with various concentrations of the electrolyte to conduct the charge-discharge tests. The charge-discharge curves are displayed in Figure 7, and performance of a single semi-V-I RFB is detailed in Table 3. The data in Table 3 verify that the energy storage capacity of the semi-V-I RFB system increases with the concentration of iodine-vitamin C solutions. The system exhibited optimal performance in the 1 M iodine solution; the coulomb efficiency (CE), voltage efficiency (VE), and energy efficiency (EE) were 96%, 81%, and 77%, respectively.
As specified in the literature [19,23,24], crystal structures of metals or alloys have different effects on the catalytic electrode. In our study, we observed that C-TiO 2 -Pd   Journal of Nanomaterials electrodes with or without sintering at a high temperature affect the battery energy storage efficiency. Figure 8 displays the charge-discharge diagram of C-TiO 2 -Pd electrodes with or without sintering in a semi-V-I RFB system, and the charge-discharge performance of the electrodes is detailed in Table 4; these data confirm that when the C-TiO 2 -Pd electrode is sintered at high temperatures, the intensity of crystalline lattice conformations of Pd (111) and Pd (200) is stronger than that without sintering; additionally, a new PdO (101) crystal compound appeared, as shown in Figure 1(c). The formation of Pd (111), Pd (200), and PdO may contribute to the electrocatalytic activity in semi-V-I-RFB systems, in turn causing the VE to increase from 81% to 87% and EE from 77% to 83%. Furthermore, a series of continued charge-discharge tests of the semi-V-I RFB system was performed at a current density of 20 mA cm −2 and is presented in Figure 9, which plots CE, VE, and EE as a function of cycle number. The average CE, VE, and EE values of the semi-V-I RFB system for 50 cycles illustrated in Figure 9 are approximately 96.52%, 84.12%, and 81.22%, respectively, revealing that the cyclic performance is excellent.

Battery Stacks.
To facilitate large-scale commercial applications, the stacks of the RFB system are often fabricated in series and parallel cross combinations for increasing the storage capacity and operating voltage of the battery module.
In the series configuration, the current supplied to the circuit is equal to the current across all elements in series (I = I 1 = I 2 = … = I N ), the total voltage is equal to the sum of each element voltage (E = V 1 + V 2 + … + V N ), and the total resistance is equal to the sum of each element resistance (R t = R 1 + R 2 + … + R N ). In the parallel configuration, the total current is equal to the sum of each element current (I t = I 1 + I 2 + … + I N ), the voltage supplied to the circuit is equal to the voltage across all elements in parallel (V t = V 1 = V 2 = … = V N ), and the reciprocal of the total resistance is equal to the sum of the reciprocal of each element resistance (1/R t = 1/R 1 + 1/R 2 + … 1/R N ). The charge and discharge curves of the all-VRFB stacks and the semi-V-I RFB stacks were composed of two cells in series and parallel, as shown in Figures 10(a) and 10(b), and the performances of the cells are summarized in Table 5. The results revealed that the charging voltage in the series stacks effectively increased from 1.8 to 3.6 V for the all-VRFB stacks, and the CE of the all-VRFB stacks significantly reduced to 73%. In the semi-V-I RFB system, a linear increase in the charging voltage from 1.2 to 2.4 V was observed. Moreover, the CE of the stacks remained at 92%, but the storage capacity of both systems was reduced. By comparison, in the series configuration, the overall efficiency value of the semi-VRFB system is higher than that of the all-VRFB system. In the parallel configuration, the capacity of the all-VRFB and semi-V-I RFB systems were altered from 139 to 137 mAh and from 147 to 162 mAh, respectively. The capacity of the semi-V-I RFB system improved slightly, but capacity of the all-VRFB system reduced. The charge-discharge performances of both systems attenuated significantly.   The capacity of the battery with the parallel combination cannot be increased; this may be due to the flow channel differences of the two batteries or due to an external line loss. If the voltage, inherent resistance, or current of the battery stacks are different, then a small voltage or battery resistance forms an electricity load and consumes part of the energy of another battery, thus decreasing the overall efficiency.

Conclusions
The C-TiO 2 -Pd composite electrode was fabricated using carbon paper as a substrate that was coated with titanium dioxide and palladium metal in sequence. The C-TiO 2 -Pd composite electrode exhibited superior electrochemical activity after high-temperature sintering. Moreover, the D o value of the iodine-vitamin C electrolyte solutions increased by approximately 60% more than that of the C-electrode. The use of C-TiO 2 -Pd as a negative electrode for the semi-V-I RFB system effectively enhanced the VE, thus improving the overall EE to 83%. In the series configuration, the charge voltage of the all-VRFB and semi-V-I RFB battery stacks linearly increased. However, performance of the all-VRFB battery stacks was significantly attenuated, and the EE decreased from 62% to 41%. By comparison, the semi-V-I RFB stacks can still be maintained at 67%. If we can use the design and assembly technology to enhance and improve the flow channel of the battery, the overall efficiency can be effectively improved. For series and parallel combinations obtained by stacking large modules, the proposed semi-V-I RFB system has high application potential to meet commercial requirements.

Data Availability
Previously reported experiment system was used to support this study and is available at https://www.hindawi.com/ journals/jchem/2017/4590952/. This prior study is cited within the text as reference [4].

Conflicts of Interest
The authors declare no conflicts of interest.