Balancing Osmotic Pressure of Electrolytes for ... - ACS Publications

Dec 6, 2016 - Balancing Osmotic Pressure of Electrolytes for Nanoporous. Membrane Vanadium Redox Flow Battery with a Draw Solute. Ligen Yan,. †...
0 downloads 0 Views 7MB Size
Research Article www.acsami.org

Balancing Osmotic Pressure of Electrolytes for Nanoporous Membrane Vanadium Redox Flow Battery with a Draw Solute Ligen Yan,† Dan Li,† Shuaiqiang Li,† Zhi Xu,‡ Junhang Dong,‡ Wenheng Jing,*,† and Weihong Xing† †

State Key Laboratory of Materials-Oriented Chemical Engineering, Nanjing Tech University, 5 Xinmofan Road, Nanjing 210009, P.R. China ‡ Department of Chemical Engineering, University of Cincinnati, Cincinnati, Ohio 45221, United States S Supporting Information *

ABSTRACT: Vanadium redox flow batteries with nanoporous membranes (VRFBNM) have been demonstrated to be good energy storage devices. Yet the capacity decay due to permeation of vanadium and water makes their commercialization very difficult. Inspired by the forward osmosis (FO) mechanism, the VRFBNM battery capacity decrease was alleviated by adding a soluble draw solute (e.g., 2-methylimidazole) into the catholyte, which can counterbalance the osmotic pressure between the positive and negative half-cell. No change of the electrolyte volume has been observed after VRFBNM being operated for 55 h, revealing that the permeation of water and vanadium ions was effectively limited. Consequently, the Coulombic efficiency (CE) of nanoporous TiO2 vanadium redox flow battery (VRFB) was enhanced from 93.5% to 95.3%, meanwhile, its capacity decay was significantly suppressed from 60.7% to 27.5% upon the addition of soluble draw solute. Moreover, the energy capacity of the VRFBNM was noticeably improved from 297.0 to 406.4 mAh remarkably. These results indicate balancing the osmotic pressure via the addition of draw solute can restrict pressure-dependent vanadium permeation and it can be established as a promising method for up-scaling VRFBNM application. KEYWORDS: vanadium redox flow batteries, capacity decay, osmotic pressure, water transfer, 2-methylimidazole



INTRODUCTION In recent years, various energy storage systems (ESS) for solar and wind technologies have been actively developed to guarantee a stable power output.1−3 Vanadium redox flow battery (VRFB) as one of ESS, had been first developed in 1985 and considered as one of the most promising large-scale energy storage devices owing to its flexible capacity design, rapid recharging, electrochemical reversibility, and long cycle life.4−6 However, the apparent decline of battery capacity caused by permeation of vanadium ions and water restricts the large-scale commercialization.7−9 Therefore, it is very importance to reduce the undesired permeation of ions and water molecules. Great efforts have been made to explore superior VRFB membranes combining low vanadium permeability with high proton conductivity, chemical stability, and low cost.10 In this sense, various types of membranes have been developed including ion-exchange, amphoteric, and nanoporous membranes.11−13 Perfluorosulfonic acid (PFSA) membranes, such as Nafion, have been employed as a separator in redox battery systems, yet they have drawbacks, such as relatively high crossover and high cost.14,15 Nanoporous membranes have recently gained increasing attention as alternative materials to replace the traditional ion exchange membranes.16 Unlike the ion-exchange membranes, the ions selectivity by nanoporous membranes is mainly ruled by pore size and Donnan exclusions.17 Since vanadium ions have larger Stokes radius © 2016 American Chemical Society

and higher valence state than that of proton, the nanoporous membranes preferably permeate protons vs vanadium ions.18 However, hydrocarbon polymer-derived membranes are at high risk from chemical oxidation due to V5+ ions.19 Ceramic nanoporous membranes are considered as more promising candidate for good chemical stability in harsh conditions in comparison with the organic nanoporous membranes. In these membranes, transportation of ions through the nanochannels dependent on the ion interactions is relevant to ionic Stokes radius and concentration/potential gradient.20 Transport mechanism of ions and water molecules through the nanochannels in membranes has been studied in order to understand the crossover of species in VRFB.21 The processes can be affected by diffusion, convection, and electric migration. Relevant osmotic pressure and operating conditions could impact on the ions transfer rate, which change the concentration and potential gradients.22 Since the crossoverinduced osmotic pressure accelerates the vanadium leakage and water transfer during the charge−discharge processes, a reverse pressure was recommended as a counterbalance force in order to weaken the capacity fade.23 However, it is very difficult to accurately control such additional pressure since the osmotic Received: September 22, 2016 Accepted: December 6, 2016 Published: December 6, 2016 35289

DOI: 10.1021/acsami.6b12068 ACS Appl. Mater. Interfaces 2016, 8, 35289−35297

Research Article

ACS Applied Materials & Interfaces

Figure 1. SEM images of the TiO2 membrane: (a) surface and (b) cross-section. (c) PEG rejection curve of the TiO2 membrane.

preferably reject vanadium ions due to the steric and electrostatic exclusion.28,29 The permeability of vanadium ions through the fabricated TiO2 membrane was measured and compared with Nafion (117) membrane (Figure 2). Concen-

pressure varies with the concentration of vanadium ions during charge−discharge processes. On the other hand, introducing an asymmetric current has been attempted to control the capacity decay yet it is at the expense of voltage efficiencies reduction.24 It is clear that developing a convenient approach to alleviate the capacity decay in VRFB is the key point now. As discussed above, the problem of battery capacity decay can be solved by narrowing the difference of osmotic pressure between the electrolytes from the two sides of the cell. It is interesting to see that a draw solute is added in forward osmosis (FO) process in order to generate an osmotic pressure which drives water passing through the membrane. Similarly, a draw solute is employed in our approach to counterbalance the osmotic pressure so that the water transfer and vanadium crossover can be lightened. A typical FO draw solute 2methylimidazole was added into the VRFB electrolytes. TiO2 nanoporous membrane with mean pore size 2.7 nm was employed as a VRFB separator and its performance was compared to that of Nafion membrane. Performance of the single cell and its long-term operative stability will be discussed in this article.



Figure 2. Concentration of VO2+ in the permeate side of Nafion and TiO2 membranes.

RESULTS AND DISCUSSION Characterization on TiO2 Nanoporous Membrane. TiO2 nanoporous thin film was synthesized on macroporous Al2O3 support with weak alkaline polymeric sol−gel route.25 The TiO2 layer had an excellent acid resistance (Supporting Information, Figure S1). Morphology of the layer was examined with scanning electron microscope (SEM) and its separation performance was tested in filtration experiment (Figure 1). As shown in Figure 1, the prepared TiO2 layer has a thickness ca. 2 μm and its surface is homogeneous and smooth. Flux of pure water through the bared support and the TiO2-coated Al2O3 membrane was reduced from 9.3 to 2.1 L·m−2·h−1·bar−1. The supported TiO2 thin film performed a cutoff to molecule weight (MW) up to 2700 Da of polyethylene glycol (PEG), which gives information that the average pore size of the layer is 2.7 nm based on the equation below.26,27

tration of vanadium ions in the permeate side of TiO2 membrane was remarkably lower than that of Nafion. Moreover, concentration of the permeated vanadium ions is almost linearly correlated to diffusion time and the permeability coefficient (P) can be calculated by the following equation:30,31 VR

dC R (t ) P = A [C L − C R (t )] dt L

(2)

where VR is the volume of the solution, CR(t) is the vanadium ion concentration at the permeate side, and CL is the vanadium ion concentration of the feed solution. A and L are the area and thickness of the membrane, respectively. The permeability coefficient of the Nafion membrane (1.47 × 10−6 cm2 min−1) was 2 orders of magnitude higher than that of the TiO2 membrane (0.006 × 10−6 cm2 min−1). As a result, it has been proved that the TiO2 membrane can serve as a suitable separator for the interested ions in the VRFB cell. Subsequently, both the TiO2 and Nafion membranes were respectively equipped in the VRFB cell (denoted as T-VRFB and N-VRFB) and their performance were compared. The schematic graph of VRFB cell is presented in Figure 3. The cell was operated at current density 30 mA cm−2 during the charge−discharge cycles. In particular, both the anolyte and the catholyte (volume of 7.5 mL each) were designed to make the

° = 0.262 × (MW(gmol−1))1/2 − 0.3 Molecular radius (A) (1)

Permeability of Vanadium Ion in TiO2 Nanoporous Membrane. TiO 2 membrane is a typical amphoteric membrane whose surface can be protonated and thus intrinsically behaves as anion exchange membrane under strong acid conditions. The prepared TiO2 membrane has average pore size 2.7 nm which is comparable to the dimension of vanadium ions. Consequently, the TiO2 membrane could 35290

DOI: 10.1021/acsami.6b12068 ACS Appl. Mater. Interfaces 2016, 8, 35289−35297

Research Article

ACS Applied Materials & Interfaces

Figure 3. Schematic diagram of the VRFB setup.

capacity decay more significant. Open circuit voltage (OCV) of the VRFB was monitored at room temperature after being charged to 1.75 V (Figure 4).

As seen in Figure 4, the OCV slightly decreased from 1.6 to 1.2 V as the T-VRFB cell being operated for 74 h. On the other hand, the same drop of OCV was obtained within shorter duration as 38 h for the N-VRFB cell. These results have further confirmed that the TiO2 membrane could maintain a relatively low leakage of vanadium ions and thus a reduced selfdischarge. Performance of VRFB with TiO2 Nanoporous Membrane. The capacity decay of the cell was recorded over many cycling processes and the results are given in Figure 5. Substantial capacity drop occurred for both cells after 70 cycles: 60.7% decay from 297 to 116.7 mAh for T-VRFB and 79.1% from 294.8 to 61.5 mAh for N-VRFB indicating a serious vanadium leakage. At the meantime, Coulombic efficiency (CE) and energy efficiency (EE) of the flow cells were measured. Both the CE and EE of VRFB remained stable over 70 cycles with current density 30 mA cm−2 (Figure 6). The TVRFB cell performed CE as 93.5% which was slightly higher than that of N-VRFB (92.7%) owing to relatively low vanadium permeability. However, T-VRFB presented lower EE in comparison to N-VRFB which might be attributed to different

Figure 4. Open circuit voltage (OCV) of VRFB cell containing TiO2 and Nafion 117 membranes.

Figure 5. Capacity decay of VRFB cells containing (a) TiO2 and (b) Nafion117 membranes. 35291

DOI: 10.1021/acsami.6b12068 ACS Appl. Mater. Interfaces 2016, 8, 35289−35297

Research Article

ACS Applied Materials & Interfaces

Figure 6. Efficiency of VRFB cell containing (a) TiO2 and (b) Nafion117 membranes.

could be suppressed as the water flux and vanadium crossover being alleviated with an extra pressure in a reverse direction. Effects of 2-Methylimidazole for VRFB. Based on the previous discussion, 2-methylimidazole was added in the catholyte as a typical draw solute to conquer the enforced osmotic pressure and then it could limit the capacity decline. Neither the presence of new substances nor any change in the electrolyte concentration was expected upon the addition of 2methylimidazole. Accordingly, UV/vis spectra were measured for V(II) and V(III) electrolytes with and without 2methylimidazole (Figure 9) with the concentration of V(II) and V(III) electrolytes as 0.02 M in 2 M H2SO4 in both situations. Characteristic absorption bands of V(II) and V(III) ions can be observed in the spectra: there is neither new absorption peak nor wavelength shift being detected. Then it excludes the formation of new substances or the potential change in negative electrolytes. Therefore, the constant property of the electrolytes has been maintained after adding the trace amounts of 2-methylimidazole. To further investigate the effect of 2-methylimidazole on electrode reaction, cyclic voltammetry (CV) measurements were carried out in standard electrolyte containing a mixture of 2 M VOSO4 and 2 M H2SO4, meanwhile, varying the concentration of 2-methylimidazole. It is clear that the peak potential separations (ΔV) has been barely influenced by the presence of 2-methylimidazole as seen in Figure 10. The V4+− V5+ and V3+−V2+ oxidation and reduction peaks appeared at approximately 1.17−0.78 V, and −0.37 to −0.75 V, respectively (vs Ag/AgCl).36 On the other hand, adding 2-methylimidazole could noticeably affect the reaction kinetics at electrode and the electrochemical reversibility of the redox couples. As a result, the peak current for V4+−V5+ has reduced from 0.035 to 0.025 A cm−2 (i.e., lower reaction activity at electrode) in the presence of 1 wt% of 2-methylimidazole. It may be attributed to the decreased diffusion coefficient of vanadium ions due to addition of 2-methylimidazole, since the electrolyte polarization has been enhanced on the surface of the electrode (confirmed by an increase of viscosity from 5.76 to 6.28 cP). The peak current first reached a maximum value and then decreased afterward after adding 2-methylimidazole. With 1 wt% of the draw solute the peak current and the electrode reaction were only slightly reduced and thus this concentration of 2methylimidazole was selected as the optimum adding amount. To follow up, the water transfer was studied when 2methylimidazole was added during the cycling process. Adding

thickness of the two membranes. This aspect will be discussed in detail as bellow. For comparison, the vanadium ion permeability of Al2O3 support (Supporting Information, Figure S2) and efficiency of VRFB with Al2O3 support (Supporting Information, Figure S3) were also investigated. The bared Al2O3 support showed a relatively high permeation rate in contrast to the TiO2-coated one. In addition, the CE and EE of battery with Al2O3 support was much lower than that of TiO2-coated membrane. It can be concluded that the improved performance of battery was mainly owing to the TiO2 coating (mean pore size 2.7 nm). Decay Mechanism. With the aim to get insight into the mechanism of capacity decay, the electrolyte volume was measured and compared. As depicted in Figure 7, the volume of

Figure 7. Volume changes of electrolytes.

catholyte was reduced by nearly 25% (from 7.5 to 5.5 mL) after 45 h charge−discharge process, while the volume of anolyte increased by the same amount (7.5 to 9.5 mL). It suggests that accumulation of water and vanadium ions taking place at the positive side causes the significant decay of the battery capacity. In general, crossover of vanadium ions is inevitable during the cycling processes and the flux of pure water also accompanies the crossover. Water flux is relevant to the number of water molecules that are bonding to vanadium ions of different valences.32,33 Both the water flux and the vanadium crossover are enlarging the osmotic pressure in the cell which aggravates the fading phenomenon (Figure 8). It is demonstrated that 75% of the transferred water contributes to the difference of osmotic pressure during the charge− discharge process.34,35 Correspondingly, the capacity decay 35292

DOI: 10.1021/acsami.6b12068 ACS Appl. Mater. Interfaces 2016, 8, 35289−35297

Research Article

ACS Applied Materials & Interfaces

Figure 8. Illustration of the draw solute balancing the osmotic pressure during the charge−discharge process.

Figure 9. UV/vis absorption spectra of electrolytes solution: (a) V(II)−H2SO4 solution, and (b) V(III)−H2SO4 with and without (pristine) 2methylimidazole.

Figure 10. Cyclic voltammetry of a standard electrolyte and mixed electrolyte solutions.

Figure 11. Volume changes of the electrolytes upon addition of 1 wt% of 2-methylimidazole.

electrolyte volume in both the half-cells was close to zero after 55 h charge−discharge process. The presence of 2-methylimidazole in the catholyte was effective for decreasing the net flux of water, hence, introducing 2-methylimidazole into the electrolyte could reduce the difference of osmotic pressure and the water transfer during the battery operation. The apparent energy capacity decay of VRFB due to vanadium ions crossover is one of the major technical barriers up-to-date that prevents these devices entering into a broader market.39 For practical use, it is important to maintain a longterm cycling stability and the additive substance was determined to be used in order to control the permeability of vanadium ions. The permeation rate of vanadium ions has been remarkably limited in the modified catholyte (Figure 12). The

1 wt% of 2-methylimidazole into the catholyte resulted in an extra driving force of 0.453 MPa based on the formula:37,38 π = βCRT

(3)

The osmotic pressure π is the osmotic pressure (bar) dependent on ionic strength, β is the Van’t Hoff factor, C is the concentration (mol L−1), and R is the universal gas constant (0.0831 L bar mol−1·K−1). This additional pressure can counterbalance the osmotic pressure that has been developed by vanadium crossover. No change of the electrolyte volume should be made when the counterbalance pressure is imposed. The volumes of electrolytes were investigated after adding 1 wt% of 2-methylimidazole, whose results are given in Figure 11. The change of the 35293

DOI: 10.1021/acsami.6b12068 ACS Appl. Mater. Interfaces 2016, 8, 35289−35297

Research Article

ACS Applied Materials & Interfaces

stability of T-VRFB were significantly enhanced upon addition of 2-methylimidazole. To further justify these findings, the capacity loss of N-VRFB was also measured. Upon addition of 2-methylimidazole, the capacity of N-VRFB after 70 cycles reduced from 407.9 to 245.1 mAh and the total loss dropped by 39.9% (Figure 14b). Its initial capacity similarly increased from 294.8 to 407.9 mAh. Based on the above analysis, it is not difficult to speculate that a more severe capacity loss would be occurred when 2-methylimidazol was added into the anolyte by aggravate the osmotic pressure difference. To verify this, the cell long-term performance with adding 2-methylimidazol into the anolyte was investigated (Supporting Information, Figure S4). A more serious capacity decay occurred after only 30 cycles (77.3%, from 297.6 to 67.5 mAh; and 78.8% from 294.2 to 62.3 mAh, for T-VRFB and N-VRFB cells, respectively), which was indicative of a serious vanadium leakage. It can be attributed to that the generated osmotic pressure by 2methylimidazol further exacerbated the osmotic pressure difference, which caused a more serious vanadium leakage and enhanced the capacity decay. In a summary, the capacity decay can be reduced using the modified the catholyte which counterbalances the osmotic pressure and reduces the vanadium leakage. At the meantime, efficiency of VRFB was also investigated in the presence of 2-methylimidazole (Figure 15). Both CE and EE remained stable for 70 cycles at a current density 30 mA cm−2. The CE of T-VRFB and N-VRFB slightly improved (95.3% and 94.7%, respectively) upon addition of 2methylimidazole. It is interesting to find that the addition of 2-methylimidazole in VRFB caused some fluctuations in CE and EE with comparison to the blank solution. The interactions between the additives and the vanadium ions in the solution require some time to equilibrate, which can explain the observed fluctuations in CE and EE during the initial cycles. After a few cycles, as a dynamic equilibrium between vanadium ions and additives is reached the Coulombic and energy efficiencies remained constant. In the situation of the cells without any additive, T-VRFB showed lower EE than that of NVRFB. To get insight into the behavior, the electrochemical impedances of T-VRFB and N-VRFB were compared. As seen in Figure 16, the ohmic resistance of the TiO2 membrane, the Al2O3 substrate, and the Nafion117 membrane containing a pristine catholyte were 1.29, 1.11, and 0.64 Ω, respectively. The ohmic resistance slightly increased after adding 2-methylimidazole (1.47 and 0.81 Ω for the cells containing TiO2 and Nafion117 membranes, respectively). The resistance increase can be explained with a higher viscosity of the modified catholyte than that of its pristine counterpart, which leads to higher internal resistance values. Therefore, VRFB containing the modified catholyte performed slightly low efficiencies as EE and VE (Supporting Information, Figure S5). Based on these results, it can be concluded that the energy capacity of VRFB can be effectively improved and maintained over many cycles upon the addition of 2-methylimidazole to the catholyte. It is very beneficial for the long-term application of VRFB devices.

Figure 12. Effect of the addition of 2-methylimidazole on the concentration of VO2+ in the permeate side.

permeability of VO2+ thorough Nafion and TiO2 membranes was both reduced (12% and 28% of their original values respectively). Taking water transfer into account, the low crossovers of vanadium ions could be caused by counterbalancing the osmotic pressure with addictive 2-methylimidazole which slowed down the water flux and inhibited the vanadium corrover. With the aim to further verify that 2-methylimidazole is effective in reducing the vanadium leakage, OCV experiments were carried out over the VRFB (Figure 13). As observed from

Figure 13. Effect of the addition of 2-methylimidazole on the OCV of VRFB containing TiO2 and Nafion117 membranes.

Figure 13, the addition of 2-methylimidazole into negative electrolyte solution lead to longer OCV duration for T-VRFB (25 h) and N-VRFB (12 h). The self-discharge in VRFB devices normally depends on the crossover of vanadium ions through the membrane.40−42 The self-discharge behavior can be managed in the presence of 2-methylimidazole as conquering the osmotic pressure and thus it reduces the vanadium leakage. As expected, the capacity of T-VRFB using the modified the catholyte decayed from 406.4 to 294.6 mAh after 70 cycles (Figure 14a). Such capacity loss was dramatically reduced by 27.5% in comparison to the original design. Moreover, it should be noted that the initial capacity of T-VRFB (i.e., after the first charging) containing 2-methylimidazole was remarkably improved from 297.0 to 406.4 mAh thanks to the vanadium crossover inhibition. The performance and long-term operating



CONCLUSIONS In a summary, we successfully used herein the FO mechanism to improve the capacity retention of VRFB by counterbalancing the osmotic pressure between the positive and negative electrolytes thanks to the addition of a draw solute (2methylimidazole). Volume change of electrolytes was near to 35294

DOI: 10.1021/acsami.6b12068 ACS Appl. Mater. Interfaces 2016, 8, 35289−35297

Research Article

ACS Applied Materials & Interfaces

Figure 14. Effect of addition of 2-methylimidazole on the capacity decay of VRFB containing (a) TiO2 and (b) Nafion117 membranes.

Figure 15. Effect of addition of 2-methylimidazole on the CE and EE of VRFB containing (a) TiO2 and (b) Nafion117 membranes.



EXPERIMENTAL SECTION

Materials. Alumina powder (Al2O3) was obtained by Nanjing Jiusi High-tech Co., Ltd. VOSO4·3H2O (analytical reagent) was supplied by Nanjing Kangmanlin Chemical Industry Co., Ltd. Sulfuric acid (98%) and magnesium sulfate (MgSO4) were provided by Shanghai Lingfeng Chemical Reagent Co., Ltd. Polyethylene glycol (PEG) was purchased from Sigma-Aldrich and used as received. The Nafion117 membrane and 2-methylimidazole were provided from Du Pont and Sinopharm Chemical Reagent Co., Ltd., respectively. Preparation of the TiO2 Membrane. The synthesized TiO2 membrane was comprised of a separation TiO2 layer and a macroporous Al2O3 substrate (thickness 1.2 mm), a diameter of 2.5 cm, an average pore size of 0.1 μm, and porosity of 32%. The separation TiO2 layer was synthesized using a weak alkaline polymer sol−gel route with a molar ratio of 1(Ti):1(acetylacetone):6(dimethylformamide):3(H2O):0.033(PEO-PPO-PEO, P123). Characterization. Surface morphology and thickness of the membranes were observed using FE-SEM (Hitachi S4800). Permeability of the membrane was measured with flux of deionized water at transmembrane pressure ca. 0.5 MPa. Separation capability of the membranes was evaluated by using 3 g L−1 PEG solution with various molecular weight as 10000, 6000, 1500, and 600 Da at transmembrane pressure ca. 0.5 MPa. The concentration and molecular weight of PEG in the feed and permeate solutions were analyzed with Gel Permeation Chromatography (Waters Co., Milford, MA). Vanadium Ion Permeation Measurements. Measurement for ion permeation was carried out using laboratory-scale experimental setup. The feed was filled with 4/7 M VOSO4 in 4/7 M H2SO4 solution, and 1 M MgSO4 in permeate solution was used to balance the ion interactions and to minimize the gap of osmotic pressure. The exposed area of membrane to solution was 3.14 cm2 and the volume of solution in both sides was 30 mL. Both solutions were continuously stirred by pumping at room temperature. MgSO4 solution was

Figure 16. EIS measurements for VRFB equipped with different membranes.

zero between the two half-cells in the presence of 2methylimidazole. Both the water transfer and the vanadium crossover were significantly suppressed correspondingly. In addition, the modified catholyte can lead to further performance improvements (e.g., excellent long-term cycling stability, higher capacity and Coulombic efficiencies and longer OCV times). The research highlights an approach with low-cost, effective, benign, and good performance to alleviate the capacity decay in VRFB devices. It can be anticipated that promising VRFB devices can be developed by means of this technology for various energy storage applications. 35295

DOI: 10.1021/acsami.6b12068 ACS Appl. Mater. Interfaces 2016, 8, 35289−35297

Research Article

ACS Applied Materials & Interfaces Notes

analyzed with ICP-AES at interval times and the concentration of vanadium ions in the solution was determined. Electrolyte Volume Change Measurements. Throughout the charge−discharge testing process, the VRFB single cell was maintained at current density 30 mA cm−2. The initial volumes of the electrolytes were both 7.5 mL and the changes of electrolytes volume were recorded as a function of time during VRFB cycles. Electrochemical Measurements. Cyclic voltammetry measurements were conducted in a three-electrode system on the electrochemical workshop (GAMRY Instruments Reference 3000) at a scan rate 5 mV s−1 in the range between −1.0 and 1.6 V at room temperature. The system consisted of a graphite felt (1 cm2) as a working electrode, a platinum sheet (4 cm2) as a counter electrode, and a saturated Ag/AgCl electrode along with a salt bridge (filled with a saturated potassium chloride solution) as a reference electrode. The graphite felt was first heated at 400 °C for 2 h and then treated in 98% sulfuric acid for 5 h before each test. Electrochemical Impedance Spectroscopy was also conducted on the electrochemical workstation (GAMRY Instruments Reference 3000). The sinusoidal excitation voltage applied to the cells was 0.01 V, whose frequency ranged from 10 to 1000 kHz. Single Cell Charge−Discharge Tests. The charge−discharge tests were performed in a VRFB single dynamic cell, which assembled a membrane as separator, two pieces of polyacrylonitrile (PAN)-based graphite felt as electrodes (area of 3.14 cm2, SGL GroupThe Carbon Company) and Ti plates as current collectors. All of the mentioned components were sealed with silicon rubber. Positive electrolyte was 7.5 mL of 2 M V(IV) in 2 M H2SO4 and negative electrolyte was 7.5 mL of 2 M V(III) in 2 M H2SO4 initially. The electrolyte was periodically pumped into the corresponding half-cell with a peristaltic pump (Baoding Longer Precision Pump Co., Ltd., China) and flowed through the electrode compartment during operation. The cell tests were conducted on BTS-5 V/6A battery test system (Shenzhen Neware Co., Ltd., China) in the working voltage between 0.7 and 1.75 V at constant current density 30 mA cm−2. OCV Measurements. The VRFB single cell was assembled as sandwiching the membrane between two pieces of graphite carbon electrodes and using 2 M V(IV)/V(V) and V(II)/V(III) in a 2 M H2SO4 solution as electrolytes in the positive and negative half cells, respectively. The Nafion and TiO2 membranes effective area were both 3.14 cm2 and the volume of the electrolyte was 7.5 mL. The battery was first charged to 1.75 V with a current density of 30 mA cm−2 and then the OCV was measured at room temperature. The experiments involving 2-methylimidazole in the negative electrolyte were also carried out using the same method.



The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work is financially supported by the National Natural Science Foundation of China (21676139), the Higher Education Natural Science Foundation of Jiangsu Province (15KJA530001), and the Project of Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.6b12068. Wide-angle X-ray diffraction (WAXRD) patterns of TiO2 and TiO2 immersed in 2.0 M H2SO4; concentration of VO2+ in the permeate side of Al2O3 support and TiO2 membrane; efficiency of the battery with Al2O3 support; and effect of addition of 2-methylimidazole on the VE of VRFB containing TiO2 and Nafion117 membranes (PDF)



REFERENCES

(1) Soloveichik, G. L. Flow Batteries: Current Status and Trends. Chem. Rev. 2015, 115, 11533−11558. (2) Yuan, Z. Z.; Duan, Y. Q.; Zhang, H. Z.; Li, X. F.; Zhang, H. M.; Vankelecom, I. Advanced Porous Membranes with Ultra-high selectivity and Stability for Vanadium Flow Batteries. Energy Environ. Sci. 2016, 9, 441−447. (3) Yuan, Z. Z.; Zhu, X. X.; Li, M. R.; Lu, W. J.; Li, X. F.; Zhang, H. M. A Highly Ion-selective Zeolite Flake Layer on Porous Membranes for Flow Battery Applications. Angew. Chem., Int. Ed. 2016, 55, 3058− 3062. (4) Skyllas-Kazacos, M. Novel Vanadium Chloride/Polyhalide Redox Flow Battery. J. Power Sources 2003, 124, 299−302. (5) Han, P. X.; Yue, Y. H.; Liu, Z. H.; Xu, W.; Zhang, L. X.; Xu, H. X.; Dong, S. M.; Cui, G. L. Graphene Oxide Nanosheets/Multi-walled Carbon Nanotubes Hybrid as an Excellent Electrocatalytic Material Towards VO2+/VO2+ Redox Couples for Vanadium Redox Flow Batteries. Energy Environ. Sci. 2011, 4, 4710−4717. (6) Ponce de Leon, C.; Frías-Ferrer, A.; González-García, J.; Szánto, D. A.; Walsh, F. C. Redox Flow Cells for Energy Conversion. J. Power Sources 2006, 160, 716−732. (7) Zhao, Y. Y.; Li, M. R.; Yuan, Z. Z.; Li, X. F.; Zhang, H. M.; Vankelecom, I. F. Advanced Charged Sponge-like Membrane with Ultrahigh Stability and Selectivity for Vanadium Flow Batteries. Adv. Funct. Mater. 2016, 26, 210−218. (8) Kim, K. J.; Park, M. S.; Kim, J. H.; Hwang, U.; Lee, N. J.; Jeong, G.; Kim, Y. J. Novel Catalytic Effects of Mn3O4 for All Vanadium Redox Flow Batteries. Chem. Commun. 2012, 48, 5455−5457. (9) Sevov, C. S.; Brooner, R. E. M.; Chenard, E.; Assary, R. S.; Moore, J. S.; Rodriguez-Lopez, J.; Sanford, M. S. Evolutionary Design of Low Molecular Weight Organic Anolyte Materials for Applications in Nonaqueous Redox Fow Batteries. J. Am. Chem. Soc. 2015, 137, 14465−14472. (10) Sun, C. X.; Chen, J.; Zhang, H. M.; Han, X.; Luo, Q. T. Investigations on Transfer of Water and Vanadium Ions across Nafion Membrane in an Operating Vanadium Redox Flow Battery. J. Power Sources 2010, 195, 890−897. (11) Kim, K. J.; Lee, S. W.; Yim, T.; Kim, J. G.; Choi, J. W.; Kim, J. H.; Park, M. S.; Kim, Y. J. A New Strategy for Integrating Abundant Oxygen Functional Groups into Carbon Felt Electrode for Vanadium Redox Flow Batteries. Sci. Rep. 2014, 4, 6906. (12) Chen, H. N.; Lu, Y. C. A High-energy-density Multiple Redox Semi-solid-liquid Flow Battery. Adv. Energy Mater. 2016, 6, 1502183. (13) Xu, Z.; Michos, I.; Wang, X. R.; Yang, R. D.; Gu, X. H.; Dong, J. H. A Zeolite Ion Exchange Membrane for Redox Flow Batteries. Chem. Commun. 2014, 50, 2416−2419. (14) Wang, W.; Sprenkle, V. Redox Flow Batteries go Organic. Nat. Chem. 2016, 8, 204−206. (15) Liu, S.; Wang, L. H.; Li, D.; Liu, B. Q.; Wang, J. J.; Song, Y. L. Novel Amphoteric Ion Exchange Membranes by Blending Sulfonated Poly(ether ether ketone)/Quaternized Poly(ether imide) for Vanadium Redox Flow Battery Applications. J. Mater. Chem. A 2015, 3, 17590−17597. (16) Zhang, H. Z.; Zhang, H. M.; Li, X. F.; Mai, Z. S.; Zhang, J. L. Nanofiltration (NF) Membranes: The Next Generation Separators for All Vanadium Redox Flow Batteries (VRBs). Energy Environ. Sci. 2011, 4, 1676−1679.

AUTHOR INFORMATION

Corresponding Author

* E-mail: [email protected]. Phone: +86-25-8358 9136. Fax: +86-25-8317 2292. ORCID

Junhang Dong: 0000-0002-4393-4968 Wenheng Jing: 0000-0002-1919-9932 35296

DOI: 10.1021/acsami.6b12068 ACS Appl. Mater. Interfaces 2016, 8, 35289−35297

Research Article

ACS Applied Materials & Interfaces (17) Zhang, H. Z.; Zhang, H. M.; Li, X. F.; Mai, Z. S.; Wei, W. P. Silica Modified Nanofiltration Membranes with Improved Selectivity for Redox Flow Battery Application. Energy Environ. Sci. 2012, 5, 6299−6303. (18) Oriji, G.; Katayama, Y.; Miura, T. Investigations on V(IV)/V(V) and V(II)/V(III) Redox Reactions by Various Electrochemical Methods. J. Power Sources 2005, 139, 321−324. (19) Mohammadi, T.; Skyllas-Kazacos, M. Evaluation of the Chemical Stability of Some Membranes in Vanadium Solution. J. Appl. Electrochem. 1997, 27, 153−160. (20) Li, D.; Jing, W. H.; Li, S. Q.; Shen, H.; Xing, W. H. Electric Field-controlled Ion Transport in TiO2 Nanochannel. ACS Appl. Mater. Interfaces 2015, 7, 11294−11300. (21) Lee, J. Y.; Wang, Y. N.; Tang, C. Y.; Huo, F. W. Mesoporous Silica Gel−based Mixed Matrix Membranes for Improving Mass Transfer in Forward Osmosis: Effect of Pore Size of Filler. Sci. Rep. 2015, 5, 16808. (22) Logan, B. E.; Elimelech, M. Membrane-based Processes for Sustainable Power Generation using Water. Nature 2012, 488, 313− 319. (23) Li, W. Y.; Zhang, Z. Y.; Tang, Y. B.; Bian, H. D.; Ng, T. W.; Zhang, W. J.; Lee, C. S. Graphene-ganowall-decorated Carbon Felt with Excellent Electrochemical Activity Toward VO2+/VO2+ Couple for All Vanadium Redox Flow Battery. Adv. Sci. 2016, 3, 1500276. (24) Park, M.; Ryu, J.; Kim, Y.; Cho, J. Corn Protein-derived Nitrogen-doped Carbon Materials with Oxygen-rich Functional Groups: A Highly Efficient Electrocatalyst for All-vanadium Redox Flow Batteries. Energy Environ. Sci. 2014, 7, 3727−3735. (25) Cao, X. P.; Li, D.; Jing, W. H.; Xing, W. H.; Fan, Y. Q. Synthesis of Visible-light Responsive C, N and Ce Co-Doped TiO2 Mesoporous Membranes via Weak Alkaline Sol-gel Process. J. Mater. Chem. 2012, 22, 15309−15315. (26) Li, B.; Nie, Z. M.; Vijayakumar, M.; Li, G. S.; Liu, J.; Sprenkle, V.; Wang, W. Ambipolar Zinc-polyiodide Electrolyte for A Highenergy Density Aqueous Redox Flow Battery. Nat. Commun. 2015, 6, 6303. (27) Puhlfürß, P.; Voigt, A.; Weber, R.; Morbé, M. Microporous TiO2 Membranes with a Cut Off < 500 Da. J. Membr. Sci. 2000, 174, 123−133. (28) Li, B.; Gu, M.; Nie, Z. M.; Wei, X. L.; Wang, C. M.; Sprenkle, V.; Wang, W. Nanorod Niobium Oxide as Powerful Catalysts for an All Vanadium Redox Flow Battery. Nano Lett. 2014, 14, 158−165. (29) Badrinarayanan, R.; Zhao, J. Y.; Tseng, K. J.; Skyllas-Kazacos, M. Extended Dynamic Model for Ion Diffusion in All-vanadium Redox Flow Battery Including the Effects of Temperature and Bulk Electrolyte Transfer. J. Power Sources 2014, 270, 576−586. (30) Park, M.; Jeon, I. Y.; Ryu, J.; Baek, J. B.; Cho, J. Exploration of the Effective Location of Surface Oxygen Defects in Graphene-Based Electrocatalysts for All-Vanadium Redox Flow Batteries. Adv. Energy Mater. 2015, 5, 1401550. (31) Lin, C. H.; Yang, M. C.; Wei, H. J. Amino-silica Modified Nafion Membrane for Vanadium Redox Flow Battery. J. Power Sources 2015, 282, 562−571. (32) Schwenzer, B.; Zhang, J. L.; Kim, S.; Li, L. Y.; Liu, J.; Yang, Z. G. Membrane Development for Vanadium Redox Flow Batteries. ChemSusChem 2011, 4, 1388−1406. (33) Wu, C. X.; Lu, S. F.; Wang, H. N.; Xu, X.; Peng, S. K.; Tan, Q. L.; Xiang, Y. A Novel Polysulfone−polyvinylpyrrolidone Membrane with Superior Proton-to-vanadium Ion Selectivity for Vanadium Redox Flow Batteries. J. Mater. Chem. A 2016, 4, 1174−1179. (34) Wei, X. L.; Xu, W.; Vijayakumar, M.; Cosimbescu, L.; Liu, T. B.; Sprenkle, V.; Wang, W. TEMPO-based Catholyte for High-energy Density Nonaqueous Redox Flow Batteries. Adv. Mater. 2014, 26, 7649−7653. (35) Pan, H. L.; Wei, X. L.; Henderson, W. A.; Shao, Y. Y.; Chen, J. Z.; Bhattacharya, P.; Xiao, J.; Liu, J. On the Way Toward Understanding Solution Chemistry of Lithium Polysulfides for High Energy Li−S Redox Flow Batteries. Adv. Energy Mater. 2015, 5, 1500113.

(36) Zhang, B. G.; Zhang, E. L.; Wang, G. S.; Yu, P.; Zhao, Q. X.; Yao, F. B. Poly(phenyl sulfone) Anion Exchange Membranes with Pyridinium Groups for Vanadium Redox Flow Battery Applications. J. Power Sources 2015, 282, 328−334. (37) Song, X. X.; Liu, Z. Y.; Sun, D. D. Nano Gives the Answer: Breaking the Bottleneck of Internal Concentration Polarization with a Nanofiber Composite Forward Osmosis Membrane for a High Water Production Rate. Adv. Mater. 2011, 23, 3256−3260. (38) Zhang, J. H.; Zhou, T.; Xia, L. P.; Yuan, C. Y.; Zhang, W. D.; Zhang, A. M. Polypropylene Elastomer Composite for the Allvanadium Redox Flow Battery: Current Collector Materials. J. Mater. Chem. A 2015, 3, 2387−2398. (39) Gong, K.; Fang, Q. R.; Gu, S.; Li, S. F.; Yan, Y. S. Nonaqueous Redox-flow Batteries: Organic Solvents, Supporting Electrolytes, and Redox Pairs. Energy Environ. Sci. 2015, 8, 3515−3530. (40) Ulaganathan, M.; Aravindan, V.; Yan, Q. Y.; Madhavi, S.; Skyllas-Kazacos, M.; Lim, T. M. Recent Advancements in Allvanadium Redox Flow Batteries. Adv. Mater. Interfaces 2016, 3, 1500309. (41) Xi, X. L.; Li, X. F.; Wang, C. H.; Lai, Q. Z.; Cheng, Y. H.; Zhou, W.; Ding, C.; Zhang, H. M. Impact of Proton Concentration on Equilibrium Potential and Polarization of Vanadium Flow Batteries. ChemPlusChem 2015, 80, 382−389. (42) Gu, S.; Gong, K.; Yan, E. Z.; Yan, Y. S. A Multiple Ion-exchange Membrane Design for Redox Flow Batteries. Energy Environ. Sci. 2014, 7, 2986−2998.

35297

DOI: 10.1021/acsami.6b12068 ACS Appl. Mater. Interfaces 2016, 8, 35289−35297