Hydrophilic Channel Alignment of Perfluoronated Sulfonic-Acid

May 31, 2018 - Hydrophilic Channel Alignment of Perfluoronated Sulfonic-Acid Ionomers for Vanadium Redox Flow Batteries. Soonyong So*† , Min Suc ...
8 downloads 0 Views 3MB Size
Research Article www.acsami.org

Cite This: ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Hydrophilic Channel Alignment of Perfluoronated Sulfonic-Acid Ionomers for Vanadium Redox Flow Batteries Soonyong So,*,† Min Suc Cha,†,‡ Sang-Woo Jo,†,§ Tae-Ho Kim,† Jang Yong Lee,† and Young Taik Hong*,† †

Membrane Research Center, Korea Research Institute of Chemical Technology, Daejeon 34114, South Korea Department of Chemical Engineering, Hanyang University, Seoul 04763, South Korea § Department of Polymer Engineering, Chungnam National University, Daejeon 34134, South Korea ‡

S Supporting Information *

ABSTRACT: It is known that uniaxially drawn perfluoronated sulfonicacid ionomers (PFSAs) show diffusion anisotropy because of the aligned water channels along the deformation direction. We apply the uniaxially stretched membranes to vanadium redox flow batteries (VRFBs) to suppress the permeation of active species, vanadium ions through the transverse directions. The aligned water channels render much lower vanadium permeability, resulting in higher Coulombic efficiency (>98%) and longer self-discharge time (>250 h). Similar to vanadium ions, proton conduction through the membranes also decreases as the stretching ratio increases, but the thinned membranes show the enhanced voltage and energy efficiencies over the range of current density, 50−100 mA/cm2. Hydrophilic channel alignment of PFSAs is also beneficial for long-term cycling of VRFBs in terms of capacity retention and cell performances. This simple pretreatment of membranes offers an effective and facile way to overcome high vanadium permeability of PFSAs for VRFBs. KEYWORDS: perfluoronated sulfonic-acid ionomers, uniaxial extension, water channel alignment, vanadium permeability, vanadium redox flow battery



INTRODUCTION A vanadium redox flow battery (VRFB) is a kind of secondary rechargeable battery involving the reduction−oxidation (redox) reactions of active species of vanadium ions (V2+, V3+, VO2+, and VO2+).1−6 The redox species are usually dissolved in a high concentration of sulfuric acid and stored in external reservoirs.7,8 The separated external reservoirs from the cell provide high flexibility not only on design,9 but also on energy storage capacity,10,11 which is decoupled with rated output power. Accordingly, VRFBs are of great interest for the applications in energy storage systems, and for uninterruptible power supply from the residential level to the industrial, electrical grid level.12−14 In a VRFB, a highly selective membrane for proton to vanadium ions is necessary to keep its performance over charge−discharge cycles by preventing the side reactions of redox couples, while conducting proton sources across the membrane.15 Because of the requirement of high proton conductivity, commercial membranes such as perfluoronated sulfonic-acid ionomers (PFSAs) (typically Nafion series) for proton exchange membrane fuel cells (PEMFCs) have been used for VRFBs.16,17 However, the low proton/vanadium selectivity and high production cost have been considered as the limitations for the further commercialization of VRFBs with © XXXX American Chemical Society

conventional PFSAs. To address this concern, various hydrocarbon-based membranes 18−20 and composite membranes15,21−23 have been proposed as promising candidates in similar ways for PEMFCs. In addition to the proton exchange membranes, bare porous membranes have been successfully applied to VRFB membranes as a size-exclusive separator, where smaller protons permeate better than bigger vanadium ions.24−27 Despite these efforts, however, expensive but chemically stable PFSAs in the VRFB electrolytes are still reliable membranes for long-term operation of the batteries without chemical degradation.8,18 Therefore, PFSAs have been integrated with other polymers or inorganic materials to mitigate the shortcomings of PFSAs, especially high vanadium permeability. At the same time, there are lots of studies to optimize PFSA itself in terms of their thickness10,28,29 or pretreatment method.16,30 For example, Jiang et al. reported the performance of Nafion series, Nafion 112 (50 μm), 1135 (88 μm), 115 (125 μm), and 117 (175 μm) over the wide range of current density from 40 to 320 mA/cm2.29 Similar to other Received: March 10, 2018 Accepted: May 22, 2018

A

DOI: 10.1021/acsami.8b03985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

immersed in 3 M sulfuric acid for 24 h and then rinsed by deionized water at room temperature for 24 h to ensure that all membranes have the H+ form. Before cell performance test and analysis, the membranes were dried at 50 °C in a convection oven and dried under vacuum at room temperature for 5 h, respectively. On the basis of the strain, the membranes were named as N0.9 (ε = 0.9), N1.9 (ε = 1.9), and N2.6 (ε = 2.6), respectively. Small-Angle X-ray Scattering. Synchrotron SAXS experiments with the dried membranes (Nafion 115, N0.9, N1.9, and N2.6) were carried out at beamline 4C of Pohang Accelerator Laboratory (PAL). Two-dimensional scattering images were recorded with a MAR CCD detector at room temperature with the X-ray beam wavelength (λ) of 1.23 Å, yielding scattering vectors q [q = 4π sin(θ/2)/λ], where θ is the scattering angle. The sample-to-detector distance was 905 mm. Proton Conductivity and Vanadium Permeability. All membranes were equilibrated in deionized water at 25 °C at least for 24 h before the proton conductivity measurement. In-plane conductivity (σx, σy) was measured using a four-point probe cell with an impedance analyzer (Solartron SI 1280B) by applying an alternating voltage of 0.1−20 kHz of 10 mV voltage amplitude. Through-plane conductivity (σz) was measured using a through-plane conductivity cell with an impedance analyzer (BioLogic, SP-300) by applying an alternating voltage of 0.1−3 MHz of 420 mV voltage amplitude to access impedance at a high frequency range. The permeation of vanadium ions (VO2+) was measured using a home-made apparatus. A membrane was placed between two cells, one was filled with 80 mL of 2 M VOSO4 in 3 M H2SO4 solution and the other was filled with the same volume of 2 M MgSO4 in 3 M H2SO4 solution. Both the cells were stirred continuously at 40 °C to prevent MgSO4 from crystallization. The transient concentration of permeated VO2+ across a membrane was measured using a UV−vis spectrometer (Agilent Technologies, Cary 8454 UV−vis); VO2+ showed the maximum absorption peak at 760 nm. The permeability (P) was calculated using the following equation

studies, as Nafion thickness increased, Coulombic efficiency (CE) increased from about 80% (Nafion 112) to over 90% (Nafion 117), whereas voltage efficiency (VE) decreased at a current density, 80 mA/cm2. Hence, in terms of energy efficiency (EE), relatively thick Nafion 115 performed better than others.29 For VRFB applications, a thin membrane with low vanadium permeation is required, both to render the system to minimize the voltage loss, and to reduce the cost of membrane, unless a thin membrane is vulnerable to vanadium permeation. Accordingly, uniaxial deformation of a PFSA membrane could be a possible solution for both thinner thickness and lower vanadium permeation. Among many studies of mechanically deformed PFSAs,31−35 Moore and Madsen groups studied the water channel alignment of drawn membranes with a molecular diffusion model in rodlike liquid crystals.36,37 Through 2H NMR, pulsed-field-gradient NMR, and small-angle X-ray scattering (SAXS) results, they concluded that the water channels of uniaxially stretched PFSAs were simply aligned without deforming their dimensions (Figure 1b); water

Figure 1. (a) Schematic illustration of a membrane for uniaxial stretching along the x axis, where the thick arrows indicate. (b) Schematic illustration of hydrophilic channel alignment after stretching a PFSA. Blue lines indicate the hydrophilic channels. (Reproduced with permission from ref 36).

P=

diffusion in transverse directions to the draw direction was suppressed, while the diffusion tensor trace was conserved.36 Here, we have applied this anisotropic diffusion channels to suppress the vanadium permeation, which is the most critical issue of PFSAs, while proton conductivity is retained. In this study, the uniaxial mechanical deformation of PFSAs suppresses vanadium permeation effectively and yields lower areal resistance, resulting in both higher CE and VE in the measured current density, longer self-discharge time, and better cyclic performance.



dC(t ) VL × AC0 dt

(1)

where V is the solution volume in a cell (80 mL), L is the thickness of a membrane, A is the effective area (2.86 cm2), C0 is the initial concentration of VO2+ in the vanadium solution (2 M), and C(t) is the permeated VO2+ concentration at time t, evaluated by the Beer− Lambert law. VRFB Single Cell Performance. For a VRFB single cell, a dried membrane was placed between two sets of copper current collector, graphite bipolar plate, and carbon felt electrode. Then, the cell was clamped at a pressure of 0.5 MPa. The active area of a membrane was 7 × 7 cm2. V3+/VO2+ (1.65 M, 1:1 mol/mol, Prudent Energy Inc.) in 4 M H2SO4 was used as a starting electrolyte for both the negative and positive electrolytes, and 108 g of the electrolyte was pumped cyclically through each electrode at a rate of 100 mL/min. A VRFB single cell was run by using a battery cycler (Scribner Associates Inc., 857 redox cell test system) under the cutoff voltages for discharge and charge at 1.0 and 1.6 V, respectively. The rate performance of VRFBs was conducted at various current densities between 50 and 100 mA/cm2. CE (%), and EE (%) were evaluated by the ratio of capacity and energy of discharging−charging cycles, respectively, and VE (%) was calculated using CE and EE [(VE) = (EE)/(CE) × 100%]. The

EXPERIMENTAL SECTION

Uniaxial Extension of PFSAs. Nafion 115 (DuPont) in the H+ form was used as received without any pretreatment for the extension. Nafion 115 was uniaxially stretched for a desired strain (ε) in the drawn direction (x as shown in Figure 1a) at a stretching rate of 0.3 mm/s, at 150 °C, which is above the α-relaxation temperature,38 using a uniaxial stretching machine (IMC-16B0, Imoto Machinery Co., Ltd) equipped with a temperature controller. All membranes were

Table 1. Thickness and Crystallinity and Ion Transportation Properties of the Membranes membrane Nafion 115 N0.9 N1.9 N2.6

thickness (dry/wet, μm) 123/141 102/117 76/87 71/81

xca (%)

σzb (S/cm)

11.8 13.2 15.6 18.9

0.052 0.044 0.035 0.031

DHc (m2/s) 6.31 5.36 4.27 3.79

× × × ×

−10

10 10−10 10−10 10−10

P (m2/s) 5.06 3.12 2.19 1.78

× × × ×

−13

10 10−13 10−13 10−13

α/αN115d 1 1.38 1.56 1.71

a

Crystallinity of the membranes. bThrough-plane proton conductivity determined by an impedance analyzer. cEstimated diffusion coefficient of protons through a membrane by eq 4. cH was assumed as 1000 mol/m3 for all membranes based on the equivalent weight of Nafion 115, and T was set as 298.15 K. dRelative ion selectivity to the selectivity of Nafion 115 (αN115), which is 1247. B

DOI: 10.1021/acsami.8b03985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces cycling stability was performed at a constant current density of 100 mA/cm2 for 230 charge/discharge cycles. For the self-discharge test, a VRFB was charged to 80% state of charge (SOC) at a current density of 50 mA/cm218 and was kept at the open circuit till the voltage was lower than 0.8 V under the constant electrolyte circulation rate of 100 mL/min.

implies that the hydrophilic channels were rotated and aligned along the stretching direction and packed closer to each other. In addition, to quantify the degree of hydrophilic channels’ orientation, Herman’s orientation factor ( f) of the ionomer peak was calculated using the following equation31,32



RESULTS AND DISCUSSION The uniaxial stretching caused the membranes to elongate in the drawing direction but to contract in the thickness and width directions. However, the total volume change of the membranes was negligible with Poisson’s ratio of 0.24 regardless of their strains because of the increased membrane area [(width) × (length)]. For example, the thickness decreased from 123 μm for Nafion 115 to 71 μm for N2.6, as listed in Table 1, while the area of N2.6 increased about 160% compared to the unstretched Nafion 115. The details of morphological anisotropy induced by the uniaxial stretching were confirmed by two-dimensional SAXS patterns, as shown in Figure 2. The as-received Nafion 115 shows the isotropic

f=

3⟨cos2 φ⟩ − 1 2

(2)

where φ is the azimuthal angle in a two-dimensional SAXS image. The average of cos2 φ was evaluated between φ = 0 and π/2 from the equation 2

⟨cos φ⟩ =

∫0

π /2

I(φ) sin φ cos2 φ dφ

∫0

π /2

I(φ) sin φ dφ

(3)

where I(φ) is the scattering intensity as a function of the azimuthal angle. For the isotropic system, f is 0, and f decreases and closes to −0.5 as scattering domains are aligned to the perpendicular direction to the drawing direction.32 As the strain of the membranes increases, f decreases from −0.035 (Nafion 115) to −0.2 (N2.6), which indicates that the uniaxial stretching led the hydrophilic channels to align along the drawing direction. In Figure S3, the orientation factors before and after the VRFB performance tests show that f values were not changed after the tests. The same orientation factors represent that the alignment of hydrophilic channels was not altered, even after the membranes were exposed to vanadium solutions. The intercrystalline domain peaks show a different behavior to the ionomer domain peaks. q* for the drawing direction decreases from 0.027 Å−1 for Nafion 115 to 0.023 Å−1 for N2.6, but q* for the perpendicular direction to the drawing direction increases from 0.027 Å−1 for Nafion 115 to 0.042 Å−1 for N2.6 (see Figure S2 for the details of d-spacing of the intercrystalline domains). These results are consistent with the study of prestretched Nafion membranes that the deformation of the ionomer domains and the crystalline domains was decoupled.31 The deformation of crystalline domains might be more complex than that of the hydrophilic domains because the crystallinity can be changed after the membrane deformation.40,41 As listed in Table 1, the crystallinity increased as the membrane was uniaxially stretched from 11.8 to 18.9% (the details of the crystallinity is depicted in the Supporting Information). The details of morphology for PFSAs are still under debate between models, such as the cylindrical-like water channel model,42 the ribbons model,43 and the locally flat water channel model.39 Therefore, the hydrophilic channels and crystallite deformation under stretching should be studied further to elucidate these decoupled changes. Nonetheless, it is evident that the SAXS data clearly indicate that the uniaxial stretching induced the morphological anisotropy depending on the direction, and the hydrophilic ionic domains were aligned along the stretching direction. As other studies of water diffusion,35,36 this morphological anisotropy largely affected the transport behavior of proton and vanadium ion through the membranes. Figure 3a shows proton conductivities (σx, σy, σz) in three orthogonal directions of the fully hydrated membranes. Proton conductivity of the stretched membranes along the drawing direction is higher than the other directions perpendicular to the drawing direction. As the strain increases, σx parallel to the draw direction (x) increases, but σy

Figure 2. Two-dimensional SAXS patterns of PFSAs at different strain, (a) Nafion 115, (b) N0.9, (c) N1.9, and (d) N2.6. The white double arrow indicates the extrusion direction for (a) and stretching direction for (b−d).

inner and outer scattering rings (Figure 2a), which correspond to the intercrystalline domain and ionomer domain spacing, respectively.39 Both intercrystalline and ionomer domain scattering patterns are strongly sensitive to the uniaxial stretching as shown in Figure 2. As the stretching ratio increases, the scattering pattern of the inner rings is changed to oblate elliptical shapes, and the outer scattering intensity along the drawn direction is diminished. 1-D SAXS spectra (intensity vs q) were extracted from the 2-D SAXS patterns by radial averaging of 30° for both equatorial (x) and meridional (y) directions. The peak maximum of the ionomer peak (q*) for both directions is slightly shifted from q* ≈ 0.14 Å−1 for Nafion 115 to q* ≈ 0.15 Å−1 for N0.9 and remains almost constant regardless of the strain. The corresponding d-spacing (d = 2π/ q*) decreases from 4.6 to 4.2 nm (see Figure S2), which C

DOI: 10.1021/acsami.8b03985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 3. (a) In-plane (σx, σy) and through-plane proton conductivity (σz) and its anisotropy (σx/σy) as a function of strain (ε). (b) Areal resistance of membranes under fully hydrated conditions.

Figure 4. (a) Transient concentration of permeated VO2+ across the membranes and (b) membrane permeability (P) as a function of strain (ε).

and σz perpendicular to the draw direction (y, z) decreases. Note that, the through-plane conductivity, σz is lower than σx, σy even for the isotropic Nafion 115 because of the considerable interfacial impedance contributions from electrode−membrane interfaces in a through-plane conductivity cell.44,45 Therefore, to clarify the effect of membrane stretching on the difference in the proton conductivity, the anisotropy of proton conductivity was defined by the ratio of σx to σy, and is shown in Figure 3a. While σx/σy is close to 1 for Nafion 115 having an isotropic hydrophilic channel connection, σx/σy of the stretched membranes increases to over 1.6 for N2.6. As the hydrophilic channels are aligned, it is much easier for protons to conduct along the aligned direction rather than the other directions, where the channels are relatively less disconnected.42 According to the bundle model made of the polymer matrix surrounded by ionic groups and water,32 the tortuosity of the hydrophilic channels along the transverse directions increases upon uniaxial stretching, while it decreases along the drawing direction.46,47 This anisotropic tortuosity changes could be attributed to the anisotropic behavior of proton conductivity. In a VRFB, areal resistance (Ω·cm2) is more directly related to the internal resistance, which affects the cell VE, than proton conductivity, as observed in other studies.9,29 The areal resistance was evaluated by L/σz, as shown in Figure 3b. The areal resistance of all stretched membranes is lower than Nafion 115 because of the reduced membrane thickness, although the σz decreases as the strain increases. The through-plane permeation of VO2+ depending on the strain is shown in Figure 4a. Similar to the proton conductivity, the stretched Nafion membranes also show the reduced vanadium crossover compared to Nafion 115, although the

thickness decreases upon stretching. As the strain increases, it appears that the vanadium permeability (P) (Figure 4b) decreases by a factor of 3 from 5.06 × 10−13 m2/s (Nafion 115) to 1.78 × 10−13 m2/s (N2.6). This reflects that the diffusion of vanadium ions in the transverse direction of the aligned hydrophilic channels was significantly suppressed. Permeability (P) is a material property defined as the product of the diffusion coefficient (D) and solubility (S).48 This means that P itself is not directly related to CE of a VRFB because D and S are not a strong function of the membrane thickness (L) in the case of conventional membranes. For example, Nafion 117, 115, and 1135 having different L, which are all extruded membranes having the same equivalent weight of 1100 g per a mole of sulfonic acid groups, showed similar P; they even showed significantly different CE under the same VRFB operating condition with low current densities.29 In this regard, permeance [p (P/d, m/s)], which is proportional to the flux of active species,49 should be considered. Interestingly, here, p of VO2+ also decreases from 4.29 × 10−9 m/s (Nafion 115) to around 3 × 10−9 m/s (2.97 × 10−9 m/s for N0.9, 2.96 × 10−9 m/s for N1.9, and 2.92 × 10−9 m/s for N2.6), which is opposite from the case of Nafion 117, 115, 1135. To evaluate the ion selectivity of a membrane for protons to vanadium ions, a dimensionless selectivity (α) was defined and calculated using the following equation30 α=

σ RT DH 1 = z2 × PV P F cH

(4)

where DH is the estimated proton diffusion coefficient using the Nernst−Einstein equation with σz,50 P is the measured vanadium permeability, R is the gas constant, T is absolute D

DOI: 10.1021/acsami.8b03985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 5. (a) CE and VE and (b) EE of Nafion 115, N0.9, N1.9, and N2.6 with various current densities from 50 to 100 mA/cm2.

temperature, F is Faraday’s constant, and cH is the proton concentration in a membrane. The estimated DH is in the reasonable range, considering the water diffusivity in Nafion membranes, is in the order of 10−10 m2/s.36 As shown in Table 1, as the strain increases, both DH and P decreases, but the corresponding selectivity increases. The enhancement of selectivity was mainly due to the better vanadium barrier property after the membrane stretching. Compared to Nafion 115, DH and P of N2.6 are lower by a factor of 1.7 and 2.9, respectively. Assuming that the solubility of vanadium in the membranes is same regardless of their strain, the diffusion coefficient of vanadium (DV) should be in the same order as P. Given these results, we might conclude that the aligned hydrophilic channels, induced by the uniaxial stretching, could block the vanadium permeation more effectively than proton diffusivity in a PFSA membrane. We think that proton conduction, via not only the vehicle mechanism but also the hopping mechanism, is less sensitive to the hydrophilic channel alignment than to vanadium ion conduction, probably only via the vehicle mechanism as other cations, such as Na+ and Li+.51 To clarify this different dependence, it is required to study the details of the transport mechanism of vanadium ions and compare it to that of protons in stretched membranes. The VRFB single cell performance at various current densities was conducted to investigate the effect of the aligned hydrophilic channels. In Figure 5, CE, VE, and EE of the membranes are shown as a function of current density. The efficiencies were averaged by running at least four consecutive charge−discharge cycles at each current density of the rate performance results (Figure S4). As expected from the vanadium permeance (p) and the areal resistance, both CE and VE were enhanced after the uniaxial extension except N0.9. The CE of N1.9 and N2.6 is maintained at around 100% over the measured current densities from 50 to 100 mA/cm2 (the left y-axis of Figure 5a), which was mainly attributed to the less vanadium crossover after the membrane stretching. The VE of all membranes (the right y-axis of Figure 5a) decreases as the current density increases because of the overpotential and ohmic resistance,24 but the more oriented N1.9 and N2.6 show less sensitive VE to the current density changes than Nafion 115 and N0.9. Hence, while the EE of N0.9 was almost identical to Nafion 115, the more stretched membranes with ε = 1.9 and 2.6 performed better in EE, about 5% higher than Nafion 115. These performance results indicate that a moderate stretching ratio is necessary to achieve both higher VE and CE than an unstretched membrane. While a membrane with too low a strain would suffer from high vanadium ion permeation, a

membrane with too high a strain would have lower VE by the hindered proton conduction through highly aligned hydrophilic channels. The charge−discharge curves of N1.9 showing the highest EE were compared to Nafion 115, as shown in Figure 6. The N1.9 membrane shows the lower overpotential and much higher charge−discharge capacity than Nafion 115 in the test range of 50−100 mA/cm2.

Figure 6. Charge−discharge curves of Nafion 115 (closed symbol) and N1.9 (open symbols) at various current densities from 50 to 100 mA/cm2.

The attractive features after stretching, afforded by the reduced areal resistance and considerable vanadium crossover rejection, led to the long-term stability of a membrane in a VRFB. To evaluate the cycling stability of the uniaxially stretched membrane, a VRFB with N1.9 was operated at a current density of 100 mA/cm2 over 200 charge/discharge cycles. During the cycles, EE, CE, VE, and discharge capacity retention of N1.9 were monitored and compared to the performance of Nafion 115, as shown Figure 7. N1.9 shows higher efficiencies than the unstretched Nafion 115 with much better discharge capacity retention. The discharge capacity of Nafion 115 dropped faster than N1.9 having a higher vanadium ion barrier property. After 230 cycles, a VRFB with N1.9 maintained the discharge capacity over 70%, while the capacity of a VRFB with Nafion 115 dropped to below 10% of the initial capacity. The cycling stability test was also performed at 50 mA/cm2, where charge/discharge time was about two times longer than at 100 mA/cm2, so that there was more time for vanadium ions to crossover a membrane during a charge/ discharge cycle. As shown in Figure S5, the long-term cycling stability of the stretched membrane was more obvious at 50 mA/cm2. The EE and CE of N1.9 were 2−3% higher than that of Nafion 115 during 90 cycles with better discharge capacity retention after 90 cycles, 63% for N1.9 and 21% for Nafion 115. E

DOI: 10.1021/acsami.8b03985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 7. (a) VE, CE, and EE for Nafion 115 and N1.9 and (b) discharge capacity retention with cycling at 100 mA/cm2.



SUMMARY In summary, we have shown the proton and vanadium transport behavior of the uniaxially stretched PFSAs. By aligning the hydrophilic channels, vanadium permeation effectively decreased along the thickness direction, which is perpendicular to the drawn direction. Proton conductivity also decreased, but the thinned membranes showed better VE because of lower areal resistance. Overall, the uniaxially stretched and thinned membranes performed better as a VFRB membrane; all efficiencies (CE, VE, and EE), selfdischarge time, and long-term cycling stability were enhanced after the hydrophilic channel alignment. This uniaxial stretching-induced conducting channel alignment offers a simple and facile approach to utilize the commercialized membranes for VRFBs with better ion selectivity by overcoming high vanadium permeability of PFSAs.

Besides the cell performance, a self-discharge test was conducted under a constant electrolyte circulation rate of 100 mL/min. As shown in Figure 8, the stretched membranes,



Figure 8. Self-discharge curves of Nafion 115, N0.9, N1.9, and N2.6 after charging a cell to 80% SOC.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.8b03985. Experimental details and results of X-ray diffraction, SEM, and FT-IR, SAXS data, rate performances of the membranes from 50 to 100 mA/cm2, and long-term cell performances of Nafion 115 and N1.9 at 50 mA/cm2 (PDF)

especially N1.9 and N2.6, show significantly enhanced selfdischarge time, which is related to vanadium crossover than that of Nafion 115. The self-discharge time to reach to the open circuit voltage (OCV) of 0.8 V was 72 h for Nafion 115 but 289 h for N2.6. The crossover of vanadium ions across the membranes is the main contribution of the self-discharge, inducing OCV decrease. It is well-known that the first drop is due to the disappearance of VO2+ in catholyte by the crossover vanadium ions and the immediate self-discharge reactions.52 The much longer self-discharge time of N1.9 and N2.6 indicate that the uniaxially stretched PFSAs could effectively keep catholyte and anolyte from cross-mixing of active vanadium species during the VRFB operations. It was found that PFSAs known as chemically stable membranes for VRFB application8,18 did not show significant physical defects after the uniaxial stretching and after the VRRB performance tests other than the fracture patterns, as shown in the scanning electron microscopy (SEM) images (Figure S6). Also, there was no observable chemical degradation after the stretching and VRFB performance tests, which was examined via Fourier-transform infrared spectroscopy (FT-IR) of Nafion 115 and N1.9 in the range of 700−1600 cm−1 (Figure S7). Note that there is shoulder at around 990 cm−1 for both membranes after the cell performance test, which was attributed to the vibration changes of the SO3− group in the presence of vanadium ions in the membranes, which were exchanged with protons in the VRFB.53



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (S.S.). *E-mail: [email protected] (Y.T.H.). ORCID

Soonyong So: 0000-0001-8677-7731 Tae-Ho Kim: 0000-0002-2130-9184 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the KRICT Core Research Program (KK-1802-C00), and GO! KRICT project (SKO1806-M04) of Korea Research Institute of Chemical Technology, South Korea. Experiments at Pohang Accelerator Laboratory (PAL) were supported in part by Ministry of Science, ICT and Future Planning of Korea and POSTECH. We acknowledge Dr. Jae Chang Lee for providing access to the F

DOI: 10.1021/acsami.8b03985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Stability for Vanadium Redox Flow Battery Applications: Comparative Study with Sulfonated Poly(ether sulfone)s and Sulfonated Poly(thioether ether sulfone)s. Electrochim. Acta 2018, 259, 427−439. (19) Xu, W.; Zhao, Y.; Yuan, Z.; Li, X.; Zhang, H.; Vankelecom, I. F. J. Highly Stable Anion Exchange Membranes with Internal CrossLinking Networks. Adv. Funct. Mater. 2015, 25, 2583−2589. (20) Jung, M.-s. J.; Parrondo, J.; Arges, C. G.; Ramani, V. Polysulfone-based Anion Exchange Membranes Demonstrate Excellent Chemical Stability and Performance for the All-Vanadium Redox Flow Battery. J. Mater. Chem. A 2013, 1, 10458−10464. (21) Zhang, F.; Zhang, H.; Qu, C. A Dication Cross-Linked Composite Anion-Exchange Membrane for All-Vanadium Flow Battery Applications. ChemSusChem 2013, 6, 2290−2298. (22) Xi, J.; Wu, Z.; Qiu, X.; Chen, L. Nafion/SiO2 Hybrid Membrane for Vanadium Redox Flow Battery. J. Power Sources 2007, 166, 531− 536. (23) Mai, Z.; Zhang, H.; Li, X.; Xiao, S.; Zhang, H. Nafion/ Polyvinylidene Fluoride Blend Membranes with Improved Ion Selectivity for Vanadium Redox Flow Battery Application. J. Power Sources 2011, 196, 5737−5741. (24) Zhang, H.; Zhang, H.; Li, X.; Mai, Z.; Zhang, J. Nanofiltration (NF) Membranes: The Next Generation Separators for All Vanadium Redox Flow Batteries (VRBs)? Energy Environ. Sci. 2011, 4, 1676− 1679. (25) Zhang, H.; Zhang, H.; Li, X.; Mai, Z.; Wei, W. Silica Modified Nanofiltration Membranes with Improved Selectivity for Redox Flow Battery Application. Energy Environ. Sci. 2012, 5, 6299−6303. (26) Chae, I. S.; Luo, T.; Moon, G. H.; Ogieglo, W.; Kang, Y. S.; Wessling, M. Ultra-High Proton/Vanadium Selectivity for Hydrophobic Polymer Membranes with Intrinsic Nanopores for Redox Flow Battery. Adv. Energy Mater. 2016, 6, 1600517. (27) Zhao, Y.; Li, M.; Yuan, Z.; Li, X.; Zhang, H.; Vankelecom, I. F. J. Advanced Charged Sponge-Like Membrane with Ultrahigh Stability and Selectivity for Vanadium Flow Batteries. Adv. Funct. Mater. 2016, 26, 210−218. (28) Jeong, S.; Kim, L.-H.; Kwon, Y.; Kim, S. Effect of Nafion Membrane Thickness on Performance of Vanadium Redox Flow Battery. Korean J. Chem. Eng. 2014, 31, 2081−2087. (29) Jiang, B.; Wu, L.; Yu, L.; Qiu, X.; Xi, J. A Comparative Study of Nafion Series Membranes for Vanadium Redox Flow Batteries. J. Membr. Sci. 2016, 510, 18−26. (30) Xie, W.; Darling, R. M.; Perry, M. L. Processing and Pretreatment Effects on Vanadium Transport in Nafion Membranes. J. Electrochem. Soc. 2016, 163, A5084−A5089. (31) Mendil-Jakani, H.; Pouget, S.; Gebel, G.; Pintauro, P. N. Insight into the Multiscale Structure of Pre-Stretched Recast Nafion Membranes: Focus on the Crystallinity Features. Polymer 2015, 63, 99−107. (32) van der Heijden, P. C.; Rubatat, L.; Diat, O. Orientation of Drawn Nafion at Molecular and Mesoscopic Scales. Macromolecules 2004, 37, 5327−5336. (33) Klein, M.; Perrin, J.-C.; Leclerc, S.; Guendouz, L.; Dillet, J.; Lottin, O. Anisotropy of Water Self-Diffusion in a Nafion Membrane under Traction. Macromolecules 2013, 46, 9259−9269. (34) Lin, J.; Wu, P.-H.; Wycisk, R.; Pintauro, P. N.; Shi, Z. Properties of Water in Prestretched Recast Nafion. Macromolecules 2008, 41, 4284−4289. (35) Elliott, J. A.; Hanna, S.; Elliott, A. M. S.; Cooley, G. E. Interpretation of the Small-Angle X-Ray Scattering from Swollen and Oriented Perfluorinated Ionomer Membranes. Macromolecules 2000, 33, 4161−4171. (36) Li, J.; Park, J. K.; Moore, R. B.; Madsen, L. A. Linear Coupling of Alignment with Transport in a Polymer Electrolyte Membrane. Nat. Mater. 2011, 10, 507−511. (37) Park, J. K.; Li, J.; Divoux, G. M.; Madsen, L. A.; Moore, R. B. Oriented Morphology and Anisotropic Transport in Uniaxially Stretched Perfluorosulfonate Ionomer Membranes. Macromolecules 2011, 44, 5701−5710.

stretcher, Dr. Ye Cheol Rho for helpful discussion about X-ray scattering, and Sung Hyun Yoon for help with the VRFB performance test.



REFERENCES

(1) Soloveichik, G. L. Flow Batteries: Current Status and Trends. Chem. Rev. 2015, 115, 11533−11558. (2) Janoschka, T.; Martin, N.; Martin, U.; Friebe, C.; Morgenstern, S.; Hiller, H.; Hager, M. D.; Schubert, U. S. An Aqueous, PolymerBased Redox-Flow Battery Using Non-Corrosive, Safe, and Low-Cost Materials. Nature 2015, 527, 78−81. (3) Park, M.; Ryu, J.; Wang, W.; Cho, J. Material Design and Engineering of Next-Generation Flow-Battery Technologies. Nat. Rev. Mater. 2016, 2, 16080. (4) Maurya, S.; Shin, S.-H.; Kim, Y.; Moon, S.-H. A Review on Recent Developments of Anion Exchange Membranes for Fuel Cells and Redox Flow Batteries. RSC Adv. 2015, 5, 37206−37230. (5) Jang, J.-K.; Kim, T.-H.; Yoon, S. J.; Lee, J. Y.; Lee, J.-C.; Hong, Y. T. Highly Proton Conductive, Dense Polybenzimidazole Membranes with Low Permeability to Vanadium and Enhanced H2SO4 Absorption Capability for Use in Vanadium Redox Flow Batteries. J. Mater. Chem. A 2016, 4, 14342−14355. (6) Lu, W.; Yuan, Z.; Zhao, Y.; Zhang, H.; Zhang, H.; Li, X. Porous Membranes in Secondary Battery Technologies. Chem. Soc. Rev. 2017, 46, 2199−2236. (7) Ding, C.; Zhang, H.; Li, X.; Liu, T.; Xing, F. Vanadium Flow Battery for Energy Storage: Prospects and Challenges. J. Phys. Chem. Lett. 2013, 4, 1281−1294. (8) Cha, M. S.; Jeong, H. Y.; Shin, H. Y.; Hong, S. H.; Kim, T.-H.; Oh, S.-G.; Lee, J. Y.; Hong, Y. T. Crosslinked Anion Exchange Membranes with Primary Diamine-Based Crosslinkers for Vanadium Redox Flow Battery Application. J. Power Sources 2017, 363, 78−86. (9) Li, X.; Zhang, H.; Mai, Z.; Zhang, H.; Vankelecom, I. Ion Exchange Membranes for Vanadium Redox Flow Battery (VRB) Applications. Energy Environ. Sci. 2011, 4, 1147−1160. (10) Reed, D.; Thomsen, E.; Wang, W.; Nie, Z.; Li, B.; Wei, X.; Koeppel, B.; Sprenkle, V. Performance of Nafion N115, Nafion NR212, and Nafion NR-211 in a 1 kW Class All Vanadium Mixed Acid Redox Flow Battery. J. Power Sources 2015, 285, 425−430. (11) Li, L.; Kim, W.; Wang, W.; Vijayakumar, M.; Nie, Z.; Chen, B.; Zhang, J.; Xia, G.; Hu, J.; Graff, G.; Liu, J.; Yang, Z. A Stable Vanadium Redox-Flow Battery with High Energy Density for Large-Scale Energy Storage. Adv. Energy Mater. 2011, 1, 394−400. (12) Alotto, P.; Guarnieri, M.; Moro, F. Redox Flow Batteries for the Storage of Renewable Energy: A Review. Renewable Sustainable Energy Rev. 2014, 29, 325−335. (13) Kondoh, J.; Ishii, I.; Yamaguchi, H.; Murata, A.; Otani, K.; Sakuta, K.; Higuchi, N.; Sekine, S.; Kamimoto, M. Electrical Energy Storage Systems for Energy Networks. Energy Convers. Manage. 2000, 41, 1863−1874. (14) Luo, X.; Wang, J.; Dooner, M.; Clarke, J. Overview of Current Development in Electrical Energy Storage Technologies and the Application Potential in Power System Operation. Appl. Energy 2015, 137, 511−536. (15) Kim, S.; Yuk, S.; Kim, H. G.; Choi, C.; Kim, R.; Lee, J. Y.; Hong, Y. T.; Kim, H.-T. A Hydrocarbon/Nafion Bilayer Membrane with a Mechanical Nano-Fastener for Vanadium Redox Flow Batteries. J. Mater. Chem. A 2017, 5, 17279−17286. (16) Jiang, B.; Yu, L.; Wu, L.; Mu, D.; Liu, L.; Xi, J.; Qiu, X. Insights into the Impact of the Nafion Membrane Pretreatment Process on Vanadium Flow Battery Performance. ACS Appl. Mater. Interfaces 2016, 8, 12228−12238. (17) Ding, C.; Zhang, H.; Li, X.; Zhang, H.; Yao, C.; Shi, D. Morphology and Electrochemical Properties of Perfluorosulfonic Acid Ionomers for Vanadium Flow Battery Applications: Effect of SideChain Length. ChemSusChem 2013, 6, 1262−1269. (18) Choi, S.-W.; Kim, T.-H.; Jo, S.-W.; Lee, J. Y.; Cha, S.-H.; Hong, Y. T. Hydrocarbon Membranes with High Selectivity and Enhanced G

DOI: 10.1021/acsami.8b03985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces (38) Osborn, S. J.; Hassan, M. K.; Divoux, G. M.; Rhoades, D. W.; Mauritz, K. A.; Moore, R. B. Glass Transition Temperature of Perfluorosulfonic Acid Ionomers. Macromolecules 2007, 40, 3886− 3890. (39) Kreuer, K.-D.; Portale, G. A Critical Revision of the NafionMorphology of Proton Conducting Ionomers and Polyelectrolytes for Fuel Cell Applications. Adv. Funct. Mater. 2013, 23, 5390−5397. (40) Lin, J.; Wu, P.-H.; Wycisk, R.; Pintauro, P. PEM Fuel Cell Properties of Pre-Stretched Recast Nafion. ECS Trans. 2008, 16, 1195−1204. (41) Hink, S.; Henkensmeier, D.; Jang, J. H.; Kim, H.-J.; Han, J.; Nam, S.-W. Reduced In-Plane Swelling of Nafion by a Biaxial Modification Process. Macromol. Chem. Phys. 2015, 216, 1235−1243. (42) Schmidt-Rohr, K.; Chen, Q. Parallel Cylindrical Water Nanochannels in Nafion Fuel-Cell Membranes. Nat. Mater. 2008, 7, 75−83. (43) Rubatat, L.; Gebel, G.; Diat, O. Fibrillar Structure of Nafion: Matching Fourier and Real Space Studies of Corresponding Films and Solutions. Macromolecules 2004, 37, 7772−7783. (44) Yun, S.-H.; Shin, S.-H.; Lee, J.-Y.; Seo, S.-J.; Oh, S.-H.; Choi, Y.W.; Moon, S.-H. Effect of Pressure on Through-Plane Conductivity of Polymer Electrolyte Membranes. J. Membr. Sci. 2012, 417−418, 210− 216. (45) Soboleva, T.; Xie, Z.; Shi, Z.; Tsang, E.; Navessin, T.; Holdcroft, S. Investigation of the Through-Plane Impedance Technique for Evaluation of Anisotropy of Proton Conducting Polymer Membranes. J. Electroanal. Chem. 2008, 622, 145−152. (46) Fischer, R. Properties of Stretched 830 EW AQUIVION. M.S. Thesis, Vanderbilt University, 2012. (47) Cable, K. M.; Mauritz, K. A.; Moore, R. B. Anisotropic Ionic Conductivity in Uniaxially Oriented Perfluorosulfonate Ionomers. Chem. Mater. 1995, 7, 1601−1603. (48) Bai, Z.; Zhao, B.; Lodge, T. P. Bilayer Membrane Permeability of Ionic Liquid-Filled Block Copolymer Vesicles in Aqueous Solution. J. Phys. Chem. B 2012, 116, 8282−8289. (49) Vijayakumar, M.; Luo, Q.; Lloyd, R.; Nie, Z.; Wei, X.; Li, B.; Sprenkle, V.; Londono, J.-D.; Unlu, M.; Wang, W. Tuning the Perfluorosulfonic Acid Membrane Morphology for Vanadium RedoxFlow Batteries. ACS Appl. Mater. Interfaces 2016, 8, 34327−34334. (50) Noda, A.; Hayamizu, K.; Watanabe, M. Pulsed-Gradient Spin− Echo 1H and 19F NMR Ionic Diffusion Coefficient, Viscosity, and Ionic Conductivity of Non-Chloroaluminate Room-Temperature Ionic Liquids. J. Phys. Chem. B 2001, 105, 4603−4610. (51) Saito, M.; Arimura, N.; Hayamizu, K.; Okada, T. Mechanisms of Ion and Water Transport in Perfluorosulfonated Ionomer Membranes for Fuel Cells. J. Phys. Chem. B 2004, 108, 16064−16070. (52) Sun, J.; Shi, D.; Zhong, H.; Li, X.; Zhang, H. Investigations on the Self-Discharge Process in Vanadium Flow Battery. J. Power Sources 2015, 294, 562−568. (53) Intan, N. N.; Klyukin, K.; Zimudzi, T. J.; Hickner, M. A.; Alexandrov, V. A Combined Theoretical-Experimental Study of Interactions between Vanadium Ions and Nafion Membrane in AllVanadium Redox Flow Batteries. J. Power Sources 2018, 373, 150−160.

H

DOI: 10.1021/acsami.8b03985 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX