Electrochimica Acta 259 (2018) 427e439

Contents lists available at ScienceDirect

Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta

Hydrocarbon membranes with high selectivity and enhanced stability

for vanadium redox flow battery applications: Comparative study with
sulfonated poly(ether sulfone)s and sulfonated poly(thioether ether
sulfone)s
So-Won Choi a, b, Tae-Ho Kim a, *, Sang-Woo Jo a, Jang Yong Lee a, Sang-Ho Cha b, **,
Young Taik Hong a, ***
a
Center for Membranes, Korea Research Institute of Chemical Technology (KRICT), 141 Gajeong-ro, Yuseong-gu, Daejeon 34114, Republic of Korea
b
Department of Chemical Engineering, Kyonggi University, 154-42 Gwanggyosan-ro, Yeongtong-gu, Suwon 16227, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: A series of sulfonated poly(ether sulfone) copolymers (SPES-Xs) with varying degrees of sulfonation were
Received 26 May 2017 prepared and investigated as ion-exchange membranes for vanadium redox flow battery (VRFB) appli-
Received in revised form cations. Sulfonated poly(thioether ether sulfone) copolymers (SPTES-Xs) were initially synthesized via
17 September 2017
polycondensation, and the SPES-Xs were then obtained by oxidation of the corresponding SPTES-Xs. The
Accepted 18 October 2017
Available online 1 November 2017
SPES-X membranes showed reduced vanadium-ion permeability, low area resistance, and, thereby, much
superior selectivity compared with the parent SPTES-X membranes and a Nafion115 membrane. In
single-cell VRFB performance tests, a SPES-50 membrane with an ion-exchange capacity of 1.80 meq/g
Keywords:
Sulfonated poly(ether sulfone)
exhibited a higher coulombic efficiency (>99%) and energy efficiency (76e89%) than the Nafion115
Sulfonated poly(thioether ether sulfone) membrane over a wide range of current densities from 40 to 100 mA/cm2 and a significantly larger
VRFB capacity retention (>62%) during 200 charge-discharge cycles. The SPES-X materials, in which every
Polymer electrolyte membrane benzene ring is deactivated by the presence of electron-withdrawing sulfone linkages, showed much
better chemical stability during ex situ and in situ tests than the SPTES-X materials, which contain an
electron-donating thioether linkage in their repeat unit.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction excellent proton conductivity and outstanding electrochemical

All-vanadium redox flow batteries (VRFBs) have received
increased attention in recent years as a possible solution for large-
scale energy storage systems due to their advantageous properties,
such as high efficiency, long cycle life, design flexibility, and safety
[1e3]. VRFBs are mainly composed of positive and negative elec-
trolytes, electrodes, and separators. Ion-exchange membranes
(IEMs) commonly serve as separators to separate the two electro-
lytes in VRFBs, allowing proton transfer to complete the circuit.
Almost all existing commercial VRFBs use perfluorosulfonic acid
(PFSA) polymer membranes, e.g. Nafion membranes, which exhibit
* Corresponding author.
** Corresponding author.
*** Corresponding author.
E-mail addresses: thkim@krict.re.kr (T.-H. Kim), sanghocha@kgu.ac.kr
(S.-H. Cha), ythong@krict.re.kr (Y.T. Hong).
stability. Nevertheless, the extremely high cost of Nafion mem-
branes (they are the most expensive component in a VRFB stack,
accounting for ~40% of the stack cost [4,5]) is the main obstacle to
their industrial application in VRFBs. Moreover, Nafion membranes
are highly permeable to vanadium species, leading to undesired
vanadium-ion transport across the membranes during charge-
discharge cycling [6]. The large vanadium-ion crossover through
the Nafion membranes results in decreased coulombic efficiency
(CE) and an imbalance of the electrolyte, which leads to capacity
loss [7,8]. Therefore, developing alternative IEMs that are inex-
pensive and highly selective is necessary for the wide commer-
cialization of the VRFB technology.
In recent years, a variety of sulfonated hydrocarbon membranes
have been investigated as replacements for PFSA membranes,
including sulfonated poly(arylene ketone)s [9e11], sulfonated
polyimides [12,13], sulfonated poly(arylene sulfone)s [14,15], and
sulfonated poly(phenylene)s [16,17]. Generally, sulfonated

https://doi.org/10.1016/j.electacta.2017.10.121
0013-4686/© 2017 Elsevier Ltd. All rights reserved.
428 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439
hydrocarbon membranes are cheaper and exhibit reduced
vanadium-ion permeability compared with Nafion membranes.
Therefore, some VRFBs assembled with hydrocarbon membranes
have exhibited comparable or higher efficiencies and larger ca-
pacity retention capabilities than those using Nafion membranes.
Nevertheless, the poor chemical stability of hydrocarbon
membranes when exposed to the VRFB electrolytes remains a
critical issue [18e21]. Hydrocarbon membranes normally suffer
from physical and/or chemical degradation under VRFB operating
conditions, which lowers the long-term reliability of batteries uti-
lizing these membranes. For example, a study on sulfonated Radel
(S-Radel) by Kim et al. showed that this polymer exhibits better
selectivity than Nafion117 and comparable energy efficiency;
however, it can only withstand a few cycles before physically
degrading [18]. Although there are a very limited number of studies
on the degradation mechanisms of hydrocarbon membranes, it is
clear that the strongly oxidizing vanadium (V) ions are responsible
for the degradation of these membranes [19]. Hickner et al. pro-
posed a degradation mechanism for S-Radel, where the vanadium
(V) oxygen species first attacks the polymer by incorporating hy-
droxyl groups into the polymer backbone and then oxidizes them
into quinine groups [20]. According to this mechanism, the authors
suggested that reducing the electron density on the aromatic
backbone might lower the risk of hydrocarbon-membrane degra-
dation. Later, the same research group applied a synthetic strategy
to incorporate partially fluorinated moieties into the polymer
backbone to enhance the chemical stability of sulfonated poly(-
arylene ether). They concluded that the electron-withdrawing po-
wer of fluorine shields the aromatic backbone, rendering the
fluorinated aromatic polymers more stable as expected [21]. In
contrast, Zhang et al. proposed a different mechanism for the
degradation of a sulfonated poly(ether ether ketone) (SPEEK)
membrane in a VRFB medium [22]. Under strong acidic conditions,
the ethereal oxygen atoms in SPEEK can be easily protonated and
become strong electron-withdrawing groups. The protonated ether
groups, together with the strong electron-withdrawing sulfonic
acid group, then induce a strongly electrophilic carbon center on
the benzene ring, which leads to further attack from the lone pair of
electrons on the vanadium oxygen species. Based on this mecha-
nism, they suggested that the chemical stability of the membrane
could be improved through introducing electron-donating groups
to the aromatic backbone, which will reduce the reactivity of the
electrophilic carbon center on the benzene ring.
Moreover, it is thought that increasing the number of ion-
exchange groups in the polymer accelerates the degradation of
sulfonated hydrocarbon membranes [23,24]. Fujimoto et al. inves-
tigated sulfonated Diels Alder poly(phenylene)s with different ion-
exchange capacities (IECs), and found that the oxidative stability of
the polymers decreased with increasing IEC [16]. However, very
recent studies on the application of polyoxadiazole-based and
polybenzimidazole-based membranes in VRFBs have demonstrated
that hydrocarbon membranes not containing conventional ion-
exchange groups also display some degradation in the harsh
VRFB medium [25e29]. Consequently, despite the publication of
some reports attempting to clarify the degradation mechanism of
hydrocarbon membranes, the understanding of their poor chemical
stability is still very limited, which is probably due to the compli-
cated medium of VRFBs (strongly acidic, oxidizing, and high electric
potential).
In this work, we aimed to understand more about the degra-
dation of hydrocarbon membranes and develop sulfonated hydro-
carbon membranes with enhanced stability for VRFB applications
through conducting comparative study with sulfonated poly(ether
sulfone) copolymers (SPES-Xs) and sulfonated poly(thioether ether
sulfone) copolymers (SPTES-Xs). SPES-Xs were chosen because
every benzene ring in their polymer backbone is adjacent to a
strongly electron-withdrawing sulfone linkage group. Meanwhile,
SPTES-Xs possess a thioether linkage in their repeat unit that has an
electron-donating nature. Although there have been a few reports
focusing on some polymer blend membranes using SPES co-
polymers for VRFB applications [30,31], we believe more detailed
study on the structure-property relationship of this kind of poly-
mers is required, especially in terms of oxidative stability and VRFB
performance. In particular, a comparative study between SPTES-Xs
and SPES-Xs can be helpful to investigate the structural effect on
the stability because their only difference in structure is whether all
benzene rings are deactivated by electron withdrawing linkage
groups or not. Since SPES-Xs are directly prepared from the cor-
responding SPTES-Xs, we can assume no difference in other
structural parameters, such as degree of sulfonation and degree of
polymerization between them.
In the synthesis, SPTES-Xs were first prepared through poly-
condensation of sulfonated and non-sulfonated difluoride mono-
mers, and the SPES-Xs were then obtained through oxidation of the
SPTES-Xs, which converted the SPTES-Xs thioether linkages to
sulfone linkages. Solvent-cast SPTES-X and SPES-X membranes
with IECs ranging from 1.68 to 2.14 meq/g were fully characterized
in terms of their various physical and electrochemical properties,
including swelling ratio, area resistance, and vanadium-ion
permeability. In particular, the effect of the electron-withdrawing
sulfone groups on the oxidative stability of the SPES-X mem-
branes was carefully analyzed by comparing their ex situ and in situ
stability with those of the SPTES-Xs membranes containing the
electron-donating thioether groups. In single-cell VRFB perfor-
mance tests, a SPES-50 membrane with an IEC of 1.80 meq/g
exhibited a higher energy efficiency (EE) (75%) and CE (>98%) than
the corresponding cell with a Nafion115 membrane, and a signifi-
cantly larger capacity retention (>93%) during 200 charge-
discharge cycles.
2. Experimental
2.1. Materials
4,40 -Thiodiphenol (TDP) was obtained from Sigma-Aldrich and
dried at 80  C prior to use. 4,40 -Difluorodiphenyl sulfone (DFDPS)
was provided by Richem and recrystallized with ethanol. Disodium
3,30 -disulfonated-4,40 -difluorodiphenylsulfone (SDFDPS) was pre-
pared from DFDPS according to the literature method [32] and
purified by recrystallization with ethanol/H2O and then dried un-
der vacuum at 120  C prior to use. Anhydrous potassium carbonate
(K2CO3) was purchased from Sigma-Aldrich and dried under vac-
uum at 80  C. N,N0 -Dimethylacetamide (DMAc), dimethyl sulfoxide
(DMSO), vanadium pentoxide (V2O5, 98%), and toluene were ob-
tained from Sigma-Aldrich. Acetic acid (glacial, 99.5%), sulfuric acid
(97%), and 30% hydrogen peroxide (H2O2) were provided by Sam-
chun, Merck, and Junsei, respectively. 2-Propanol (IPA) was ob-
tained from Samchun and used without further purification. Nafion
115 was acidified in a 1.5 M H2SO4 aqueous solution at room tem-
perature for 24 h, and then washed with distilled water and dried
on a vacuum plate for 1 h.
2.2. Synthesis of sulfonated poly(thioether ether sulfone)
copolymers (SPTES-Xs)
Sulfonated poly(thioether ether sulfone) copolymers (SPTES-Xs;
X denotes the molar feed ratio of SDFDPS relative to the total
difluoride-monomer content) were prepared via direct copoly-
merization of SDFDPS and DFDPS with TDP (Fig. 1). The typical
procedure for preparing SPTES-50 is as follows. A 100 ml four-neck
 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439
Fig. 1. Preparation of SPTES-X and SPES-X copolymers.
round-bottomed flask was equipped with an argon inlet, a me-
chanical stirrer, a Dean Stark trap, and a condenser. The flask was
charged with TDP (5.6983 g, 0.0261 mol), K2CO3 (4.3298 g,
0.0313 mol), 40 ml of DMAc, and 37.5 ml of toluene. The temper-
ature of the flask was raised to 150  C for 2 h in an oil bath and then
slowly increased from 150  C to 160  C over 4 h to azeotropically
remove water with refluxing toluene. After the toluene was
removed, DFDPS (3.3188 g, 0.0131 mol), SDFDPS (5.9829 g,
0.0131 mol), and 20 ml of DMAc were added to the flask. Finally, the
mixture temperature was slowly raised to 175  C and held at this
temperature until the viscosity had increased. The viscous
yellowish solution was precipitated in H2O and then washed
several times with H2O. The product was dried under vacuum at
80  C for at least 24 h.
2.3. Synthesis of sulfonated poly(ether sulfone) copolymers (SPES-
Xs)
Sulfonated poly(ether sulfone) copolymers (SPES-Xs; X is the
same as in Section 2.2) were obtained by oxidizing the corre-
sponding SPTES-Xs (Fig. 1) [33]. Typically, the synthesis of SPES-50
was conducted as follows. 2 g of SPTES-50 was dispersed in a
mixture of 40 ml of acetic acid and 3 ml of sulfuric acid, and
hydrogen peroxide (2.6 ml of 30% H2O2 in water) was then slowly
added to the reaction mixture. The reaction was kept at 30  C for 2
days. The resulting powder was filtered and stirred overnight in a
0.5 wt% K2CO3 aqueous solution. The product was then washed
several times with H2O and dried under vacuum at 80  C.
2.4. Membrane preparation
The SPTES-Xs and SPES-Xs were dissolved in DMSO (8% w/v)
and the polymer solutions were then filtered with 5-mm Teflon
syringe filters, cast onto dust-free glass plates, and dried at 60  C for
18 h. The membranes were acidified in a 1.5 M H2SO4 aqueous
solution at room temperature for 24 h, and then washed with
distilled water and dried on a vacuum plate for 1 h. The thickness of
the membranes was 75e85 mm.
429
2.5. Characterization
The Proton nuclear magnetic resonance (1H NMR) spectra of the
copolymers were obtained using a Bruker 400 MHz spectrometer
and deuterated dimethyl sulfoxide (DMSO-d6) as the solvent. The
DMSO signal at 2.5 ppm was used as the chemical-shift reference.
The inherent viscosity of the copolymers in their potassium form
was measured in N-methyl pyrrolidinone (NMP) (0.05 dl g1) at
25  C using a Cannon Ubbelohde viscometer.
The theoretical ion-exchange capacity (IECtheo) of the mem-
branes was calculated from the following equation:
2n
IECtheo ¼ ; (1)
MWm  m þ MWn  n
where MWm and m are the molecular weight and mole fraction,
respectively, of the non-sulfonated unit, while MWn and n are those
of the sulfonated unit.
The experimental ion-exchange capacity values were deter-
mined from titration method (IECtitra) and 1H NMR measurements
(IECNMR). The IECtitra was determined via titration with an auto-
titrator (Metrohm 794 Basic Titrino). The membranes were dried
after washing with 1.5 M H2SO4 and the dried membranes were
then weighed and stirred in 100 ml of a 0.01 M NaCl solution for
24 h to deprotonate the sulfonic acid groups. The solutions were
then titrated with a 0.01 M NaOH solution to pH 7. IECtitra values
were calculated using Eq. (2).
CNaOH  VNaOH
IECtitra ¼ ; (2)
Wdry
where CNaOH and VNaOH are the concentration (mol$l1) and volume
(ml), respectively, of the consumed NaOH solution and Wdry is the
weight (g) of the dried membrane. The IECNMR was calculated by
comparing the areas of the peaks assigned to the sulfonated and
unsulfonated diphenyl sulfone groups. From these values, the
average number of sulfonic acid groups per repeat unit was
determined and the number was divided by the calculated average
formula weight per repeat unit to give the IECNMR.
The water uptake and swelling ratio of the membranes were
determined by assessing the weight and volume change between
430 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439
the dry and swollen membranes. The wet membranes (soaking in
deionized water at 25  C) were cut into squares with one side
measuring 2 cm, and then their weight, area, and thickness were
precisely measured as soon as wiping. Thereafter, the membranes
were dried in the vacuum oven at 80  C for 24 h and their weight,
area, and thickness were re-measured in the same way as before.
The water uptake, swelling ratio, and density were calculated from
the following equations:
Wwet  Wdry
Water uptakeð%Þ ¼  100;
Wdry
Vwet  Vdry
Swelling ratioð%Þ ¼  100;
Vdry
where Wwet/Vwet and Wdry/Vdry are the weight (g)/volume (mm3) of
the wet and dry membranes, respectively.
2.6. Area resistance
The area resistance was measured according to the method re-
ported in the literature [34]. Two compartments of a conductivity
cell were filled with a 0.05 M H2SO4 aqueous solution. Two plat-
inum electrodes with an effective area (S) of 0.20 cm2 were placed
at a fixed separation distance. The resistances of the conductivity
cell with a membrane (r1 ) and without a membrane (r2 ) were
determined through electrochemical impedance spectroscopy (EIS)
measurements using an SP-300 instrument (Bio Logic) over a fre-
quency range of 0.1 Hze4 MHz at room temperature. The area
resistance (U$cm2) was calculated from the following equation:


Area resistance Ucm2 ¼ ðr1  r2 Þ  S:
The proton conductivity was calculated by dividing the mem-
brane thickness by the area resistance.
2.7. Vanadium-ion permeability
For the VO2þ permeability measurements, the membranes were
located between two separate diffusion cells with an effective area
of 2.84 cm2. One cell was filled with 80 ml of 1.2 M VOSO4 in a
3 M H2SO4 solution, while the other cell was filled with 80 ml of
1.2 M MgSO4 in a 3 M H2SO4 solution. The two solutions in the cells
were stirred continuously at room temperature to ensure uniform
distribution of the ions. Test samples were taken from the MgSO4
solution at certain time intervals, and their VO2þ concentration was
measured using a Ultraviolet-visible (UV-vis) spectrometer (Agilent
Technologies Cary 8454). The VO2þ permeability (P) was calculated
from the following equation:
dCb ðtÞ P
Vb ¼ A ðCa  Cb ðtÞÞ;
dt L Coulombic efficiency ðCEÞ ¼ Z
where Vb is the solution volume of MgSO4 in the H2SO4 solution and
Ca and Cb (t) represent the VO2þ concentration in the VOSO4 in the
H2SO4 solution and in the MgSO4 in the H2SO4 solution, respec-
tively. A is the effective area of the sample membrane and L is the
membrane thickness.
2.8. Ex situ chemical stability
The ex situ chemical stability of the SPES-X and SPTES-X mem-
branes was evaluated by immersion tests in two VOþ 2 solutions with
different concentration [19]. The electrolyte solutions were (1)
dilute solution with 0.1 M VOþ 2 þ 5.0 M H2SO4 solutions (0.1 M VO2
þ
solution) and (2) concentrated 1.5 M VOþ 2 þ 3.0 M H 2 SO 4 solutions
(1.5 M VOþ þ
2 solution). The 0.1 M VO2 solution was prepared by
dissolving V2O5 with a 5.0 M H2SO4 aqueous solution, and the 1.5 M
VOþ 2 solution was prepared by the flow cell operation. At first, the
immersion tests in 0.1 M VOþ 
2 solution at 40 C were conducted for
the membranes. The amount of reduced VO2þ in the experiment
was standardized by measuring the absorbance of the mixed so-
lutions, with various ratios of VO2þ/VOþ 2 (total vanadium
concentration ¼ 0.1 M), at 760 nm using a UV-vis spectrometer.
(3)
From the linear relationship between VOþ 2 concentration and
absorbance, we obtained a standard equation (Fig. S1). A
5.0 M H2SO4 aqueous solution was used as the reference for these
(4) measurements. Membrane samples of 2 cm  3 cm area (about
0.03 ± 0.01 g) were immersed in 15 ml of the 0.1 M VOþ 2 solution at
40  C for 1440 h. The change in absorbance at 760 nm over im-
mersion time was then monitored for each test solution using a UV-
vis spectrometer. The degree of degradation was estimated for each
membrane from the increase in the VO2þ concentration (reduced
from VOþ 2 ) of the test solution, which was calculated using the as-
determined standard equation. In the same way, 0.07 ± 0.01 g of the
membrane were also immersed in 1.5 M VOþ 
2 solution at 40 C for
320 h. The change in absorbance at 760 nm over immersion time
was monitored for each test solution using a UV-vis spectrometer.
In this case, the sample was diluted with 3 M sulfuric acid solution
before the measurement. Inherent viscosities of the post-test
membranes were measured and compared with those of the pris-
tine polymers.
2.9. VRFB single-cell performance and in situ chemical stability
A VRFB single cell was manufactured by sandwiching a mem-
(5) brane with an effective area of 49 cm2 between two carbon-felt
electrodes (heat-treated in air at 500  C for 4 h to improve their
electrochemical activity and hydrophilicity). Two flow frames,
graphite bipolar plates (as current collectors), and copper plates
were then assembled in that order. An 80 ml volume of 1.65 M V3.5þ
(V3þ:VO2þ ¼ 1:1) in 4 M H2SO4 was used as both the positive and
negative electrolyte. The electrolytes were cycled using a flow-cell
system including two peristaltic pumps (Scribner Associates Inc.,
857 redox cell test system) at a flow rate of 75 ml min1 and varying
current densities of 40, 60, 80, and 100 mA cm2. The flow cell was
charged to 1.6 V and then discharged to 1.0 V, in which the ranges
correspond to the state of charge (SOC) levels from 24 to 82%. The
SOC levels at various voltages were determined from the standard
curve prepared from the absorbance at 760 nm of the mixed VO2þ/
VOþ2 solutions (total 1.5 M) with various ratios using a UV-vis
spectrometer (Fig. S2). The coulombic efficiency (CE), energy effi-
ciency (EE), and voltage efficiency (VE) of the charge-discharge
cycle were calculated according to the following equations:
Z
(6) Id dt
 100%; (7)
Ic dt
Z
Vd Id dt
Energy efficiencyðEEÞ ¼ Z  100%; (8)
Vc Ic dt
EE
Voltage efficiencyðVEÞ ¼  100%; (9)
CE
where I, V, and t are current, voltage, and time, respectively; the
 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439
subscripts c and d represent the charge and discharge processes,
respectively.
The charge-discharge cycling tests were also performed up to
400 cycles at a current density of 100 mA cm2. After the cycling
tests, the in situ chemical stability of the membrane was analyzed in
detail as follows. The chemical structure of the post-test membrane
was determined by 1H NMR measurements (Bruker 500 MHz
spectrometer). The cross-sectional morphology of the membranes
was measured using scanning electron microscopy (SEM) (Tescan
Mira 3 LMU FEG) with an acceleration voltage of 20 kV. The cross-
sections were obtained using a freeze-fracture technique in liquid
nitrogen and then coating the samples with platinum prior to
observation.
3. Results and discussion
3.1. Copolymer synthesis and characterization
SPTES-Xs were synthesized by nucleophilic polycondensation of
SDFDPS and DFDPS with TDP (Fig. 1). The value of X is the degree of
sulfonation and was varied between 45, 50, and 55 by controlling
the molar feed ratio of SDFDPS to DFDPS. By measuring the mass of
the synthesized polymers, we found out that the total mass yields
are over 96% for all SPTES-Xs, implying almost complete conversion
of the monomers. The chemical structures of the obtained SPTES-Xs
were confirmed by 1H NMR spectroscopy, as shown in Fig. S3 (a)-
(c). The peaks at 8.3 ppm and 7.4 ppm were assigned to protons
located on the sulfonated phenyl rings and diphenyl thioether
groups, respectively. It can be seen that the peak at 8.3 ppm in-
creases relatively with increasing SDFDPS feed ratio. From two
dimensional NMR (1H-13C HMQC) analysis, we confirmed that
SPTES-Xs possess random copolymer architecture (Fig. S4).
SPES-Xs were prepared from the corresponding SPTES-Xs by
selectively oxidizing the thioether groups to sulfone groups using
hydrogen peroxide as an oxidant. The quantitative conversion to
sulfone groups in SPES-Xs was confirmed from their 1H NMR
spectra (Fig. S3 (d)-(f)), with the peak at 7.4 ppm absent and the
intensity of the broad peaks from 7.85 to 7.93 ppm increased. The
proton peaks of the phenyl rings adjacent to the thioether linkages
in the SPTES-Xs were largely shifted downfield after the oxidation
reaction due to the strong electron-withdrawing character of sul-
fone groups.
The IECtitra values (1.74e2.05 mequiv$g1) of SPTES-Xs were
very close to the theoretical values (IECtheo, 1.78e2.11 mequiv$g1).
The IECNMR values calculated by comparing the peak areas at
8.3 ppm and 7.8e8.0 ppm were also in good agreement with the
IECtheo. Likewise, both IECtitra and IECNMR of SPES-Xs were also close
to their IECtheo values. The IEC values of the SPES-Xs became slightly
lower than those of the SPTES-Xs because converting thioether to
sulfone results in an increase in the average molecular weight of the
copolymer repeat units.
All SPTES-Xs and SPES-Xs had high inherent viscosities ranging
from 1.29 to 2.31 dl g1 in NMP solutions, indicating their high
molecular weights. The inherent viscosities of the SPES-Xs were not
much different from those of the SPTES-Xs, which indicates the
absence of any side reactions, such as chain degradation, during the
oxidation process. The inherent viscosities and IEC values of the
SPTES-Xs and SPES-Xs copolymers are summarized in Table 1.
3.2. Water absorption behavior and density
The water uptake and swelling ratio of the SPTES-X, SPES-X, and
Nafion115 membranes were measured at room temperature under
immersion in deionized water. As shown in Table 1, the Nafion115
membrane, which had the lowest IEC (0.87 meq$g1), exhibited the
431
lowest water uptake (20%) and swelling ratio (40%) among the test
membranes. For both the SPTES-X and SPES-X membranes, the
water uptake and swelling ratio increased with increasing IEC. For
example, SPTES-55 and SPES-55 exhibited about 76% and 64%
higher water uptake than SPTES-45 and SPES-45, respectively. To
analyze the water absorption behavior of the membranes more
carefully, we prepared SPTES-X and SPES-X membranes with
different thicknesses (35, 50, 80, and 100 mm). As seen in Fig. S5, the
water uptake and swelling ratio of the membranes strongly depend
on their thickness, which indicates they show the volume of mixing
effects [35e40]. For example, the swelling ratio of SPES-55
increased from 107% to 154% with decreasing the membrane
thickness from 100 to 35 mm. The volume of mixing effect became
larger with increasing the degree of sulfonation of the polymers.
For example, the difference in swelling ratio between the SPTES-45
membranes with thickness of 35 and 100 mm was about 25%.
However, the difference between the 35 and 100 mm thick SPTES-
55 membranes was more than 50%. It should also be noted that
the volume of mixing effect became negligible for all SPTES-X and
SPES-X membranes when the thickness of membrane was larger
than 80 mm.
It is interesting that the SPES-X membranes always showed
slightly higher swelling ratios than the SPTES-X membranes. Since
the SPES-Xs have lower IEC values than the corresponding SPTES-
Xs, the increase in swelling ratio cannot be explained by their IEC
differences. We believe that the higher concentration of diphenyl
sulfone groups in the SPES-Xs, which have a more rigid but curved
structure than the diphenyl thioether groups in SPTES-X, results in
a higher free volume and, consequently, a larger swelling ratio in
water [41,42].
The higher free volume of SPES-Xs was roughly verified by the
density measurement. The densities of SPTES-Xs and SPES-Xs were
calculated by measuring weight and volume of the membranes in
dry state (Fig. S6). It is important to note that SPES-Xs exhibited
smaller densities than their corresponding SPTES-Xs, implying
their relatively smaller free volume. Moreover, the density of the
polymers increased as the degree of sulfonation increased probably
due to the enhanced inter- and/or intra-chain interactions by sul-
fonic acid groups [43].
3.3. Vanadium-ion permeability
Fig. 2 shows the change in VO2þ ion concentration over diffusion
time, as measured by UV-Vis spectroscopy, for the SPTES-X, SPES-X,
and Nafion115 membranes. The slope reflects the diffusion rate of
the VO2þ ions and the permeability of the membranes was calcu-
lated from the slope using Eq. (6). As seen in Table 2, Nafion115
showed the highest VO2þ permeability (7.936  107 cm2 min1),
whereas the SPTES-X and SPES-X membranes exhibited much
lower permeabilities, which is probably due to their more rigid
aromatic structure and smaller ion-conducting channel size
[14,44,45]. The slope in Fig. 2 increases with increasing IEC, which
indicates that a high degree of sulfonation leads to high VO2þ
permeability across the membranes. The VO2þ permeability of the
SPTES-X and SPES-X membranes increased from 1.227  107 to
2.958  107 cm2 min1 and 0.203  107 to
0.614  107 cm2 min1, respectively, with increasing sulfonation
degree (the value of X). It is important to note that there are large
differences between the VO2þ permeability of the SPTES-X and
SPES-X membranes. For example, the SPES-55 membrane exhibits
only 1/5 of the VO2þ permeability of SPTES-55. The SPES-Xs have
slightly lower IECNMR values than the corresponding SPTES-Xs, as
discussed in Section 3.1; therefore, the IEC difference does not
explain the lower permeability of the SPES-Xs. We believe that the
flexibility of the polymer chains induces this difference. The
432 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439

Table 1
IECtheo, IECtitra, IECNMR and inherent viscosity results for SPTES-Xs, SPES-Xs, and Nafion115.

Membrane IECtheo (mequiv$g1) IECtitra (mequiv$g1) IECNMR (mequiv$g1) a

[h] (dl$g1)

SPTES-45 1.78 1.74 1.68 2.19

SPTES-50 1.95 1.96 1.82 1.81
SPTES-55 2.11 2.05 2.14 1.29
SPES-45 1.68 1.67 1.57 2.31
SPES-50 1.83 1.80 1.71 1.83
SPES-55 1.99 1.95 2.02 1.49
b
Nafion115 e 0.87 ± 0.02
a
Inherent viscosity measured in NMP at 25  C.
b
The reference of [6]; Titration method.

noteworthy that the SPTES-X and SPES-X membranes, with the

Fig. 2. VO2þ permeability across SPTES-Xs, SPES-Xs, and Nafion115 membranes. Inset
shows expansion of the low concentration range.
exception of SPTES-45, have lower area resistances than Nafion115.
We believe that both the higher IECs and lower thickness of the
SPTES-X and SPES-X membranes are the origin of such low area
resistances.
The ion selectivity of the polymer membranes, defined as the
ratio of proton conductivity to VO2þ permeability, is summarized in
Table 2. While Nafion115 showed poor selectivity
(0.65  105 S min$cm3), the SPTES-X and SPES-X membranes
displayed much improved selectivity. As expected, SPES-Xs, which
possess both very low VO2þ selectivity and good proton conduc-
tivity, showed excellent ion selectivity values of higher than
16  105 S min$cm3. Among the SPES-X membranes, SPES-50
exhibited the highest ion selectivity (21.20  105 S min$cm3),
which indicates that this polymer has the most well-balanced
combination of proton-conduction and vanadium-ion-rejection
properties.

3.5. Ex situ chemical stability

presence of thioether linkage groups in the repeat unit of SPTES-X
offers large flexibility for the polymer chains and decreases the
glass transition temperature, which allows VO2þ ions to move
across the membranes more easily [41,46,47]. Since the oxidation
reaction converts these thioether groups to rigid sulfone groups,
the SPES-Xs have a much-decreased permeability. Generally a low
VO2þ permeability leads to a lower self-discharge rate and higher
coulombic efficiency, as will be discussed in the following Sections
3.6.
3.4. Area resistance
Table 2 lists the area resistances of the SPTES-X, SPES-X, and
Nafion115 membranes. We controlled the thickness of the SPTES-X
and SPES-X membranes to be 80 ± 5 mm, which is lower than that of
the 128-mm-thick Nafion115 membrane, because these polymers
exhibit significantly lower vanadium-ion permeability than
Nafion115. It was found that the area resistance of SPTES-X and
SPES-X decreased from 0.324 to 0.089 U cm1 and 0.227 to
0.075 U cm1, respectively, with increasing sulfonation degree. It is
Ex situ chemical stability tests on the polymer membranes were
performed in 0.1 M and 1.5 M VOþ 2 solutions. The color of a pure
0.1 M VOþ 2 solution is light yellow, but this changes to green when
the amount of blue-colored VO2þ increases through the reduction
of VOþ 2þ
2 to VO . Optical images of the color change of the test so-
lutions over measuring time are shown in Fig. S7. The solution
containing Nafion115 did not show any color change after 1400 h,
which is the same behavior as that of the blank solution (without a
membrane) and confirms the outstanding chemical stability of
Nafion115. For the SPTES-X membranes, the color of the test solu-
tions gradually changed to dark green over time, which indicates
that the strongly oxidizing VOþ 2 ions degrade the SPTES-X mem-
branes. Interestingly, the test solutions containing the SPES-X
membranes did not show any notable color change even after
1400 h. These results imply that the SPES-Xs possess significantly
improved chemical stability compared with the SPTES-Xs. To
quantitatively evaluate the membrane degradation, the concen-
tration of VO2þ in the test solutions was monitored by UV-Vis
spectroscopy (Fig. 3(a)). The Nafion115 membrane showed

Table 2
Basic properties, area resistances, vanadium ions permeabilities, and selectivities of SPTES-X, SPES-X, and Nafion115 membranes.

Membrane Thickness Swelling ratio (%) Water uptake (%) Area resistance Vanadium ion permeability Ion selectivity
(mm) (U$cm2) (  107 cm2 min1) (  105$S min$cm3)

SPTES-45 75 51 45 0.324 1.227 1.89

SPTES-50 77 72 56 0.175 2.596 1.69
SPTES-55 82 97 79 0.089 2.958 3.11
SPES-45 84 62 44 0.227 0.203 18.23
SPES-50 79 78 54 0.162 0.230 21.20
SPES-55 75 107 72 0.075 0.614 16.29
Nafion115 128 40 20 0.249 7.936 0.648
 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439 433

Fig. 3. Ex situ chemical stability test results; the VO2þ concentration change in (a) 0.1 M and (b) 1.5 M VOþ
2 solutions over soaking time.

negligible change in the VOþ2 concentration over the test period, as
expected. For the solutions containing the SPTES-X membranes,
sharp increases in the VO2þ concentration were observed during
the first 300 h, indicating their rapid degradation by VOþ 2 ions.
However, the VO2þ concentration in the solutions containing the
SPES-X membranes was very low and remained stable during the
testing period, which is the same behavior exhibited for the
Nafion115 membrane.
The soaking tests were further conducted for SPTES-X and SPES-
X membranes to evaluate their chemical stability in more severe
oxidative condition using 1.5 M VOþ 2 solution, which is the typical
concentration for VRFB operation. The concentration of VO2þ
reduced from VOþ 2 and the optical images of the color change in the
test solutions over measuring time are displayed in Fig. 3(b) and
Fig. S8, respectively. The results have a tendency similar to those
using the 0.1 M VOþ þ
2 solution. A small change in the VO2 concen-
tration over the test period was observed for Nafion115 membrane,
as expected. The increase of VO2þ concentration in the solutions
containing the SPES-X membranes was also very small, but slightly
higher than that with the Nafion115 membrane. However, the so-
lutions containing the SPTES-X membranes exhibited sharp in-
creases in the VO2þ concentration, confirming their rapid
degradation by VOþ 2 ions. Finally, we measured the inherent vis-
cosities of the post-test membranes and the results are
434 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439

summarized in Table 3. It can be seen that the inherent viscosities

of SPES-Xs decreased 10% or less, while the amount of decrease
became larger with increasing the degree of sulfonation. On the
contrary, large decrease in inherent viscosity from 49.7 to 55.6%
were observed for SPTES-X membranes, indicating extensive chain
scission in 1.5 M VOþ 2 solution.
We believe that the difference in chemical stability between the
SPTES-X and SPES-X membranes originates from their structural
differences. In the SPES-Xs, the adjacent electron-withdrawing
sulfone groups deactivate every benzene ring in the polymer
backbone. However, the thioether groups in the SPTES-X repeat
unit acts as an electron-donating group that increases the electron
density on the neighboring benzene rings. These results indicate
that electron deficiency in the hydrocarbon polymer chains is
necessary to achieve enhanced chemical stability in strongly
oxidizing VRFB electrolytes, which will be further discussed in
Section 3.7.
Fig. 4. VRFB single cell performance of SPES-X and Nafion115 membranes at various
current densities (40e100 mA cm2).

3.6. VRFB single-cell performance tests

were circulated continuously during the course of this experiment.
Fig. 4 compares the single-cell performances of VRFBs equipped
The OCV decay over time is shown in Fig. 5. The OCV of VRFBs with
with Nafion115 and SPES-X membranes operated under various
both SPES-50 and Nafion115 membranes initially decreased slowly
current densities (40, 60, 80, and 100 mA cm2). It has been re-
and then dropped rapidly to 0.8 V. The time taken for complete OCV
ported that the CE of VRFBs, defined as the ratio of the cell
drop was significantly longer for the cell with the SPES-50 mem-
discharge capacity to its charge capacity, is closely related to the
brane (303 h) than that with the Nafion115 membrane (86 h). This
diffusion of vanadium ions across the membrane [48,49]. Nafion115
is attributed to the much lower vanadium-ion permeability of the
exhibited the largest vanadium-ion permeability and the lowest CE
SPES-50 membrane and indicates that SPES-50 can effectively
of the test membranes over the entire range of test current den-
retain charge for longer than Nafion115 in VRFBs by hindering the
sities, while the CE gradually climbed from 91% to 94% with
crossover of vanadium ions.
increasing current density from 40 to 100 mA cm2 due to the
shorter crossover time of the vanadium species. On the contrary, all
the SPES-X membranes exhibited high CEs (99.3 ± 0.2%) over the 3.7. VRFB cycling tests and in situ chemical stability
entire range of current densities studied; these values are signifi-
cantly larger than those for the Nafion115 membrane. The excellent VRFB cycling tests were conducted to investigate the stability of
CEs of the SPES-X membranes can be attributed to their substan- the SPES-50 and SPTES-50 membranes under VRFB operating
tially low vanadium-ion permeability as described in Section 3.3. conditions, where the strong acidic and oxidizing electrolyte is
It was found that the EE varied inversely with current density for present. Fig. 6 shows the efficiencies during continuous charge-
all membranes due to the increased ohmic polarization [11,50,51]. discharge cycles (up to 400 cycles) at a current density of
The EE of the VRFB assembled with the Nafion115 decreased from 100 mA cm2 for VRFBs with SPES-50, SPTES-50, and Nafion115
86% to 75% upon increasing the current density from 40 to membranes. The initial CEs (EEs) for the SPES-50 and Nafion115
100 mA cm2. Although increasing the test current density also membranes were 99% (76%), and 97% (75%), respectively; these
decreased the EE of the VRFBs using SPES-X membranes, the SPES- values remained stable, with only a slight decline during the entire
X membranes showed comparable or better EEs compared with 400 cycling test. For the battery with the SPTES-50 membrane, the
those for Nafion115 at all measured current densities. Moreover, efficiencies were stable during the initial period of 150 cycles;
increasing the IEC improved the EE of the VRFBs using SPES-X
membranes, which is attributed to the lower area resistance. In
particular, the VRFBs with SPES-50 and SPES-55 membranes
showed 75 and 77% higher EEs, respectively, than SPES-45 at a
current density of 100 mA cm2.
Self-discharge of the VRFBs assembled with SPES-50 and
Nafion115 due to cross-mixing of the anolyte and catholyte was
studied by monitoring the open circuit voltage (OCV) at ambient
temperature after the cell was fully charged. Electrolyte solutions

Table 3
Inherent viscosity decrease of SPTES-X and SPES-X membranes after ex situ
soaking test in 1.5 M VOþ
2 solution.

Membrane Decrease in inherent viscosity(%)

SPTES-45 49.69
SPTES-50 50.28
SPTES-55 55.56
SPES-45 3.03
SPES-50 3.64
Fig. 5. Open circuit voltage curve of VRFB single cells with SPES-50 and Nafion115
SPES-55 10.11
membranes.
 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439
Fig. 6. Efficiency of VRFB single cells with SPTES-50, SPES-50 and Nafion115 for 400
charge-discharge cycles at 100 mA cm2.
however, large variation and fluctuation in both the CE and EE
occurred after the 150th cycle, and the test was stopped at the 200th
cycle due to the failure of the SPTES membrane. These results are
consistent with what we have observed in the ex situ chemical
stability tests. SPES-50, in which the electron-withdrawing sulfone
linkages deactivate every benzene ring, showed good chemical
stability in the VOþ 2 electrolyte, while SPTES-50, which contains
electron-donating thioether linkages in the repeat unit, degraded
gradually under the same conditions.
The variation in VRFB capacity with the different membranes is
plotted in Fig. 7. Since electrolyte rebalancing was required for
Nafion 115 after 200 cycles, it is valid to compare the capacity re-
sults until 200 cycles. The VRFB assembled with Nafion115 suffers
from serious capacity loss (below 43% of the initial value after 200
cycles), mainly due to significant vanadium-ion crossover through
the membrane. Conversely, the battery with SPES-50 showed much
better capacity retention (93%) over 200 cycles, further confirming
the low vanadium-ion crossover. The decrease in charge and
discharge times as a function of cycle number for each membrane is
also demonstrated in Fig. S9. The vanadium ion-crossover across a
membrane causes the imbalance of the amount of vanadium active
species between catholyte and anolyte, which results in a gradual
decrease in charge and discharge time. Nafion 115 showed a fast
decrease of both times, and marked difference between charge and
discharge time of each cycle, implying again its high vanadium ion
Fig. 7. Capacity retention of VRFB single cells with SPTES SPTES-50, SPES-50 and
Nafion115 over 200 cycles at 100 mA cm2.
435
permeability and low CE. In contrary, SPTES-50 (CE 98e99%) and
SPES-50 (CE 99e100%) exhibited not only smaller gap between
charge and discharge times but also much slower decay in charge/
discharge time than Nafion 115. Moreover, it can be seen that SPES-
50 with the lowest vanadium ion permeability among three
membrane shows the slowest decline. These results are in good
agreement with those of vanadium ion permeability test. The slow
rate of capacity decay is beneficial for long-term VRFB operation,
minimizing the need to rebalance the electrolyte repeatedly. In the
case of VRFB with SPTES-50, the capacity seemed stable during the
initial testing stages; however, a capricious change in the capacity
was observed after 100 cycles. Some membrane degradation due to
the poor chemical stability of SPTES-50 may be the origin of this
behavior.
After charge-discharge cycle tests, the cells with the SPES-50
and SPTES-50 membranes were disassembled and post-mortem
tests were performed on the membranes to further verify their in
situ chemical stability by measuring their cross-section SEM pro-
files, 1H NMR spectra, inherent viscosity, and IECtitra. For the
detailed analysis, “active area” and “edge area” of the post-test
membrane were separated and then washed thoroughly with a
1.5 M H2SO4 solution. Herein, the “active area” is the central section
of the membrane (7  7 cm2) that is in surface contact with the
carbon-felt electrodes and, thus, fully exposed to the electrolytes,
while the “edge area” refers to the edge section of the membrane
covered by the gasket and not directly exposed to the electrolyte.
Fig. 8 compares the 1H NMR spectra of the active and edge areas
of each membrane. The spectra from the edge area of both the post-
test SPTES-50 and SPES-50 membranes were identical with those of
the pristine samples (before testing, shown in Fig. S3). Their SEM
images (Fig. 9(a), (c)) showed dense cross-sections without any
pinholes and cracks were found on the edge area. Moreover, no
changes in the inherent viscosity and IECtitra values before and after
testing were found for the edge area of both membranes (Table 4).
From these results, we conclude that the edge areas of the mem-
branes are not degraded during the cycling test because they are
not in direct contact with the electrolyte.
Nevertheless, the active area of the post-test SPTES-50 mem-
brane showed a significant decrease in inherent viscosity (58%)
after 200 cycles. In the 1H NMR spectrum (Fig. 8(b)), the intensity of
the peaks at 7.4 ppm and 7.0e7.1 ppm (assigned as “d” and “c, e0 ”,
respectively, in the Fig. 8(b)) decreased significantly after the
cycling test. This indicates that the number of protons near the
thioether linkages, and the ether linkages in the sulfonated part,
decreased in the active area of the post-test SPTES-50 membrane,
which is attributed to extensive chain scission by the VOþ 2 ions.
Conversely, new aromatic peaks of relatively high intensity
appeared around 7.2e7.3 ppm, 7.68 ppm, and 7.75 ppm. From the
model compound analysis using oxybisbenzenethiol (Fig. S10(a)),
the new peak at 7.2e7.3 ppm was assigned to the protons located at
the ortho position of the thiol (-SH) groups. Moreover, from pre-
vious literature reports, we expect that the new peaks at 7.68 ppm
and 7.75 ppm correspond to aromatic protons adjacent to the
sulfoxide groups, which are likely produced by oxidation of the
thioether groups [52e55]. The strongly oxidizing VOþ 2 ions in the
electrolyte oxidize some of the thioether groups to sulfoxide
groups. Nevertheless, it is important to note that the IECtitra value of
the SPTES-50 active area rarely changed after the cycling tests, as
shown in Table 4. This indicates that the degradation did not
accompany the loss of sulfonic acid groups. From these findings, we
suggest that the degradation of SPTES-50 is mainly triggered by the
cleavage of thioether linkages and, thereby, the formation of thiol
groups as depicted in Fig. 10(a). Consequently, many pinholes
accompanying sharp cracks were observed across the entire cross
section of the SPTES-50 active-area SEM image (Fig. 9(b)), which
436 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439

Fig. 8. 1H NMR spectra of the membranes after cycling test: (a) edge area and (b) active area of SPTES-50 after 200 cycles, and (c) edge area and (d) active area of SPES-50 after 400
cycles.

Fig. 9. Cross-section SEM images of the membranes after cycling test: (a) edge area and (b) active area of SPTES-50 after 200 cycles, and (c) edge area and (d) active area of SPES-50
after 400 cycles.

again confirms the severe degradation of the SPTES-50 membrane

under the VRFB operating conditions.
Table 4 However, no big differences were observed between the active
IECtitra, and inherent viscosities of SPTES-50 and SPES-50 membranes after charge-
and edge area 1H NMR spectra (Fig. 8(c), (d)) for the post-test SPES-
discharge 200 cycles.
50 membrane, which indicates that SPES-50 possesses much
Membrane a
IECtitra (mequiv$g1) b
[h] (dl$g1) enhanced stability compared with SPTES-50. The IECtitra values of
SPTES-50 Edge area 1.96 1.81 both the active and edge areas also remained unchanged after
Active area 1.94 0.76 testing (Table 4). Furthermore, the active area SEM image (Fig. 9(d))
SPES-50 Edge area 1.81 1.83
displayed a fine, smooth, cross-section structure without any sign
Active area 1.81 0.98
of degradation. Nevertheless, it was found that the SPES-50
a
Titration with 0.01 M NaOH aq. membrane also showed some chemical degradation in the VRFB
b
Inherent viscosity measured in NMP at 25  C.
 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439 437

Fig. 10. Possible degradation products of SPTES-X and SPES-X in VRFB.

operating conditions because new small peaks at 6.93 ppm and linkages or not, we believe that using a polymer design strategy to
7.75 ppm appeared in the 1H NMR spectrum of the active area. To increase the electron deficiency is highly necessary for improving
assign these new peaks, we prepared an oligo(phenylene sulfone) the chemical stability of these membrane in the strong oxidizing
model compound containing hydroxyl (-OH) end groups and environment of VRFBs. Further enhancing the chemical stability
analyzed its 1H NMR spectrum (Fig. S10(b)). The aromatic protons while maintaining a low vanadium-ion permeability and high
located at the ortho position of the hydroxyl end groups have proton conductivity remain important research topics for future
similar chemical shifts of 6.9e7.0 ppm. From this finding, we pro- studies.
pose that the new small peaks in the 1H NMR spectrum of active
area SPES-50 are originate from the hydroxyl end groups that are
4. Conclusions
expected to form by chain scission near the ether linkages, as
shown in Fig. 10(b). This is further confirmed by the inherent vis-
With the aim of understanding more about the degradation of
cosity analysis, in which 75.7% decrease of inherent viscosity in
hydrocarbon membranes and developing sulfonated hydrocarbon
active-area was detected after the 400 cycling test (Fig. S11). This is
membranes with enhanced stability, we investigated a series of
somewhat unexpected because the SPES-50 membrane exhibited
sulfonated poly(ether sulfone) copolymers (SPES-Xs, where X ¼ 45,
much smaller decrease (3.64%) of inherent viscosity at the ex situ
50, and 55) and their application as IEMs for VRFBs. Sulfonated
soaking test using 1.5 M VOþ 2 solution. This result imply that the poly(thioether ether sulfone) copolymers (SPTES-Xs) containing
degradation behavior for SPES-50 in the in situ cycling tests is
thioether linkages were prepared via the direct copolymerization of
rather different from that observed in the ex situ tests.
sulfonated and non-sulfonated difluoride monomers, and SPES-Xs
We believe the differences between the ex situ test (soaking in
were then obtained by oxidizing the SPTES-Xs, which converted
VOþ 2 solution) and in situ VRFB operation test (presence of both the thioether linkages in SPTES-Xs to sulfone linkages. It was found
oxidizing VOþ 2 species and large cell voltages) conditions lead to that the SPES-X membranes, based on a rigid diphenyl sulfone
this phenomenon [56e58]. To verify this assumption, we carried
structure, possess a much-reduced vanadium-ion permeability
out another inherent viscosity analysis for the SPES-50 membrane
than their parent SPTES-X membranes and also Nafion115 mem-
used in the self-discharge test. It should be noted that, in the self-
branes. Owing to this low vanadium-ion permeability and good
discharge test, one side of the membrane is exposed to the fully
proton conductivity, the SPES-X membranes showed excellent
charged positive electrolyte (high SOC > 90%) for a long period of
selectivity, with selectivity values more than ten times higher than
test time as in the ex situ 1.5 M soaking test. In the result, 64.2%
those of Nafion115. The SPES-Xs, in which every benzene ring is
decrease in inherent viscosity was observed for the post-test SPES-
deactivated by the presence of electron-withdrawing sulfone
50 (Fig. S11), which value is much higher than that after 1.5 M
linkages, showed much better ex situ chemical stability than the
soaking test (3.64%). The 1H NMR analysis, where new small
SPTES-Xs, which contain an electron-donating thioether linkage in
peaks at 6.93 ppm and 7.75 ppm were observed, further confirms
their repeat unit. Consequently, the single-cell performance of a
the degradation of SPES-50 in the self-discharge condition
VRFB with a SPES-Xs membrane possesses higher coulombic effi-
(Fig. S12). Considering that the main difference of the environ-
ciencies, longer self-discharge time, and lower capacity loss over
ments between the self-discharge test and soaking test is whether
200 charge-discharge cycles as well as energy efficiencies
an electric potential difference is present across the membrane or
(40e100 mA cm2) comparable with that of the equivalent cell
not, the degradation of SPES-50 is believed to be accelerated by the
using a Nafion115 membrane. After the cycling tests, the SPTES-50
presence of cell voltage.
membrane showed many signs of degradation, such as a decrease
Nevertheless, it is important to note that the SPES-50 mem-
in inherent viscosity, pin-hole and crack formation, and the
brane, with increased electron deficiency, shows much more stable
appearance of new peaks in the 1H NMR spectrum. Although some
performance than the SPTES-50 membrane, with relatively
chemical degradation was also observed for the post-test SPES-50
electron-rich aromatic nuclei in the actual VRFB cycling test due to
membrane, it is evident that SPES-50 is much more stable than the
its enhanced chemical stability. We believe that the chemical sta-
parent SPTES-50. Finally, we conclude that using a polymer design
bility of SPES-Xs is also much better than that of the S-Radel
strategy to increase electron deficiency is necessary to improve the
membrane previously reported by Chen et al. [20]. In their results,
chemical stability of membranes under VRFB operating conditions.
the S-Radel membrane suffered large degradation in ex situ soaking
tests and withstood only a few tens of cycles in in situ VRFB cycling
tests. Considering that S-Radel has structural similarities with Acknowledgements
SPES-X, with the exception of whether every benzene ring in the
polymer repeat unit has adjacent electron-withdrawing sulfone This work was supported by the Program through the National
Research Foundation of Korea (NRF) Grant funded by the Korean
438 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439
government (MSIP:Ministry of Science, ICT and Future Planning)
(NRF-2015M1A2A2056722) and the New & Renewable Energy Core
Technology Program of the Korea Institute of Energy Technology
Evaluation and Planning (KETEP), granted financial resources from
the Ministry of Trade, Industry & Energy, Republic of Korea (no.
20153030031670).
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.electacta.2017.10.121.
References
[1] M. Skyllas-Kazacos, M. Rychcik, R.G. Robins, A.G. Fane, M.A. Green, New all-
vanadium redox flow cell, J. Electrochem. Soc. 133 (1986) 1057e1058.
https://doi.org/10.1149/1.2108706.
[2] K.J. Kim, M.-S. Park, Y.-J. Kim, J.H. Kim, S.X. Dou, M. Skyllas-Kazacos,
A technology review of electrodes and reaction mechanisms in vanadium
redox flow batteries, J. Mater. Chem. A 3 (2015) 16913e16933. https://doi.org/
10.1039/C5TA02613J.
[3] A. Parasuraman, T.M. Lim, C. Menictas, M. Skyllas-Kazacos, Review of material
research and development for vanadium redox flow battery applications,
Electrochimica Acta 101 (2013) 27e40. https://doi.org/10.1016/j.electacta.
2012.09.067.
[4] G. Kear, A.A. Shah, F.C. Walsh, Development of the all-vanadium redox flow
battery for energy storage: a review of technological, financial and policy
aspects, Int. J. Energy Res. 36 (2012) 1105e1120. https://doi.org/10.1002/er.
1863.
[5] S.-W. Choi, S.-H. Cha, T.-H. Kim, Nanostructured membranes of vanadium
redox flow batteries, Nanosci. Nanotech.-Asia 5 (2015) 109e129. https://doi.
org/10.2174/2210681205666150903213628.
[6] B. Jiang, L. Wu, L. Yu, X. Qui, J. Xi, A comparative study of Nafion series
membranes for vanadium redox flow batteries, J. Membr. Sci. 510 (2016)
18e26. https://doi.org/10.1016/j.memsci.2016.03.007.
[7] M. Ulaganathan, V. Aravindan, Q. Yan, S. Madhavi, M. Skyllas-Kazacos,
T.M. Lim, Recent advancements in all vanadium redox flow batteries, Adv.
Mater. Interfaces 3 (2016), 1500309. https://doi.org/10.1002/admi.
201500309.
[8] S.-H. Cha, Recent development of nanocomposite membranes for vanadium
redox flow batteries, J. Nanomater. 2015 (2015), 207525. https://doi.org/10.
1155/2015/207525.
[9] W. Dai, L. Yu, Z. Li, J. Yan, L. Liu, J. Xi, X. Qiu, Sulfonated poly(ether ether ke-
tone)/graphene composite membrane for vanadium redox flow battery,
Electrochimica Acta 132 (2014) 200e207. https://doi.org/10.1016/j.electacta.
2014.03.156.
[10] J. Xi, W. Dai, L. Yu, Polydopamine coated SPEEK membrane for a vanadium
redox flow battery, RSC Adv. 5 (2015) 33400e33406. https://doi.org/10.1039/
C5RA01486G.
[11] B. Yin, Z. Li, W. Dai, L. Wang, L. Yu, J. Xi, Highly branched sulfonated poly(-
fluorenyl ether ketone sulfone)s membrane for energy efficient vanadium
redox flow battery, J. Power Sources 285 (2015) 109e118. https://doi.org/j.
jpowsour.2015.03.102.
[12] J. Li, Y. Zhang, S. Zhang, X. Huang, Sulfonated polyimide/s-MoS2 composite
membrane with high proton selectivity and good stability for vanadium redox
flow battery, J. Membr. Sci. 490 (2015) 179e189. https://doi.org/j.memsci.
2015.04.053.
[13] Y. Zhang, S. Zhang, X. Huang, Y. Zhou, Y. Pu, H. Zhang, Synthesis and prop-
erties of branched sulfonated polyimides for membranes in vanadium redox
flow battery application, Electrochimica Acta 210 (2016) 308e320. https://
doi.org/j.electacta.2016.05.116.
[14] S. Kim, J. Yan, B. Schwenzer, J. Zhang, L. Li, Z.G. Yang, M.A. Hickner, Cycling
performance and efficiency of sulfonated poly(sulfone) membranes in vana-
dium redox flow batteries, Electrochem. Commun. 12 (2010) 1650e1653.
https://doi.org/j.elecom.2010.09.018.
[15] L. Semix, N.D. Sankir, M. Sankir, Influence of the basic membrane properties of
the disulfonated poly(arylene ether sulfone) copolymer membranes on the
vanadium redox flow battery performance, J. Membr. Sci. 468 (2014)
209e215. https://doi.org/j.memsci.2014.06.019.
[16] C. Fujimoto, S. Kim, R. Stains, X. Wei, L. Li, Z.G. Yang, Vanadium redox flow
battery efficiency and durability studies and sulfonated Diels Alder poly(-
phenylene)s, Electrochem. Commun. 20 (2012) 48e51. https://doi.org/j.
elecom.2012.03.037.
[17] M.S. Cha, J.Y. Lee, T.-H. Kim, H.Y. Jeung, H.Y. Shin, S.-G. Oh, Y.T. Hong, Prep-
aration and characterization of crosslinked anion exchange membrane (AEM)
materials with poly(phenylene ether)-based short hydrophilic block for use in
electrochemical applications, J. Membr. Sci. 530 (2017) 73e83. https://doi.org/
j.memsci.2017.02.015.
[18] S. Kim, T.B. Tighe, B. Schwenzer, J. Yan, J. Zhang, J. Liu, Z. Yang, M.A. Hickner,
Chemical and mechanical degradation of sulfonated poly(sulfone) membranes
in vanadium redox flow batteries, J. Appl. Electrochem 41 (2011) 1201e1213.
https://doi.org/.
[19] T. Sukkar, M. Skyllas-Kazacos, Membrane stability studies for vanadium redox
cell applications, J. Appl. Electrochem 34 (2004) 137e145. https://doi.org/10.
1007/s10800-011-0313-0.
[20] D. Chen, M.A. Hickner, V5þ degradation of sulfonated Radel membranes for
vanadium redox flow batteries, Phys. Chem. Chem. Phys. 15 (2013)
11299e11305. https://doi.org/10.1039/C3CP52035H.
[21] D. Chen, S. Kim, L. Li, G. Yang, M.A. Hickner, Stable fluorinated sulfonated
poly(arylene ether) membranes for vanadium redox flow batteries, RSC Adv. 2
(2012) 8087e8094. https://doi.org/10.1039/C2RA20834B.
[22] Z. Yuan, X. Li, J. Hu, W. Xu, J. Cao, H. Zhang, Degradation mechanism of sul-
fonated poly(ether ether ketone) (SPEEK) ion exchange membranes under
vanadium redox flow battery medium, Phys. Chem. Chem. Phys. 16 (2014)
19841e19847. https://doi.org/10.1039/C4CP03329A.
[23] J. Xi, Z. Li, L. Yu, B. Yin, L. Wang, L. Liu, X. Qiu, L. Chen, Effect of degree of
sulfonation and casting solvent on sulfonated poly(ether ether ketone)
membrane for vanadium redox flow battery, J. Power Sources 285 (2015)
195e204. https://doi.org/j.jpowsour.2015.03.104.
[24] X. Huang, S. Zhang, Y. Zhang, H. Zhang, X. Yang, Sulfonated polyimide/chi-
tosan composite membranes for a vanadium redox flow battery: influence of
the sulfonation degree of the sulfonated polyimide, Polym. J. 48 (2016)
905e918. https://doi.org/10.1038/pj.2016.42.
[25] Z. Yuan, X. Li, Y. Duan, Y. Zhao, H. Zhang, Application and degradation
mechanism of polyoxadiazole based membrane for vanadium flow batteries,
J. Membr. Sci. 488 (2015) 194e202. https://doi.org/j.memsci.2015.04.019.
[26] J.-K. Jang, T.-H. Kim, S.J. Yoon, J.Y. Lee, J.-C. Lee, Y.T. Hong, 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 4 (2016) 14342e14355. https://doi.
org/10.1039/C6TA05080H.
[27] X.L. Zhou, T.S. Zhao, L. An, L. Wei, C. Zhang, The use of polybenzimidazole
membranes in vanadium redox flow batteries leading to increased coulombic
efficiency and cycling performance, Electrochimica Acta 153 (2015) 492e498.
https://doi.org/j.electacta.2014.11.185.
[28] S. Maurya, S.H. Shin, J.Y. Lee, Y. Kim, S.H. Moon, Amphoteric nanoporous
polybenzimidazole membrane with extremely low crossover for a vanadium
redox flow battery, RSC Adv. 6 (2016) 5198e5204. https://doi.org/10.1039/
C5RA26244E.
[29] T. Luo, O. David, Y. Gendel, M. Wessling, Porous poly (benzimidazole) mem-
brane for all vanadium redox flow battery, J. Power Sources 312 (2016)
45e54. https://doi.org/j.jpowsour.2016.02.042.
[30] X. Ling, C. Jia, J. Liu, C. Yan, Preparation and characterization of sulfonated poly
(ether sulfone)/sulfonated poly (ether ether ketone) blend membrane for
vanadium redox flow battery, J. Membr. Sci. 415 (2012) 306e312. https://doi.
org/j.memsci.2012.05.014.
[31] W. Xu, X. Li, J. Cao, Z. Yuan, H. Zhang, Morphology and performance of pol-
y(ether sulfone)/sulfonated poly(ether ether ketone) blend porous mem-
branes for vanadium flow battery application, RSC Adv. 4 (2014)
40400e40406. https://doi.org/10.1039/C4RA05083E.
[32] M. Ueda, H. Toyota, T. Teramoto, Synthesis and characterization of aromatic
poly(ether sulfone)s containing pendant sodium sulfonate groups, J. Polym.
Sci. Part A Polym. Chem. 31 (1993) 853e858. https://doi.org/10.1002/pola.
1993.080310402.
[33] M. Schuster, K.-D. Kreuer, H.T. Andersen, J. Maier, Sulfonated poly(phenylene
sulfone) polymers as hydrolytically and thermooxidatively stable proton
conducting ionomers, Macromolecules 40 (2007) 598e607. https://doi.org/10.
1021/ma062324z.
[34] G.-J. Hwang, H. Ohya, Preparation of cation exchange membrane as a sepa-
rator for the all-vanadium redox flow battery, J. Membr. Sci. 120 (1996)
55e67. https://doi.org/10.1016/0376-7388(96)00135-4.
[35] X. Wu, X. Wang, G. He, J. Benziger, Differences in water sorption and proton
conductivity between Nafion and SPEEK, J. Polym. Sci. B Polym. Phys. 49
(2011) 1437e1445. https://doi.org/10.1002/polb.22326.
[36] X. Yan, G. He, X. Wu, J. Benziger, Ion and water transport in functionalized
PEEK membranes, J. Membr. Sci. 429 (2013) 13e22. https://doi.org/10.1016/j.
memsci.2012.11.026.
[37] A. Kusoglu, D. Kushner, D.K. Paul, K. Karan, M.A. Hickner, A.Z. Weber, Impact
of substrate and processing on confinement of Nafion thin films, Adv. Funct.
Mater 24 (2014) 4763e4774. https://doi.org/10.1002/adfm.201304311.
[38] S.G. Jo, T.-H. Kim, S.J. Yoon, S.-G. Oh, M.S. Cha, H.Y. Shin, J.M. Ahn, J.Y. Lee,
Y.T. Hong, Synthesis and investigation of random-structured ionomers with
highly sulfonated multi-phenyl pendants for electrochemical applications,
J. Membr. Sci. 510 (2016) 326e337. https://doi.org/j.memsci.2016.03.018.
[39] K.S. Yoon, J.Y. Lee, T.-H. Kim, D.M. Yu, S.-K. Hong, Y.T. Hong, Multiblock co-
polymers based on poly(p-phenylene)-co-poly(arylene ether sulfone ketone)
with sulfonated multiphenyl pendant groups for polymer electrolyte fuel cell
(PEMFC) application, Eur. Polym. J. 66 (2015) 1e11. https://doi.org/j.
eurpolymj.2015.01.021.
[40] J.Y. Lee, D.M. Yu, T.-H. Kim, S.J. Yoon, Y.T. Hong, Multi-block copolymers based
on poly(p-phenylene)s with excellent durability and fuel cell performance,
J. Membr. Sci. 492 (2015) 209e219. https://doi.org/j.memsci.2015.04.013.
[41] M.S. Jung, T.-H. Kim, Y.J. Yoon, C.G. Kang, D.M. Yu, J.Y. Lee, H.-J. Kim, Y.H. Hong,
Sulfonated poly(arylene sulfone) multiblock copolymers for proton exchange
membrane fuel cells, J. Membr. Sci. 459 (2014) 72e85. https://doi.org/j.
 S.-W. Choi et al. / Electrochimica Acta 259 (2018) 427e439
memsci.2014.01.072.
[42] S. Niemela €, J. Leppa
€nen, F. Sundholm, Structural effects on free volume dis-
tribution in glassy polysulfones: molecular modelling of gas permeability,
Polymer 37 (1996) 4155e4165. https://doi.org/10.1016/0032-3861(96)
00241-8.
[43] G. Chowdhury, K. Boguslaw, M. Takeshi, Polyphenylene Oxide and Modified
Polyphenylene Oxide Membranes: Gas, Vapor and Liquid Separation, Ch. 3,
Springer Science & Business Media, USA, 2001, p. 71.
[44] M.A. Hickner, C.H. Fujimoto, C.J. Cornelius, Transport in sulfonated poly(-
phenylene)s: proton conductivity, permeability, and the state of water,
Polymer 47 (2006) 4238e4244. https://doi.org/j.polymer.2006.02.034.
[45] Z. Li, W. Dai, L. Yu, J. Xi, X. Qiu, L. Chen, Sulfonated poly(ether ether ketone)/
mesoporous silica hybrid membrane for high performance vanadium redox
flow battery, J. Power Sources 257 (2014) 221e229. https://doi.org/j.
jpowsour.2014.01.127.
[46] J.-G. Liu, Y. Nakamura, Y. Shibasaki, S. Ando, M. Ueda, Synthesis and charac-
terization of highly refractive polyimides from 4,40 -thiobis[(p-phenyl-
enesulfanyl)anilie] and various aromatic tetracarboxylic dianhydrides,
J. Polym. Sci. Part A Polym. Chem. 45 (2007) 5606e5617. https://doi.org/10.
1002/pola.22308.
[47] S. Masaki, N. Sato, A. Nishichi, S. Yamazaki, K. Kimura, Fluorinated poly(aryl
thioether)s and poly(aryl sulfone)s derived from 2,3,4,5,6-penrafluorobenzoic
acid, J. Appl. Polym. Sci. 108 (2008) 498e503. https://doi.org/10.1002/app.
27347.
[48] X. Li, H. Zhang, Z. Mai, H. Zhang, I. Vankelecom, Ion exchange membranes for
vanadium redox flow battery (VRB) applications, Energy Environ. Sci. 4 (2011)
1147e1160. https://doi.org/10.1039/C0EE00770F.
[49] S. Xiao, L. Yu, L. Wu, L. Liu, X. Qiu, J. Xi, Broad temperature adaptability of
vanadium redox flow battery-Part 1: electrolyte research, Electrochimica Acta
187 (2016) 525e534. https://doi.org/j.electacta.2015.11.062.
[50] Y.J. Leng, S.H. Chan, S.P. Jiang, K.A. Khor, Low-temperature SOFC with thin film
439
GDC electrolyte prepared in situ by solid-state reaction, Solid State Ionics 170
(2004) 9e15. https://doi.org/j.ssi.2004.02.026.
[51] D. Aaron, Z. Tang, A.B. Papandrew, T.A. Zawodzinski, Polarization curve
analysis of all-vanadium redox flow batteries, J. Appl. Electrochem 41 (2011)
1175e1182. https://doi.org/10.1007/s10800-011-0335-7.
[52] R.E. Hardy, P.R.H. Speakman, P. Robson, The vanadium pentoxide-catalysed
oxidation of thio-compounds with hydrogen peroxide, J. Chem. Soc. C. Org.
18 (1969) 2334e2336. https://doi.org/10.1039/J39690002334.
[53] V. Conte, A. Coletti, B. Floris, G. Licini, C. Zonta, Mechanistic aspects of vana-
dium catalysed oxidations with peroxides, Coord. Chem. Rev. 255 (2011)
2165e2177. https://doi.org/j.ccr.2011.03.006.
[54] D. Zhao, J. Li, M.-K. Song, B. Yi, H. Zhang, M. Liu, A durable alternative proton-
exchange membranes: sulfonated poly(benzoxazole thioether sulfone)s, Adv.
Energy Meter 1 (2011) 203e211. https://doi.org/10.1002/aenm.201000062.
[55] P. Veerakumar, Z.-Z. Lu, M. Velayudham, K.-L. Lu, S. Rajagopal, Alumina sup-
ported nanoruthenium as efficient heterogeneous catalyst for the selective
H2O2 oxidation of apliphatic and aromatic sulfides to sulfoxides, J. Mol. Catal.
A Chem. 332 (2010) 128e137. https://doi.org/j.molcata.2010.09.008.
[56] S. Yun, J. Parrondo, V. Ramani, Derivatized cardo-polyetherketone anion ex-
change membranes for all-vanadium redox flow batteries, J. Mater. Chem. A 2
(2014) 6605e6615. https://doi.org/10.1039/C4TA00166D.
[57] A.M. Pezeshki, R.L. Sacci, G.M. Veith, T.A. Zawodzinski, M.M. Mench, The cell-
in-series method: a technique for accelerated electrode degradation in redox
flow batteries, J. Electrochem. Soc. 163 (2016) A5202eA5210. https://doi.org/
10.1149/2.0251601jes.
[58] I. Derr, M. Bruns, J. Langner, A. Fetyan, J. Melke, C. Roth, Degradation of all-
vanadium redox flow batteries (VRFB) investigated by electrochemical
impedance and X-ray photoelectron spectroscopy: Part 2 electrochemical
degradation, J. Power Sources 325 (2016) 351e359. https://doi.org/j.
jpowsour.2016.06.040.