Journal of The Electrochemical

Society

OPEN ACCESS

Electrochemical Aging and Characterization of Graphite-Polymer Based

Composite Bipolar Plates for Vanadium Redox Flow Batteries
To cite this article: Gaurav Gupta et al 2022 J. Electrochem. Soc. 169 080503

View the article online for updates and enhancements.

This content was downloaded from IP address 103.27.10.101 on 02/02/2023 at 06:02

 Journal of The Electrochemical Society, 2022 169 080503

Electrochemical Aging and Characterization of Graphite-Polymer

Based Composite Bipolar Plates for Vanadium Redox Flow
Batteries
Gaurav Gupta,1 Barbara Satola,1 Lidiya Komsiyska,1,a Corinna Harms,2,z
Thorsten Hickmann,3 and Alexander Dyck1
1
German Aerospace Center, Institute of Networked Energy Systems, Carl-von-Ossietzky-Str. 15, 26129 Oldenburg, Germany
2
German Aerospace Center, Institute of Engineering Thermodynamics, Carl-von-Ossietzky-Str. 15, 26129 Oldenburg,
Germany
3
Eisenhuth GmbH & Co. KG, Friedrich-Ebert-Str. 203, 37520 Osterode, Germany

Three bipolar plates (BPP) comprised of a composite of polypropylene or polyvinylidene fluoride polymer and varying average
graphite particle size were studied for application in a vanadium redox flow battery (VRFB). The BPPs were electrochemically
aged via 3000 cyclic voltammetry curves in 1.8 M VOSO4 + 2.0 M H2SO4 electrolyte. After every 500th cycle the aging
progression was determined by performing cyclic voltammetry on the bipolar plates in 0.1 M H2SO4 solution where the double
layer capacitance, the quinone/hydroquinone and the vanadium species redox activity were quantitatively evaluated. Prior to the
aging, the composite plates were extensively characterized using various physical methods. The performed studies reveal that the
wettability, surface roughness and accessible porosity of the bipolar plates significantly influence their electrochemical stability.
Cycling tests in vanadium redox flow single cells at a constant current density of 60 mA cm−2 revealed a close correlation of the
cell efficiencies to the electrochemical stability of the bipolar plates. Thus, the proposed electrochemical characterization method
can be an effective foresight to predict the applicability of a bipolar plate in a vanadium redox flow battery.
© 2022 The Author(s). Published on behalf of The Electrochemical Society by IOP Publishing Limited. This is an open access
article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/
by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/
1945-7111/ac8240]

Manuscript submitted December 22, 2021; revised manuscript received June 27, 2022. Published August 2, 2022.

The world is transitioning from non-renewable to renewable Overall reaction:

energy sources such as wind and solar power. These sources are
environmentally friendly and cost efficient, however, their inability VO 2+ + V 2 + + 2H + ↔ VO 2 + + V3 +
to generate power constantly hinders their adaptability.1 This issue + H2 O U° = +1.26 V vs SHE
can be overcome by deployment of energy storage systems such as
redox flow batteries (RFBs). The significant advantage of RFBs is The BPPs conduct electrical current and substantially influence the
their ability to scale the power and energy capacity independently, power capability of the VRFB, while separating electrolyte solutions
i.e. higher power can be achieved by increasing the electrode area or between the adjacent cells and also providing mechanical integrity to
number of cells in series while higher electrolyte volume or molar the VRFB stack.4 Hence, it is essential that the BPPs have high
concentrations would raise the capacity.2–4 Moreover, as the mechanical strength, low ion permeability, high electrical conductivity
electrolytes are water-based, fire hazard is reduced compared to and low contact resistance to the porous electrode. Apart from these, for
lithium-ion battery systems.5 application in a VRFB other desired characteristics of a BPP is their
The all vanadium redox flow battery (VRFB) owed to its high chemical and electrochemical stability4,9 and high overpotential to
cycling life, quick response time4 and deep-discharge capability6 has hydrogen and oxygen evolution.9
attracted a lot of attention for grid-scale energy storage system3 and BPPs can be generally divided into three categories: metallic,
provides power at peak time.7 As vanadium is the redox couple in graphitic and composite type.4,8 The high electrical conductivity and
both half-cells, rebalancing of the electrolyte can restore the capacity easy formability make the metallic BPPs quite attractive,8 but low
loss caused by vanadium species crossover.2,3 The VRFB stack stability in an acidic medium impedes their use in VRFBs.4,8,9 The
consists of porous electrodes, bipolar plates (BPP), membranes and graphitic BPPs have high chemical stability accompanied with good
current collectors.4,8 The positive (VO2+/VO2+) and negative electrical conductivity,4 however, their poor processability, low
(V2+/V3+) electrolytes are composed of dissolved vanadium ions mechanical strength, high cost4,8,9 and porous structure9 make
in a sulfuric acidic solution and stored in separate reservoirs. The them unattractive for the application in RFBs. The composite
electrolytes are pumped through each half-cell during battery BPPs owed to their high chemical stability, mechanical strength,
operation and the half-cells are separated by an ion exchange low production cost4,6 and light weight6 are a viable option. The
membrane.8,9 The electrochemical reactions occur on the surface polymer, such as thermoplastics (e.g. polyethylene) and
of the porous electrodes as shown below thermosets,4,9 provide the matrix for the mechanical integrity and
Positive electrode: carbon-based fillers like graphite, carbon black and carbon
nanotubes4,9 impart the conductive properties. It has been reported
VO 2+ + 2H + + e− ↔ VO 2 + + H2 O E° = +1.0 V vs SHE that the type6,7 and particle size8,10 of graphite can have a significant
effect on the electrical and electrochemical properties of the BPP.
Negative electrode: Even though the vanadium redox reactions as well as side
reactions, such as oxygen and hydrogen evolution, occur on the
V 2 + ↔ V3 + + e− E° = −0.26 V vs SHE
graphite felt electrodes during battery operation, some working
groups4,9,10 have reported that occasionally electrochemical reac-
tions may also occur on the BPP surface. This can happen if an
abnormal current passes through the VRFB electrolyte causing the
a sulfuric acid in the electrolyte to oxidize the BPP. This phenomenon
Present address: Technische Hochschule Ingolstadt, Esplanade 10, 85049
Ingolstadt, Germany may affect the integrity of the BPP and negatively influence the
z
E-mail: [email protected] VRFBs performance.7
 Journal of The Electrochemical Society, 2022 169 080503

A few studies have been published describing various electro- Table I. Investigated graphite-polymer composite BPPs.
chemical characterization methods and VRFB cycling tests. Cyclic
voltammetry (CV) on a carbon-polythene composite BPP was Constituents wt. [%]
performed in a 2 M VOSO4 + 2 M H2SO4 at 20 °C by
potentiodynamic polarization from 0 to 2.5 V vs SCE at a scan BPP PP PVDF GS GL
rate of 5 mV s−1.6 It was demonstrated that the carbon-polythene
composite BPP had higher gas evolution potentials and better PP-GS 14 — 86 —
electrochemical stability compared to a graphite felt and a graphite PVDF-GS — 20 80 —
electrode. Despite the suppression of gas evolution, the new BPP PVDF-GL — 20 — 80
showed severe structural erosion and decline in electrical conduc-
tivity. In another study, a graphite-polymer composite BPP con-
sisting of large flake type major filler with 80 μm and minor filler
electrochemical behaviour of a BPP.6,8,10,11 Furthermore, surficial
with 50 nm, was characterized electrochemically in a 0.1 M VOSO4
properties of a BPP such as hydrophilicity and surface roughness
+ 3 M H2SO4 electrolyte solution. The CV measurement was
could be an indication of their electrochemical stability.12 In the
performed by sweeping between −0.2 and +1.6 V vs Ag/AgCl at a
current contribution three different graphite-polymer composite
scan rate of 5 mV s−1.8 The developed composite BPP showed
BPPs were subjected to identical electrochemical characterization
suppressed oxygen evolution. Single cell VRFB tests with the
and compared to single cell VRFB tests in order to determine if
developed and commercial BPPs at current densities of 40, 60, 80,
preliminary electrochemical investigations might forecast the per-
100 mA cm−2 were also performed. The new BPP exhibited an
formance of BPPs in real battery application. The work correlates
energy efficiency decay of 0.87% compared to 2.5% for the
the influence of surficial properties to the electrochemical behaviour
commercial one. While Zhang et al. a carbon-polypropylene
and to the VRFB efficiencies.
thermoplastic elastomer (PP-elastomer) composite BPP was com-
pared to a graphite BPP in a 1.5 M VOSO4 + 2.0 M H2SO4, here CV
Experimental
was performed by sweeping between −0.1 and +2.0 V vs SCE at a
scan rate of 20 mV s−1.10 The CV of the graphite BPP showed In this study three types of compress molded graphite-polymer
significant vanadium redox activity (VO2+ ⇄ VO2+ and V3+ ⇄ composite BPPs (Eisenhuth GmbH & Co. KG, Germany) were
V2+). In contrast, the PP-elastomer composite BPP exhibited only a investigated (Table I). The studied composite BPPs constitute either
cathodic peak (VO2+ → VO2+). Furthermore, a 5-cell VRFB stack polypropylene (PP) or polyvinylidene fluoride (PVDF) polymers and
test at a current density of 50 mA cm−2 with the prepared PP- different graphite types with small graphite (GS) or large graphite
elastomer composite BPP achieved a stable performance for 2300 (GL) particles sizes (PS).
cycles with an average energy efficiency of 75% while the graphite As a pre-treatment all as received BPPs were cleaned with
BPP broke down after 1450 cycles and the test was discontinued. In ethanol solution and dried with low coarse tissue paper (Kimtech,
another investigation by Lee et al. carbon polymer (epoxy resin) USA). These BPPs are referred as pristine.
composite BPP (15 wt% carbon black) was prepared and character- In order to study the structural and physical morphology of GS
ized by comparing it against a conventional BPP by potentiodynamic and GL particles the two raw graphite powders were imaged and
cycling in 2.5 M H2SO4 by sweeping from +0.1 to +2.0 V vs SCE their average graphite PS was quantified using X-ray computed
at a scan rate of 5 mV s−1.11 The composite BPP exhibited lower micro tomography (mCT, SkyScan 1172, Bruker, Belgium). For
corrosion current densities compared to the conventional BPP. the mCT scan the raw graphite powders were introduced onto the
Though, single cell VRFB test at current density of 40 mA cm−2 sample holder which was then placed onto the measuring stage. The
revealed similar energy efficiencies for both BPPs. scan resolution was set to 1 μm pixel−1 with a source voltage of
Apart from the reported potentiodynamic methods, galvanody- 80 kV, rotation step of 0.18° and rotation angle from 0° to 180°.
namic characterization has also been reported.12 A graphite-poly- Averages of six frames were recorded to generate a set of cross-
propylene composite BPP was galvanodynamically aged in a sections of images in gray scale. The recorded gray scale data was
positive vanadium electrolyte with 1.6 M VO2+/VO2+ + 2 M loaded onto the NRecon software (SkyScan, Bruker, Belgium)
H2SO4 + 0.05 M H3PO4at varying states of charge (10, 50 and where the 3D images were reconstructed. The range of the histogram
90%). The CV was recorded by sweeping the current density from the gray scale, value of ring artefact reduction and beam
between +100 mA cm−2 and −100 mA cm−2 at a scan rate of 20 hardening were kept the same for each graphite powder. A volume
mA s−1. The aged BPPs were examined in a diluted H2SO4 solution of interest (VOI) of 2000 μm × 2000 μm × 1000 μm was defined
to assess the effect of aging by sweeping between 0 V and +1.05 V for quantification of the PS size (CTAn, SkyScan, Bruker, Belgium).
vs Hg/Hg2SO4 at a scan rate of 5 mV s−1. Here, three aging A series of operations were defined for analysis in the following
indicators, the double layer capacitance (Cdl), the quinone/hydro- order (i) filtering and despeckling was done to eliminate the
quinone (QQ/HQ) and the vanadium redox activity (QV4+/V5+) were background noise, (ii) binary imaging was performed by bitmap
identified. They evaluated these aging indicators and related them to operation through qualitative determination of the minimum and
increased hydrophilicity of the BPP and oxidation of the graphite maximum values on the threshold scale to segment the image
compound.12 between graphite particles (white) and background (black) followed
From above it can be derived that the constituents, the graphite by (iii) morphological and quantitative 3D analysis to determine the
particle size and wt% composition can significantly affect the average graphite PS.

Figure 1. (a) In-plane and (b) through-plane electrical conductivity measurement setup for BPPs.
 Journal of The Electrochemical Society, 2022 169 080503

Figure 2. Example and determination of electrochemical aging features of the BPP in 0.1 M H2SO4 solution (red line indicates the integration of the curve done
on originlab software).

The accessible porosity was determined through intrusion of Corporation, DMM 4050, USA). A constant pressure of 1 N
mercury into the accessible pores of the BPPs. The BPPs were mm−2 was applied. Usually, after roughly 1 min the resistance value
introduced into a dilatometer and placed onto a mercury porosimetry was stable and it was noted after 3 min. Three measurements were
device (Pascal 140, Pascal 400, Thermo Fisher Scientific S.p.A., done per BPP and the average through-plane electrical conductivity
Italy). The accessible porosity percentage was estimated using SOL. (σth) was calculated.
I.D. software (Thermo Fisher Scientific S.p.A., Italy).
The density of each BPP was defined by measuring the mass on a σ in =1/p σ th = L/RA
lab weighing scale and dividing it by the volume.
Where σin: in—plane electrical conductivity [S cm−1].
D = M /V σth: through—plane electrical conductivity [ S cm−1].
L: thickness of BPP [cm], R: resistance [Ω], A: area of BPP[cm2],
Where, {D: density [g cm−3], M: mass[g] and V: Volume[cm−3]} p; resistivity.
In-plane electrical resistivity was measured on a four-point probe The aging and analysis of the BPPs was done in two separate
testing unit (RM3-AR, Jandel, UK) (Fig. 1a). Five measurements three-electrode setups, each consisting of the investigated BPP as
were performed on different locations of each BPP surface and the working electrode (WE), Ag/AgCl in 3 M KCl (+0.210 V vs SHE)
average electrical conductivity (σin) was calculated. The through- as reference electrode and the counter electrode was a platinum mesh
plane electrical conductivity of the BPP was measured on an adapted attached to a thermally treated porous graphite felt (GFD4.6EA,
strain/compression testing machine (ZwickRoell, Ulm, Germany) SGL Carbon group, Germany). The test setups were placed in a
(Fig. 1b). The BPP was placed between two gold plated copper Faraday cage. The potential sweeps were applied by a potentiostat/
stamps. In order to improve the contact between the BPP and stamp galvanostat (VersaSTAT 3F, Ametek, USA). The BPP was placed in
surfaces and for better comparison between samples with different a sample holder (Metrohm Autolab B. V., Netherlands) made of PP
surface topology, gas diffusion layer sheets (GDL, Freudenberg and sealed with a polytetrafluoroethylene seal locking ring. The
Group, Germany) were placed in between. The current was applied, active area of the BPP was 1 cm2. Prior to each measurement the
and voltage drop was measured on a multimeter (Tektronix aging and analysis solutions were purged with nitrogen gas for at

Figure 3. mCT 2D images of (a) GS and (b) GL graphite powders at 1 μm/pixel resolution.
 Journal of The Electrochemical Society, 2022 169 080503

Table II. Structural properties of the composite BPPs and the PS of the raw graphite particles used in production of the BPPs.

Properties Average graphite PS in raw graphite powders [μm] Accessible-porosity [%] Sq [μm] Density [g cm−3]

BPP
PP-GS 25 2.7 0.45 ± 0.2 1.8
PVDF-GS 25 2.4 0.73 ± 0.4 2.1
PVDF-GL 75 3.1 0.99 ± 0.3 2.1

least 15 min and during the measurement nitrogen gas was main-
tained in the gas compartment above the liquid phase in the three-
electrode setup. The electrochemical aging and analysis routine was
performed as follows.

(i) Pre-aging: CV was performed by cycling the pristine BPP in

0.1 M H2SO4 solution between +1 and −0.2 V vs Ag/AgCl at
a scan rate of 5 mV s−1. Two CV cycles were performed per
BPP. The second cycle was evaluated.
(ii) Electrochemical aging: the CV were performed by cycling the
BPP in 1.8 M VOSO4 + 2 M H2SO4 electrolyte solution
between +1.7 and +0.7 V vs Ag/AgCl at a scan rate of 200
mV s−1. A total of 3000 CVs were performed. After every
500th aging cycle post-aging (iii) analysis was performed.
(iii) Post-aging: after every 500th aging cycle the BPP was
removed from the vanadium electrolyte, rinsed and immersed
in millipore water after which the BPP was dried carefully
using compressed air and low coarse paper. The post-aging
analysis was done in 0.1 M H2SO4 solution with the same test
conditions as for the pre-aging analysis. For every trial a fresh
Figure 4. In-plane and through-plane electrical conductivity of pristine
0.1 M H2SO4 solution and thermally treated porous graphite composite BPPs.
felt for the counter electrode was used. Two CV cycles were
performed per BPP. The second cycle was evaluated to
determine the aging progression. collector. The cell was operated at a constant current density of 60
mA cm−2 which was applied through a potentiostat/galvanostat
The quantitative evaluation of the post-aging CVs was performed (VersaSTAT 4, Ametek, USA). The electrolyte 1.8 M VOSO4 in 2
on the basis of Cdl,, QQ/HQ and QV4+/V5+ of vanadium ion traces M H2SO4 was prepared by dissolving anhydrous VOSO4xH2O
(Fig. 2). For the electrochemical characterization measurement three (Sigma-Aldrich, USA) in diluted H2SO4 (Sigma-Aldrich, USA).
trials were done per BPP. Prior to each single cell test the oxidation state of the electrolyte had
For Cdl the anodic and cathodic current density was selected at to be balanced. This was done by filling the tank of the positive side
points where no redox activity was observed (dashed line in Fig. 2a). with 80 ml and of the negative side with 40 ml electrolyte. The
QQ/HQ was calculated in the range between +0.2 to +0.4 V vs Ag/ electrolytes were pumped through their respective half-cell at a flow
AgCl (Fig. 2a) and as reported12,13 QV4+/V5+ was calculated in the rate of 25 ml min−1 (KNF, Germany). The cell was charged to +1.7
range between +0.7 to +0.9 V vs Ag/AgCl (Fig. 2a). For the V. Following the first charging step 40 ml of electrolyte was
integration and evaluation of QQ/HQ, QV4+/V5+ the respective, Cdl removed from the positive side and then the cell was discharged
was subtracted. with the positive and negative tank each containing 40 ml of
The static surface wettability of BPPs was measured with contact electrolyte. After which the single cell test was initiated by
angle system OCA 15plus (Data physics, Germany). The device was performing 200 charge/discharge cycles between +1.7 to +1 V.
equipped with a Hamilton 500 μl syringe, a charge-couple device During the electrolyte balancing and single cell test the negative
camera and a SCA20 software (version 2.0.0, Data physics, electrolyte was continuously purged with nitrogen gas. One trial was
Germany). Using the sessile drop method 4 μl of deionized water done per BPP.
was introduced onto the surface of the BPP and the contact angle The single cell VRFB tests were evaluated as follows:
was measured. Six measurements were done for each BPP to Voltage efficiency,
t2
determine the average value.
∫ Vdischarge (t ) dt
The BPP’s surface roughness was measured using a confocal t1
microscope (Sensofar, PLu neox, Spain). The microscope was set to ηV = t2 [%]
100x magnification and on the SensoScan software (Sensofar, PLu ∫ Vcharge (t ) dt
t1
neox, Spain) a vertical step size of 32 μm was used to measure an Columbic efficiency,
area of 127.32 μm × 95.45 μm. The root mean square surface t2
roughness (Sq) was calculated and five measurements were done for Qdischarge
∫ Idischarge (t ) dt
t1
each BPP to estimate the average Sq. ηAh = Qcharge
= t2 [%]
A single flow cell (CNL Energy co, Republic of Korea) was ∫ Icharge (t ) dt
assembled for galvanostatic charge and discharge tests. Thermally t1

treated (in air at 400 °C, 18 h) XF30A carbon felts (4 mm, Toyobo, Energy efficiency,
Japan) with an active area of 5 cm × 5 cm were used as electrodes. ηWh = ηV × ηC [%]
The half-cell was separated by Nafion 117 (DuPont, USA) cation- Discharge capacity,
N
conducting membrane (pre-treated in 10% H2O2 solution, 80 °C, Dc = tdischarge × Idischarge [mAh]Where V: voltage [V], I: current
30 min). The polymer-graphite composite BPP(s) were placed [A], Q: charge [C], during charge and discharge cycles t: time [h]
between the carbon felt electrode and the gold coated nickel current and N is the discharge cycle number.
 Journal of The Electrochemical Society, 2022 169 080503

Figure 5. Comparing CV curves from BPPs in 0.1 M H2SO4 solution at a scan rate of 5 mV s−1. (a) Pre-aged BPPs and (b) PP-GS, (c) PVDF-GS and (d) PVDF-
GL BPPs pristine and after aging.

All the above-mentioned measurements were done at room on the influence of the PP or PVDF on the surface roughness cannot
temperature. be drawn in the current contribution. Nevertheless, it is worth
mentioning, that the PP-GS BPP showed rather a smooth and
Results probably homogenous surface.
Another important parameter of the composite materials is their
As explained in the experimental section three BPPs, named PP-
density. In general, the density of PVDF it is 1.65–1.97 g cm−314
GS, PVDF-GS and PVDF-GL, were studied with each BPP
while of graphite is 1.9–2.2 g cm−3 and of PP it is 0.9 g cm−3.15 The
composed of GL or GS. The mCT 2D images of the pure GL and
lower density in conjunction with the lower polymer content of PP in
GS powders (Fig. 3) revealed that the graphite particles vary in
comparison to the other two PVDF BPPs would explain the low-
shape, PS and size distribution. The corresponding average graphite
density value for PP-GS (Table II). Whereas, for the higher density
PS values are listed in Table II.
values of PVDF-GS and PVDF-GL the primary influencer is the
density of PVDF.
Discussion The varied PS of the graphite powders and the selection of
polymer types also influence the electrical conductivity behaviour of
The accessible porosity was quantified using mercury porosi- each BPP. Both PVDF based composite BPPs with 80% graphite
metry and confocal microscopy provided Sq of each BPP high- content each exhibit higher in— and through-plane electrical
lighting the influence of the constituents on the composite material conductivities compared to the PP-GS (Fig. 4), even though the
characteristics (Table II). It can be inferred that the larger graphite PP based BPP is composed of higher graphite weight percentage of
PS in PVDF-GL lead to larger or more accessible pores and higher 86%. It is assumed that the properties of the PVDF, such as higher
surface roughness probably due to protruding graphite particles on density, might provide better connected pathways between the
the surface. In comparison, the PVDF-GS and PP-GS BPPs are conductive graphite particles within the composite structure.
comprised of smaller graphite particles exhibiting smoother surfaces Hence, the PVDF-GS and PVDF-GL exhibit higher and comparable
and lower accessible porosity. Here it should be noted that not only electrical conductivity values. The higher standard deviation for
the graphite filler-type/PS, but also the manufacturing process and PVDF-GL and PVDF-GS is probably due to the higher Sq.
the used polymer influence the surface properties of the BPPs. To assess the influence of the surface properties and composite
However, since the PP based BPP constitutes different filler to composition of the BPPs on their electrochemical stability, the
polymer ratio to that of the PVDF based BPPs, a direct conclusion pristine BPPs were subjected to continuous potential sweeping in
 Journal of The Electrochemical Society, 2022 169 080503

Figure 6. (a) Double layer capacitance, (b) quinone/hydroquinone activity and (c) vanadium redox activity in 0.1 M H2SO4 solution of each composite BPP after
aging in vanadium electrolyte.

Figure 7. (a) Water contact angle and (b) surface roughness of pristine and electrochemically aged composite BPPs after 3000 CV cycles in vanadium
electrolyte.

high concentrated 1.8 M VOSO4 + 2 M H2SO4 electrolyte. The pristine PVDF-GL has the most discernible Cdl and QQ/HQ features.
aging progression of the BPPs was monitored via intermittent CV The Cdl is influenced by BPP surface-electrolyte interaction and can
measurements in 0.1 M H2SO4 solution. The pre-aged CV curves be related to surface hydrophilicity and roughness.12,16 It is prob-
(Fig. 5a) in 0.1 M H2SO4 solution of the studied BPPs show that the able, that the PVDF-GL exhibits larger graphite-electrolyte interface
 Journal of The Electrochemical Society, 2022 169 080503

PVDF-GL could be attributed to a higher amount of oxygen

functional groups as reported by Satola et al.12
Comparing the aging progression for each individual BPP
(Figs. 5b–5d), the expected aging behaviour was observed, i.e.
increasing Cdl, rising QQ/HQ activity and presence of vanadium
species redox activity QV4+/V5+. All features were quantitatively
evaluated and plotted (Fig. 6). In general, the increasing Cdl (Fig. 6a)
is induced primarily by increased hydrophilicity and high surface
area leading to higher BPP surface-electrolyte interaction area.12,16
Whereas, the QQ/HQ activity (Fig. 6b) is related to the oxidation of
the graphite components in the BPPs4,12 and is further increasing the
hydrophilicity. Moreover, even though prior to post-aging analysis
each BPP was cleaned with Millipore water, the accessible porosity
of the BPP could have caused entrapment of vanadium species
within pores causing vanadium redox activity QV4+/V5+.4,12,13
The PVDF-GL exhibits the highest values for Cdl, QQ/HQ and
QV4+/V5+ (Fig. 6d). The measured water contact angle and surface
roughness for the PVDF-GL (Fig. 7) indicate that it had the highest
hydrophilicity and surface roughness in pristine state which would
Figure 8. Comparison of charge/discharge cycles of VRFB single cell tests
with each BPP at current density of 60 mA cm−2. lead to significantly higher BPP-electrolyte contact with a progres-
sive increase of electrochemical aging. Whereas, the pristine PVDF-
in good agreement with the highest observed surface roughness. GS and PP-GS have lower hydrophilicity, surface roughness values,
Whereas, the pristine PP-GS with the lowest electrical conductivity and lower accessible porosity which led to reduced electrochemical
and lowest surface roughness values, shows the least current aging. Thus, the aging progression with highest to lowest values is
response in terms of Cdl. In addition, the QQ/HQ characteristics of PVDF-GL > PP-GS > PVDF-GS.

Figure 9. (a) Voltage, (b) coulombic, (c) energy efficiency and (d) discharge capacity of VRFB single cell test with each composite BPP at current density of 60
mA cm−2.
 Journal of The Electrochemical Society, 2022 169 080503

The change in surficial properties can be further explained by the As the cycling progresses the discharge capacity declines for all
introduction of oxygen functional groups during electrochemical VRFB tests, most likely because of vanadium ion crossover and side
aging which made all the BPPs more hydrophilic (Fig. 7a). Though, reactions causing a difference in the state-of-charge between the two
the change in surface roughness (Fig. 7b) was hard to distinguish, half-cell electrolytes17 and aging of the battery components (Fig. 9d)
most likely due to surface inhomogeneity of the BPPs causing high Moreover, the continuous increase of the cell resistance would
standard deviation. But the surface roughness changes could be decrease the rate capability and the capacity that can be discharged
owed to slight material abrasion or moderate gas evolution occurring at the given current density. The lowest electrochemical stability of
at the BPP surface during CV cycling in the vanadium electrolyte. PVDF-GL leads to higher drop of SoH, whereas the VRFB cells
Thus from above it can be inferred that the surfacial properties with BPPs containing smaller graphite PS, show less SoH decline.
siginificantly influence the electrochemcial behaviour/stability of
each BPP. Conclusions
To validate the findings of the electrochemical experiments and
The composition of each individual BPP influences its surficial,
to evaluate the electrochemical behaviour of the BPPs, single cell
electrical and electrochemical behaviour. The PVDF-GL with large
VRFB tests with each composite BPP were performed. Various
average graphite PS (75 μm) exhibits higher surface roughness and
factors, such as vanadium ion precipitation, resistance change of
higher accessible porosity compared to composites with smaller
membrane, felt electrode and electrolyte as well as change in
graphite PS i.e., PP-GS (25 μm) and PVDF-GS (25 μm). The
electrode polarization might actively contribute to a VRFB perfor-
electrical properties of the examined BPPs were influenced by the
mance decline, but they are difficult to define and distinguish from
polymer type. The electrochemical aging of each BPP was asserted
each other.17 Thus, this characterization is beyond the scope of this
through quantification of Cdl, QQ/HQ and QV4+/V5+. The PVDF-GL
publication. However, since the single cell test parameters were the
had the lowest electrochemical stability compared to PP-GS and
same for the measurements with only the type of BPP being
PVDF-GS owing to its higher hydrophilicity, surface roughness and
changed, it is presumed that any change in the evaluated VRFB
accessible porosity. Thus, these surficial properties have a significant
performance was related to the applied BPP. The 1st and 200th
influence on the electrochemical behaviour. The influence of the
charge/discharge curves from each VRFB test were plotted and
BPPs on the single cell VRFB performance was evaluated by ηV , C , E ,
compared (Fig. 8). In the initial cycle all three cells show the typical
charge/discharge behaviour of a VRFB. The VRFB cell containing and SoH. A similar trend was observed in electrochemical aging and
PP-GS exhibits the lowest capacity and reaches the cut off voltage analysis, i.e., PVDF-GL showed the worst performance followed by
earliest probably due to its lower electrical conductivity compared to PP-GS and PVDF-GS which is in good agreement with the single
the other both BPPs. Whereas the VRFB with PVDF-GL had better cell evaluation.
charge/discharge behaviour during the 1st cycle compared to PVDF- Hence, a good electrochemical stability of the BPPs is essential
GS and PP-GS probably because of its high electrical conductivity. in order to reduce side reactions and to obtain a long BPP and
However, due to its low electrochemical stability, its VRFB electrolyte cycle life for battery operation. The proposed potentio-
performance significantly deteriorated, which is in good agreement dynamic electrochemical characterization method can help to
with the electrochemical CV test results. determine the electrochemical stability of BPPs by identification
The efficiencies and discharge capacity from the single cell and quantification of the aging features. Also, it can be an effective
experiments were plotted in Fig. 9. In general, the voltage efficiency foresight to predict the applicability of a BPP in a VRFB.
is influenced by the activation overpotential, ohmic resistances17,18
and the electrolyte crossover.17,18 It is assumed that the voltage Acknowledgments
efficiency is mainly affected by the exchange of the three different The authors are thankful to the German Federal Ministry of
types of BPPs since all other components and their influence on the Economic Affairs and Energy for funding the project “Top Level
ohmic resistance remain unchanged within the single cell battery.7 Redox Flow” Grant Agreement Number ZF4090701ZG5.
As observed from the evaluated charge/discharge curves, the poor
PVDF-GL VRFB voltage efficiency (Fig. 9a) could be due to low ORCID
electrochemical stability caused by poor surficial properties, as
Barbara Satola https://orcid.org/0000-0002-5807-1763
explained under the electrochemical CV measurement discussion,
Corinna Harms https://orcid.org/0000-0001-5916-3224
leading to the high ohmic resistance. Initially, the VRFB with the
PP-GS had higher voltage efficiency compared to that with PVDF- References
GS, presumably due to its lower surface roughness which might
have reduced the contact resistance. But owed to the relatively 1. B. Caglar, “Conductive polymer composites and coated metals as alternative
bipolar plate materials for all-vanadium redox-flow batteries.” Advanced Materials
poorer electrochemical stability as charge-discharge cycles pro- Letters., 5, 299 (2014).
gressed its performance eventually declined. 2. P. C. Ghimire, A. Bhattarai, R. Schweiss, G. G. Scherer, N. Wai, and T. M. Lim,
The coulombic efficiency in a VRFB is generally influenced by “Investigation of reactant conversion in the vanadium redox flow battery using
mass transport losses and vanadium ion crossover through the spatially resolved state of charge mapping.” Batteries., 5, 2 (2019).
3. J. L. Barton and F. R. Brushett, “A one-dimensional stack model for redox flow
membrane.19,20 In addition, it also provides information about battery analysis and operation.” Batteries., 5, 25 (2019).
faradic losses caused by possible side reactions.17 As can be seen in 4. B. Satola, C. N. Kirchner, L. Komsiyska, and G. Wittstock, “Chemical stability of
Fig. 9b the coulombic efficiencies of the cells, containing PP-GS and graphite-polypropylene bipolar plates for the vanadium redox flow battery at resting
PVDF-GS are comparable. The decreased coulombic efficiency of state.” J. Electrochem. Soc., 163, A2318 (2016).
5. M. Jung, W. Lee, N. Nambi Krishnan, S. Kim, G. Gupta, L. Komsiyska, C. Harms,
the PVDF-GL VRFB can be attributed to side reactions such as Y. Kwon, and D. Henkensmeier, “Porous-Nafion/PBI composite membranes and
continuous oxidation of the graphite surface as indicated by the Nafion/PBI blend membranes for vanadium redox flow batteries.” Appl. Surf. Sci.,
rising QQ / HQ activity quantified in the electrochemical CV measure- 450, 301 (2018).
ments. 6. H. Liu, L. Yang, Q. Xu, and C. Yan, “Corrosion behavior of a bipolar plate of
carbon–polythene composite in a vanadium redox flow battery.” RSC Adv., 5, 5928
The energy efficiency is the product of the coulombic and voltage (2015).
efficiency. The performance of the VRFBs with the PVDF-GS and 7. J. W. Lim and D. G. Lee, “Carbon fiber/polyethylene bipolar plate-carbon felt
PP-GS have a similar efficiency but the latter gradually declines electrode assembly for vanadium redox flow batteries (VRFB).” Compos. Struct.,
slightly more than PVDF-GS after 120 cycles. In general, PVDF-GL 134, 483 (2015).
8. M. Park, Y.-J. Jung, J. Ryu, and J. Cho, “Material selection and optimization for
provides the lowest efficiency (Fig. 9c). The electrochemical aging highly stable composite bipolar plates in vanadium redox flow batteries.” J. Mater.
and analysis of the BPPs show the same trend. Chem. A, 2, 15808 (2014).
 Journal of The Electrochemical Society, 2022 169 080503

9. B. Caglar, J. Richards, P. Fischer, and J. Tuebke, “Conductive polymer composites 15. S. E. Selke and R. J. Hernandez, “Packaging: polymers in flexible packaging.”,
and coated metals as alternative bipolar plate materials for all-vanadium redox-flow ed. K. H. J. Buschow et al. Encyclopedia of Materials: Science and Technology.
batteries.” Adv Mat Lett., 5, 299 (2017). (Oxford)(Elsevier) p. 6652 (2001).
10. J. Zhang, T. Zhou, L. Xia, C. Yuan, W. Zhang, and A. Zhang, “Polypropylene 16. J. Chmiola, G. Yushin, R. K. Dash, E. N. Hoffman, J. E. Fischer, M. W. Barsoum,
elastomer composite for the all-vanadium redox flow battery: current collector and Y. Gogotsi, “Double-layer capacitance of carbide derived carbons in sulfuric
materials.” J. Mater. Chem. A, 3, 2387 (2015). acid.” Electrochem Solid State Lett., 8, A357 (2005).
11. N. J. Lee, S.-W. Lee, J.-H. Kim, M.-S. Park, G. Jeong, Y.-J. Kim, and D. Byun, 17. S.-K. Park, J. Shim, J. H. Yang, C.-S. Jin, B. S. Lee, Y.-S. Lee, K.-H. Shin, and
“Development of carbon composite bipolar plates for vanadium redox flow J.-D. Jeon, “The influence of compressed carbon felt electrodes on the performance
batteries.” Bull. Korean Chem. Soc., 33, 3589 (2012). of a vanadium redox flow battery.” Electrochim. Acta, 116, 447 (2014).
12. B. Satola, L. Komsiyska, and G. Wittstock, “Bulk aging of graphite-polypropylene 18. L. Yang, Y. Zhou, S. Wang, Y. Lin, T. Huang, and A. Yu, “A novel bipolar plate
current collectors induced by electrochemical cycling in the positive electrolyte of design vanadium redox flow battery application.” Int. J. Electrochem. Sci., 7031
vanadium redox flow batteries.” J. Electrochem. Soc., 164, A2566 (2017). (2017).
13. E.-M. Hammer, B. Berger, and L. Komsiyska, “Improvement of the performance of 19. S. Kim, “Vanadium redox flow batteries: electrochemical engineering.” Energy
graphite felt electrodes for vanadium-redox-flow-batteries by plasma treatment.” Int Storage Devices, ed. M. T. Demirkan and A. Attia (IntechOpen, London, United
Journal of Renewable Energy Development., 3, 6 (2014). Kingdom) p. 20 (2019).
14. S. Ebnesajjad, “2 - Fluoropolymers: properties and structure.”, ed. S. Ebnesajjad 20. Y. Ashraf Gandomi, D. S. Aaron, and M. M. Mench, “Coupled membrane transport
Melt Processible Fluoroplastics. (Norwich, NY)(William Andrew Publishing) p. 9 parameters for ionic species in all-vanadium redox flow batteries.” Electrochim.
(2003). Acta, 218, 174 (2016).