An oxo-verdazyl radical for a symmetrical non-aqueous redox flow battery

Article abstract of DOI:10.1039/d0ta07891c

Verdazyl free radical compounds are promising candidates for symmetrical all-organic redox flow batteries (RFBs) due to their redox stability, the ease with which their chemical structure can be varied, and their unique bipolar nature. The present work re

Full text of DOI:10.1039/d0ta07891c

View Article Online
View Journal
Journal of
Materials Chemistry A
Materials for energy and sustainability
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: A. Korshunov, M.
J. Milner, M. Grünebaum, A. Studer, M. Winter and I. Cekic-Laskovic, J. Mater. Chem. A, 2020, DOI:
10.1039/D0TA07891C.
Volume
Number 12
28 March 2018
Pages 4883-5230
6
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Journal of
Materials Chemistry A
Materials for energy and sustainability
rsc.li/materials-a
Accepted Manuscripts are published online shortly after acceptance,
before technical editing, formatting and proof reading. Using this free
service, authors can make their results available to the community, in
citable form, before we publish the edited article. We will replace this
Accepted Manuscript with the edited and formatted Advance Article as
soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes to the
text and/or graphics, which may alter content. The journal’s standard
Terms & Conditions and the Ethical guidelines still apply. In no event
shall the Royal Society of Chemistry be held responsible for any errors
or omissions in this Accepted Manuscript or any consequences arising
from the use of any information it contains.
ISSN 2050-7488
COMMUNICATION
Zhenhai Wen et al.
An electrochemically neutralized energy-assisted low-cost
acid-alkaline electrolyzer for energy-saving electrolysis
hydrogen generation
rsc.li/materials-a

Please do not adjust margins
Journal of Materials Chemistry A
Page 1 of 13
View Article Online
DOI: 10.1039/D0TA07891C
ARTICLE
An oxo-verdazyl radical for a symmetrical non-aqueous redox
flow battery
Aleksandr Korshunov,a Matthew James Milner,b Mariano Grünebaum,c Armido Studer,b Martin
Received 00th January 20xx,
Accepted 00th January 20xx
Winter*a,c and Isidora Cekic-Laskovic*c
DOI: 10.1039/x0xx00000x
Verdazyl free radical compounds are promising candidates for symmetrical all-organic redox flow batteries (RFBs) due to
their redox stability, the ease with which their chemical structure can be varied, and their unique bipolar nature. The present
work reports a preliminary screening of a selection of oxo-verdazyl compounds for key redox electrolyte parameters. Of the
considered candidates, the 1,5-diphenyl-3-isopropyl-6-oxo-verdazyl radical performed best and is investigated in extensive
RFB experiments to compare its electrochemical behavior in cyclic voltammetry (CV) to that within an actual battery. The
symmetrical oxo-verdazyl non-aqueous electrolyte RFB provides a mean voltage of 1.42 V and demonstrates good stability
as well as high Coulombic (>97%) and energy efficiencies over more than 100 charge/discharge cycles. The redox electrolyte
is characterized at different stages within a single cycle (‘state of charge’ experiments) independently for each half-cell. To
address the specifics of the electrolyte transition to RFB cell setup an ‘in-cell’ CV flow-enabled electrochemical study are
conducted, introduced here as a new step towards standardization of the electrochemical description of RFB electrolytes.
The electrochemical performance results highlight oxo-verdazyls as versatile materials for energy applications and indicate
great promise for their further development and optimization in the field of RFBs.
to, for instance, the excessively high cost of essential cell
elements (e.g. separators, pumps) or commercial electrolytes
Introduction
(in particular, those based on vanadium) relative to the
competitive cost of the energy carrier units ($100 per kWh).6 In
order to bypass the practical issues of inorganic electrolytes and
follow modern ecological trends, the scientific community is
directing considerable attention towards organic redox
electrolyte formulations. Organic molecules represent an
almost unlimited framework in which theory meets target
application for material science.7-9 The numerous properties
crucial for consistent RFB energy storage, such as exceptional
bulk electrolyte stability, high solubility for each oxidative state,
and prevention of irreversible capacity loss, can be reverse
engineered into organic redox materials with the desired
tailored properties. It is hardly surprising, then, that redox-
active organic molecules (ROMs) are frequently purported to be
The contemporary speed of technological advance gives rise to
an increasing demand for energy supply worldwide. In contrast,
the natural stock of available fossil fuels is declining steadily,
making the path towards alternative sources of less convenient
but practically limitless energy more attractive. However,
harvested energy output from most of the promising renewable
sources (e.g. solar or wind) still struggles to satisfy the key
requirement of consistent supply.1-3 Redox flow batteries (RBFs)
offer an effective static grid-scale measure to manage and ease
the energy delivery procedure. Their high adaptivity to frequent
and unregulated energy inputs and outputs over unbound
periods of time, as well as their safety and scalability, make RFBs
a promising technology for buffering (or so-called ‘peak
shaving’) and, thereafter, effectively conveying energy
generated from inherently intermittent sources to
consumers.4,5
a
promising and sustainable scaffold for modern RFB
electrolytes.10-12
While aqueous redox flow batteries are still dominant among
those reported to date – with the vanadium redox flow battery
standing out as the ‘uncontested leader’13,14 – there has
recently emerged a rising interest in non-aqueous systems.15-18
To an extent, the trend is that the development of suitable
commercially available separators, and specific membrane
research more generally, will be the determining factor of the
rate at which non-aqueous redox flow batteries (NARFBs) will
evolve.19-22 One of the greatest challenges for contemporary
Despite their potential advantages, widespread application of
RFBs to actual grid-scale solutions encounters limitations due
a. MEET Battery Research Center, University of Münster, 48149 Münster, Germany.
b. Organisch-Chemisches Institut, Westfälische Wilhelms-Universität, Corrensstraße
40, 48149 Münster, Germany.
c. Helmholtz Institute Münster, IEK-12, Forschungszentrum Jülich GmbH, 48149
Münster, Germany
Electronic Supplementary Information (ESI) available: [details of any supplementary
information available should be included here]. See DOI: 10.1039/x0xx00000x
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 2 of 13
ARTICLE
Journal Name
non-aqueous RFBs lies in combining selective separation of two have no tendency to dimerize.36 The tunability of relevant
View Article Online
DOI: 10.1039/D0TA07891C
half-cell compartments to prevent irreversible co-mixing of physicochemical properties and chemical stability even greater
electrolyte constituents with sufficient resilience against than that of the well-known nitroxide radicals37,38 has fed a high
swelling and external mechanical action. Even the state-of-the- demand for these compounds in the fields of molecular
art separators for NARFBs offer only a trade-off between ionic magnetochemistry39 and radical probing40. Verdazyls have long
and molecular selectivity and the high conductivity necessary to been foreseen to be promising materials for battery
complete the cell electric circuit and operate at higher current applications owing to their bipolar redox activity.41,42 Even
densities. To date, ion-selective membranes and porous within those synthetic frameworks already developed,
separators from companies like Fumatech GmbH,9,23,24 verdazyls can be engineered to bear substituents beneficially
Daramic®25,26 or Astom Corp.27,28 have been successfully affecting electrochemical potentials and reversibility in redox
employed for non-aqueous laboratory-scale static and flow processes, targeting a higher energy capability of the final
battery systems.
electrolyte product.
R4 R5
Despite the posed technical challenges, a transition from
aqueous to non-aqueous redox electrolytes offers expanded
scope to select electrolytes offering higher capacity and higher
voltages. Simultaneously, the wide electrochemical stability
window of organic solvents, their compatibility with modern
RFB cell elements, their wide working temperature range, and
beneficial hydrodynamic properties explain the modern trend
of utilizing organic solvents in flow battery applications.7 Among
the common aprotic electrolyte solvents, acetonitrile plays an
important role since it provides an exceptionally wide
electrochemical stability window (6.1 V, from −2.6 V to 3.5 V vs.
the standard hydrogen electrode (SHE)), low dynamic viscosity
(d = 0.341 mPa·s) and density ( = 786 kg·m-3) as well as high
conductance (σ) of the resulting supporting salt solutions (for
O
R1
R2 R1
R2
N
N
N
N
N
N
N
N
R3
R3
Figure 1. General structure of verdazyl free radicals: A) Kuhn verdazyl; B) oxo-
verdazyl.
Despite their promise, it was only recently that Dyker and co-
workers demonstrated the potency of verdazyl radicals as a
redox electrolyte for RFBs using a static 2032 coin cell battery
setup.43 Their battery made use of 3-phenyl-1,5-p-tolylverdazyl,
a Kuhn-type verdazyl, to generate a cell voltage of 0.97 V. A
small study by Hicks and co-workers characterized the redox
potentials of a selection of verdazyls and found that the oxo-
verdazyls possessed significantly higher cell voltages than the
Kuhn-type verdazyls, and essentially complete electrochemical
reversibility.38 However, to the best of our knowledge there has
been no report of an oxo-verdazyl radical being used as a redox
electrolyte, nor has any verdazyl ever been studied under actual
flow conditions. In the present work we report a systematic
study on oxo-verdazyl-based electrolytes as a core electrolyte
element of a symmetrical non-aqueous RFB. The main
discussion will revolve around 1,5-diphenyl-3-isopropyl-6-oxo-
verdazyl (isoV), identified as the most promising redox material
for further investigation by means of a preselection process.
instance, σBu4NClO4, 1M = 27.0 mS·cm-1 and σEt4NBF4,
= 55.5
1M
mS·cm-1).29 Overall, the use of acetonitrile has a positive impact
on energy density, allowing an NARFB electrolyte to eventually
deliver a higher energy density than aqueous analogues.
Besides the choice of a suitable membrane separator, a further
strategy to preclude irreversible capacity fading is to utilize the
same redox material for both negolyte and posolyte (where a
negolyte is a redox electrolyte characterized by a lower half-cell
reaction potential in comparison to the posolyte). Such a
requirement introduces a considerable design limitation as any
selected substance must possess three or more oxidative states
at sufficiently separated redox potentials. Naturally, other
requirements for ROMs – such as high solubility in a given
solvent, or electrochemical (or at least chemical) reversibility
for each of the half-cell reactions – still apply. To our knowledge,
only a limited number of pure ROM cases (i.e. not including
metal-coordinated complexes) have been reported as being
adopted as an electrolyte for non-aqueous RFBs. Among them
are diaminoanthraquinone (DAAQ),30 croconate violet,31
Although the nature of a symmetrical electrolyte introduces
undeniable simplification and unification to the setup, there is
no guarantee of a similar electrochemical response from each
of the two oxidative state transitions. As such, a more detailed
description of electrode processes must be continuously
assessed and updated for each of the half-cell compartments in
order to determine where the redox electrolyte tends to fail
most severely within the RFB lifespan. Incorporating a reference
electrode into the common redox flow battery cell setup
introduces a powerful tool for decoupling and monitoring of the
corresponding states of electrolytes.4,44 Modern technologies
incessantly offer state-of-the-art solutions to make common
and well-described reference electrode chemistries more
versatile and thus to introduce a new tool of comparison,
necessary for advanced battery systems.45,46 For a complete
electrochemical picture, one should therefore take advantage
of the possibility to integrate electrochemical techniques
common in RFB research (e.g. cyclic voltammetry (CV) and
nitronyl
nitroxides
(e.g.
2-phenyl-4,4,5,5-
tetramethylimidazoline-1-oxyl-3-oxide
(PTIO)),32,33
N-(α-
ferrocenyl)ethylphthalimide,34 and porphyrin.35 For the sake of
diversification and better understanding of how symmetrical
ROM-based electrolytes fundamentally function, novel
examples are indispensable.
Verdazyl radicals (Figure 1) have grown as a group in recent
decades to become one of the largest families of stable free
radicals. They are, with only a few exceptions, stable indefinitely
both as solids and in solution, are air- and water-stable, and
2 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 3 of 13
Journal of Materials Chemistry A
Please do not adjust margins
Journal Name
ARTICLE
electrochemical impedance spectroscopy (EIS)) directly into the alumina particles. Considered redox electrolyte solutions
View Article Online
DOI: 10.1039/D0TA07891C
applied battery setup. Gathering these data within the frame of consisted of 2 mM of each given oxo-verdazyl radical and 0.1 M
battery cell conditions opens a way to include and evaluate a TBAP as the supporting salt in 2 mL of anhydrous acetonitrile.
multivariate influence of the setup on redox electrochemical The chosen potential scan rate range for CV measurements was
properties; for example, through the type and geometry of set from −1 V to 1 V with a starting point of 0 V vs. Ag|AgCl(sat.).
current collector,47-49 the given kind of ion exchange membrane The set of P-EIS measurements was collected by applying an
or porous separator,50,51 and the velocity of the flow.52 Such an alternating voltage of magnitude 3 mV at a frequency in the
approach aids in deciphering the true nature of existing range 0.1 – 100000 Hz, with a number of starting potentials in
overvoltages for operational RFBs.44,53 Although the the vicinity of the corresponding half-cell reaction potential (vs.
physicochemical impact of RFB cell components on Ag|AgCl (sat.)). Before each individual CV or P-EIS experiment,
electrochemical performance is not doubted, it is rarely taken the filled electrochemical cell was purged with nitrogen gas for
into account and described. Here we address the issue by two minutes to remove oxygen. Electrochemical data were
extending the standard redox electrolyte characterization collected using a connected stack of an SI 1287 Electrochemical
routine, thereby demonstrating how the transition to a battery Interface (Solartron) and an SI 1260 Impedance/Gain-phase
cell affects the electrochemical properties of a given studied Analyzer (Solartron) together with dedicated software
material.
(Scribner, USA).
Experimental
Flow cell experiments
Galvanostatic and potentiodynamic cycling experiments in
static and flow conditions were carried out with a custom-made
zero-gap flow cell. Conductive end plates consisted of two
mirrored graphite blocks (Alfa Aesar) with CNC-processed flow
frames of predefined exposed area 0.9 cm2, and were liquid-
tight inserted to polypropylene flow-through supporting blocks
with inlet and outlet for electrolyte access as well as additional
inlets for the connection of the reference electrode. The sealing
set for the flow cell included handmade silicon (MaTecK,
Germany) gaskets with 0.9 cm2 exposed area and commercial
Nylon O-rings (Arcus, Germany) of appropriate size. Each
experiment involving RFBs employed 6 mm thick carbon felt
electrodes (Alfa Aesar, uncompressed), a leakless miniature
Ag|AgCl(sat.) reference electrode (ET072-1, eDAQ, Australia)
and a fresh piece of anion exchange membrane (AEM)
Chemicals and Reagents
Tetrabutylammonium perchlorate (TBAP, >98%) and Al2O3
nanopowder (>98%, <50 nm TEM size) were purchased from
Sigma Aldrich, anhydrous acetonitrile (MeCN, ROTIDRY®Sept,
>99.9%, < 10 ppm H2O) from Carl Roth, and ferrocene (>99.5%,
high purity) from Alfa Aesar. Dimethylformamide (DMF, 99.8%,
Extra Dry) was purchased from Acros Organics. Reagents and all
other chemicals required for synthesis were purchased from
ABCR, Acros Organics, Alfa Aesar, Sigma Aldrich, Fisher
Scientific, Fluorochem, or TCI, and were used as received.
Synthesis of oxo-verdazyl radicals
The synthesis of all screened radicals was reported previously.54
Each is easily prepared in three steps, beginning with a Goldberg
reaction between carbohydrazide and an aryl iodide, followed
by condensation with an aldehyde to form the tetrazinanone
precursors (Figure S1). Oxidation to the radicals is
straightforward with e.g. benzoquinone. Further details, a
general procedure, and full synthetic details and
characterization data for isoV can be found in the Supporting
Information.
Fumasep®FAP-375-PP (FUMATECH, Germany) as
a cell
separator. Carbon felt electrodes were preliminarily treated
several times with acetonitrile in order to wash away any
organic contaminants present, and the membrane was pre-
dried for 20 minutes under reduced pressure at room
temperature. All measurements were carried out within a
nitrogen-filled SEKUROKA® glove-bag (Carl Roth) under a
constant stream of gas. Redox electrolyte preparation and final
RFB cell assembly proceeded within the inflated glove-bag after
prior continuous N2 gas purging for 30 minutes. The pumping
line system consisted of a peristaltic dual TVP15 pump
(Medorex, Germany) providing a constant flow rate (2 mL·min-
1) and stretchable PharMed®BPT tubing (Saint-Gobain
Performance Plastics, France) jointed with PTFE tubing
(Bohlender, Germany) to ensure secure and prolonged pumping
Preliminary electrochemical studies
Cyclic voltammetry (CV) and potentiostatic electrochemical
impedance spectroscopy (P-EIS) experiments were carried out
within a custom three-electrode electrochemical cell, with a
volume of 2 mL, using platinum disc electrodes (eDAQ, 3 mm
diameter) as the working and counter electrodes, and a leakless
miniature Ag|AgCl(sat.) reference electrode (ET072-1, eDAQ,
Australia). Calibration of the leakless Ag|AgCl (sat.) reference
electrode was carried out within the described three-electrode
setup with the help of 10 mM ferrocene in a 0.1 M acetonitrile
solution of TBAP as a supporting salt (Figure S2 in SI). Working
and counter electrodes were polished prior to each
measurement session with fine sandpaper (1200 grit) to
remove coarse contaminants from the electrode surface, with
fresh alumina paste on a glass substrate to reduce the
roughness of the surface, and were then sequentially sonicated
in distilled water and acetonitrile to remove residuals of
through the flow-cell.
A 2 mL solution of 0.3 M
tetrabutylammonium perchlorate and 10 mM of the redox
material 1,5-diphenyl-3-isopropyl-6-oxo-verdazyl in anhydrous
acetonitrile was used as a standard electrolyte formulation for
all RFB cell experiments. Galvanostatic charge/discharge cycling
of the RFB cell proceeded between −0.20 V and 2.00 V threshold
potentials at current density of 2.22 mA·cm-2.
For ‘in-cell’ cyclic voltammetry investigations in static mode (no
pump-induced flow), a potential scan rate range of 2 mV·s-1
–
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 3
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 4 of 13
ARTICLE
Journal Name
30 mV·s-1 was employed. On a working pumping line a
potentiostatic pre-electrolysis step for 200 s preceded every
measurement, at −0.20 V and 0 V vs. Ag|AgCl(sat.) for negolyte
and posolyte half-cells, respectively, to minimize the initial
current response from a potential step. For flow mode (pump-
induced flow) a potential scan rate range of 0.2 mV·s-1 – 0.8
mV·s-1 was employed without preliminary treatment.
An exploratory preselection process was first carried out on a
DOI: 10.1039/D0TA07891C
set of previously synthesized oxo-verdazyl radicals, to obtain
View Article Online
initial insight into their electrochemical properties. The
structure of the radicals varied by the identity of a single
substituent at the 3-position, including both aliphatic and
aromatic groups of varying polarity (Figure 2a). Each considered
redox species, within supporting media closely resembling a
desirable RBF electrolyte formulation, was subjected to a
general screening protocol: solubility determination, three-
electrode cyclic voltammetry (CV) measurements and, in the
case of reasonably satisfying results, potentiostatic
electrochemical impedance spectroscopy (P-EIS) experiments.
Further synthetic details and full electrochemical data are given
in the Supporting Information (Figure S4, S5).
State of charge studies
Open Circuit Potential (OCP) measurements
After each galvanostatic step of ±1.1 mA·cm-2, corresponding to
a change of 10% of the theoretical state of charge, continuous
OCP measurement was applied until the OCP no longer deviated
for more than 5 mV over 30 s.
UV-Vis measurements
Among the screened verdazyl radicals, the most optimal
performance was observed for 1,5-diphenyl-3-isopropyl-6-oxo-
verdazyl (isoV). 1,5-diphenyl-3-methyl-6-oxo-verdazyl (4)
behaved similarly and even marginally outperformed the
former, due primarily to its inherently lower molecular mass.
However, application of isoV was expected to be more
reasonable as the isopropyl group introduces additional steric
bulk in comparison to the methyl analogue and thus higher
resistance to envisaged undesirable crossover action through
an ion-exchange membrane9,20 and due to the high solubility in
acetonitrile (up to 2.4 M) as shown in Figure S6. Furthermore,
After galvanostatic steps of 1.1 mA·cm-2, corresponding to 20%,
40%, 60%, 80% and 100% changes in the state of charge (SOC)
and, in order to ascertain complete mixing inside the half-cell,
continuous measurement of the OCP until it no longer deviated
for more than 5 mV over 30 s, 0.1 mL aliquots of the negolyte
and posolyte were taken and diluted up to 3 mL with anhydrous
acetonitrile. Spectra of the resulting samples were recorded in
the 350–800 nm wavelength region with 1 nm resolution, using
a Shimadzu UV-2450 spectrometer and a quartz High Precision
Cell with a pathlength of 10 mm (Hellma Analytics).
Linear Sweep Voltammetry (LCV) measurements
Initially each analyzed RFB cell was completely pre-charged.
After galvanostatic steps of ±1.1 mA·cm-2 and a rest step to
allow adequate mixing of the resulting half-cell electrolyte, the
pumps were switched off and static LSV data were recorded at
a 50 mV·s-1 potential sweep rate and, starting from the value of
OCP, down to −1.75 V vs. Ag|AgCl (sat.) for negolytes and up to
1.40 V vs. Ag|AgCl (sat.) for posolytes. Since an LSV scan
inherently changes the ratio of oxidized and reduced forms of a
given redox couple due to the Faradaic process of interest, a
capacity balance was updated after every scan.
a
O
O
O
N
N
N
N
N
N
N
N
N
N
N
N
F
NO2
N
1
2
3
Galvanostatic Electrochemical Impedance Spectroscopy (G-EIS)
measurements
O
O
O
N
N
N
N
N
N
N
N
N
N
N
N
Initially each analyzed RFB cell was completely pre-charged. For
this measurement no reference electrode was applied. After
galvanostatic steps of 1.1 mA·cm-2, and a rest step required for
the stabilization of open circuit voltage (OCV) upon continuous
convective mixing of redox electrolyte, galvanostatic
electrochemical impedance spectra were recorded.
Measurement was then conducted by applying an alternating
current of magnitude 0.2 mA in the absence of direct current,
with frequency range 5 – 100000 Hz and starting voltage of 0 V
vs. OCV. During each GEI scan, peristaltic pumps provided a
continuous flow of electrolyte through the cell.
4
5
isoV
b
O
O
O
N
N
N
N
N
N
N
N
N
N
- e
- e
N
N
isoV-
isoV
isoV+
Figure 2. a) Oxo-verdazyl radicals for preselection: 1,5-diphenyl-3-4-nitrophenyl-
6-oxo-verdazyl (1), 1,5-diphenyl-3-4-N,N´-dimethylphenyl-6-oxo-verdazyl (2), 1,5-
diphenyl-3-4-fluorophenyl-6-oxo-verdazyl (3), 1,5-diphenyl-3-methyl-6-oxo-
verdazyl (4), 1,5-diphenyl-3-ethyl-6-oxo-verdazyl (5), 1,5-diphenyl-3-isopropyl-6-
oxo-verdazyl (isoV); b) Oxidative states of 1,5-diphenyl-3-isopropyl-6-oxo-
verdazyl.
Results and discussion
Electrochemical Studies
4 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 5 of 13
Journal of Materials Chemistry A
Please do not adjust margins
Journal Name
ARTICLE
some verdazyls with methyl groups at the 3-position are known of the appropriate fitting of the result for the intermediate
View Article Online
to exhibit reduced long-term stability.36
frequency region, where mass-transfer impedance and
DOI: 10.1039/D0TA07891C
potential-dependent charge-transfer impedance have
a
comparable impact on total impedance, to a mathematical
model derived specifically for Faradaic processes governed by
semi-infinite planar diffusion, shown in Equation 2:
The
1,5-diphenyl-3-isopropyl-6-oxo-verdazyl
radical
participates in two different electrochemical redox processes
(Figure 2b) at the surface of the platinum electrode,
accompanied by either electron abstraction (oxidation to isoV+)
at 0.70 V vs. Ag|AgCl (sat.), or electron capture at −0.72 V vs.
Ag|AgCl (sat.) (reduction to isoV−) (Figure 3a). The half-cell
reaction potentials of the observed processes are separated by
a difference of 1.42 V, providing a theoretical cell voltage for a
symmetrical NARFB based on this radical. Observed anodic and
cathodic peaks near the mentioned potentials have similar
shapes of equal heights (the ratios of the corrected current
peaks are close to unity) and a separation of ~70 mV for both
oxidative and reductive processes, with only slight variation
upon increasing the potential sweep rate, thus indicating
diffusion-limited electrochemically reversible redox reactions.
The aforesaid, together with a reasonable assumption of
established semi-infinite type diffusion to a planar electrode
surface, permits Randles-Sevcik analysis for each of the
considered redox couples isoV+/isoV and isoV/isoV− (Figure
3b), through the related analytical expression, shown in
Equation 1:55
1/2
훼
퐷푂1 ― 훼퐷푅
휔1/2
푘0
푐표푡 휑
= 1 +
(2)
퐸1/2
(
)
2
where Do is the diffusion coefficient of oxidized form
(in cm2·s-1), Dr is the diffusion coefficient of reduced form (in
cm2·s-1), α is the current exchange coefficient, ω is the
frequency (in Hz), and k0 is the standard rate constant
(in cm·s-1).
Additional precautions must be first considered to diminish any
influence from the inherent impedance of the reference
electrode, even though it barely affects the frequency region of
interest. To this end, a sequence of spectra was recorded for
each redox reaction near the corresponding half-wave
potentials, converted to an appropriate format and, finally, a
dataset with the highest cotangent values related to the
frequency band of interested was selected (Figure 3c,d). Further
fitting to the mathematical model (Equation 2) provides an
estimate for the standard electrochemical rate constant of k0 =
0.0081 cm·s-1 for the redox couple isoV+/isoV, which is on a
reasonable level typical for an NARFB. Unfortunately, the total
electrochemical impedance is extremely sensitive to
background processes such as potential drifts or electrode
contamination,56,57 sometimes making an appropriate fitting
impossible. This was observed for the redox couple isoV/isoV−,
for instance.
0.5
푛퐹휐퐷
푖 = 0.4463푛퐹퐴퐶
p
(1)
(
)
푅푇
where n is the number of transferred electrons, A is the
electroactive surface area (in cm2), D is the diffusion coefficient
(in cm2·s-1), С is the concentration of the redox compound (in
mol·cm−3), υ is the potential scan rate (in V·s-1), T is the
temperature (in K), and F is the Faraday constant (96485 C·mol-
1).
As an actual RFB setup has many structural differences
compared to a three-electrode setup, additional efforts must be
made to characterize the transition between them in order to
have a better overview of processes occurring inside the RFB
during operation. The geometry of the mechanically ‘zero-gap’
attached flow-through ‘carbon felt-on-graphite plate’ electrode
combination differs significantly from that of a disc electrode.
Thereby, for potentiodynamic studies on porous electrodes in
Derived diffusion coefficients for isoV+/isoV and isoV/isoV−
amount
1.8 · 10−7 cm2·s-1, respectively, at 298 K. These values lie within
the typical range for redox materials used in symmetrical
NARFBs.32-35
to
2.8
·
10−7
cm2·s-1
and
The P-EIS method provides a rapid tool to treat the kinetics of
heterogeneous electrochemical processes quantitively and,
more importantly, conveniently, without changing the three-
electrode CV setup. The approach described (see Experimental)
allows for decoupling of charge-transfer resistance from mass
transfer resistance and for the evaluation of a standard
static
mode
(flow
rate
=
0 mL·min-1), two diffusion regimes prevail for the most part:
external linear semi-infinite and, predominantly, external finite
cylindrical. To date there is only a limited number of works
where such problems are treated quantitively as there is no
simple analytical description as for redox processes governed by
semi-infinite linear diffusion.58-60 In light of this, effects of the
transition from a three-electrode setup to an RFB cell on a 10
mM oxo-verdazyl-based redox electrolyte were studied.
electrochemical rate constant given
a known diffusion
coefficient, similar to ‘De Levie-Husovsky’ analysis.56 To apply
this approach, it is necessary to convert the resulting P-EIS
spectra to a tantamount representation of a Bode phase
diagram, where the cotangent of the registered phase angle and
correspondent square root of frequency are used in place of
phase angle and frequency, respectively. The final goal consists
For the posolyte half-cell with the isoV+/isoV redox couple
(Figure 3f), apparent separation between cathodic and anodic
peaks increased up to ~200 mV, accompanied by noticeable
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 5
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 6 of 13
ARTICLE
Journal Name
broadening and an additional resolved anodic peak at 1.10 V vs. scan rates. Obviously, a redox retardation negatively influences
View Article Online
DOI: 10.1039/D0TA07891C
Ag|AgCl (sat.). Observed signal broadening corresponds to the the final Coulombic efficiency of potentiodynamic bulk
expected ‘in pore’ external finite cylindrical diffusion response electrolysis. Nevertheless, related coulometric analysis of the
for the carbon felt electrode since it can be treated as a cathodic process at −0.60 to −1.40 V vs. Ag|AgCl(sat.) suggests
microelectrode network with randomly distributed distances a relatively low dependency on the potential sweep rate as
between fibers. The seemingly irreversible oxidation process at similar values of total charge for both lowest or highest scan
1.10 V vs. Ag|AgCl (sat.) relates to an unresolved cathodic peak rate are registered, whilst the potential shift for corresponding
at ~0.45
V vs. Ag|AgCl (sat.), implying a very slow current peak does not exceed 200 mV.
electrochemical response with regard to linear diffusion to the In the case of the posolyte, the resulting picture (Figure 3g) is
surface of the conductive graphite end plate or the external generally more complex. Sequential potentiodynamic cycling at
surface of the carbon felt block.
various sweep rates induced additional peak splitting for the
For the negolyte half-cell with the isoV/isoV− redox couple single-electron oxidation process. Interestingly, the same
(Figure 3e), only a slight amount of lateral splitting at −0.60 V electrochemical process seems to proceed at different
vs. Ag |AgCl (sat.) is present for the anodic process, although electrode domains and has a predominant dependence on
the cathodic peak potential strongly depends on the potential diffusional mode, producing anodic current peak splitting at
scan rate whereas the connected potential of the anodic peak 0.60 – 0.80 V and 0.80 – 1.00 V vs. Ag|AgCl (sat.), where each
rests mostly unaffected at −0.72 V vs. Ag|AgCl (sat.). Marked peak group exhibits a different dependence on the potential
interaction with an associated induced broad peak separation sweep rate. Regarding the cathodic process (‘discharge’),
makes the negolyte part of the symmetrical electrolyte a corresponding groups of current peaks appear almost
plausible overvoltage source, which may preclude efficient RFB completely overlapped, giving a sharp response at 0.30 – 0.40 V
charging at high current densities.
vs. Ag|AgCl, which from coulometric analysis does not scale up
As a further step, the symmetrical oxo-verdazyl-based redox with increasing potential sweep rate. This, together with a
electrolyte was subjected to potentiodynamic bulk electrolysis. subsequent fairly constant but sluggish mass transfer current
By adding convection to an analyzed system, the observed response, suggests that oxidized isoV+ has a high propensity for
electrolyte redox behavior closely approaches that of the actual reduction back to the initial oxidative state, although the total
battery of interest. Characteristics like ‘in-flow’ overvoltage as a overvoltage in comparison to the negolyte is higher. Another
function of potential sweep rate or efficiency of mass important aspect to highlight is the average maintained current
convection under mass transfer limits pertain to the final throughout the potentiodynamic cycle: it is systematically lower
performance of any RFB system. It is no secret that the flow of for the posolyte compared to the negolyte for each applied scan
redox electrolyte through a porous hydrodynamic electrode rate. This can be considered responsible for premature
significantly reshapes the dimensional distribution of the overvoltage development during the late stage of charge and
concentration of redox-active species, supplying more efficient discharge processes.
flux of redox material near the electrode surface beyond Fickian
limits and, as
a consequence, enabling higher current
densities.4, 61 In comparison to the static case, porous electrodes
subjected to a flow of redox material paint a decisively different
picture within the scope of otherwise technically identical
potentiodynamic studies.
For the negolyte, a similar profile was observed for both
cathodic (‘charge’) and anodic (‘discharge’) processes with peak
splitting near −1.2 and −0.35 V vs. Ag|AgCl (sat.), respectively,
due to delayed ‘convectional’ mixing within the electrolyte tank
and the same electrochemical reaction occurring on the surface
of graphite end plate (Figure 3h). The observed symmetry of the
current maximum peaks for the acquired set of cyclic
voltammograms suggests that the observed anodic
electrochemical process possesses a lower current exchange
coefficient in terms of Butler-Volmer theory.58 This
phenomenon strongly affects the discharge time required for
reaching the maximum current value and, furthermore, makes
the concentration gradient less pronounced under the constant
electrolyte flow. As a result, a so-called diffusional tail of
secondary ‘convectional’ current peak is observed at higher
6 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 7 of 13
Journal of Materials Chemistry A
Please do not adjust margins
Journal Name
ARTICLE
a
View Article Online
DOI: 10.1039/D0TA07891C
b
isoV / isoV+
isoV- / isoV
0.00015
0.00010
0.00005
0.00000
-0.00005
-0.00010
1.42 V
100 mV/s
50 mV/s
0.00010
0.00005
0.00000
-0.00005
-0.00010
20 mV/s
10 mV/s
5 mV/s
100 mV/s
50 mV/s
20 mV/s
10 mV/s
5 mV/s
isoV / isoV-
isoV+ / isoV
0.05
0.10
0.15
0.20
0.25
0.30
0.35
(Potential sweep rate)0.5 [V0.5s-0.5
]
-1.25
-1.00 -0.75
-0.50
0.50
0.75
Potential [V vs. Ag | AgCl (sat.)]
1.00
1.25
c
d
Region of interest
Region of interest
5
0.701 V
0.704 V
0.707 V
0.709 V
0.71 V
2.0
1.5
1.0
4
3
2
-0.723V
-0.727V
-0.729V
-0.732V
-0.725V
inappropriate for model
50
100
150
200
50
100
150
200
e
Square Root from Frequency [Hz0.5
]
f
Square Root from Frequency [Hz0.5
2 mV/s
]
2 mV/s
5 mV/s
0.005
0.004
0.003
0.002
0.001
0.000
-0.001
-0.002
-0.003
-0.004
0.008
0.006
0.004
0.002
0.000
-0.002
-0.004
-0.006
-0.008
-0.010
5 mV/s
7.5 mV/s
10 mV/s
15 mV/s
20 mV/s
30 mV/s
7.5 mV/s
10 mV/s
15 mV/s
20 mV/s
30 mV/s
-1.8
-1.6
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Potential [V vs. Ag | AgCl(sat.)]
Potential [V vs. Ag | AgCl (sat.)]
h
g
0.2 mV/s, 'in Flow'
0.4 mV/s, 'in Flow'
0.6 mV/s, 'in Flow'
0.8 mV/s, 'in Flow'
0.2 mV/s 'in Flow'
0.4 mV/s 'in Flow'
0.6 mV/s 'in Flow'
0.8 mV/s 'in Flow'
0.003
0.002
0.001
0.000
0.002
0.001
0.000
-0.001
-0.002
-0.003
-0.001
-0.002
-0.003
-0.004
-1.47 C
-1.42 C
-1.36 C
-1.33 C
-1.43 C
-1.41C
-1.17 C
-0.92 C
-1.8
-1.6
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
Potential [V vs. Ag | AgCl (sat.)]
Potential [V vs. Ag | AgCl (sat.)]
Figure 3. a) Cyclic voltammograms for isoV at different potential scan rates (5 mV/s – 100 mV/s) in the potential range −1 V vs. Ag|AgCl (sat.) to 1 V vs. Ag|AgCl (sat.). b) Randles-Sevcik
analysis for each electrochemical process of isoV. c) Potentiostatic electrochemical impedance spectra for isoV−/isoV recorded close to the corresponding half-reaction potential (−0.723 to
−0.732 V vs. Ag|AgCl (sat.)) and converted for cotangent analysis. d) Potentiostatic electrochemical impedance spectra for isoV/isoV+ recorded close to the corresponding half-reaction
potential (0.701 to 0.710 V vs. Ag|AgCl (sat.)) and converted for cotangent analysis. e) Cyclic voltammograms for isoV-based negolyte recorded within the static RFB setup at different
potential scan rates (2.5 mV/s – 30 mV/s) in the potential range from −0.2 V to 1.4 V vs. Ag|AgCl (sat.). f) Cyclic voltammograms for isoV-based posolyte recorded within the static RFB setup
at different potential scan rates (2.5 mV/s – 30 mV/s) in the potential range from −0.2 V to −1.6 V vs. Ag|AgCl (sat.). g) Potentiodynamic bulk electrolysis series for isoV-based posolyte
recorded within the ‘in-flow’ RFB setup for different potential scan rates (0.2 mV/s – 0.8 mV/s) in the potential range −0.6 V to 1.1 V vs. Ag|AgCl (sat.). h) Potentiodynamic bulk electrolysis
series for isoV-based negolyte recorded within the ‘in flow’ RFB setup at different potential scan rates (0.2 mV/s – 0.8 mV/s) in the potential range −0.2 V to −1.6 V vs. Ag|AgCl (sat.).
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 7
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 8 of 13
ARTICLE
Journal Name
State of Charge Studies
frequency band, is observed and attributed to limited mass
View Article Online
DOI: 10.1039/D0TA07891C
In order to assess battery cycling behavior, electrochemical transfer. As in the previous OCP measurements, where limited
properties of the symmetrical redox electrolyte were studied ex performance at a late stage in the discharge process was
situ within a single arbitrary galvanostatic cycle via a sequential governed by the posolyte, a corresponding early cell voltage
cycle interruption. Multidimensional state of charge (SOC) drop correlates with a significant increase in the real part of the
analysis can help to explain the observed asymmetry of mass transfer impedance, supporting the hypothesis of specific
electrochemical responses from both half-cells.
interactions with the ion-exchange membrane. In order to
As a starting point, OCP measurements were conducted at achieve a theoretical 0% charge state, a potentiostatic step at 0
various SOCs for each half-cell (Figure 4a). Surprisingly, even at V vs. OCV was employed and a respective electrochemical
this stage the resulting asymmetric OCP profile revealed impedance measurement provided, as expected, the highest
inconsistency in the simultaneous processes of oxidation and values of Ωre for the low frequency region.
reduction. For the electrochemical oxidation of 1,5-diphenyl-3- In order to characterize the electrochemical properties of the
isopropyl-6-oxo-verdazyl, the determined OCP for starting and symmetrical verdazyl-based electrolyte more comprehensively,
finishing iterations does not follow classical Nernstian behavior, linear sweep voltammograms (LSV) were then recorded for
reaching the formal plateau by 20% SOC for the charging both negolyte and posolyte. A series of LSVs extend and
process and leaving it by 30% SOC for the discharge process, supplement the prior explicit OCP measurements, as the
thus hindering achievement of the theoretical capacity. Since starting point for every individual voltammetry experiment lies
this is an indication that the oxidized form of the verdazyl (isoV+) exactly where the cell current equals zero – or, in other words,
does not reach thermodynamic (Nernstian) equilibrium near at the OCP. The following approach allows for surveying how
the working porous electrode in the expected way for a given instantaneous electrolyte constitution affects the resulting
SOC, a plausible explanation for the posolyte issues is a specific electrochemical response on porous electrodes. However, an
but reversible interaction with the membrane (membrane SOC assignment gives the best result only for potential plateaus
fouling) (Figure S9). In further galvanostatic RFB cycling sessions near corresponding half-wave potentials and is barely
higher applied current densities were therefore favored, applicable for regions with an SOC of 80% or more, where the
causing the active area of membrane to be exposed to apparent current response from residual electroactive redox
intermediate concentrations of isoV+ for less time. On the other form is naturally low. For the sake of visibility, a contour map
hand, the negolyte shows no sign of a negative effect on final representation of an arbitrary galvanostatic cycle of interest
performance of the symmetric battery cell, corroborating the was recorded starting from the discharge of the RFB. Although
preceding potentiodynamic ‘in flow’ experiments.
the resulting graphs for the negolyte (Figure 4d) and posolyte
To further address the issue with the discharge process, the pre- (Figure 4c) are essentially similar, featuring a high apparent
charged (1.60 V) RFB was subjected to stepwise discharge with degree of electrochemical reversibility and a gradual current
a set of electrochemical impedance spectra recorded after each peak value shift towards more negative or positive potentials,
step. Since it was intended that the RFB should be studied at respectively, some details remain contrasted. For instance, each
intermediate states of charge, even a minor change of voltage electrolyte behaves differently when the completely charged
amplitude would result in an inconsistent non-linear current form of the electrolyte is considered. In particular, the
response, thus necessitating the choice of a galvanostatic mode electrochemical response of the posolyte exhibits a systematic
of EIS instead of potentiostatic. The real part of the current peak value upshift (50 mV) immediately after the
electrochemical impedance, Ωre, of both frequency- charging process is switched to discharging. The negolyte does
independent charge transfer resistance and frequency- and not do this to such an extent. Presumably, this discrepancy once
symmetry-dependent mass transfer impedance, comprising a again stems from the assumed susceptibility of the posolyte
total Faradaic impedance, has an essential dependency on the towards the membrane. Another important differing aspect is a
composition of the electrolyte at a given instant.56 When a pronounced lateral diffusional current observed for the
stable OCV is present, an established thermodynamic posolyte, where the current response on porous electrodes is
equilibrium near the surface of a given porous electrode systematically ‘smeared’ towards higher potentials; the
provides an opportunity to interrelate the OCV to a snapshot of negolyte does not exhibit this trend, instead demonstrating
the redox electrolyte, as according to the Nernst equation. more rapid current decay after reaching the corresponding
Therefore, mapping a set of the real part of impedance spectra reduction current peak.
at different OCVs helps to trace deviations from ideal Since solutions of each of the verdazyl oxidative states isoV+,
thermodynamic behavior. As expected, the absolute minima of isoV and isoV− possess rather distinguishable colors, an
the observed real impedances present over the whole additional approach applied was to carry out UV-Vis studies to
frequency band appear close to the OCV (Figure 4b). Relatively determine the corresponding negolyte (Figure 4f) and posolyte
low Ωre values are observed in the region 1.25–1.50 V, where (Figure 4e) compositions. Analyzed aliquots were sequentially
the expected galvanostatic discharge plateau should emerge; extracted after each galvanostatic step-charge and rest-step
otherwise, a steady increase of Ωre, especially for the low sequence and diluted with dry acetonitrile under an inert
8 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 9 of 13
Journal of Materials Chemistry A
Please do not adjust margins
Journal Name
ARTICLE
View Article Online
Real part of impedance [Ω ]
b
re
a
DOI: 10.1039/D0TA07891C
0.8
0.6
46.10
Posolyte
1.50
1.25
1.00
0.75
0.50
0.25
Theoretical
42.31
OCV
0.4
38.53
0.2
SOC 30%
34.74
0.0
SOC 20%
30.95
-0.2
-0.4
-0.6
-0.8
-1.0
27.16
23.38
19.59
15.80
SOC 10%
SOC 0%
Negolyte
Potentiostatic
5 SOC 0%
0.00
0
20
40
60
80 100 80
60
40
20
0
1
2
3
4
Logarithm from frequency [lg(rad/s)]
State of the charge [%]
I [A]
I [A]
c
d
80
0.009400
0.000
80
-0.001394
-0.002987
-0.004581
-0.006175
-0.007769
-0.009362
-0.01096
0.008583
0.007358
0.006133
0.004908
0.003683
0.002458
0.001233
0.000
60
40
60
40
20
20
0
0
-20
-40
-60
-80
-20
-40
-60
-80
-0.01255
0.6
0.8
1.0
1.2
1.4
-1.50
-1.25
-1.00
-0.75
Potential [V vs. Ag | AgCl (sat.)]
Potential [V vs. Ag | AgCl (sat.)]
e
0.30
f
0.25
0.20
0.15
0.10
0.05
0.00
0.4
0.25
0.20
0.15
0.10
0.05
0.00
0.25
100%
0.20
0.15
80%
0.3
0.2
0.1
0.0
100%
0.10
60%
20
40
60
80
100
0.05
State of charge [%]
80%
0.00
20
40
60
80
100
State of charge, %
40%
20%
60%
40%
350
400
450
500
550
600
650
350
400
450
500
550
600
Wavelength [nm]
Wavelenght [nm]
Figure 4.a) Open circuit half-cell potential vs. state of charge dependency for the posolyte and negolyte of the symmetrical isoV RFB. b) Contour map representation of the real part
of the galvanostatic electrochemical impedance evolution upon discharge of isoV RFB relative to the cell open circuit voltage and applied AC frequency. c) Linear sweep
voltammograms for different states of charge for the posolyte. d) Linear sweep voltammograms for different states of charge for the negolyte. e) Difference UV-Vis spectra for
different states of charge for the posolyte. f) Difference UV-Vis spectra for different states of charge for the negolyte.
atmosphere. It is worth noting that the rest step is essential for is observed at 380 nm (Figure S7). The broad absorption band
SOC measurements, as it enables redox electrolytes to at 490 nm hinders spectroscopic analysis of the posolyte. For
effectively mix throughout the half-cell space and thus prevents this reason, each set of UV-Vis absorption spectra was
undesirable mass gradients for both oxidative states of the transformed to a set of difference spectra by subtracting a
redox material present within. Obviously, for any symmetrical dataset recorded previously of the initial ‘discharged’ state. The
electrolyte, the initial ‘discharged’ state remains the same no subsequent general observation regarding the posolyte is an
matter which half-cell is considered.
A
characteristic initial absence of isoV+ related signal development until 40% of
absorbance peak maximum for the isoV radical electrolyte the theoretical state of charge (SOC), which does not contradict
solution appears at 490 nm, and a broad but minor absorption the previous OCP series of measurements. However, after this
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 9
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 10 of 13
ARTICLE
Journal Name
point characteristic signals from both isoV+ and isoV− appear at
547 nm and 388 nm respectively. Linear behavior of the
absorbance indicates compliance with the Beer–Lambert law,
making UV-Vis spectroscopy a method acceptable for SOC
monitoring of the symmetrical verdazyl electrolyte.
to the evolution of the voltage efficiency and, therefore, energy
View Article Online
efficiency upon cycling,62 they naturally converge as the
DOI: 10.1039/D0TA07891C
Coulombic efficiency stabilizes. Moreover, the voltage
efficiency (Figure 5d) had decreased by the 80th cycle from an
initial 77% down to 52%, demonstrating reasonable cyclability,
while after this point the rate of decay considerably increases
and the energy efficiency approached 10% by the final cycle of
Galvanostatic Cycling in Battery
The RFB, filled with the non-aqueous verdazyl-based (isoV) the test. Most likely, a steady change of the voltage drop arises
electrolyte, was exposed to galvanostatic cycling at various from the irreversible loss of redox active isoV through a specific
current densities (Figure 5a). The choice of the lower limit (‘cut-
off’) voltage is based on the fact that the initial state of any
symmetrical redox flow electrolyte must have an OCV close to
0 V, although mass transfer limitations dictate a shift down to
an optimal value of −0.20 V. The total voltage range for each
galvanostatic cycling experiment conducted was set between
−0.20 V and 2.00 V. Increasing the current density from 0.55
mA·cm-2 to 2.2 mA·cm-2 had a strong influence on the part of
the cycle governed by mass transport. Notably, close to the cell
voltage (1.42 V, as inferred from CV measurements) the well-
defined charge and discharge plateaus appear, being mostly
affected only by Ohmic voltage drop. The bottleneck for
performance of the RFB cell setup lies primarily in the discharge
process as at the lowest current density the applied Coulombic
efficiency reaches 88%, while the resulting charge capacity
fulfils theoretical expectations. As previously mentioned, in-
flow potentiodynamic behavior for the negolyte suggests a
considerable shift in the secondary ‘convectional’ anodic
current peak at high potential sweep rates, resulting in an
incomplete bulk discharge process. Besides this, a persistent
but minor current response for the posolyte remains down to
very low discharge potentials, until right after a particularly
sharp, yet narrow, Faradaic reduction process. Evidently, in
terms of galvanostatic cycling, the prior current peaks must be
interpreted as a discharge plateau, flat yet short, and the
aforementioned minor current response as a mass transfer
governed voltage drop. The same reasoning applied for the
charge profile interpretation provides a link to the barely
defined plateau at ~1.20 V, taking into account a set of broad
current peaks in the range of 0.25 – 0.40 V vs. Ag|AgCl(sat.) for
the posolyte.
sorption and from a developing of the membrane’s areal
resistance. Unsurprisingly, a similar trend is observed in the
capacity fading (Figure 5c). In the 40th galvanostatic cycle only
~50% of the theoretical discharge capacity is extracted from the
battery cell, and by the 80th cycle, which yields only 36% of the
theoretical discharge capacity, an abrupt linear decay begins.
The redox electrolyte instability mostly originates from an
intrinsic flaw of ion-exchange membranes used for symmetrical
systems like isoV. The initial oxidative state of isoV is charge
neutral and hence both oxidized and reduced forms of the
redox-active energy carrier possess opposite molecular charges.
Since the sign of the charge carried by one of the forms
inevitably matches the conductivity type of the chosen ion-
exchange membrane (e.g. anionic or cationic), the result is that
undesirable mutual interactions such as irreversible membrane
fouling can occur (Figure S9). This leads to a constant rise in
areal resistance and interferes with the cycling protocol used,
counteracting the pre-set voltage limits and eventually resulting
in poor apparent battery performance.
Conclusions
After initial screening of oxo-verdazyl radical compounds for
electrochemical properties relevant for battery applications, we
introduced the best-performing candidate, the 1,5-diphenyl-3-
isopropyl-6-oxo-verdazyl radical (isoV), as
a redox-active
material with two reversible oxidative states for a 1.42 V
symmetrical non-aqueous redox flow battery. A non-aqueous
redox electrolyte formulation with this species is the first
example of an oxo-verdazyl RFB and exhibits reasonable
capacity retention of 36% of theoretical capacity after over 80
charge/discharge cycles, with a related average Coulombic
efficiency of 97% at 2.2 mA·cm−2 and an average discharge
energy density of 0.147 Wh/L (a theoretical value of energy
density – 0.380 Wh/L). These encouraging results constitute a
successful proof of concept that will serve as the basis for
further research comprising: an optimization of the oxy-
verdazyl structure to achieve better compatibility with a chosen
type of separator by, for example, introduction of the constant
molecular charge, additional steric hindrance. Regarding the
supporting electrolytes, a proper selection of promising
supporting salts might enable enhanced stability of every
oxidative state and, therefore, extend the electrolyte life.
Moreover, appropriate ion-exchange membranes for bipolar
compounds require specific modifications as well.
Finally, the major requirement for any redox-active material
intended for use as part of an RFB electrolyte is that of high
long-term cycling stability. To evaluate this, a 10 mM solution of
the isoV electrolyte was subjected to continuous galvanostatic
charge/discharge sequences for 150 cycles at 2.2 mA·cm-2. For
this RFB setup as a battery separator a FAP-375-PP anion-
exchange membrane was selected as recently it demonstrated
a decent for NARFB compromise between mechanical stability,
selectivity and conductivity.9,24 Selected cycles from the
experiment (Figure 5b) suggest fairly reproducible
charge/discharge profiles in the vicinity of the theoretical open
circuit voltage of 1.42 V, although they do show a steady
moderate performance decay, meaning the duration of cycles
constantly declines in time. Despite a low observed Coulombic
efficiency of 85% for the first cycle, it rapidly increases during
following cycles and averages of ~97% thereafter. With regard
10 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 11 of 13
Journal of Materials Chemistry A
Please do not adjust margins
Journal Name
ARTICLE
View Article Online
b
DOI: 10.1039/D0TA07891C
a
0.55 mA/cm2
1.1 mA/cm2
1.65 mA/cm2
2.2 mA/cm2
2.0
1.5
1.0
0.5
0.0
2.0
1.5
Theoretical open circuit voltage
1.0
0.5
0.0
-0.5
0.5
1.0
7.5
14.5
20.5
0.00
0.05
0.10
0.15
0.20
0.25
Time [h]
Volumetric Capacity [Ah/L]
с
d
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
100
90
80
70
60
50
40
30
20
10
0
Coulombic Efficiency
Voltage Efficiency
Energy Efficiency
Normalized charge capacity
Normalized discharge capacity
0
25
50
75
100
125
150
0
25
50
75
100
125
150
Number of Cycles
Number of Cycles
Figure 5. a) Cycling profiles for the 10 mM symmetrical isoV battery at various current densities. b) Voltage profiles: evolution of charge/discharge curves for the symmetrical isoV
battery. c) Normalized (to the value of the theoretical capacity) capacity retention vs. number of cycles for the 10 mM isoV battery. d) Coulombic efficiency, voltage efficiency, and
energy efficiency vs. number of cycles for the 10 mM isoV battery.
be demonstrated that the posolyte is the source of the major
limiting impact on the battery’s performance, especially for
marginal states of charge. We expect the insights gained from
this level of analysis to inform our future optimization of
verdazyl-based batteries.
Introducing integrable reference electrodes to the RFB setup
allowed more sophisticated and electrode-local
electrochemical analysis compared to the usual three-electrode
setup. This enabled deeper understanding of the performance
imperfections related to specific electrolyte interactions with
accessible interfaces and other pathways of capacity losses. An
applied in-flow CV approach allowed the description of the
electrochemical processes occurring at the surface of the
porous hydrodynamic electrode of the working RFB and to
accomplish it separately for each half-cell. For instance, it
enabled the tracing of a mass transfer influence and explanation
of the galvanostatic discharge behavior of the RFB. We propose
this approach as a promising tool for the future evaluation of
battery design, and particularly to assess, explain, and even
quantify the impact of individual RFB optimizations and
enhancements.
To our knowledge, the highlighted approaches to
electrochemical analysis have not been previously employed for
the description of RFBs and we consider the described protocol
a good step towards method standardization in the field of
modern redox flow battery research.
Conflicts of interest
There are no conflicts to declare.
In the case of our isoV symmetrical electrolyte, the posolyte and
negolyte both affect battery performance, but in different ways,
rendered distinguishable and resolvable by our setup. It could
Acknowledgements
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 11
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 12 of 13
ARTICLE
Journal Name
26 C. Zhang, Y. Qian, Y. Ding, L. Zhang, X. Guo, Y. Zhao and G. Yu,
Authors thank the Ministry of Economic Affairs, Innovation,
Digitalization and Energy of the State of North Rhine-
Westphalia for funding this research work in the frame of the
“GrEEn” project (313-W044A) and the Deutsche
Forschungsgemeinschaft (DFG) for funding through the SFB 858
(project B11).
View Article Online
Angew. Chem. Int. Ed., 2019, 58, 7045–7050.
DOI: 10.1039/D0TA07891C
27 J.M. Stauber, S. Zhang, N. Gvozdik, Y. Jiang, L. Avena, K.J.
Stevenson and C.C. Cummins, J. Am. Chem. Soc., 2018, 140,
538–541.
28 Q. Liu, A.A. Shinkle, Y. Li, C.W. Monroe, L.T. Thompson and
A.E.S. Sleightholme, Electrochem. Commun., 2010, 12, 1634–
1637.
29 K. Gong, Q. Fang, S. Gu, S.F.Y. Li and Y. Yan, Energy Environ.
Sci., 2015, 8, 3515–3530.
30 R.A. Potash, J.R. McKone, S. Conte and H.D. Abruña, J.
Electrochem. Soc., 2016, 163, A338-A344.
31 C.G. Armstrong, R.W. Hogue and K.E. Toghill, J. Power Sources,
2019, 440, 227037.
References
32 W. Duan, R.S. Vemuri, J.D. Milshtein, S. Laramie, R.D. Dmello,
J. Huang, L. Zhang, D. Hu, M. Vijayakumar, W. Wang, J. Liu,
R.M. Darling, L. Thompson, K. Smith, J.S. Moore, F.R. Brushett
and X. Wei, J. Mater. Chem., 2016, A 4, 5448–5456.
33 T. Hagemann, J. Winsberg, B. Häupler, T. Janoschka, J.J.
Gruber, A. Wild and U.S. Schubert, NPG Asia Mater, 2017, 9,
e340.
34 S. Hwang, H.-s. Kim, J.H. Ryu and S.M. Oh, J. Power Sources,
2019, 421, 1–5.
35 T. Ma, Z. Pan, L. Miao, C. Chen, M. Han, Z. Shang and J. Chen,
Angew. Chem. Int. Ed., 2018, 57, 3158–3162.
36 R. G. Hicks, in Stable Radicals: Fundamental and Applied
Aspects of Odd-Electron Compounds, Wiley, Chichester,
2010, 250–251
1
2
3
4
5
6
7
8
9
S. Dühnen, J. Betz, M. Kolek, R. Schmuch, M. Winter and T.
Placke, Small Methods, 2020, 4, 2000039.
T. Placke, R. Kloepsch, S. Duehnen and M. Winter, J. Solid
State Electrochem., 2017, 21, 1939-1964.
M. Winter, B. Barnett and K. Xu, Chem. Rev., 2018, 118,
11433–11456.
L.F. Arenas, C. Ponce de León and F.C. Walsh, Curr. Opin.
Electrochem., 2019, 16, 117–126.
A. Lucas and S. Chondrogiannis, Int. J. Elec. Power, 2016, 80,
26–36.
R.M. Darling, K.G. Gallagher, J.A. Kowalski, S. Ha and F.R.
Brushett, Energy Environ. Sci., 2014, 7, 3459–3477.
Y. Ding, C. Zhang, L. Zhang, Y. Zhou and G. Yu, Chem. Soc. Rev.,
2018, 47, 69–103.
R.P. Fornari, M. Mesta, J. Hjelm, T. Vegge and P. de Silva, ACS
Mater. Lett., 2020, 2, 239–246.
M. Sanford, A. Shrestha, K. Hendriks, M. Sigman and S.
Minteer, Chem. Eur. J., 2020, 26, 5369-5373.
37 P.V. Petunin, E.A. Martynko, M.E. Trusova, M.S. Kazantsev,
T.V. Rybalova, R.R. Valiev, M.N. Uvarov, E.A. Mostovich and
P.S. Postnikov, Eur. J. Org. Chem., 2018, 2018, 4802–4811.
38 J.B. Gilroy, S.D.J. McKinnon, B.D. Koivisto and R.G. Hicks, Org.
Lett., 2007, 9, 4837–4840.
10 P. Leung, A.A. Shah, L. Sanz, C. Flox, J.R. Morante, Q. Xu, M.R.
Mohamed, C. Ponce de León and F.C. Walsh, J. Power Sources,
2017, 360, 243–283.
11 J. Luo, B. Hu, M. Hu, Y. Zhao and T.L. Liu, ACS Energy Lett.,
2019, 4, 2220–2240.
12 J. Winsberg, T. Hagemann, T. Janoschka, M.D. Hager and U.S.
Schubert, Angew. Chem. Int. Ed., 2017, 56, 686–711.
13 H. R. Jiang, J. Sun, L. Wie, M. C. Wu, W. Shyy and T. S. Zhao,
Energy Storage Mater., 2020, 24, 529-540.
14 H. R. Jiang, W. Shyy, M. C. Wu, R. H. Zhang and T. S. Zhao,
Appl. Energy, 2019, 233, 105-113.
15 Y. Yan, S.G. Robinson, M.S. Sigman and M.S. Sanford, J. Am.
Chem. Soc., 2019, 141, 15301–15306.
16 W. Wang and V. Sprenkle, Nat. Chem., 2016, 8, 204–206.
17 H. Chen, G. Cong and Y.-C. Lu, J. Energy Chem., 2018, 27,
1304–1325.
18 C.S. Sevov, R.E.M. Brooner, E. Chénard, R.S. Assary, J.S.
Moore, J. Rodríguez-López and M.S. Sanford, J. Am. Chem.
Soc., 2015, 137, 14465–14472.
19 S.-H. Shin, S.-H and Yun and S.-H. Moon, RSC Adv., 2013, 3,
9095–9116.
20 J. Chai, A. Lashgari, Z. Cao, C.K. Williams, X. Wang, J. Dong and
J.J. Jiang, ACS Appl. Mater. Inter., 2020, 12, 15262–15270.
21 S.E. Doris, A.L. Ward, A. Baskin, P.D. Frischmann, N.
Gavvalapalli, E. Chénard, C.S. Sevov, D. Prendergast, J.S.
Moore and B.A. Helms, Angew. Chem. Int. Ed., 2017, 56,
1595–1599.
22 S. Peng, L. Zhang, C. Zhang, Y. Ding, X. Guo, G. He and G. Yu,
Adv. Energy Mater., 2018, 8, 1802533.
23 Y. Huang, S. Gu, Y. Yan and S.F.Y. Li, Curr. Opin. Chem. Eng.,
2015, 8, 105–113.
24 S.G. Robinson, Y. Yan, K.H. Hendriks, M.S. Sanford and M.S.
Sigman, J. Am. Chem. Soc., 2019, 141, 10171–10176.
25 X. Wei, W. Duan, J. Huang, L. Zhang, B. Li, D. Reed, W. Xu, V.
Sprenkle and W. Wang, ACS Energy Lett., 2016, 1, 705–711.
39 B.D. Koivisto and R.G. Hicks, Coord. Chem. Rev., 2005, 249,
2612–2630.
40 D. Matuschek, S. Eusterwiemann, L. Stegemann, C.
Doerenkamp, B. Wibbeling, C.G. Daniliuc, N.L. Doltsinis, C.A.
Strassert, H. Eckert and A. Studer, Chem. Sci., 2015, 6, 4712–
4716.
41 G.-H. Ning, B. Tian, L.-M. Tan, Z. Ding, T.S. Herng, J. Ding and
K.P. Loh, Inorg. Chem., 2016, 55, 9892–9897.
42 S. Muench, A. Wild, C. Friebe, B. Häupler, T. Janoschka and
U.S. Schubert, Chem. Rev., 2016, 116, 9438–9484.
43 G.D. Charlton, S.M. Barbon, J.B. Gilroy and C.A. Dyker, J.
Energy Chem., 2019, 34, 52–56.
44 J. Langner, J. Melke, H. Ehrenberg and C. Roth, ECS Trans.,
2014, 58, 1–7.
45 J.M. Wrogemann, S. Künne, A. Heckmann, I.A.
Rodríguez-Pérez, V. Siozios, B. Yan, J. Li, M. Winter, K. Beltrop
and T. Placke, Adv. Energy Mater., 2020, 10, 1902709.
46 R.-S. Kühnel, D. Reber, A. Remhof, R. Figi, D. Bleiner and C.
Battaglia, Chem. Commun. (Cambridge, U.K.), 2016, 52,
10435–10438.
47 M. Cecchetti, M. Messaggi, A. Donazzi, A. Facibeni, V. Russo,
C.S. Casari, A.L. Bassi, A. Casalegno and M. Zago, Electrochim.
Acta, 2020, 329, 135143.
48 R.E.G. Smith, T.J. Davies, N.d.B. Baynes and R.J. Nichols, J.
Electroanal., 2015, 747, 29–38.
49 D. Menshykau and R.G. Compton, Electroanalysis, 2008, 20,
2387–2394.
50 V.V. Nikonenko and A.E. Kozmai, Electrochim. Acta, 2011, 56,
1262–1269.
51 L. Gubler, Curr. Opin. Electrochem., 2019, 18, 31–36.
52 H. Sun, M. Yu, Q. Li, K. Zhuang, J. Li, S. Almheiri and X. Zhang,
Energy, 2019, 168, 693–701.
53 A.M. Pezeshki, R.L. Sacci, F.M. Delnick, D.S. Aaron and M.M.
Mench, Electrochim. Acta, 2017, 229, 261–270.
12 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 13 of 13
Journal of Materials Chemistry A
Please do not adjust margins
Journal Name
ARTICLE
54 S. Eusterwiemann, D. Matuschek, L. Stegemann, S. Klabunde,
C.C. Doerenkamp, C.G. Daniliuc, N.L. Doltsinis, C.A. Strassert,
H. Eckert and A. Studer, Chimia, 2016, 70, 172–176.
55 R.G. Compton, C.E. Banks, in Understanding voltammetry,
Reprinted, World Scientific, New Jersey, 2009, 120–121.
56 A. Lasia, in Electrochemical Impedance Spectroscopy and its
Applications, Springer New York, New York, NY, s.l., 2014,
100–101.
View Article Online
DOI: 10.1039/D0TA07891C
57 S.N. Victoria and S. Ramanathan, Electrochim. Acta, 2011, 56,
2606–2615.
58 T. Tichter, J. Schneider, D. Andrae, M. Gebhard and C. Roth,
ChemPhysChem, 2020, 21, 428–441.
59 T. Tichter, D. Andrae, J. Mayer, J. Schneider, M. Gebhard and
C. Roth, Phys. Chem. Chem. Phys., 2019, 21, 9061–9068.
60 I. Streeter, G.G. Wildgoose, L. Shao and R.G. Compton, Sensor.
Actuat. B-Chem., 2008, 133, 462–466.
61 C. Yin, Y. Gao, S. Guo and H. Tang, Energy, 2014, 74, 886–895.
62 P. Meister, H. Jia, J. Li, R. Kloepsch, M. Winter and T. Placke,
Chem. Mater, 2016, 28, 7203-7217.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 13
Please do not adjust margins