Expanding the potential of redox carriers for flow battery applications

Article abstract of DOI:10.1039/d0ta04511j

Using theoretical modeling to guide our approach, substituents were selected to improve the negative and positive potentials associated with a representative metal-based redox carrier, a phenyl spaced nickel bispicolinamide complex, [Ni(bpb)],1. To broade

Full text of DOI:10.1039/d0ta04511j

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: G. A. Andrade, I.
A. Popov, C. R. Federico, P. Yang, E. R. Batista, R. Mukundan and B. L. Davis, J. Mater. Chem. A, 2020, DOI:
1
0.1039/D0TA04511J.
Volume
Number 12
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.
28
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

Page 1 of 23
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/D0TA04511J
Article type: Full
Title: Expanding the Potential of Redox Carriers for Flow Battery Applications
‡
‡
‡
Author(s), and Corresponding Author(s)*, Andrade, Gabriel A. ; Popov, Ivan. A. ; Federico,
Celia R.; Yang, Ping*; Batista, Enrique R.*; Mukundan, Rangachary; Davis, Benjamin L.*
Dr. G. A. Andrade, Celia R. Federico, Dr. R. Mukundan, Dr. B. L. Davis
MS K763
MPA-11: Materials Synthesis and Integrated Devices
Los Alamos National Laboratory
Los Alamos, New Mexico, USA, 87544
*
E-mail: bldavis@lanl.gov
Dr. I. A. Popov, Dr. P. Yang
MS B221
T-1: Physics and Chemistry of Materials
Los Alamos National Laboratory
Los Alamos, New Mexico, USA, 87544
*
E-mail: pyang@lanl.gov
Dr. E. R. Batista
MS B258
T-CNLS: Center for Nonlinear Studies
Los Alamos National Laboratory
Los Alamos, New Mexico, USA, 87544
*
E-mail: erb@lanl.gov
‡These authors contributed equally.
Keywords: redox flow battery, expanded window, non-aqueous
TOC: Small synthetic changes, Big Voltage Impact
EWG
O
N
N
N
N
O
3
V – 4V?
Ni
O
N
N
N
N
O
Ni
EDG
EDG
2

Journal of Materials Chemistry A
Page 2 of 23
View Article Online
DOI: 10.1039/D0TA04511J
Abstract
Using theoretical modeling to guide our approach, substituents were selected to improve the
negative and positive potentials associated with a representative metal-based redox carrier, a
phenyl spaced nickel bispicolinamide complex, [Ni(bpb)], 1. To broaden the cell potential,
electron donating groups were selected and installed on the pyridyl moieties, [Ni(bpb-
(
NMe ) )], 2, while electron withdrawing groups were incorporated on the phenyl linker,
2 2
[
Ni(bpb-R)] (R = -F, -CF , and -NO ), 3-5. Our model predicts an increase of ~300mV and
3 2
~
500mV for the first and second negative waves of Ni(bpb-NMe ), 2, and up to ~330mV and
2
~210 mV increase to the first and second positive waves in 3-5. The modeled complexes were
synthesized in good yields using a modification of the literature procedure. Due to limited
solubility, the differential pulse voltammetry was measured on complexes 2-5 using a drop cast
technique. Comparison of the observed and predicted potentials is discussed in acetonitrile
(MeCN) and dimethylformamide (DMF) solvents.
2

Page 3 of 23
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/D0TA04511J
Introduction
In the continuing quest to find new ways to store renewable energy that interface with our ever-
growing demand, we focused our acumen on nonaqueous redox flow batteries (RFB). RFBs
stand apart from traditional batteries, where the electrode sets the limit of stored capacity, by
using molecular or oligomeric/polymeric redox carriers stored outside the cell.[1–5] Thus, in a
non-hybrid forms, power (cell stack) and energy (tanks) are fully decoupled and allow for
flexible end-user implementations. Of course, such promise has hefty demands – the redox
carriers are asked to perform at extreme potentials, form concentrated solutions (> 1M) that
aren’t too viscous, not embark on irreversible side reactions, and do this over a 20-year
lifecycle. Very few manifestations even flirt with all of these criteria,[6,7] and only a few
aqueous systems have been tested that may meet these lifecycle requirements.[8] Recent work
on novel electrode designs may facilitate aqueous vanadium RFBs to achieve the lifetime
[9]
[2]
requirements – many more hurdles are required for nonaqueous systems. Indeed, solubility
of active species also plays an important role for the performance of RFBs. Various strategies
on solubility tuning of charge carriers have been reported elsewhere.[10,11] Other studies
demonstrated an effective way to significantly improve the reaction kinetics and redox
capacity[12] and computationally explored stability and reversibility of redox species for
RFBs.[13]
Rather than try to find a redox carrier that meets all of these criteria, we endeavored to
explore whether energy density could be improved by tuning the redox waves. This has already
been examined in aqueous systems, most notably with quinones[14] demonstrating some
variation in potential. In the nonaqueous realm, there have been modifications to posolytes
(
ferrocene,[15] cyclopropenium[16]) and negolytes (cobaltocene,[17] metal clusters[18]). Some
systems, such as the polyoxovandate-alkoxide clusters,[19] can be tuned with metal and ligand
substitutions, although the ligand changes only exhibit a minor impact.[18] A summary of
2

Journal of Materials Chemistry A
Page 4 of 23
View Article Online
DOI: 10.1039/D0TA04511J
selected redox carrier potential changes is shown in Table 1.
Redox Carrier
Base Potential
V)
Substitution
New
Potential
(
(Change)
Fc-
3
.4 (0.23)
3
.26 (vs.
+
CH N EtMe2
2
Fc[15]
Li/Li )
+
0
.3 (vs. SHE)
C H C=OMe
0.65 (0.35)
1.33 (0.5)
5
4
Cyclopropenium[16]
AQDS
0.83 (vs. Fc)
Thioether
0
.21 (vs. SHE)
DHAQDMS
0.02 (0.19)
-1.65 (0.55)
[8]
Aqueous
[17]
Cp Co
-1.1 (vs. SHE)
-0.72, -0.22
Cp* Co
2
2
Lindqvist POV[19]
V O (OMe)
-2.1, -1.58
(1.38, 1.36)
Ti V
4
+
2
(vs. Ag )
6
7
12
0
.46, 1.00
-0.67, -0.11
0.21-0.24)
Lindqvist POV[18]
V O (OEt)
0.22, 0.79
-0.88, -0.34
-
NMe variant
3
6
7
12
(
Table 1. Redox Carriers and modified variations used in RFBs.
Bold potential changes are to highlight large magnitude changes
AQDS = anthraquinone-2,7-disulfonic acid
DHAQDMS = 1,4-dihydroxyanthraquinone-2,3-dimethylsulfonic acid
Cp = C , Cp* = C Me , Fc = ferrocene, Cp Fe; POV = polyoxovanadate
5
H
5
5
5
2
In order for any RFB to be deployed on a significant scale, the cost must come down. One
mechanism for that goal to be achieved is through improving the cell potential above 2V; better
yet 3V and above.[20] As Table 1 indicates, changes in metal center provide the largest impact
for a particular system (polyoxometalate) with functional group changes only providing modest
impact of 0.2-0.5 V. Of course, if both the posolyte and negolyte can be modified readily, the
cell could be improved by ~1V.
We previously reported nickel systems, which displayed promising voltammograms and
cycled well through the initial negative wave.[ 21] While one of the complexes studied, Ni(bpb),
1
, displayed low currents due to low solubility, we found that 1 would be a good model
2

Page 5 of 23
Journal of Materials Chemistry A
View Article Online
compound to study because it contained qualitatively reversible positive and negative waves
DOI: 10.1039/D0TA04511J
that might be influenced with judicious incorporation of functional groups, vide infra.
Theoretical studies of the phenyl spaced bispicolinamide nickel complex, 1, found that the
negative waves corresponded to a metal-based reduction, followed by reduction of the ligand
at the pyridine ring. The addition of electron donating groups (EDG) to the pyridyl periphery
should make these events more difficult to reduce, resulting in a more negative reduction
potential. Similarly, the positive waves corresponded to oxidation at the metal followed by
oxidation of the ligand, although at the phenyl linker. Addition of electron withdrawing groups
(EWG) on the phenyl backbone should make these events harder to oxidize and thus lead to
more positive oxidation waves (Eq. 1).
O
O
N
N
N
N
O
O
N
N
N
N
O
O
N
N
N
N
-e-
Ni
+e-
Ni
Ni
EDG
EDG
EDG
EDG
EDG
EDG
less stable
more potential
EWG
EWG
EWG
O
N
N
N
O
+e-
O
N
N
N
O
O
N
N
N
N
O
-
e-
Ni
Ni
Ni
N
N
less stable
more potential
Equation 1. Generalized equations showing Ni(bpb) equipped with electron donating groups (EDGs, green, above) or
electron withdrawing groups (EWGs, red, below) and the associated oxidized/reduced complexes.
Due to the simplicity in synthesis of this class of compounds, we sought commercially available
starting materials to accomplish these modifications and test our hypothesis. Using theoretical
guidance, and availability of starting materials, NO , F, and CF groups were targeted as
2
3
EWGs, and NMe as the EDG. The synthesis of the proligands was accomplished using a
2
common one-pot peptide coupling strategy and metalation was also easily accomplished with
good overall yields. The results indicated that the target compounds were successful in altering
2

Journal of Materials Chemistry A
Page 6 of 23
nd
View Article Online
DOI: 10.1039/D0TA04511J
the redox potential of the 2 positive and negative waves by ~0.5V.
Results and Discussion
Potential Change Predictions.
Using previously described methods,[22] complexes 1-5 were modeled to predict their redox
potentials and the localization of the electron upon oxidation/reduction in DMF (Table 3) and
in MeCN (Table S3). Regardless of the solvent, for complexes 2-5 with an EDG or EWG, the
oxidations follow intuition with EDG making the process more facile (complex 2) and more
difficult with EWG substitutions (complexes 3-5). The changes predicted in these complexes
are worth highlighting specifically: an increase of ~300mV and ~500mV for the first and
second negative waves of [Ni(bpb-(NMe ) )], 2 and up to ~330mV and ~210 mV increases to
2
2
the first and second positive waves in 3-5, respectively. These increases, if combined in the
same cell, would result in a significant increase in cell potential. Similar potential changes are
found when the solvent medium is MeCN (Table S3).
Synthesis of Materials.
The synthetic pathway to prepare the phenyl bridged di-picolinamide nickel complexes is
shown in Scheme 1. The preparation began with coupling of o-phenlyenediamine derivatives
with picolinic acid derivatives using HATU (1-[Bis(dimethylamino)methylene]-1H-1,2,3-
triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate) to generate the target diamide
proligands. HATU was used as the conditions and reagents are much milder than previously
reported in the literature and worked as a general method for all derivatives.[23] The respective
metal complexes were generated rapidly by deprotonation of the amide protons with
trimethylamine and adding nickel(II) tetrafluoroborate hexahydrate to precipitate the target
complexes in high yields (79 to 95%). Using this pathway, we were able to synthesize a group
2

Page 7 of 23
Journal of Materials Chemistry A
View Article Online
of complexes with the desired electron withdrawing and donating groups (Scheme 1), labelled
DOI: 10.1039/D0TA04511J
as [Ni(bpb)], 1, [Ni(bpb-(NMe ) )], 2, [Ni(bpb-F)], 3, [Ni(bpb-CF )], 4 and [Ni(bpb-NO )], 5.
2
2
3
2
R2
R2
R2
H2N
NH2
O
OH
N
O
O
Ni(H2O)6(BF4)2
NEt3
O
N
N
N
N
O
HATU
NH HN
DIEA
Ni
DMF
MeCN
R1
N
N
R1
R1
R1
R1
1
. [Ni(bpb)] R1=H, R2=H
. [Ni(bpb-(NMe2)2)] R1=NMe2, R2=H
. [Ni(bpb-F)] R1=H, R2=F
. [Ni(bpb-CF3)] R1=H, R2=CF3
. [Ni(bpb-NO2)] R1=H, R2=NO2
2
3
4
5
Scheme 1. Synthesis of bispicolinamide variants
Examining the IR stretch of the C=O in each complex, a predictable trend is observed
of greater double bond character for stronger EWG on the phenyl linker. NO has the strongest
2
-1
withdrawing power in complex 5, yielding a stretch of 1653 cm ; -CF is comparable with
3
-1
-1
1
648 cm in complex 4 and -F is weaker still with 1632 cm in complex 3. In the case of
-1
complex 2, without any phenyl substitution, the 1604 cm stretch indicates the most donation
into the amide functionality resulting in the least double bond character. The reduced donating
power of the amide nitrogen is a portent for the changes in reduction potentials (vide infra).
Electrochemical Analysis.
Evaluation of the electrochemical impact of substitution on 1 was initially attempted using
standard cyclic voltammetric methods: acetonitrile (MeCN) solvent, hexafluorophosphate
supporting electrolyte. Due to poor solubility in MeCN, several solvents were tried and DMF
was found to dissolve complexes 1-4 sufficiently for CV measurement (Fig 1, Table 2). The
small currents observed precluded quantitative analysis of the waves, but qualitatively a few
trends are apparent. When an EDG is appended to the pyridyl moieties (complex 2), the
resulting waves associated the ligand’s reduction are shifted to move negative potentials, with
2nd event at the periphery of the solvent window. Conversely, when EWGs are attached to the
phenyl linker (complexes 3 and 4), the positive waves are shifted more positive and separated
2

Journal of Materials Chemistry A
Page 8 of 23
View Article Online
from each other. In attempt to get more confidence in our E assignments, we explored drop
DOI: 10.1039/D0TA04511J
1
/2
casting complexes 1-5 onto the carbon electrodes. Control experiments with 1 (Fig S1 and S2)
indicate excellent agreement between the E1/2 observed by CV, validating this approach. The
differential pulse voltammetry (DPV) of 1 and 2 (Fig 2a) shows the donating group -NMe2
shift the negative waves by ~0.5V. The impact of EWGs is also now easier to visualize (Fig
2
b), with both positive waves shifting 0.3-0.5V upon substitution with -CF or -NO . The
3 2
influence of -F is less clear, as the prominent positive feature is at nearly the same potential as
with small feature at slightly larger potential (~1V).
1
a
a
Figure 1. Voltammograms of 1 in DMF (red solid line), 1 in MeCN (black dotted line), 2 in DMF (purple), 3 in DMF
rd
(
orange), and 4 in DMF (green). Data shown is obtained from 3 sweep.
Figure 2. DPV for drop cast complexes 1 and 2 (left), and complexes 1, 3-5 (right).
2

Page 9 of 23
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/D0TA04511J
Complex
I
p
2nd Reduction
V)
I
p
1st Reduction (V)
I
p
1st Oxidation (V)
I
p
2nd Oxidation (V)
(
[Ni(bpb)], 1
–2.35 V
–1.96
0.55
0.68
[
Ni(bpb-NMe
Ni(bpb-F)], 3
Ni(bpb-CF )], 4
Ni(bpb-NO )], 5
2
)], 2
–2.86 (–0.51)
–2.39 (–0.04)
–2.35 (0.0)
–2.33 (–0.37)
–1.96 (0.0)
0.48 (–0.07)
0.72 (+0.17)
0.79 (+0.11)
0.92 (+0.24)
[
[
3
–2.17 (–0.21)
0.85 (+0.30)
1.07 (+0.39)
[
2
–2.34 (+0.01)
–2.00 (–0.04)
0.93 (+0.38)
1.15 (+0.47)
Table 2. Summary of potentials (V) observed by DPV experiments with dropcast complexes (diff from complex 1).
Corrected to Fc as a reference (-0.087V)
To assess whether the incorporation of an EWG or EDG had a desirable impact on the
electrochemical properties, we examined how the E1/2 of the waves changed relative to each
other and the spacing between the most negative/positive waves. The results (Table S1)
indicate, with the exception of negative wave differences in complexes 4/5 (negative
percentage changes), that the spacing increased by 0.07-0.18V relative to complex 1 in nearly
all cases. Additionally, the widest spacing (right most column) between the most
negative/positive waves improved by 0.07-0.62V, equating to a 3-20% increase. These data
show that judicious installation of an EWG or EDG substituent can improve the voltaic
qualities of a redox carrier.
2

Journal of Materials Chemistry A
Page 10 of 23
View Article Online
DOI: 10.1039/D0TA04511J
Comparison with the Computed Reduction Potentials
Figure 3. Computed vs. experimental reduction potentials (V) of complexes 1-5 in DMF in the {2+ →
−} range of charge states. Theoretical reduction potentials corresponding to the metal-centered events
are depicted in the 2+ → 1+ and 0 → 1− steps, and to the ligand-centered events in the 1+ → 0 and 1− →
− steps.
2
2
The theoretical calculation of the redox potentials yielded insight into the nature of the
redox events at each step (metal- or ligand-based reduction/oxidation) in the {2+ → 2−} range
of charge states. Most of the calculated reduction potentials are found within 0.2V of the
corresponding experimental values (Fig 3, Table 3), with only a few of them within 0.3V, in
accordance with the error bar, previously proposed for the calculation of reduction potentials
within the revised Born-Haber approximation.[22] This approach has previously been
successfully applied to various Ni- and Fe-based charge carriers as described elsewhere.[21,24,25]
For clarity in the following discussion, the electrochemical waves are described relative to 0V,
e.g. the ‘first’ wave is closest to 0, followed by the second, etc. The charge state of 0 ([1-5]0)
2
+
2−
corresponds to neutrally charged complexes (Ni , L ), while the 1+/2+ and 1−/2− charge
states correspond to their oxidized and reduced forms, respectively. Comparing the
experimental and computed reduction potentials for the 1+ → 0 step, it is concluded that the
2

Page 11 of 23
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/D0TA04511J
[Ni(bpb-CF )], 4 [Ni(bpb-NO )], 5
3 2
Comple
x/
reductio
n step
[Ni(bpb-
[
Ni(bpb)], 1
[Ni(bpb-F)], 3
(NMe
2
)
2
)], 2
Theor.
0.99 [L]
Exp.
0.68
Theor.
Exp.
Exp.
0.92
Theor.
Exp.
1.07
Theor.
Exp.
1.15
Theor.
1
.05 [L]
1.06 [L]
0.93 [M]
1.25 [L]
1.12 [M]
1.38 [L]
1.02 [M]
2
+ → 1+
0.79
0.48
0
.91 [M]
0.78 [M]
0
0
.42 [M]
.56 [L]
0.23 [M]
0.43 [L]
0.44 [M]
0.58 [L]
0.57 [M]
0.71 [L]
0.53 [M]
0.89 [L]
1
+ → 0
→ 1−
0.55
−1.96
−2.35
0.72
−1.96
−2.39
0.85
−2.17
−2.35
0.93
−2.00
−2.34
−
2.13
−2.43
[M]
−2.09
[M]
−2.14
[M]
−1.98
[M]
0
−2.33
−2.86
[M]
−
2.56
−3.06
[L]
−2.53
[L]
−2.48
[L]
−2.05
[L]
1
− → 2−
[L]
Table 3. Redox potentials experimentally measured and modeled for the [Ni(bpb-R)] series in DMF. All
potentials reported are vs. Fc and are reported as E1/2 unless the peak is irreversible, in which case the E is
p
reported. [L] or [M] indicates ligand- or metal-centered events, respectively. Theoretical values shown in bold
correspond to the redox potentials based on the most thermodynamically stable electronic configurations. For
the 1+ → 0 and 2+ → 1+ steps, two values of reduction potentials are shown based on the two possible
1
+
localization sites in the [Ni(bpb-R)] complexes for comparison. Theoretical reduction potentials of
complexes 1-5 obtained in MeCN are summarized in Table S3.
first oxidation of all 1-5 complexes in DMF should be regarded as ligand-based due to the
removal of an electron from the aryl moiety and amide linker (Tables S9, S13, S17, S25).
However, the first oxidation of complex 1 in MeCN solvent was previously[21] suggested to be
3
+
metal-centered (Ni ) due to the analogy to the redox potentials in similar Ni(bpen) and
Ni(bppn) as well as Co- and Fe-coordination complexes featuring aryl linkers. From the free
1
+
3+
energy calculations of the [1] complex in DMF, removal of an electron from the metal (M )
1
−
or from the ligand (L ) is energetically indistinguishable (within the errors expected for these
calculations), although slightly favoring the metal oxidation, i.e. by ~0.14 eV (Table S5). In
MeCN, this difference is even smaller, i.e. 0.05-0.07 eV, but in favor of the ligand oxidation
(Table S4). Hence, the assignment of where the electron comes from is equivocal for complex
1
+
[
1] . In any scenario, values of the reduction potentials for the 1+ → 0, and consequently, the
2
+ → 1+ events are very similar to each other regardless of the assignment (Table 3). For the
1
+
EWG- or EDG-substituted complexes [2-5] , the energy difference between metal- and
ligand-centered oxidation gets larger in DMF, in favor of the metal-based oxidation (Table S5),
while the ligand-centered first oxidations are favored in MeCN based on the most
2

Journal of Materials Chemistry A
Page 12 of 23
1
+
View Article Online
DOI: 10.1039/D0TA04511J
thermodynamically stable [2-5] configurations (Table S3). Comparing the experimental and
computed reduction potentials in DMF for all 1-5 complexes, it is clear that the theoretical
ligand-centered events show better agreement with the experimental data for the 1+ → 0 step
(
0.01-0.14 eV deviation for the ligand-centered events vs. 0.13-0.40 eV for the metal-centered
events).
Assuming ligand-based event for the first oxidation, the 2+ → 1+ step corresponding
to the second oxidation is assigned to the metal-based event resulting in Ni3+ species.
Specifically, the 2+ state is characterized as a broken-symmetry singlet state with an unpaired
electron on the metal antiferromagnetically coupled with the one on the 1,2-diaminobenzene
moiety (Tables S8, S9, S12, S13, S16, S17, S20, S21, S24, S25), similar to the case of the
III
ox1
[26]
[
Fe (bpb) (CN) ] complex. The assignment of the electron localization is in accordance
2
with the experimental reduction potential data, demonstrating a better agreement for the metal-
based second oxidation than for the ligand-based event (0.01-0.23 eV deviation for the metal-
centered events vs. 0.14-0.37 eV for the ligand-centered events).
Similar to complex 1, in the negative waves of complexes 2-5, the first reduction places
1
+
the electron on the metal (Ni ), whereas in the second reduction pyridyl moieties accept an
electron antiferromagnetically coupling with the metal. Thus, it is concluded that the nature of
the redox events for complexes 2-5 is identical to that of complex 1 in the entire range of {2+
→
2−} charge states, with the reduction potentials getting more negative for the EDG-
substituted complex 2 in the negative waves, and more positive for the EWG-substituted
complexes 3-5 in the positive waves. It is worth noting that according to our calculations the
same conclusions are drawn for both DMF and MeCN solvents (Tables S3-S27). As one would
expect, both solvents are redox innocent throughout the {2+ → 2−} charge states in all
complexes.
2

Page 13 of 23
Journal of Materials Chemistry A
View Article Online
It is also worthy to note that upon functionalization geometries of the EWG- or EDG-
DOI: 10.1039/D0TA04511J
substituted complexes stay close to the original complex 1[21] at their respective charge states,
regardless of a solvent (Figs S3-S6, Tables S6, S7, S10, S11, S14, S15, S18, S19, S22, S23).
0
1+
While the neutral [Ni(bpb-R)(DMF) ] and ligand-based oxidized form, [Ni(bpb-R)(DMF) ] ,
2
2
of complexes 1-5 exhibit square-planar geometries (Fig S4) with weakly bound DMF
3
+
molecules, forming Ni produces complexes with two DMFs directly coordinating to Ni in
axial positions (R(Ni-ODMF)=2.10-2.12 Å in all 1-5. Similar coordination modes are found in
the [Ni(bpb-R)(MeCN)2]0/1+/2+ complexes (Fig S3), where MeCN molecules also form the
shortest bonding with the Ni center at the 2+ charge state (R(Ni-NMeCN)=2.15-2.18 Å).
Conclusion
The modification of redox carriers used in RFBs can have a profound impact on their
st
nd
performance metrics. For complex 2, the change predicted for its 1 and 2 negative waves,
300mV and 500mV, were very close to the measured changes of 370mV and 510mV;
~
likewise, for complexes 3-5, predictions of up to 330mV and 210mV were also reasonable to
the measured changes of 380mV and 470mV. The resulting potential changes for the posolyte
are modest compared to the state of the art (cyclopropenium,[16] Tab 1), but comparable to the
best negolyte (cobaltocene,[17] Tab 1). As such, at 100% state-of-charge and utilizing the
nd
nd
positive 2 wave of complex 5 and 2 negative wave of complex 2, the initial cell voltage
would be ~4V, almost 1V over the unsubstituted Ni(bpb). Although these particular complexes
have insufficient solubility to be used in practical applications, several published methods to
improve solubility could be utilized to offset this characteristic.[6,27] Additionally, the described
approach can be applied to other redox carriers in the literature to enhance the cell potential for
which they are a part.
2

Journal of Materials Chemistry A
Page 14 of 23
View Article Online
DOI: 10.1039/D0TA04511J
Experimental
All syntheses were conducted under atmospheric conditions unless otherwise noted. The ligand
H bpb was prepared according to literature procedure. All remaining reagents were purchased
2
from commercial sources and used as received, including HATU (C H F N OP, CAS
1
0
15
6
6
1
48893-10-1). DMSO-d was dried over activated molecular sieves before use. NMR spectra
6
were obtained on a Bruker DRX 500 or Neo 500 spectrometer at room temperature, then
processed and analyzed with MestReNova software (v11.0.4-18998). H and 13C chemical
shifts are reported in parts per million (ppm) relative to TMS, with the residual solvent peak
used as an internal reference. NMR multiplicities are reported as follows: singlet (s), doublet
1
(d), triplet (t), quartet (q), multiplet (m), broad signal (br). Coupling constants (J) are reported
in hertz (Hz). Infrared spectroscopy was performed on a Thermo Scientific Nicolet iS50
spectrometer using an ATR attachment (diamond crystal). All electrochemical analyses were
carried out in an argon-filled glovebox or in an argon saturated solution. The supporting
electrolyte, tetrabutylammonium hexafluorophosphate (Sigma, ≥98%), was recrystallized from
absolute ethanol 3 times and evacuated prior to use. The DMF used for analysis was distilled
over CaH and stored over activated 3 Å molecular sieves prior to use. Cyclic voltammetry was
2
performed with a Metrohm PGSTAT204 potentiostat in a three-electrode electrochemical
electrochemical cell consisting of a glassy carbon disk working electrode (0.07 cm2, BASi), a
+
Ag/Ag reference electrode (BASi) containing 0.01 M AgNO or AgBF in MeCN, and a
3
4
platinum wire counter electrode (23 cm, ALS). Drop cast technique was adapted from the
literature.[28,29] All experiments were run at scan rates ranging from 20 to 300 mV/s in an MeCN
(unless otherwise specified) electrolyte containing 1mM of the Ni complex and 0.1 M TBAPF6.
High resolution mass spectrometric data was obtained on a Waters Synapt G2, usingMassLynx
software. Elemental analysis was performed at Los Alamos National Laboratory on the
following instrument: CHNS-O Analyzer, Model No.: EA1112.
2

Page 15 of 23
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/D0TA04511J
Synthesis of 4-(Dimethylamino)picolinic acid
The title compound was prepared from a modified literature procedure. 0.402 g (2.55 mmol)
of 4-Chloro-2-pyridinecarboxylic acid and 4 ml of a 40 % by wt solution of dimethylamine in
water was added to a Lab-Crest pressure reaction vessel (Andrews Glass Co.). The reaction
was magnetically stirred and heated to 150 °C for 3 hr. After the reaction cooled to room
temperature, the solvent was removed under vacuum yielding an off white solid. The solid was
washed with 3 ml of EtOH and 5 ml of diethyl ether, collected on a fritted filter and dried,
1
yielding the title compound (0.372 g, 72 % yield) as the HCl salt. H NMR in D O was
2
1
consistent with the literature. H NMR (500 MHz, D O) δ 7.94 (d, J = 7.3 Hz, 1H), 7.19 (d, J
2
=
2.7 Hz, 1H), 6.82 (dd, J = 7.2, 2.7 Hz, 1H), 3.20 (s, 6H).
Synthesis of H bpb-(NMe )
2
2 2
The proligand was prepared using standard HATU coupling conditions. In a 50 ml round
bottom flask, which was dried in an oven at 150 °C overnight prior to use, was added 4-
(dimethylamino)picolinic acid (0.277 g, 1.67 mmol), HATU (0.634 g, 1.67 mmol), and 5 ml
of dry DMF. To the mixture, N,N-diisopropylethylamine (0.66 ml, 3.79 mmol) was added and
the suspension was stirred at room temperature for 30 min. Next, o-phenylenediamine (0.081
g, 0.76 mmol) was added and the mixture was stirred at room temperature overnight. The DMF
was then removed to yield a dark solid which was washed with DCM. The resulting off white
material was dissolved in a 1:1 solution (10 ml) of 0.1 M aqueous KOH and DCM. This
biphasic solution was stirred for 5 min then extracted with DCM (3 x 12 ml). The organic layer
was dried and the solvent was removed. The solid was then purified on silica using a 10 %
1
MeOH in DCM solution to elute the title compound (0.120 g, 40 % yield). H NMR (500 MHz,
CDCl ) δ 10.33 (s, 2H), 8.15 (d, J = 5.8 Hz, 2H), 7.92 (dd, J = 5.9 Hz, 2H), 7.57 (d, J = 2.6 Hz,
3
1
3
2
H), 7.26 (d, J = 9.5 Hz, 2H), 6.56 (dd, J = 3.1 Hz, 2H), 3.08 (s, 12H). C NMR (126 MHz,
CDCl ) δ 164.08, 155.48, 150.18, 148.44, 130.43, 125.87, 124.48, 108.38, 105.56, 39.49.
3
2

Journal of Materials Chemistry A
Page 16 of 23
View Article Online
Calculated for (M+Na): C H N O Na, 427.1859; found 427.1854. Calculated for (M+H):
DOI: 10.1039/D0TA04511J
2
2
24
6
2
C H N O , 405.2039; found 405.2031 (found both ions)
2
2
25
6
2
Synthesis of [Ni(bpb-(NMe ) )], 2
2
2
The proligand (0.120 g, 0.297 mmol) was suspended in MeCN (10 ml) and triethylamine was
added (0.188 ml, 1.35 mmol) yielding a solution. Nickel(II) tetrafluoroborate hexahydrate
(0.091 g, 0.270 mmol) was added to the solution, immediately forming an orange precipitate.
This mixture was stirred at room temperature for 5 hours and the precipitate was collected on
a fritted filter, washed with DCM (10 ml) and dried, yielding the desired nickel complex (0.118
g, 95 % yield) IR (ATR): 1604 (b s, C=O), 1564, 1544, 1525, 1474, 1451, 1434, 1383 (b s),
1
273, 1070, 1028, 950, 936, 858, 794, 776, 767, 610, 573, 527, 468. HR-ESI-MS: m/z
calculated for (M+Na) C H N O NiNa: 483.1055; found: 483.0999
2
2
22
6
2
Synthesis of H bpb-F
2
Preparation of the fluorinated proligand followed the same protocol for H bpb-(NMe ) with
2
2 2
the only modification being the reaction was performed in an argon filled glovebox. 2-Picolinic
acid (0.305 g, 2.48 mmol), HATU (0.942 g, 2.48 mmol), and DMF (5 ml) were added to a dry
5
5
0 ml round bottom flask equipped with a stir bar and N,N-diisopropylethylamine (0.981 ml,
.63 mmol) was added. This solution was stirred at room temperature for 30 min. 4-fluoro-1,2-
phenylenediamine (0.142 g, 1.13 mmol) was then added and the solution was stirred at room
temperature for 24 hours. The majority of the solvent was then removed under reduced pressure
leaving roughly 1 ml of the crude solution. Water was then added to precipitate the desired
1
product which was collected on a fritted filter and dried (0.326 g, 86 % yield). H NMR (500
MHz, CDCl ) δ 10.40 (s, 1H), 10.01 (s, 1H), 8.60 (d, J = 4.3 Hz, 1H), 8.49 (d, J = 4.3 Hz, 1H),
3
8
.30 (dd, J = 18.3, 7.8 Hz, 2H), 7.90 (dt, J = 16.0, 7.7 Hz, 3H), 7.74 – 7.61 (m, 1H), 7.55 – 7.47
1
3
(
m, 1H), 7.48 – 7.39 (m, 1H), 6.97 (t, J = 9.5 Hz, 1H). C NMR (126 MHz, CDCl3) δ 163.39,
1
62.79, 161.86, 159.91, 149.58, 148.38, 137.72, 132.97, 132.87, 126.82, 124.81, 122.72,
2

Page 17 of 23
Journal of Materials Chemistry A
View Article Online
1
3
12.49, 112.31, 110.96, 110.74. Calculated for (M+Na): C H FN O Na, 359.0920; found
18 13 4 2
DOI: 10.1039/D0TA04511J
59.0922
Synthesis of [Ni(bpb-F], 3
Preparation of the complex was identical to the dimethylamine derivative. The proligand (0.091
g, 0.271 mmol) was suspended in MeCN (10 ml) and trimethylamine was added (0.075 ml,
0
.541 mmol) yielding a solution. Nickel(II) tetrafluoroborate hexahydrate (0.084 g, 0.246
mmol) was added to the solution, immediately forming an orange precipitate. This mixture was
stirred at room temperature for 5 hours and the precipitate was collected on a fritted filter,
washed with DCM (10 ml) and dried, yielding the desired nickel complex (0.076 g, 79 % yield).
1
8
H NMR (500 MHz, DMF-d ) δ 8.66 – 8.53 (m, 2H), 8.29 (q, J = 7.4 Hz, 2H), 8.22 (dd, J =
7
.8, 6.1 Hz, 1H), 7.82 (dd, J = 12.4, 7.7 Hz, 2H), 7.78 – 7.70 (m, 2H), 6.62 (dt, J = 10.2, 5.1
Hz, 1H). IR (ATR): 1652, 1632 (s, C=O), 1602, 1576, 1486, 1469, 1435, 1382, 1300, 1244,
1
4
4
176, 1142, 1097, 1077, 1026, 996, 930, 889, 869, 796, 772, 676, 652, 621, 603, 552, 523, 487,
70, 450. HR-ESI-MS: m/z calculated for (M+Na) C H FN O NiNa: 415.0117; found
1
8
11
4
2
15.0074
Synthesis of H bpb-CF
2
3
2
-Picolinic acid (0.215 g, 1.75 mmol), HATU (0.665 g, 1.75 mmol), and DMF (5 ml) were
added to a dry 50 ml round bottom flask equipped with a stir bar and N,N-
diisopropylethylamine (0.692 ml, 3.97 mmol) was added. This solution was stirred at room
temperature for 30 min. 4-trifluoromethyl-o-phenylenediamine (0.140 g, 0.795 mmol) was then
added and the solution was stirred at room temperature for 24 hours. Next, water was added to
the solution and the crude product was extracted with DCM (2x20 ml). The organic fraction
was washed again with water (10 ml) and dried with Na SO . The crude material was then
2
4
purified on silica using 8 % MeOH in DCM to elute the desired product (0.169 g, 55 % yield).
1
H NMR (500 MHz, CDCl ) δ 10.45 (s, 1H), 10.27 (s, 1H), 8.57 (dd, J = 21.6, 4.1 Hz, 2H),
3
2

Journal of Materials Chemistry A
Page 18 of 23
View Article Online
8
1
1
1
.33 (t, J = 8.0 Hz, 2H), 8.17 (d, J = 10.2 Hz, 2H), 7.93 (q, J = 7.8 Hz, 2H), 7.57 (d, J = 8.4 Hz,
DOI: 10.1039/D0TA04511J
1
3
H), 7.54 – 7.42 (m, 2H). C NMR (126 MHz, CDCl ) δ 163.26, 163.05, 149.43, 149.32,
3
48.42, 148.40, 137.86, 137.82, 133.73, 129.78, 127.00, 126.94, 124.35, 123.36, 123.34,
22.85, 122.17, 122.14. Calculated for (M+Na): C H F N O Na, 409.0888; found 409.0878
1
9
13
3
4
2
Synthesis of [Ni(bpb-CF ], 4
3
Preparation of the complex was identical to the dimethylamine derivative. The proligand
(0.100 g, 0.259 mmol) was suspended in MeCN (10 ml) and trimethylamine was added
(0.072 ml, 0.518 mmol) yielding a solution. Nickel(II) tetrafluoroborate hexahydrate
(0.080 g, 0.235 mmol) was added to the solution, immediately forming an orange
precipitate. This mixture was stirred at room temperature for 5 hours and the precipitate
was collected on a fritted filter, washed with DCM (10 ml) and dried, yielding the
desired nickel complex (0.091 g, 88 % yield). IR (ATR): 1648 (b s, C=O), 1607, 1589,
1
8
562, 1526, 1497, 1472, 1437, 1382, 1314, 1259, 1238, 1141, 1094, 1056, 1026, 985,
86, 822, 802, 753, 680, 653, 642, 521, 451, 425. HR-ESI-MS: m/z calculated for
(
M+Na) C H F N O NiNa: 465.0085; found 465.0038.
19 11 3 4 2
Synthesis of H bpb-NO
2
2
2
-Picolinic acid (0.155 g, 1.26 mmol), HATU (0.480 g, 1.26 mmol), and DMF (5 ml) were
added to a dry 50 ml round bottom flask equipped with a stir bar and N,N-
diisopropylethylamine (0.500 ml, 2.87 mmol) was added. This solution was stirred at room
temperature for 30 min. 4-nitro-o-phenylenediamine (0.088 g, 0.575 mmol) was then added
and the solution was stirred at room temperature for 24 hours. Next, water was added to the
solution and the crude product was extracted with DCM (2x20 ml). The organic fraction was
washed again with water (10 ml) and dried with Na SO . The crude material was then purified
2
4
on silica using 10 % MeOH in DCM to elute the desired product (0.125 g, 60 % yield).
Analytical data was identical to previously reported data.[30]
2

Page 19 of 23
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/D0TA04511J
Synthesis of [Ni(bpb-NO ], 5
2
H bpb-NO (0.117g, 0.322 mmol) was placed in 8 ml of DMSO and triethylamine was added
2
2
(0.204 ml, 1.46 mmol) yielding a yellow mixture. This mixture was heated to 100 °C and the
mixture became a solution. Nickel(II) tetrafluoroborate hexahydrate (0.100 g, 0.293 mmol) was
then added to the solution and an orange precipitate immediately formed. This mixture was
stirred at 100 °C for an additional 3 hours. The precipitate was then collected on a fritted filter
and washed with room temperature DMSO and DCM. The solid was then collected and dried
yielding the desired product. IR (ATR): 3076, 1653 (b s, C=O), 1605, 1562, 1476, 1378, 1316,
1
135, 1060, 994, 964, 892, 822, 803, 752, 731, 706, 680, 653, 616, 508, 467. EA: Calculated:
C 51.47, %H 2.64, %N 16.67, Found: %C 51.36, %H 2.53, %N16.23
%
Computational Details
Geometry optimizations of complexes 1-5 were performed in a gas phase by considering all
possible spin multiplicities at particular charge state. Single-point energies in DMF and MeCN
were obtained using self-consistent reaction field approach based on the integral equation
formalism of the polarized continuum model (PCM).[31–33] All calculations were carried out
with Gaussian 16 software package (Version B.01).[34] Frequency calculations were performed
to ensure that the optimized structures represented minima on the potential energy surface for
each species. Hybrid PBE0 functional was utilized since it was previously shown to produce
the best correlation between calculated and experimental values of reduction potentials of Ni-
coordination complexes.[22] For some of the complexes considered in this study in the 0 charge
2
+
state, triplet spin (Ni ) states were found to be lower in energy by up to ~0.37 eV than the
2
+
singlet spin (Ni ) state configurations (Tables S26, S27). Overstabilization of the high spin
(HS) states as compared to the low spin (LS) configurations has been reported for hybrid
functionals.[35–38] Based on the diamagnetic H NMR spectra of the neutral charge state species
1
1
-5 as well as the recent studies utilizing PBE0 functional for similar linked picolinamide Ni
2

Journal of Materials Chemistry A
Page 20 of 23
complexes,[21] LS configurations are accepted as the ground electronic states and their energies
View Article Online
DOI: 10.1039/D0TA04511J
are taken for the calculations of redox properties. Dispersion corrections (D3)[39] were applied
to account for possible intramolecular noncovalent interactions that can be important for the
correct description of solvent-complex interaction. As reported previously for various Ni‐based
catalysts,[40] PBE0‐D3 was found to be the best functional in the complete benchmark set
relative to estimated CCSD(T)/CBS reference data, with a mean absolute deviation from the
reference values of 1.1ꢀkcal/mol. LANL2DZ effective core potential (ECP) basis set[41] was
employed for Ni, 6-311G(d)[42] basis set for C, N, O and 6-31G(d) for H atoms.[43]
The reduction potentials were computed using the revised Born-Haber cycle,[22] in
which all reduction steps are referenced to a calculated absolute half-cell potential of a
0
n
+
n–1
ferrocene couple, i.e., Fc + A → Fc + A , as opposed to an experimental standard reference
electrode, like the standard hydrogen electrode (SHE) or the saturated calomel electrode (SCE).
Mulliken spin density analysis was applied to determine the oxidation states of the Ni center
and nature of redox events (metal- or ligand-centered) at each charge transfer step. Mulliken
spin densities of the studied complexes were found to have similar values in both environments
(gas-phase or solvent) at each charge state.
Similar to the previous study on complex 1[21] reporting a coordination of two MeCN
molecules to the Ni center in the 2+/1+/0 charge states (Fig S3), two DMF molecules were
considered in the [Ni(bpb-R)]2+/1+/0 complexes in this study (Fig S4). In agreement with that,
the Ni-solvent contacts are significantly longer (R(Ni-O
)
=2.60-3.36 Å) in the 0 charge
DMF aver
state than in the 2+ charge state with direct Ni-O bonding (R(Ni-ODMF) aver=2.10-2.12 Å).
Migration of the two DMFs from the metal center in the 0 charge state results in the formation
of hydrogen bonds in lieu of the Ni-O bonding, i.e. ODMF-Hpyridine, HDMF-Oamide, and HDMF
-
Namide (Fig S4). Hence, to avoid overestimating the impact of the hydrogen bonding interactions
on the reduction potentials associated with the 1– and 2– charge states (Figs S5, S6), two DMF
2

Page 21 of 23
Journal of Materials Chemistry A
View Article Online
molecules were considered in the 2+ and 1+ charge states as well as in 0 charge state as an
DOI: 10.1039/D0TA04511J
intermediate one between the solvent-non-coordinating and the solvent-coordinating
complexes exhibiting the Ni-ODMF contacts.[21] For consistency, the same approach that was
used for the calculations of the reduction potentials in 1 in MeCN[21] was employed for
complexes 1-5 in DMF in this work.
Supporting Information
Supporting Information (solution state CV and DVP measurements of [Ni(bpb)]; DPV
experiments of [Ni(bpb)] drop cast onto a glassy carbon disk electrode; geometrical parameters,
Mulliken spin densities, spin states and Cartesian coordinates of the computed complexes by
charge state; free energies in MeCN and DMF of the LS and HS state configurations of the
0
0
[
Ni(bpb-R)(MeCN) ] and [Ni(bpb-R)(DMF) ] complexes; free energies in MeCN and DMF
2 2
1
+
1+
of the [Ni(bpb-R)(MeCN)2] and [Ni(bpb-R)(DMF)2] complexes) is available from the
Wiley Online Library or from the author.
Acknowledgements
This work was supported by the Laboratory Directed Research and Development (LDRD)
program of Los Alamos National Laboratory under project number 20170046DR. I.A.P.
acknowledges the support from J. Robert Oppenheimer Distinguished Postdoctoral Fellowship.
C.R.F. would like to thank the Department of Energy’s Summer Undergraduate Laboratory
Internship program for funding. High resolution mass spectral data was acquired at the Center
for Integrated Nanotechnologies, an Office of Science User Facility operated for the U.S.
Department of Energy (DOE) Office of Science (agreement 2018BU0048). Los Alamos
National Laboratory is operated by Triad, LLC, for the National Nuclear Security
Administration of U.S. Department of Energy (Contract No. 89233218NCA000001).
Received: ((will be filled in by the editorial staff))
Revised: ((will be filled in by the editorial staff))
Published online: ((will be filled in by the editorial staff))
2

Journal of Materials Chemistry A
Page 22 of 23
View Article Online
DOI: 10.1039/D0TA04511J
References
[1]
[2]
[3]
G. L. Soloveichik, Chem. Rev. 2015, 115, 11533.
R. W. Hogue, K. E. Toghill, Curr. Opin. Electrochem. 2019, 18, 37.
R. Chen, D. Bresser, M. Saraf, P. Gerlach, A. Balducci, S. Kunz, D. Schröder, S.
Passerini, J. Chen, ChemSusChem 2020, 1.
[4]
[5]
[6]
Y. Ding, C. Zhang, L. Zhang, Y. Zhou, G. Yu, Chem. Soc. Rev. 2018, 47, 69.
Y. Ding, C. Zhang, L. Zhang, Y. Zhou, G. Yu, Chem 2019, 5, 1964.
C. S. Sevov, S. L. Fisher, L. T. Thompson, M. S. Sanford, J. Am. Chem. Soc. 2016,
1
38, 15378.
[
[
[
[
7]
8]
9]
L. E. Vangelder, A. M. Kosswattaarachchi, P. L. Forrestel, T. R. Cook, E. M. Matson,
Chem. Sci. 2018, 9, 1692.
M. A. Goulet, L. Tong, D. A. Pollack, D. P. Tabor, S. A. Odom, A. Aspuru-Guzik, E.
E. Kwan, R. G. Gordon, M. J. Aziz, J. Am. Chem. Soc. 2019, 8.
P. Huang, W. Ling, H. Sheng, Y. Zhou, X. Wu, X. X. Zeng, X. Wu, Y. G. Guo, J.
Mater. Chem. A 2017, 6, 41.
10] C. Zhang, Y. Qian, Y. Ding, L. Zhang, X. Guo, Y. Zhao, G. Yu, Angew. Chem. Int.
Ed. 2019, 58, 7045.
[
[
11] Y. Ding, C. Zhang, L. Zhang, H. Wei, Y. Li, G. Yu, ACS Energy Lett. 2018, 3, 2641.
12] G.-M. Weng, B. Yang, C.-Y. Liu, G.-Y. Du, E. Y. Li, Y.-C. Lu, Energy Environ. Sci.
2
019, 12, 2244.
[13] Y. Ding, Y. Li, G. Yu, Chem 2016, 1, 790.
[
14] M. R. Gerhardt, L. Tong, R. Gómez-bombarelli, Q. Chen, M. P. Marshak, C. J. Galvin,
A. Aspuru-guzik, R. G. Gordon, M. J. Aziz, Adv. Energy Mater. 2017, 1601488.
15] X. Wei, L. Cosimbescu, W. Xu, J. Z. Hu, M. Vijayakumar, J. Feng, M. Y. Hu, X.
Deng, J. Xiao, J. Liu, V. Sprenkle, W. Wang, Adv. Energy Mater. 2015, 5, 1.
16] Y. Yan, S. G. Robinson, M. S. Sigman, M. S. Sanford, J. Am. Chem. Soc. 2019, 141,
[
[
[
[
1
5301.
17] Y. Ding, Y. Zhao, Y. Li, J. B. Goodenough, G. Yu, Energy Environ. Sci. 2017, 10,
91.
4
18] B. E. Schurr, O. Nachtigall, L. E. Vangelder, J. Drappeau, W. W. Brennessel, E. M.
Matson, B. E. Schurr, O. Nachtigall, L. E. Vangelder, J. Drappeau, J. Coord. Chem.
2
019, 72, 1267.
[
[
19] L. E. VanGelder, E. M. Matson, J. Mater. Chem. A 2018, 6, 13874.
20] R. M. Darling, K. G. Gallagher, J. A. Kowalski, S. Ha, F. R. Brushett, Energy Environ.
Sci. 2014, 7, 3459.
[
[
[
[
[
[
[
21] T. Chu, I. A. Popov, G. A. Andrade, S. Maurya, P. Yang, E. R. Batista, B. L. Scott, R.
Mukundan, B. L. Davis, ChemSusChem 2019, 12, 1304.
22] L. E. Roy, E. Jakubikova, M. G. Guthrie, E. R. Batista, J. Phys. Chem. A 2009, 113,
6
745.
23] D. J. Barnes, R. L. Chapman, R. S. Vagg, E. C. Watton, J. Chem. Eng. Data 1978, 23,
49.
3
24] I. A. Popov, N. Mehio, T. Chu, B. L. Davis, R. Mukundan, P. Yang, E. R. Batista, ACS
Omega 2018, 3, 14766.
25] I. A. Popov, B. L. Davis, R. Mukundan, E. R. Batista, P. Yang, Front. Phys. 2019, 6,
1
.
26] S. K. Dutta, U. Beckmann, E. Bill, T. Weyhermuller, K. Wieghardt, Inorg. Chem.
000, 39, 3355.
2
27] J. A. Suttil, J. F. Kucharyson, I. L. Escalante-Garcia, P. J. Cabrera, B. R. James, R. F.
Savinell, M. S. Sanford, L. T. Thompson, J. Mater. Chem. A 2015, 3, 7929.
2

Page 23 of 23
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/D0TA04511J
[28] A. Choi, H. Jeong, S. Kim, S. Jo, S. Jeon, Electrochim. Acta 2008, 53, 2579.
[29] R. McGuire, D. K. Dogutan, T. S. Teets, J. Suntivich, Y. Shao-Horn, D. G. Nocera,
Chem. Sci. 2010, 1, 411.
[30] A. F. Abdel-Magied, W. A. A. Arafa, T. M. Laine, A. Shatskiy, M. D. Kärkäs, B.
Åkermark, E. V. Johnston, ChemCatChem 2017, 9, 1583.
[
[
[
[
31] E. Cancès, J. Chem. Phys. 1997, 107, 3032.
32] B. Mennucci, E. Cancès, J. Tomasi, J. Phys. Chem. 1997, 101, 10506.
33] E. Cancès, B. Mennucci, J. Math. Chem. 1998, 23, 309.
34] M. J. Frisch, E. Al, Gaussian 16 (Revision B.01), Gaussian, Inc., Wallingford CT,
2
016.
[35] J. N. Harvey, Principles and Applications of Density Functional Theory in Inorganic
Chemistry I, Springer, Berlin, 2004.
[
[
[
36] M. Swart, Int. J. Quantum Chem 2013, 113, 2.
37] M. Reiher, O. Salomon, B. A. Hess, Theor. Chem. Acc. 2001, 107, 48.
38] B. Pinter, A. Chankisjijev, P. Geerlings, J. N. Harvey, F. D. Proft, Chem. Eur. J. 2018,
2
4, 5281.
[
[
[
[
[
39] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132, 154104.
40] M. Steinmetz, S. Grimme, ChemistryOpen 2013, 2, 115.
41] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270.
42] A. D. McLean, G. S. Chandler, J. Chem. Phys. 1980, 72, 5639.
43] P. C. Hariharan, J. A. Pople, Theor. Chim. Acta 1973, 28, 213.
2