Zn-Templated synthesis of substituted (2,6-diimine)pyridine proligands and evaluation of their iron complexes as anolytes for flow battery applications

Article abstract of DOI:10.1039/d0dt00543f

Pseudo-octahedral iron complexes supported by tridentate N^N^N-binding, redox 'non-innocent' diiminepyridine (DIP) ligands exhibit multiple reversible ligand-based reductions that suggest the potential application of these complexes as anolytes in redox flow batteries (RFBs). When bearing aryl groups at the imine nitrogens, substitution at the 4-position can be used to tune these redox potentials and impact other properties relevant to RFB applications, such as solubility and stability over extended cycling. DIP ligands bearing electron-withdrawing groups (EWGs) in this position, however, can be challenging to isolate via typical condensation routes involving para-substituted anilines and 2,6-diacetylpyridine. In this work, we demonstrate a high-yielding Zn-templated synthesis of DIP ligands bearing strong EWGs. The synthesis and electrochemical characterization of iron(ii) complexes of these ligands is also described, along with properties relevant to their potential application as RFB anolytes.

Full text of DOI:10.1039/d0dt00543f

Dalton
Transactions
View Article Online
View Journal
PAPER
Zn-Templated synthesis of substituted
(2,6-diimine)pyridine proligands and evaluation
of their iron complexes as anolytes for flow
battery applications†
Cite this: DOI: 10.1039/d0dt00543f
Jason D. Braun,
David E. Herbert
Paul A. Gray, ‡ Baldeep K. Sidhu, ‡ Dion B. Nemez
*
and
Pseudo-octahedral iron complexes supported by tridentate N^N^N-binding, redox ‘non-innocent’ diimi-
nepyridine (DIP) ligands exhibit multiple reversible ligand-based reductions that suggest the potential
application of these complexes as anolytes in redox flow batteries (RFBs). When bearing aryl groups at the
imine nitrogens, substitution at the 4-position can be used to tune these redox potentials and impact
other properties relevant to RFB applications, such as solubility and stability over extended cycling. DIP
ligands bearing electron-withdrawing groups (EWGs) in this position, however, can be challenging to
isolate via typical condensation routes involving para-substituted anilines and 2,6-diacetylpyridine. In this
work, we demonstrate a high-yielding Zn-templated synthesis of DIP ligands bearing strong EWGs. The
synthesis and electrochemical characterization of iron(II) complexes of these ligands is also described,
along with properties relevant to their potential application as RFB anolytes.
Received 13th February 2020,
Accepted 9th April 2020
DOI: 10.1039/d0dt00543f
rsc.li/dalton
their cost-competitiveness with aqueous RFBs has been ques-
tioned.6 The increased solubility of many metal coordination
Introduction
Redox flow batteries (RFBs) have received renewed attention of complexes (MCCs) in organic solvents, however, does mean a
late owing to their potential for the scalable, inexpensive larger library of candidate anolytes and catholytes based on
storage of the growing proportion of power generation devoted MCCs is available (Fig. 1).7
to renewable but intermittent resources such as wind and
While simple coordination complexes (e.g., A; acac = acetyl-
solar.1 In an RFB, electrical energy is converted to chemical acetonate) exhibit reversible reductions and oxidations that
energy through the electrochemical interconversion of redox can enable use as both anolyte and catholyte in symmetric
pairs serving as the electrolyte.2 The key contrast with conven- RFBs,8 the introduction of redox ‘non-innocent’ ligands9,10 can
tional batteries is that these redox pairs can be spatially separ- in principle augment the performance of MCCs in RFBs11 by
ated from the electrode.3 If both oxidized and reduced providing additional sites for electron-transfer. MCCs of redox
members of the pairs are stable and soluble in the flow battery non-innocent 2,2′-bipyridine (bpy; B)12 and (bipyridylimino)
medium, scalability will depend in part on the abundance of isoindoline (BPI; C)13 ligands, for example, have been shown
the materials employed as catholyte/anolyte, and to a more sig-
nificant extent than on that of the electrode material.4 With
respect to the solvent medium, nonaqueous solvents with
wider windows of electrochemical stability can boost the
energy density output of an RFB compared to water,5 though
Department of Chemistry and the Manitoba Institute for Materials, University of
Manitoba, 144 Dysart Road, Winnipeg, Manitoba, R3T 2N2, Canada.
E-mail: david.herbert@umanitoba.ca
Fig. 1 Selected examples of metal coordination complexes (MCCs) that
†Electronic supplementary information (ESI) available: A PDF containing all
spectra, ESI tables of bond distances and angles, and additional electrochemical
plots. CIFs for 1b–d and 3c–d. CCDC 1983239–1983243. For ESI and crystallo-
graphic data in CIF or other electronic format see DOI: 10.1039/d0dt00543f
‡These authors contributed equally to this work.
have been evaluated for use in non-aqueous redox flow batteries.6
n+
A: M(acac)3, M = V, Cr, Mn, Fe, Ru; acac = acetylacetonate. B: M(bpy)3
,
M = Ru, Fe, Ni, Cr, Co; bpy = 2,2’-bipyridine. C: M(BPI)2, M = Mg, Mn, Fe,
Co, Ni, Zn; BPI = (bipyridylimino)isoindoline. D: M(DIP)2n+, M = Fe; DIP =
2,6-diiminepyridine.
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.

View Article Online
Paper
Dalton Transactions
to have properties favourable to RFB applications including cations where these ligands are utilized.24 We furthermore
(for BPI MCCs) multiple electron transfers per molecule, high report on the electrochemical evaluation of properties of their
solubility, and long-term stability towards charge/discharge pseudo-octahedral Fe(II) complexes which speak to potential
cycles (∼200) with very little capacity fade.13 For comparison, use as anolytes in RFB applications.
one of the more commercially promising systems is the all-
vanadium flow battery, which utilizes four different oxidation
state vanadium and oxyvanadium ions and shows very
Results and discussion
Synthesis and characterization of ligands and complexes
impressive long-term cyclability of over 1000 stable cycles.14
The high system costs comprised to a large extent of the cost
DIP proligand synthesis is generally accomplished by the acid-
catalyzed condensation of 2,6-diacetylpyridine with two equiva-
lents of the appropriately substituted aniline. To avoid forcing
of the redox-active material, however, still far exceed the U.S.
Department of Energy’s target.15
In this context, we have been interested in the application
conditions required for appreciable conversion using anilines
of iron coordination complexes of a popular class of redox
non-innocent scaffold, 2,6-diiminepyridines (DIP),16 which are
substituted in the 4-position with electron-withdrawing groups
(EWGs), we turned to templated synthesis. Templated ligand
able to accommodate up to three additional electrons in s-,17
f-18 and d-block19 metal complexes of triply reduced DIPs.
syntheses have long been used to overcome competing side
reactions,25 but can also be used to drive to completion ligand
Pseudo-octahedral iron complexes of N-aryl DIPs bearing elec-
formation reactions that would suffer from less favourable
tron-releasing substituents (D, R = tBu or OMe, M = Fe, n = 2),
for example, exhibit good solubility (0.1–0.3 M) in CH3CN and
thermodynamics in the absence of coordination to a templat-
ing metal ion.26 The templated synthesis of imine-based
two reversible reductions at negative potentials beyond the
water voltage window.16 Such molecular geometries also make
ligands, for example, has enabled construction of complex
scaffolds including chiral P^N^N and P^N^N^P multidentate
use of tridentate DIP chelation which limits ligand dis-
architectures.27 In comparison, reports of demetallation and
sociation and promotes higher cyclability compared to biden-
subsequent use of the liberated proligands appear far less
tate ligand environments; 2,6-unsubstituted aryl groups simi-
often28 than the targeted assembly of coordination complexes
larly reduce ligand hemilability associated with steric conges-
tion.20 We have also found that, in addition to the position of
templated by a metal ion of choice. For imines, attempts to
demetallate can lead to ligand hydrolysis. In the case of ArBIAN
the reduction potentials, the solubility and stability towards
ligands constructed around Zn2+ as a templating ion, using
cycling were also impacted by the identity of the substituent in
the 4-position of N-phenyl rings. Complexes of DIP ligands
bearing tBu substituents showed improved capacity retention
oxalate (C2O42−) as a displacing ligand produces insoluble
ZnC2O4 and drives demetallation reactions to completion
without evidence of hydrolysis.23
attributed to enhanced solubility of both reduced and oxidized
species.16 As tBu and OMe are both electron-releasing, we
Four (DIP)ZnCl2 complexes (1a–d, Fig. 2) were prepared via
the Zn2+-templated condensation of two equivalents of aniline
thought to explore the impact electron-withdrawing groups
with 2,6-diacetylpyridine, broadly following the protocol for
(EWGs) might have on reduction potentials and cycling stabi-
incorporating strong EWGs in ArBIANs outlined by Ragaini.23
lity in the context of RFB anolytes.
Complex 1a (R = H) has been previously reported.29 Each of
1a–d was prepared in this way by heating slightly more than
Installation of para-EWGs on the flanking aryl substituents
in DIP frameworks, however, is potentially more problematic
than electron-donating groups (EDGs). DIPs are typically pre-
2,6-diacetylpyridine and excess ZnCl2 in CH3OH (acetic acid
pared by condensation of 2,6-diacetylpyridine and the corres-
two equivalents of the appropriate aniline (15% excess) with
for 1d). Unlike complexes based on the acenaphthoquinone
ponding anilines. Substitution of anilines in the 4-position
with EWGs can significantly reduce their nucleophilicity, ham-
pering conversion.21 For example, condensation syntheses of
(para-fluoro)phenyl and (para-bromo)phenyl-substituted DIPs
are reported to give isolated yields of only 24% and 41%,
respectively, despite azeotropic removal of water.21
Furthermore, the para-nitro analogue could only be isolated,
and in similarly low yields, following a five day, acid-catalyzed
Dean–Stark reflux in high-boiling p-xylene.22 In seeking to
overcome these challenges, we considered a report on the
inclusion of EWGs on ArBIAN-type diimine ligands via a Zn-
templated synthesis (ArBIAN = bis (aryl)acenaphthenequinone-
diimine).23 Here, we describe the application of this approach
to the synthesis of DIP ligands with strong EWGs in relatively
high yields under mild conditions. This methodology opens
Fig. 2 Synthesis of (a) (DIP)ZnCl2 complexes 1a–d; (b) free proligands
2a–d via de-zincation, and [bis(DIP)2Fe][PF6]2 complexes 3a–d with iso-
the chemical space for the synthesis of DIP ligands that incor-
lated yields in parentheses. Complexes 3e (ref. 32) and 3f (ref. 16) have
porate strong EWGs and so may be useful for the many appli- been previously reported.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Paper
(ArBIAN) backbone,23 using acetic acid as the solvent leads to
full solubilization of the (DIP)ZnCl2 fluoro- and bromo-deriva-
tives, but partial solubilization of the cyano-derivative.
Changing the reaction solvent to methanol, the crude (DIP)
ZnCl2 product precipitates as a yellow powder upon cooling to
room temperature and can be filtered to easily separate 1a–c
from excess reactants. The (DIP)ZnCl2 complexes can then be
recrystallized by dissolving in hot acetonitrile, followed by slow
cooling to −20 °C. 1H NMR spectra for 1a–d show a single,
pseudo C2v-symmetric magnetic environment in solution for
all four Zn complexes. Diagnostic resonances for the formation
of the imine arms can also be observed by 13C{1H} NMR at δ =
∼165 ppm. The molecular formula for each compound was
confirmed by high-resolution mass spectrometry (HR-MS).
To confirm the structures of the (DIP)ZnCl2 complexes
suggested by solution NMR and HR-MS, solid-state structures
of 1b–d were also determined by single-crystal X-ray diffraction
(XRD; Fig. 3). In each case, a single DIP ligand is bound to Zn
in a meridional, tridentate fashion and forms part of what is
best described as a distorted square-based pyramid with τ5
values30 ranging from 0.30 (1c) to 0.40 (1d). This deviation
from ideal geometry arises from different Npyr–Zn–Cl angles.
The two chlorides cant asymmetrically away from the pyridine
ring, opening one Npyr–Zn–Cl angle wider than the other.
The relatively long Zn–Npyr and Zn–Nimine distances, typical
of (N-heterocycle/imine)-Zn2+ coordination,31 are consistent
across the series 1b–d at ∼2.08 Å and 2.21–2.26 Å, respectively.
As a result, the Nimine–Zn–Nimine angles are quite pulled back
(147–148°). The narrow range of bond distances and angles
observed for the set of (DIP)ZnCl2 complexes is consistent with
the distal placement of the R groups on the ligand periphery.
The free proligands (2a–d) can be displaced from the ZnCl2
unit and isolated in good to excellent yields (67–98%) by
mixing dichloromethane solutions of 1a–d with aqueous solu-
tions containing three equivalents of potassium oxalate, then
extracting and drying the organic layer. Multinuclear NMR
confirmed the molecular composition of the resulting off-
white/yellow solids as demetallated ligand, with shifts
observed to all signals including those attributed to the diag-
nostic imine carbon nuclei (δCvN = ∼168 ppm). Iron MCCs 3a–
d were subsequently prepared by reaction of the desired proli-
Fig. 4 Solid-state X-ray diffraction structures of 3c and 3d, shown with
ellipsoids at the 50% probability level. Hydrogen atoms, counterions and
solvent molecules are omitted for clarity. Selected bond lengths and
angles reported in Tables S1 and S2.†
gand with 0.5 equivalents of anhydrous FeCl2 and two equiva-
lents of NaPF6 in CH3OH. These complexes were isolated as
air- and moisture-stable, deep purple solids, with molecular
formulae again confirmed by HR-MS. Multinuclear NMR
suggests a symmetric environment around each Fe(II) centre
with 13C resonances for the diagnostic imine carbon centers
shifted considerably downfield to δCvN ∼180 ppm. The most
downfield resonance is observed for 3d (δCvN = 182 ppm), con-
sistent with 3d having the most deshielded imine carbon. The
dark purple colour of 3a–d is reflected in strong and broad low
energy features in the steady-state UV-Vis absorption spectra
(Fig. S37†). These transitions are assigned as MLCT in charac-
ter, as is typical of [(DIP)2Fe]2+ complexes.16,32 Within the
series, 3d showed a marked hypsochromic shift in the tran-
sitions observed beyond 400 nm, but otherwise, a similar
absorption profile is observed for all four complexes. The
pseudo-octahedral coordination environment around Fe was
confirmed by XRD studies of 3c and 3d (Fig. 4). A much tighter
coordination environment is seen for the Fe complexes com-
pared with the Zn congeners (Tables S1 and S2†), with closer
M–N distances ranging from 1.86–1.87 Å for Fe–Npyr and
1.96–2.00 Å for Fe–Nimine. As a result, the intraligand Nimine
Fe–Nimine angles are much larger at ∼160°.
–
Cyclic voltammetry (CV) and differential pulse voltammetry
(DPV; Fig. 5 and Table 1) revealed identical redox behaviour
for the series 3a–d, with two reversible 1e− reductions evident
between −1.0 and −2.0 V and a 1e− oxidation observed near
+1.0 V vs. FcH0/+ (FcH = ferrocene). Based on previous analysis
of the redox behaviour of DIP complexes of Fe, we assign the
cathodic events as ligand centered.32 The oxidation is attribu-
ted to a metal-centered Fe2+/3+ couple. Each of the three redox
events observed for 3a–d are shifted anodically compared to
those observed for [(DIP)Fe]2+ analogs bearing electron-releas-
ing OMe (3e) or tBu (3f) substituents.16 This highlights the
ability to tune redox potentials using distal substitution of the
N-arene rings. In accordance with their Hammett para-
meters,33 the largest anodic shift is observed for R = CN, fol-
lowed by R = Br and R = F, consistent with inductive removal
of electron density from the imine-based LUMO. The reversi-
bility of the ligand-based reductions observed by CV does not
seem to be as adversely affected by the introduction of EWGs
Fig. 3 Solid-state X-ray diffraction structures of 1b, 1c and 1d shown
with ellipsoids at the 50% probability level. Hydrogen atoms and solvent
molecules are omitted for clarity. Selected bond lengths and angles
reported in Tables S1 and S2.†
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 6 (a and b) Total cell voltage, (c) coulombic efficiency (% CE), and
(d) capacity retention for 3a. Anodic and cathodic current set to 7 mA
with a charging rate of 1 C. Voltage limits set according to previously
obtained CVs in order to limit accessing irreversible redox events; in 0.3
M nBu4PF6 acetonitrile solution. Plateaus are marked by arrows.
Fig. 5 CV/DPVs of 3a–d (0.6 mM of analyte, 0.1 M nBu4PF6 in CH3CN,
GCE, scan rate = 100 mV s−1).
Table 1 Electrochemical and RFB parameters for 3a–fa
E1/2/V
Δptp/mV
ired/iox
CE/%
95.2
FE/%
93.4
e−/mol
1.9
ensure capture of the second reversible redox event without
going so far as to irreversibly reduce the complexes.16 The
charge/discharge cycling experiments for 3a are shown in
Fig. 6. While less pronounced than those observed for 3e–f,16
plateaus corresponding to the reduction events observed by
CV/DPV can be discerned in the charging segment. In the dis-
charge segment, two plateaus corresponding to the reverse oxi-
dation events can similarly be noted (Fig. 6b). These decrease
in prominence after extended cycling.
The cycling stability of RFB electrolytes is one of a number
of parameters important for evaluating the potential for use in
a commercial battery.36 Through 25 cycles, 3a exhibits a cou-
lombic efficiency (% CE) of 95.2% (Fig. 6c), comparable to
reported values for aqueous Li/I RFBs over 20 cycles,37 but
passing an average of ∼1.9 electrons per molecule out of a
theoretical maximum of 2 (Fig. 6d and Table 1). Zn–I flow bat-
teries, in comparison have been reported to exhibit extended
stability over 300 cycles, with a one-electron redox couple.38 In
comparison to complexes bearing donating groups (3e–f),16 3a
surpasses the faradaic efficiency (% FE) and number of elec-
trons per molecule observed for 3e–f without any evidence of
degradation. For 3b–d, over the course of extended cycling, the
cycles become narrower resulting in reduced overall efficiency
parameters (Table 1 and Fig. S42–S44†). We note the potentials
for reduction of organic aryl halides and nitriles39 fall close to
those observed for the reductions of 3a–d. Over time, irrevers-
ible chemical reduction of these groups may contribute an
underlying degradation pathway for the ligands that include
these substituents. Indeed, slightly less Nernstian peak para-
meters are observed by CV for 3c–d (Table 1).
3a
−1.55
−1.27
0.97
−1.49
−1.20
1.03
−1.43
−1.15
1.05
−1.26
−1.01
1.16
−1.60
−1.30
0.86
−1.59
−1.32
0.90
67
65
82
69
66
158
65
59
103
68
62
138
75
61
77
60
60
69
0.88
1.14
0.86
0.90
1.16
0.23
0.86
1.23
0.68
0.87
1.31
0.68
0.97
0.99
0.97
1.04
0.95
0.97
3b
3c
84.8
95.2
61.2
24.6
42.6
70.0
82.1
1.2
0.5
0.9
1.4
1.6
3d
3eb
3fb
85.0
94.3
>99.9
a Average coulombic efficiency (CE), average faradaic efficiency (FE)
taken by averaging charging and discharging FE, and average number
of electrons cycled taken by averaging number of electrons cycled upon
charging and discharging. b Values taken from ref. 16.
as the Fe2+/3+ couples, which are much more reversible for 3e–f
likely as a result of occurring at less positive potentials.16 In
particular, both cathodic events show peak current ratios near
to unity and narrow peak-to-peak separations close to the
Nernstian limit of 59 mV.34 Moreover, the potentials surpass
the voltage limits of water (ca. −1.2 V vs. FcH0/+ 35
and the mul-
tiple electron transfers possible with a single complex suggest
the possibility of high energy storage capacity.13 We therefore
proceeded to evaluate the suitability of 3a–d as RFB anolytes.
)
Charge/discharge measurements
Cycling measurements were performed on 3a–d using a reticu-
lated vitreous carbon (RVC) working electrode in a bulk elec-
trolysis cell to examine the viability of these compounds as
Conclusions
RFB anolytes. The maximum cathodic potentials were set This work presents a facile synthetic route for the incorpor-
according to the CV/DPV collected for each compound to ation of strong EWGs into the para position of the flanking
Dalton Trans.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Paper
aryl substituents on DIP ligands. This Zn-templated method- 6H; NvCMe). 13C{1H} NMR (CD3CN, 125 MHz, 25 °C): δ
ology offers higher yields, shortened reaction times and 164.9 (MeCvN), 149.7 (CAr), 147.6 (CAr), 144.9 (CAr), 129.6
requires considerably milder conditions than previously (CAr), 128.1 (CAr), 126.9 (CAr), 122.8 (CAr), 17.4 ppm (CH3). MS
reported for analogous DIPs,21,22 expanding the chemical (ESI-TOF/MS, m/z) calcd for C21H19ClN3Zn [M]+, 412.0553;
scope available in the construction of these increasingly widely found 412.0550.
used ligands.40 Isolation of DIP ligands bearing strong EWGs
has also enabled the high-yielding synthesis of new 2,6-Bis(4-fluorophenylimino)pyridine zinc dichloride (1b)
[(DIP)2Fe]2+ salts 3b–d, along with novel, structurally character-
ized examples of (DIP)ZnCl2 complexes 1a–d. Regarding the
Procedure as for 1a using: ZnCl2 (0.250 g, 1.83 mmol), 2,6-dia-
cetylpyridine (0.103 g, 0.631 mmol), methanol (2.5 mL), and
4-fluoroaniline (0.164 g, 1.48 mmol; 0.14 mL). Light-yellow
potential application of 3a–d in RFBs, 3a is shown to be a
viable candidate as an RFB anolyte, displaying stability over at
powder. Isolated yield: 0.230 g (75%). 1H NMR (CD3CN,
least 25 cycles while maintaining access to multiple electrons
500 MHz, 25 °C): δ 8.54 (m, 1H; PyrHp), 8.41 (m, 2H; PyrHm),
equivalents per molecule (∼1.9) over this range, surpassing
7.25 (m, 4H; ArH), 7.19 (m, 4H; ArH), 2.47 ppm (s, 6H;
previously reported examples 3e–f.16 The use of halides or
NvCMe). 13C{1H} NMR (CD3CN, 125 MHz, 25 °C, ppm): δ
pseudohalides at the ligand periphery, however, was found to
165.6 (MeCvN), 161.8 (d, JCF = 241.3 Hz; CAr), 149.7 (CAr),
144.9 (CAr), 143.7 (d, JCF = 3.75 Hz; CAr), 128.2 (CAr), 124.8 (d,
be detrimental to cycling performance. This observation
should help guide future ligand designs for MCCs for RFB
applications.
JCF = 7.5 Hz; CAr), 116.3 (d, JCF = 23.8 Hz; CAr), 17.5 ppm (CH3).
19F{1H} NMR (CD3CN, 282 MHz, 25 °C): δ −118.9 ppm (s). MS
(ESI-TOF/MS, m/z) calcd for C21H17ClN3ZnF2 [M]+, 448.0365;
found 448.0367.
Experimental section
2,6-Bis(4-bromophenylimino)pyridine zinc dichloride (1c)
Unless otherwise specified, all air sensitive manipulations
Procedure as for 1a using: ZnCl2 (1.00 g, 7.32 mmol), 2,6-dia-
cetylpyridine (0.398 g, 2.44 mmol) methanol (8 mL) and
4-bromoaniline (0.880 g, 5.12 mmol). Light brown powder.
Isolated yield: 1.10 g (74%). 1H NMR (CD3CN, 300 MHz,
25 °C): δ 8.54 (m, 1H; PyrHp), 8.42 (m, 2H; PyrHm), 7.60 (m,
4H; ArH), 7.15 (m, 4H; ArH), 2.47 ppm (s, 6H; NvCMe). 13C
{1H} NMR (CD3CN, 125 MHz, 25 °C): δ 165.6 (MeCvN),
149.3 (CAr), 146.4 (CAr), 144.8 (CAr), 132.5 (CAr), 128.1 (CAr),
124.7 (CAr), 119.7 (CAr), 17.4 ppm (CH3). MS (ESI-TOF/MS,
were carried out either in a N2 filled glove box or using stan-
dard Schlenk techniques under Ar. ZnCl2 (Alfa Aesar), 4-fluor-
oaniline (Combi Blocks), 4-bromoaniline (Acros Organics),
aniline (Sigma Aldrich), 4-aminobenzonitrile (Combi Blocks),
2,6-diacetylpyridine (Combi Blocks), K2[C2O4] (Alfa Aesar),
FeCl2 (Acros Organics), NaPF6 (Alfa Aesar), acetic acid and
methanol (Fisher Scientific) were purchased and used without
any further purification. Organic solvents used for electro-
chemical tests were dried and distilled using appropriate
drying agents prior to use. 1- and 2D NMR spectra were
recorded on Bruker Avance 300 MHz or Bruker Avance – III
500 MHz spectrometers. 1H and 13C{1H} NMR spectra were
referenced to residual solvent peaks. Mass spectrometry
(ESI-TOF/MS), was performed at the University of Manitoba on
a Bruker Compact LC-ESI-TOF/MS analyzer. Electronic absorp-
tion spectra were recorded on an Agilent Technologies Cary
5000 Series UV-Vis-NIR spectrophotometer in dual beam
mode.
m/z) calcd for
567.9041.
C
21H17ClN3ZnBr2 [M]+, 567.8764; found
2,6-Bis(4-cyanophenylimino)pyridine zinc dichloride (1d)
In a thick-walled flask, ZnCl2 (1.00 g, 7.34 mmol) and 2,6-
diacetylpyridine (0.440 g, 2.70 mmol) were combined with
acetic acid (8 mL) giving a colourless precipitate. This
mixture was heated to 60 °C and 4-aminobenzonitrile
(0.744 g, 6.3 mmol) added to the hot solution. The flask
was then sealed with a Teflon stopper and heated to 130 °C
for 3 h behind a blast shield. A yellow precipitate was
2,6-Bis(phenylimino)pyridine zinc dichloride (1a)
In a thick-walled flask, ZnCl2 (0.250 g, 1.83 mmol) and 2,6- formed upon cooling to room temperature and collected by
diacetylpyridine (0.102 g, 0.627 mmol) were combined with filtration. The solid was suspended in diethyl ether and
methanol (2.5 mL) giving a colourless precipitate. This stirred for 10 min. This suspension was filtered and the
mixture was heated to 60 °C, and aniline (0.134 g, collected solid washed with an additional 3 × 15 mL of
1.43 mmol; 0.13 mL) added to the hot solution. The flask diethyl ether, then dried in vacuo. Yellow solid. Isolated
was then sealed with a Teflon stopper and heated to 90 °C yield: 0.836 g (62%). 1H NMR (CD3CN, 300 MHz, 25 °C): δ
for 4 h behind a blast shield. A yellow precipitate was formed 8.60 (m, 1H; PyrHp), 8.48 (m, 2H; PyrHm), 7.81 (m, 4H;
upon cooling to room temperature and collected by fil- ArH), 7.32 (m, 4H; ArH), 2.46 ppm (s, 6H; NvCMe). 13C{1H}
tration. The solid product was recrystallized from acetonitrile NMR (CD3CN, 75 MHz, 25 °C): δ 167.4 (MeCvN), 151.0
and washed with chloroform to give a light-yellow powder. (CAr), 149.0 (CAr), 145.5 (CAr), 134.1 (CAr), 128.8 (CAr), 123.6
Isolated yield: 0.220 g (78%). 1H NMR (CD3CN, 500 MHz, (CAr), 119.4 (CuN), 110.4 (CAr) 17.9 ppm (CH3). MS
25 °C): δ 8.53 (m, 1H; PyrHp), 8.41 (m, 2H; PyrHm), 7.45 (m, (ESI-TOF/MS, m/z) calcd for C23H17ClN5Zn [M]+, 462.0458;
4H; ArH), 7.29 (m, 2H; ArH), 7.21 (m, 4H; ArH), 2.46 ppm (s, found 462.0447.
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.

View Article Online
Paper
Dalton Transactions
General procedure for the isolation of decoordinated
proligands
ArH), 2.56 ppm (s, 6H; NvCMe). 13C{1H} NMR (CD3CN,
75 MHz, 25 °C): δ 179.6 (MeCvN), 160.2 (CAr), 144.3 (CAr),
136.8 (CAr), 130.8 (CAr), 128.7 (CAr), 127.9 (CAr), 120.2 (CAr),
19.6 ppm (CH3). 19F NMR (CD3CN, 282 MHz, 25 °C): δ
−72.9 ppm (d, JPF = 705.4 Hz; PF6). 31P{1H} NMR (CD3CN,
121 MHz, 25 °C): δ −144.7 (q, JFP = 706.7 Hz; PF6). Anal. calcd
for C42H38N6F12P2Fe: C, 51.87; H, 3.94. Found: C, 52.26; H,
4.02. MS (ESI-TOF/MS, m/z) calcd for C42H38N6Fe [M + H]+,
681.2424; found 681.2439.
Zn complexes 1a–d (1 mmol) were each suspended in CH2Cl2
(50 mL) in a separatory funnel. An aqueous solution (20 mL) of
potassium oxalate (3 mmol) was then added and the mixture
was shaken for 5 min, giving a cloudy aqueous layer over a
yellow organic layer. The organic layer was washed with an
additional 2 × 30 mL of water, stirred over Na2SO4, and the
volatiles removed under reduced pressure.
Bis[2,6-bis(4-fluorophenylimino)pyridine] iron(II) hexafluoro-
phosphate (3b). Procedure as for 3a using: 2b (0.14 g,
0.40 mmol) and FeCl2 (0.25 g, 0.20 mmol). Isolated yield:
2,6-Bis(phenylimino)pyridine (2a). Isolated yield: 0.298 g
1
(95%). H NMR (CDCl3, 300 MHz, 25 °C): δ 8.35 (d, 2H, JHH
=
7.8 Hz; PyrHm), 7.88 (t, 1H, JHH = 7.8 Hz; PyrHp), 7.39 (t, 4H,
JHH = 7.9 Hz; ArH), 7.13 (m, 2H; ArH), 7.85 (m, 4H; ArH),
2.41 ppm (s, 6H; NvCMe). 13C{1H} NMR (CDCl3, 75 MHz,
25 °C): δ 167.5 (MeCvN), 155.6 (CAr), 151.4 (CAr), 137.0 (CAr),
129.2 (CAr), 123.8 (CAr), 122.4 (CAr), 119.4 (CAr), 16.4 ppm
(CH3).
2,6-Bis(4-fluorophenylimino)pyridine (2b). Isolated yield:
0.342 g (98%). 1H NMR (CDCl3, 300 MHz, 25 °C): δ 8.32 (d, 2H,
JHH = 7.8 Hz; PyrHm), 7.88 (t, 1H, JHH = 7.8 Hz; PyrHp), 7.08 (m,
4H; ArH), 6.81 (m, 4H; ArH), 2.41 ppm (s, 6H; NvCMe). 13C
{1H} NMR (CDCl3, 75 MHz, 25 °C): δ 168.2 (MeCvN), 159.7 (d,
JCF = 241.2 Hz; F–CAr), 155.5 (CAr), 147.3 (d, JCF = 2.8 Hz; CAr),
137.0 (CAr), 122.5 (CAr), 120.9 (d, JCF = 7.9 Hz; CAr), 115.8 (d, JCF
= 22.5 Hz; CAr), 16.4 ppm (CH3). 19F{1H} NMR (CDCl3,
282 MHz, 25 °C): δ −120.5 ppm (s).
2,6-Bis(4-bromophenylimino)pyridine (2c). Isolated yield:
0.443 g (94%). 1H NMR (CDCl3, 300 MHz, 25 °C): δ 8.32 (d, 2H,
JHH = 7.8 Hz; PyrHm), 7.88 (t, 1H, JHH = 7.8 Hz; PyrHp), 7.49 (m,
4H; ArH), 6.73 (m, 4H; ArH), 2.39 ppm (s, 6H; NvCMe). 13C
{1H} NMR (CDCl3, 75 MHz, 25 °C): δ 168.1 (MeCvN), 155.3
(CAr), 150.3 (CAr), 137.1 (CAr), 132.2 (CAr), 122.7 (CAr), 121.3
(CAr), 116.8 (CAr), 16.4 ppm (CH3).
2,6-Bis(4-cyanophenylimino)pyridine (2d). Isolated yield:
0.243 g (67%). 1H NMR (CDCl3, 300 MHz, 25 °C): δ 8.34 (d, 2H,
JHH = 9.0 Hz; PyrHm), 7.93 (t, 1H, JHH = 7.5 Hz; PyrHp), 7.68 (d,
4H, JHH = 6.0 Hz; ArH), 6.92 (d, 4H, JHH = 9.0 Hz; ArH),
2.39 ppm (s, 6H; NvCMe). 13C{1H} NMR (CDCl3, 75 MHz,
25 °C): δ 168.0 (MeCvN), 155.3 (CAr), 154.8 (CAr), 137.3 (CAr),
133.5 (CAr), 123.2 (CAr), 120.0 (CAr), 119.3 (CuN), 107.2 (CAr),
16.7 ppm (CH3).
1
0.203 g (97%). H NMR (CD3CN, 300 MHz, 25 °C): δ 8.22 (m,
3H; PyrHm and PyrHp), 6.92 (m, 4H; ArH), 6.18 (m, 4H; ArH),
2.58 ppm (s, 6H; NvCMe). 13C{1H} NMR (CD3CN, 75 MHz,
25 °C): δ 180.5 (MeCvN), 162.1 (d, JCF = 246.7 Hz; CAr), 160.0
(CAr), 140.3 (d, JCF = 3.1 Hz; CAr), 137.2 (CAr), 128.6 (CAr), 122.5
(d, JCF = 8.8 Hz; CAr), 117.5 (d, JCF = 23.5 Hz; CAr), 19.8 ppm
(CH3). 19F{1H} NMR (CD3CN, 282 MHz, 25 °C): δ −72.9 (d, JPF
=
707.0 Hz; PF6), −114.5 (s; Ar–F). 31P{1H} NMR (CD3CN,
121 MHz, 25 °C): δ −144.7 (q, JFP = 707.1 Hz; PF6). MS
(ESI-TOF/MS, m/z) calcd for C42H34F4N6Fe [M + H]+, 753.2047;
found 753.2069.
Bis[2,6-bis(4-bromophenylimino)pyridine]
iron(II)
hexa-
fluorophosphate (3c). Procedure as for 3a using: 2c (0.188 g,
0.40 mmol) and FeCl2 (0.25 g, 0.20 mmol). Isolated yield:
1
0.252 g (98%). H NMR (CD3CN, 300 MHz, 25 °C): δ 8.24 (m,
3H; PyrHm and PyrHp), 7.33 (m, 4H; ArH), 6.07 (m, 4H; ArH),
2.59 ppm (s, 6H; NvCMe). 13C{1H} NMR (CD3CN, 75 MHz,
25 °C): δ 180.9 (MeCvN), 160.2 (CAr), 143.3 (CAr), 137.3 (CAr),
133.8 (CAr), 129.0 (CAr), 122.3 (CAr), 122.0 (CAr), 19.9 ppm
(CH3). 19F NMR (CD3CN, 282 MHz, 25 °C): δ −72.8 (d, JPF
=
706.4 Hz; PF6). 31P{1H} NMR (CD3CN, 121 MHz, 25 °C): δ
−144.6 (q, JFP 706.8 Hz; PF6). Anal. calcd for
=
C42H34N6F12P2FeBr4: C, 39.16; H, 2.66. Found: C, 39.07; H,
2.79. MS (ESI-TOF/MS, m/z) calcd for C42H34Br4N6Fe [M + H]+,
992.8850; found 992.8894.
Bis[2,6-bis(4-cyanophenylimino)pyridine] iron(II) hexafluoro-
phosphate (3d). Procedure as for 3a using: 2d (0.145 g,
0.40 mmol) and FeCl2 (0.25 g, 0.20 mmol). Isolated yield:
1
0.208 g (97%). H NMR (CD3CN, 300 MHz, 25 °C): δ 8.32 (m,
3H; PyrHm and PyrHp), 7.57 (d, 4H, JHH = 8.4 Hz; ArH), 6.31 (m,
4H, JHH = 8.4 Hz; ArH), 2.64 ppm (s, 6H; NvCMe). 13C{1H}
Synthesis of iron complexes
Bis[2,6-bis(phenylimino)pyridine] iron(II) hexafluorophosphate NMR (CD3CN, 75 MHz, 25 °C): δ 182.0 (MeCvN), 160.0 (CAr),
(3a). A 100 mL flask was charged with 2a (0.125 g, 0.40 mmol) 147.2 (CAr), 138.1 (CAr), 135.2 (CAr), 129.9 (CAr), 121.6 (CAr),
and FeCl2 (0.25 g, 0.20 mmol) under N2. Degassed methanol 118.4 (CuN), 112.5 (CAr), 20.3 ppm (CH3). 19F NMR (CD3CN,
(30 mL) was added via cannula, immediately forming a dark 282 MHz, 25 °C): δ −72.9 (d, JPF = 706.7 Hz; PF6). 31P{1H} NMR
purple solution. The solution was stirred for 30 min and (CD3CN, 121 MHz, 25 °C): δ −144.6 (q, JFP = 706.5 Hz; PF6). MS
solid NaPF6 (0.101 g, 0.60 mmol) was added. The solution (ESI-TOF/MS, m/z) calcd for C46H34N10Fe [M + H]+, 783.2391;
was stirred for an additional 30 min, and the volatiles were found 783.2053.
removed under reduced pressure. Water (20 mL) was added
and the mixture was triturated, filtered, and washed with
Electrochemical methods
an addition 3 × 5 mL of water, leaving a dark purple solid. Cyclic voltammetry (CV) and differential pulse voltammetry
This solid was collected and dried in vacuo. Isolated yield: (DPV) experiments were conducted using 0.6 mM of analyte
0.165 g (85%). 1H NMR (CD3CN, 300 MHz, 25 °C): δ 8.11 dissolved in 15 mL dry CH3CN containing 0.1 M (nBu4N)PF6
(m, 3H; PyrHm and PyrHp), 7.18 (m, 6H; ArH), 6.21 (m, 4H; and purged with Ar for 20 minutes prior to analysis. A CHI
Dalton Trans.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Paper
760c bipotentiostat was employed, using a 3 mm diameter
Crystal structure parameters for 1c (CCDC 1983241†): X-ray
glassy carbon working electrode, a Ag/Ag+ quasi-non-aqueous quality single crystals were grown by cooling a concentrated
reference electrode separated by a Vycor tip, and a Pt wire CH3CN/DMSO (10 : 1) solution to −20 °C overnight. Yellow
counter electrode. CV experiments were conducted using scan rods, C23H18.5Br2Cl2N3.5Zn, 627.99 g mol−1, triclinic, space
rates of 50–800 mV s−1. DPV experiments were carried out group P1; a = 12.9613(7) Å, b = 14.5938(6) Å, c = 15.1144(7) Å, α
ˉ
using a 5 mV increment, 50 mV amplitude, 0.1 s pulse width, = 92.855(2)°, β = 110.335(2)°, γ = 102.818(2)°, V = 2588.7(2) Å3;
0.0167 s sample width, and 0.5 s pulse period. Following ana- Z = 4, ρcalcd = 1.611 g cm−3; crystal dimensions 0.70 × 0.24 ×
lysis, ferrocene (FcH) was added to each solution as an internal 0.18 mm; diffractometer Bruker D8 QUEST ECO CMOS; Mo Kα
standard, and potentials are reported versus the FcH0/+ redox radiation, 150.0 K, 2θmax = 61.442°; 34 726 reflections, 15 756
couple.41
independent (Rint = 0.0702), intrinsic phasing; absorption
Charging/discharging experiments were conducted via a coeff (μ = 4.257 mm−1), absorption correction semi-empirical
2
chronopotentiometry protocol under an N2 atmosphere using from equivalents (SADABS); refinement (against Fo ) with
a
reticulated vitreous carbon (RVC) working electrode SHELXTL V6.1, 555 parameters, 0 restraints, R1 = 0.0708 (I >
(∼700 cm2) in a glass cylindrical chamber (85 mL) containing 2σ) and wR2 = 0.1345 (all data), Goof = 1.037, residual electron
an acetonitrile solution of both analyte and nBu4PF6 (0.3 M), density 1.61/−1.16 e Å−3. One CH3CN solvent molecule was
and a Teflon-coated stirbar. A graphite rod counter electrode modeled successfully, however due to difficulties modeling
immersed in a 0.3 M nBu4PF6 solution was placed in a fritted remaining solvents, the SQUEEZE protocol imbedded in
tube (10 mL) separating the working and counter electrode PLATON44 was used to remove a solvent void of 249 Å3 contain-
chambers, and a fritted Ag/AgCl quasi-reference electrode ing 101 e−.
placed into the working electrode chamber. Potential cut-offs
Crystal structure parameters for 1d (CCDC 1983240†): X-ray
were set to voltages at which the reversible couples for each quality single crystals were grown by layering a concentrated
analyte was observed to start and finish, according to CV CH3CN solution with Et2O and cooling to −5 °C. Yellow
experiments. Cycling experiments were executed at various blocks; C25H20Cl2N6Zn 540.74 g mol−1, monoclinic, space
anodic and cathodic currents, with a (dis)charge time of 3600 group P21/c; a = 7.7284(7) Å, b = 14.8601(13) Å, c = 22.4323(19)
s, which corresponds to a 1C (dis)charging rate assuming a Å, α = γ = 90°, β = 97.628(4)°, V = 2553.4(4) Å3; Z = 4, ρcalcd
=
2e− reduction process.
1.407 g cm−3; crystal dimensions 0.280 × 0.160 × 0.110 mm;
2θmax = 56.11°; 66 904 reflections, 6121 independent (Rint
0.0966, intrinsic phasing; absorption coeff (μ = 1.196 mm−1),
=
X-Ray crystallography
Crystal structure data was using collected from a multi-faceted absorption correction semi-empirical from equivalents
2
crystal of suitable size and quality selected from a representa- (SADABS); refinement (against Fo ) with SHELXTL V6.1, 310
tive sample of crystals of the same habit using an optical parameters, 0 restraints, R1 = 0.0656 (I > 2σ) and wR2 = 0.1495
microscope. Each crystal was mounted on a MiTiGen loop and (all data), Goof
data collection carried out in a cold stream of nitrogen (150 K; 0.85/−1.04 Å−3
Bruker D8 QUEST ECO; Mo Kα radiation). All diffractometer Crystal structure parameters for 3c (CCDC 1983242†): X-ray
=
1.219, residual electron density
.
manipulations were carried out using Bruker APEX3 soft- quality single crystals were grown by layering isopropyl ether
ware.42 Structure solution and refinement was carried out in over an acetonitrile solution and placing it in the freezer.
the OLEX243 program using XS, XT and XL software, as well as Purple blocks, C44H37Br4F12FeN7P2, 1329.23 g mol−1, monocli-
the Bruker SHELXTL suite.42 For each structure, the absence of nic, space group Cc; a 17.9244(7) Å, b = 17.5570(7) Å, c =
additional symmetry was confirmed using ADDSYM incorpor- 15.9880(6) Å, α = γ = 90°, β = 102.194(2)°, V = 4917.9(3) Å3; Z =
ated in the PLATON program.44
4, ρcalcd = 1.795 g cm−3; crystal dimensions 0.370 × 0.270 ×
Crystal structure data for 1b (CCDC 1983239†): X-ray quality 0.210 mm; diffractometer Bruker D8 QUEST ECO CMOS; Mo
single crystals were grown by cooling a concentrated CH3CN Kα radiation, 150.0 K, 2θmax = 61.234°; 79 707 reflections,
solution to −20 °C overnight. Yellow plates, C23H20Cl2F2N4Zn, 14 968 independent (Rint = 0.0501), intrinsic phasing; absorp-
526.70 g mol−1, triclinic, space group P1; a = 8.8832(3) Å, b = tion coeff (μ = 3.708 mm−1), absorption correction semi-
ˉ
2
12.6990(5) Å, c = 21.7752(8) Å, α = 105.6760(10)°, β = 95.8720 empirical from equivalents (SADABS); refinement (against Fo )
(10)°, γ = 94.3010(10)°, V = 2339.09(15) Å3; Z = 4, ρcalcd = 1.496 g with SHELXTL V6.1, 637 parameters, 2 restraints, R1 = 0.0356 (I
cm−3; crystal dimensions 0.27 × 0.24 × 0.07 mm; diffract- > 2σ) and wR2 = 0.0590 (all data), Goof = 1.048, residual elec-
ometer Bruker D8 QUEST ECO CMOS; Mo Kα radiation, tron density 0.50/−0.51 e Å−3. A CH3CN solvent molecule was
150.0 K, 2θmax = 55.132°; 56 870 reflections, 10 800 indepen- modeled successfully within the asymmetric unit.
dent (Rint = 0.0757), intrinsic phasing; absorption coeff (μ =
Crystal structure parameters for 3d (CCDC 1983243†): X-ray
1.312 mm−1), absorption correction semi-empirical from quality single crystals were grown by layering isopropyl ether
2
equivalents (SADABS); refinement (against Fo ) with SHELXTL over an acetonitrile solution and placing it in the freezer.
V6.1, 583 parameters, 0 restraints, R1 = 0.0618 (I > 2σ) and wR2 Purple plates. C50H40N12F12P2Fe, 1154.73 g mol−1, monoclinic,
= 0.1038 (all data), Goof = 1.111, residual electron density space group P21/n; a = 11.5217(6) Å, b = 32.6571(16) Å, c =
0.63/−0.53 e Å−3. Two CH3CN solvent molecules were success- 13.8662(7) Å, α = γ = 90°, β = 98.732(2)°, V = 5156.9(5) Å3; Z = 4,
fully modeled within the asymmetric unit.
ρcalcd = 1.487 g cm−3; crystal dimensions 0.23 × 0.22 ×
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.

View Article Online
Paper
Dalton Transactions
0.07 mm; diffractometer Bruker D8 QUEST ECO CMOS; Mo Kα 14 C. Ding, H. Zhang, X. Li, T. Liu and F. Xing, J. Phys. Chem.
radiation, 150.0 K, 2θmax = 49.700°; 121 300 reflections, 8890 Lett., 2013, 4, 1281–1294.
independent (Rint = 0.1056), intrinsic phasing; absorption 15 A. Crawford, V. Viswanathan, D. Stephenson, W. Wang,
coeff (μ = 0.447 mm−1), absorption correction semi-empirical
from equivalents (SADABS); refinement (against Fo ) with
E. Thomsen, D. Reed, B. Li, P. Balducci, M. Kintner-
Meyer and V. Sprenkle, J. Power Sources, 2015, 293, 388–
399.
2
SHELXTL V6.1, 700 parameters, 0 restraints, R1 = 0.0717 (I >
2σ) and wR2 = 0.1611 (all data), Goof = 1.087, residual electron 16 G. M. Duarte, J. D. Braun, P. K. Giesbrecht and
density 1.29/−0.97 e Å−3. Two CH3CN solvent molecules were
modeled successfully within the asymmetric unit.
D. E. Herbert, Dalton Trans., 2017, 46, 16439–16445.
17 D. Enright, S. Gambarotta, G. P. A. Yap and
P. H. M. Budzelaar, Angew. Chem., Int. Ed., 2002, 41, 3873–
3876.
18 J. J. Kiernicki, P. E. Fanwick and S. C. Bart, Chem.
Commun., 2014, 50, 8189–8192.
Conflicts of interest
There are no conflicts to declare.
19 A. M. Tondreau, S. C. E. Stieber, C. Milsmann,
E. Lobkovsky, T. Weyhermüller, S. P. Semproni and
P. J. Chirik, Inorg. Chem., 2013, 52, 635–646.
20 B. M. Wile, R. J. Trovitch, S. C. Bart, A. M. Tondreau,
E. Lobkovsky, C. Milsmann, E. Bill, K. Wieghardt and
P. J. Chirik, Inorg. Chem., 2009, 48, 4190–4200.
Acknowledgements
We are grateful to the Natural Sciences Engineering Research
Council of Canada for a Discovery Grant to DEH (RGPIN-2014- 21 A. A. Antonov, N. V. Semikolenova, V. A. Zakharov,
03733); the Canadian Foundation for Innovation and Research
W. Zhang, Y. Wang, W.-H. Sun, E. P. Talsi and
Manitoba for an award in support of an X-ray diffractometer
K. P. Bryliakov, Organometallics, 2012, 31, 1143–1149.
(CFI #32146); and the University of Manitoba for GETS support 22 A. S. Ionkin, W. J. Marshall, D. J. Adelman, A. L. Shoe,
(JDB, DBN) and
P. K. Giesbrecht are thanked for helpful discussions.
a
VP-RI USRA (BKS). C. Kuss and
R. E. Spence and T. Xie, J. Polym. Sci., Part A: Polym. Chem.,
2006, 44, 2615–2635.
23 M. Gasperini, F. Ragaini and S. Cenini, Organometallics,
2002, 21, 2950–2957.
24 B. L. Small, Acc. Chem. Res., 2015, 48, 2599–2611.
25 M. C. Thompson and D. H. Busch, J. Am. Chem. Soc., 1962,
84, 1762–1763.
26 D. H. Busch and N. A. Stephenson, Coord. Chem. Rev.,
1990, 100, 119–154.
References
1 P. V. Kamat, K. S. Schanze and J. M. Buriak, ACS Energy
Lett., 2017, 2, 1368–1369.
2 W. Wang and V. Sprenkle, Nat. Chem., 2016, 8, 204–206.
3 Y. Huang, S. Gu, Y. Yan and S. F. Y. Li, Curr. Opin. Chem. 27 A. A. Mikhailine, E. Kim, C. Dingels, A. J. Lough and
Eng., 2015, 8, 105–113. R. H. Morris, Inorg. Chem., 2008, 47, 6587–6589.
4 M. Arbabzadeh, J. X. Johnson, G. A. Keoleian, 28 T. Simler, A. A. Danopoulos and P. Braunstein, Dalton
P. G. Rasmussen and L. T. Thompson, Environ. Sci.
Technol., 2016, 50, 1046–1055.
Trans., 2017, 46, 5955–5964.
29 E. C. Alyea and P. H. Merrell, Synth. React. Inorg. Met.-Org.
Chem., 1974, 4, 535–544.
30 A. W. Addison, T. N. Rao, J. Reedijk, J. van Rijn and
G. C. Verschoor, J. Chem. Soc., Dalton Trans., 1984, 1349–
1356.
31 I. B. Lozada, T. Murray and D. E. Herbert, Polyhedron, 2019,
161, 261–267.
5 S. M. Laramie, J. D. Milshtein, T. M. Breault, F. R. Brushett
and L. T. Thompson, J. Power Sources, 2016, 327, 681–692.
6 G. L. Soloveichik, Chem. Rev., 2015, 115, 11533–11558.
7 R. W. Hogue and K. E. Toghill, Curr. Opin. Electrochem.,
2019, 18, 37–45.
8 J. A. Suttil, J. F. Kucharyson, I. L. Escalante-Garcia,
P. J. Cabrera, B. R. James, R. F. Savinell, M. S. Sanford and 32 B. de Bruin, E. Bill, E. Bothe, T. Weyhermüller and
L. T. Thompson, J. Mater. Chem. A, 2015, 3, 7929–7938.
9 J. I. van der Vlugt, Chem. – Eur. J., 2019, 25, 2651–2662.
K. Wieghardt, Inorg. Chem., 2000, 39, 2936–2947.
33 L. P. Hammett, J. Am. Chem. Soc., 1937, 59, 96–103.
10 T. Storr and R. Mukherjee, Inorg. Chem., 2018, 57, 9577– 34 A. J. Bard and L. R. Faulkner, Electrochemical methods: fun-
9579.
damentals and applications, Wiley, New York, 2nd edn,
11 P. J. Cabrera, X. Yang, J. A. Suttil, K. L. Hawthorne,
2001.
R. E. M. Brooner, M. S. Sanford and L. T. Thompson, 35 J. Noack, N. Roznyatovskaya, T. Herr and P. Fischer, Angew.
J. Phys. Chem. C, 2015, 119, 15882–15889. Chem., Int. Ed., 2015, 54, 9776–9809.
12 P. J. Cabrera, X. Yang, J. A. Suttil, R. E. M. Brooner, 36 X. Wei, W. Pan, W. Duan, A. Hollas, Z. Yang, B. Li, Z. Nie,
L. T. Thompson and M. S. Sanford, Inorg. Chem., 2015, 54,
10214–10223.
J. Liu, D. Reed, W. Wang and V. Sprenkle, ACS Energy Lett.,
2017, 2, 2187–2204.
13 C. S. Sevov, S. L. Fisher, L. T. Thompson and M. S. Sanford, 37 Y. Zhao and H. R. Byon, Adv. Energy Mater., 2013, 3, 1630–
J. Am. Chem. Soc., 2016, 138, 15378–15384. 1635.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Paper
38 C. Xie, H. Zhang, W. Xu, W. Wang and X. Li, Angew. Chem., 42 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found.
Int. Ed., 2018, 57, 11171–11176. Crystallogr., 2008, 64, 112–122.
39 H. Roth, N. Romero and D. Nicewicz, Synlett, 2015, 27, 43 O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard
714–723.
and H. Puschmann, J. Appl. Crystallogr., 2009, 42, 339–341.
44 A. L. Spek, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2009,
65, 148–155.
40 Z. Flisak and W.-H. Sun, ACS Catal., 2015, 5, 4713–4724.
41 N. G. Connelly and W. E. Geiger, Chem. Rev., 1996, 96, 877–910.
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.