Structure–Activity and Stability Relationships for Cobalt Polypyridyl-Based Hydrogen-Evolving Catalysts in Water

Article abstract of DOI:10.1002/cssc.201701511

A series of eight new and three known cobalt polypyridyl-based hydrogen-evolving catalysts (HECs) with distinct electronic and structural differences are benchmarked in photocatalytic runs in water. Methylene-bridged bis-bipyridyl is the preferred scaffold, both in terms of stability and rate. For a cobalt complex of the tetradentate methanol-bridged bispyridyl–bipyridyl complex [Co^IIBr(tpy)]Br, a detailed mechanistic picture is obtained by combining electrochemistry, spectroscopy, and photocatalysis. In the acidic branch, a proton-coupled electron transfer, assigned to formation of Co^III?H, is found upon reduction of Co^II, in line with a pK_a(Co^III?H) of approximately 7.25. Subsequent reduction (?0.94 V vs. NHE) and protonation close the catalytic cycle. Methoxy substitution on the bipyridyl scaffold results in the expected cathodic shift of the reduction, but fails to change the pK_a(Co^III?H). An analysis of the outcome of the benchmarking in view of this postulated mechanism is given along with an outlook for design criteria for new generations of catalysts.

Full text of DOI:10.1002/cssc.201701511

Accepted Article
Title: Structure-Activity and Stability Relationships for Cobalt
Polypyridyl based Hydrogen Evolving Catalysts in Water
Authors: Stephan Schnidrig, Cyrill Bachmann, Peter Müller, Nicola
Weder, bernhard Spingler, Evelyne Joliat-Wick, mathias
Mosberger, johannes windisch, roger alberto, and Benjamin
Probst
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.201701511
Link to VoR: http://dx.doi.org/10.1002/cssc.201701511
A Journal of
www.chemsuschem.org

10.1002/cssc.201701511
ChemSusChem
Structure-Activity and Stability Relationships for
Cobalt Polypyridyl based Hydrogen Evolving
Catalysts in Water
Stephan Schnidrig, Cyril Bachmann, Peter Müller, Nicola Weder, Bernhard Spingler,
Evelyne Joliat-Wick, Mathias Mosberger, Johannes Windisch, Roger Alberto* and
Benjamin Probst*
University
of
Zürich,
Department
of
Chemistry,
Winterthurerstrasse
190,
CH-8057 Zürich Switzerland
Abstract: A series of eight new and three known cobalt polypyridyl based hydrogen evolving
catalysts (HEC), with distinct electronic and structural differences are benchmarked in
photocatalytic runs in water. Methylene-bridged bis-bipyridyl is the preferred scaffold, both in
terms of stability and rate. For a cobalt complex of the tetradentate, methanol bridged bis-
pyridyl-bipyridyl [CoIIBr(tpy)]Br (1a) a detailed mechanistic picture is obtained by combining
electrochemistry, spectroscopy and photocatalysis. In the acidic branch, a proton coupled
electron transfer (PCET), assigned to formation of CoIII-H, is found upon reduction of CoII, in
line with a pKa(CoIII-H) of ~7.25. Subsequent reduction (-0.94 V vs NHE) and protonation close
the catalytic cycle. Methoxy substitution on the bipyridyl scaffold results in the expected
cathodic shift of the reduction, but fails to change the pKa(CoIII-H). An analysis of the outcome
of the benchmarking in view of this postulated mechanism is given along with an outlook for
design criteria for new generations of catalysts.
Introduction:
Research on artificial photosynthesis and the production of solar fuels[1-5] produced numerous
molecular catalysts for both the water oxidation and proton reduction reaction over the last
years.[6-12] For proton reduction, a multitude of hydrogen evolving catalysts (HEC) based on
earth abundant metals like cobalt,[9, 13-22] nickel,[15, 19, 23-25] molybdenum[26] and iron[27] have been
reported. The performance of catalysts is described in organic solvents or in buffered aqueous
solution for electro- or photocatalytic H2 production. In photocatalysis, the presence of a
sacrificial electron donor (SED) and a photosensitizer (PS; see [5-8, 13-15, 28-31] and refs. therein)
is mandatory. However, different conditions render an accurate and predictive comparison of
catalysts difficult. For instance, quantities are strongly influenced even by subtle variations of
the different components (SED, PS, pH, concentrations/ratios, light source, etc.).[32] There is a
great deal of information in the literature on the initial steps of photocatalysis in homogenous
solutions, namely the study of the processes that follow excitation of the photosensitizer
(reductive / oxidative quenching, first electron transfer to cobalt, follow up reactions of oxidised
electron donors), both in organic[33-37] and aqueous media,[38-45] or mixtures.[46-48] There are,
from a spectroscopic point of view, only few studies focusing on the follow-up reactions of the
50]
reduced cobalt species in water[49,
or organic solvents.[51] The picture is however
complemented by some very careful electrochemical studies on proton and CO2 reduction by
cobalt and other catalysts, in organic media[24, 25, 52-62] and in water,[21, 22, 43, 63-69] or mixtures.[70,
71] The qualities of HECs are characterized by i) high complex stability, e.g. turnover numbers
1
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
(TON, H2/HEC), ii) high catalytic activity, e.g. turnover frequency (TOF, H2/HEC/h) and iii) low
overpotentials, e.g. catalytic proton reduction potential close to the thermodynamic limit.[17, 19,
51, 72-79]
For the latter two criteria, catalytic tafel plots (TOF-η relationship) have emerged for
benchmarking of molecular electrocatalysts,[29, 54, 74] but their general applicability is hampered
by their dependence on the acid source and concentration, making comparison situation
dependant.[28] Only few reports of catalytic tafel plots in water exist for the case of water
oxidation catalysts.[80]
In 2010, Chang and co-workers presented a new class of molecular HECs, namely cobalt
complexes with poly-pyridyl ligand frameworks.[81] They exhibit excellent stabilities even in
purely aqueous solutions (several 1’000 or even 10’000 H2/Co, see [32, 68, 82, 83] and refs. therein).
In 2015, Artero and co-workers reviewed known poly-pyridyl based cobalt HECs in a
comprehensive, comparative article.[32] Some interesting trends between the nature of the
ligand backbone and the photocatalytic or electrocatalytic performance could be identified: i)
poly-pyridyl based complexes gave higher TONs as standard [Co(bpy)3]2+; ii) complexes with
purely pyridine based ligands are less stable than catalysts with bpy unit(s); iii) for tetradentate
poly-pyridyl complexes cis open coordination sites were superior over trans open sites; iv) no
clear trends by the electronic nature of the substituents could be revealed on catalysis.[32]
Cobalt complexes with bpy units are prone to ligand-assisted reductions; these are not
necessarily cobalt centred. Thus, electronic variation on the bpy units were predicted to have
a more pronounced influence on performance, as compared to variations on pyridine units.
They clearly concluded the importance of investigating catalysts under the same – or, at least,
similar - conditions. Understanding the influence of the ligand’s electronics on the cobalt centre
would allow specific modifications to improve their performance.
Inspired by this pioneering work, we aimed at getting deeper insights into poly-pyridyl based
HECs in water for extending the scope of these complexes. Even though cobalt polypyridyl
catalysts are amongst to most stable and fastest HECs in photocatalysis, they typically display
rather high overpotentials as compared to cobaloxime.[17, 19, 74, 76] Lowering the overpotential,
but maintaining the stability of cobalt poly-pyridyl catalysts would certainly advance the field of
molecular HECs, and allow an important step towards sustainable molecular water splitting.
In this article, we compare various poly-pyridyl cobalt HECs for photocatalytic proton reduction
in water. Photochemical benchmarking over a range of HEC concentrations is followed by an
electrochemical analysis of selected HEC’s in water. We previously reported different acyclic
and cyclic molecular Co poly-pyridyl HECs (1a, 2a, 8 and 9, Scheme 1), which all exhibited
excellent stabilities in photocatalytic H2 formation (9 - 30 kTON) with rhenium and/or ruthenium
based photosensitizers (RuPS and RePS, Scheme 1).[84-88] We complement this series with a
set of new HECs with ligands of different electronic and steric properties. We follow the concept
of highly distorted structures leading to lower reduction potentials (2, 3 and 4), investigate the
effect of acyclic vs cyclic ligands (5 and 6 vs 8 and 9) and the influence of free coordination
sites on tetradentate catalysts (1a, 1b, 5, 6, 8 and 9). Also the effect of the third substituent on
the methylene-bis-bipyridine moiety for catalysts 2a, 2b, 3, 4 and 5 on structure and reactivity
is assayed. Two ligands with amine bridged bpy units (complex 6 and 7, Scheme 1) were
synthesized for comparison with similar, but carbon bridged structures (5 and 3). Finally, we
tuned the electronic properties of complexes 1a and 2a by distinct substitutions of the
bipyridine ligands (complex 1b and 2b, Scheme 1) to establish the respective influence on
catalysis and the overpotentials as proposed by Artero et al.[32] Stabilizing the reduced form of
the HEC (CoI) leads to a reduced basicity at CoI for cobaloximes.[89] In turn, further reduction
or strong acids are then required to form a cobalt hydride species (CoIII-H or CoII-H) for
2
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
subsequent H2 evolution. Electron donating groups on the bpy units will thus render CoI more
basic and H2 production should be facilitated at a lower pH value, but with cathodically shifted
potentials, as proposed for cobaloximes.[89] To corroborate this hypothesis, we studied H3CO-
substituted versions of 1a and 2a (1b and 2b, Scheme 1). Based on the respective physico-
chemical results, we correlate stabilities and activities in the photochemical assay with selected
electrochemical insights from pH dependent measurements of the HEC with their ligand
framework.
Results and Discussion
Synthesis and Characterization: Following the procedures for L1a (tpy)[87] and L2a (appy),[88]
all ligands were synthesized from bpy-Br as depicted in Scheme 2. The corresponding
methoxy-substituted tpy and appy ligands - (MeO)2tpy (L1b) and (MeO)4appy (L2b) - were
prepared by this strategy from the respective (MeO)2bpy-Br derivative. 6,6'-dibromo-4,4'-
dimethoxy-2,2'-bipyridine ((MeO)2bpyBr2) was synthesized according to a modified procedure
(experimental part in supplementary information).[90] The mono-bromo derivative (MeO)2bpy-
Br was obtained by lithiation of (MeO)2bpyBr2 with n-BuLi, followed by MeOH quenching.
Nucleophilic attack with lithiated (MeO)2bpy-Br to bis(2-pyridyl)methanone and ethyl picolinate
gave L1b and the ketone (MeO)2bpyCOpy, respectively. The latter was converted to
(MeO)4appy (L2b) by repeated reaction with lithiated (MeO)2bpy-Br. Tetradentate bpy2CO (L5)
and hexadentate hpy (L4) have been synthesized by Singh et al.[91, 92] We present herein a
simpler pathway: Reaction of lithiated bpy-Br with 0.5 eq. ethyl-chloroformate (ClCOOEt) gave
bpy2CO (L5). Further reaction with lithiated Me-Py or bpy-Br resulted in pentadentate C-appy
(L3) and hexadentate hpy (L4), respectively. Amine bridged ligand bpy2NH (L6) was obtained
by reacting bpy-Br with NH3 in an autoclave at high temperature. Substitution of the amine of
L6 with PyCH2Br gave N-appy (L7). Ligands L1a – L9 were characterized by LC-MS and NMR
and subjected to complexation with Co salts. All Co complexes (1a - 9) were obtained by
stirring a small excess of ligand with CoBr2 or Co(OTf)2 in MeOH. Full complexation was
achieved within a few minutes (HPLC monitoring). Precipitation by adding excess Et2O to
concentrated MeOH solutions gave the respective complexes in good yield. Crystallization
allowed for further purification. Single crystals for x-ray diffraction were obtained by vapor
diffusion.[93] All complexes were characterized by MS, elemental analysis and complexes 1a,[87]
1b, 2a,[88] 2b, 4, 5, a dimer of 5, 6, 7, 8 and 9[84] additionally by single crystal x-ray diffraction.
3
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
Scheme 1: Structure of the complexes in this work with the rhenium and ruthenium based photosensitizers.[84, 86-88]
4
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
Scheme 2: Schematic representation of the bpy based ligand syntheses starting from bpy-Br or (MeO)2bpy-Br. i)
1. THF, n-BuLi, -90°C, 5 min, 2. MeOH, ii) For R = OMe: 1. Et2O, n-BuLi, -90°C, 5 min, 2. bis(2-pyridyl)methanone;
for R = H: as reported earlier[87]; iii) NH3, 8 bar, 265°C, 44 h; iv) 1. DMA, NaH, 100°C, 15 min, 2. 2-
(bromomethyl)pyridine, 100°C, 15 min; v) For R = OMe: 1. Et2O, n-BuLi, -90°C, 5 min, 2. ethyl picolinate; for R = H:
as reported earlier[88]; vi) For R = OMe: 1. (MeO)2bpyBr, Et2O, n-BuLi, -90°C, 5 min, 2. (MeO)2bpyCOpy. For R = H:
as reported earlier[88]; vii) 1. Et2O, n-BuLi, -78°C, 25 min, 2.: ClCOOEt, viii) 1. Et2O, bpy-Br, n-BuLi, -78°C, 30 min,
2. bpy2CO (L5), -78°C → 20°C, 4 h; ix) 1. THF, 2-methylpyridine, LiTMP, -78°C, 45 min, 2. bpy2CO (L5), -78°C →
-40°C, 2 h; Detailed procedures are shown in the experimental part and the synthesis of (MeO)2bpyBr2 is described
in the supplementary information.
All crystal structures and crystallographic details are given in Figure 1, Figure SI 1 – Figure SI
4 and Table SI 1. Complexes 6 and 7, based on the N-bridged ligands L6 (bpy2NH) and L7
(N-appy) showed both a very similar coordination geometry. The two bpy subunits occupy the
equatorial plane and two bromides the axial positions (Figure SI 1). The bridging nitrogen of
L7 appears to have strong sp2 character as the three substituents form an almost perfect
trigonal plane, with a pendent but non-coordinating α-picolyl. Ligand L4 offers a sterically
restricted coordination pocket. Consequently, the corresponding complex 4 exhibits a trigonal
prismatic geometry (Figure SI 2, left), in agreement with the reported Mn and Zn complexes.[92]
Based on LC-MS, both [CoBr(C-appy)]Br (3) and [Co(bpy2CO)(Sol)2]OTf2 (5) showed an
equilibrium in solution between the monomer depicted in Scheme 1, and a dimeric form, as
depicted for 5 in Scheme SI 1. Based on the comparison of UV/Vis absorption of the monomer
species 3, we hypothesize a similar structure as 2a, e.g. distorted trigonal prismatic. Similar
absorptions were observed for the two dimeric species, and X-ray analysis of suitable crystals
5
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
of [Co2(bpy2CO(OMe))2](ClO4)2, grown from a solution of 5 in MeOH, did confirm the dinuclear
species with two ligand molecules (L5) in the hemi-ketal form (Scheme SI 1 and Figure SI 3).
Crystals grown from an acetonitrile solution, however, gave the mononuclear species of 5
(Figure SI 2, right).The interconversion of the mono- to the dinuclear complex in protic solvents
was reported for zinc and iron. The dimerization was shown to be reversible by NMR for Zn2+.[91]
The electron donating effect of the methoxy substitution with ligand (MeO)2tpy (L1b) and its
complex 1b becomes evident, when comparing to six-coordinate 1a. The structure of 1a shows
a distorted trigonal prismatic geometry with one weakly bound Br-, whereas 1b is five
coordinate with a trigonal bipyramidal structure (Figure 1, left).[87] The structure of 2b strongly
resembles the one of 2a, i.e. a distorted trigonal prismatic geometry with a labile bromide ligand
cis to the pyridyl unit (Figure 1, right). An ORTEP together with an overlay 2a/2b is given in
Figure SI 4.
Figure 1: ORTEP drawing of [CoBr((MeO)2tpy)]Br (1b, left) and [CoBr((MeO)4appy)]Br (2b, right) at 50% probability
level. Hydrogen atoms, the counter ion and solvent molecules are omitted for clarity.
Photocatalysis is a very powerful method to benchmark different catalysts under light
irradiation in pure aqueous media. We compared the performance of complexes 1 – 9 using
visible light, RuPS and ascorbic acid at pH 5 to assess the long-term stability and initial rates
as a function of catalyst concentration. To prevent system stability limitations through
accumulation of dehydroascorbic acid (DHA),[39] TCEP (tris(2-carboxyethyl)phosphine) was
added to regenerate oxidised ascorbic acid, as outlined in Scheme 3.[86] Complete turnover in
TCEP (0.1 M) was thus possible for stable catalysts at higher concentrations, whereas the
stability of the HECs, resp. the reduced PS becomes limiting at lower concentrations. The
overall rate of H2 formation is determined by i) the yield of reductive quenching (~0.6 for 0.1 M
ascorbic acid, pH = 5),[77, 94] ii) the rate of electron transfer to the catalyst, iii) the rate for
decomposition of PS-,[50, 95-97] iv) the rate of back-electron transfer to ascorbic acid radical, resp.
DHA[39] in ratio to v) the total rate for H2 formation on the cobalt catalyst. The fastest cobalt
catalysts in this work achieved an external quantum yield of 0.13 (H/hν) at high cobalt
concentration, but a decrease was observed in all cases below 5 µM, likely due to electron
transfer competing with PS- decomposition. Typical rate constants for electron transfer from
PS- to cobalt are in the range of 108-109 M-1s-1,[38, 49, 50, 87, 98] whereas back-electron transfer
rates to the ascorbate radical anion are typically in the range of 109 M-1s-1.[49, 50] Thus, a fast
catalyst should show complete conversion in electron donor and high H2 formation rates at
high concentration, but high turnover per catalyst at low concentration.
6
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
HO
decomp.
PS
O
Φ<
H2
0.13
O
asc
asc
O
OH
OH
H+
II-H
OH
CoII
DHA
TCEPH+
Co
asc
Φ
=0.6
asc
PS*
e
-
PS CoI
III-H
Co
Hasc
H+
H2
ν
h
TCEPO
net:
+
+
ν
+
+
+
TCEPO
+
H
h
H
TCEPH
2O
Scheme 3: Representation of photocatalytic reaction scheme for hydrogen production using [RuII(bpy)3]2+ (PS),
ascorbate (Hasc-), TCEP and 1a as cobalt based HEC. DHA
dehydroascorbic acid, TCEPO
triscarboxyethylphosphine oxide at pH 5, 100 mM NaHasc / H2asc. Unproductive side reactions in gray.
=
=
Blank experiments assess interference of experimental outcomes by background activities,
especially at very low HEC concentrations. Photocatalytic experiments without HEC reveal H2
evolution by the PS only, and catalysis with equal amounts of cobalt salts instead of the
complexes[45, 77, 84] are relevant since the CoII cations might also form active species in the
catalytic environment (donor, electrolyte, decomposition products). Both blank tests must be
performed to substantiate the performance of the supposed catalysts. As shown in Table SI 2,
both blank experiments gave only very small amounts of H2 (~5 µmol H2) at rates below
1 nmol/s. Thus, all complexes can be denoted as catalytically active, but caution is required
especially for slow catalysts at low concentrations. Furthermore, dynamic light scattering (DLS)
was used to assess the molecular nature of the catalytic systems at hand (see Table SI 4 and
Figure SI 30). To ensure that the setup provides enough sensitivity to probe positively for
nanoparticles at the low concentrations used in this study, reference experiments were
conducted.[99] Comparison of the auto correlation function of the photocatalytic runs to the
reference experiment clearly demonstrates that the catalysts are of molecular nature.
Figure 2: HEC concentration dependent TON in cobalt (left graph) and maximal hydrogen evolution rates (right
graph) of photocatalysis in 10 mL water with 0.1 M TCEP, 0.1 M NaAscO (pH 5), 0.5 mM [Ru(bpy)3]Cl2 (RuPS) and
1, 5, 10 or 20 µM of different cobalt polypyridyl complexes (450 nm LED, 0.3 µE/s).
The two limiting scenarios outlined before are recognised in the comprehensive overview from
photocatalytic benchmarking. Table SI 2 summarizes the results of the catalytic experiments
which are depicted in Figure 2: on the left axis, the total turnover in cobalt (TONCo; H2/Co) is
plotted, and the maximal rate of H2 formation (nmol H2/s) is given on the right one, both versus
the HEC concentration. TONCo increase towards low cobalt concentrations, with more than 20
kTONCo for 2a and 5, and more than 10 kTONCo for essentially all catalysts based on the
7
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
methylene-bis-bipyridyl scaffold (2 – 5, 8 and 9). On the other hand, both the catalysts based
on the methylene-mono-bipyridyl scaffold (1a and 1b), and the catalysts based on the amino-
bis-bipyridyl scaffold (6 and 7) show rather poor stability: a constant TONCo around 750 at all
concentrations clearly indicates that these HECs limit long-term stability. Moreover, the latter
four catalysts also fail to achieve significant conversion of the SED even at high concentration,
as opposed to all other catalysts. In terms of rate of H2 formation, again 2a and 5, but
surprisingly also 1b stand out with around 20 nmol H2/s at high catalyst concentration (Φext
~
0.13), and a TOF of around 1 H2/Co/s at low concentration. The two amino-bis-bipyridyl based
catalysts 6 and 7, along with 1a, display a poor performance in terms of rates, matching well
with their low stability as obvious from the low TONs. Both macrocyclic compounds (8 and 9)
show rather low rates as well, despite their high turnover stabilities. The catalysts derived from
the very robust and fast [CoBr(appy)]Br (2a), namely 2b, 3 and 4, clearly do not improve in
terms of rates and turnovers as compared to their parent compound. Thus, it seems that i)
acyclic compounds featuring the methylene-bis-bipyridyl scaffold are very fast and stable
catalysts, ii) replacement of the bridging methylene by an amine greatly diminishes both
stability and rate, iii) macrocyclic catalysts are reasonably stable, despite their lower rates, and
iv) substituents on the bipyridyl can have a substantial effect on the rate but little on the stability
of the catalyst.
Figure 3: Summary of photocatalytic performance of 1a, 2a and 5 as a function of pH (left scale, solid: maximal
rate; right scale, dashed: turnover number; see Figure SI 6 - Figure SI 8 and Table SI 3). 0.5 mM RuPS, 0.1 M
NaAscO, 0.1 M TCEP, pH as indicated, 5 µM HEC, 10 ml, 450 nm LED (from below), 0.3 µE/s.
To substantiate the performance of the methylene-bis-pyridyl-bipyridyl and the methylene-bis-
bipyridyl scaffolds further, pH dependencies of photocatalytic H2 production were measured
for 1a, 2a and 5 (Figure 3). As evident, rates observed for the two binding motives differ
substantially above pH 4: a decrease is observed for the methylene-bis-pyridyl-bipyridyl
scaffold (1a) with increasing pH, whereas rates increase for the methylene-bis-bipyridyl
scaffold (2a and 5) up to pH 5. The apparent peak in performance at pH 5 might be owed to
the reduced buffer capacity of the system at pH > 5 (pKa H2asc: 4.1 and 11.8; pKa TCEP: 2.9,
3.6, 4.3 and 7.7; pKa TCEPO: 3.5, 4.2, 4.9).[46, 100] Below pH 4, rates decrease in all cases,
consistent with a reduced availability of ascorbate, and thus a diminished supply of reducing
equivalents through the reductive quench cycle.[46] The increase in rate for the methlyene-bis-
pyridyl-bipyridyl scaffold from pH 5.5 to pH 4.0 indicates a proton dependent mechanism for
H2 formation. Electrochemical analysis (see below) confirmed that both tetradentate catalysts
1a and 1b undergo a PCET step from the resting CoII state to give CoIII-H, followed by reduction
to CoII-H. In both cases, these processes occur at potentials easily accessible by RuPS-. The
8
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
rate of protonation of CoII-H (2.2 × 107 M-1s-1) is known to be the rate limiting step for H2
formation for 1a.[49] This could explain the increase in H2 formation rate, as observed from pH
5.5 to 4.0. Turnovers for the latter scaffold remain low for all pH, with a slight increase towards
the acidic branch (~0.5 to 1.0 kTON), indicating a limited catalyst stability. The behaviour of
the penta-coordinate [CoIIBr(appy)]Br (2a) and the planar, tetracoordinate [CoII(bpy2CO)]OTf2
(5) is strikingly different: a clear correlation between rate and turnover as a function of pH is
observed.
A cis-open coordination site might be favorable over a trans-open one in the case of
tetradentate bipyridyl-based complexes, both in terms of stability and rate of catalyses as
previously reported.[32, 77] From the selection of tetradentate, bipyridyl-based catalysts studied
in here (1a, 1b, 5, 6, 7, 8 and 9), such a trend cannot unambigously be established. At pH 5,
the two structures with a cis-open coordination site (1a and 1b) are clearly inferior in terms of
stability (TONs) to the tetradentate structures with a trans-open coordination site (5, 8 and 9),
with the exception of the amine bridged 6 and 7. Moreover, the acyclic trans-compound
[CoII(bpy2CO)]OTf2 (5) is one of the fastest and most stable catalysts of this series, followed
by the pyrphyrin catalysts [CoBr(pyr-bu)MeOH] (8) and [CoBr2(cis-pyr-bu2)] (9), which have
two open coordination sites in trans-position, but are cyclic. Care should be taken when
comparing these results with catalyst 5, as we found evidence of an equilibrium with a dinuclear
structure (Scheme SI 1) when crystallized from protic solvents (Figure SI 3). Dinuclear cobalt
complexes have already been reported to reduce protons under photocatalytic conditions,[101,
102]
thus, we cannot exclude that 5 is active in its dimeric form and should not be directly
compared to the other catalysts since probably representing a class by its own. As noted
before, the cyclic nature of 8 and 9 increases the stability of the catalyst, but only at
intermediate rates. α-picolyl substituted [CoBr2(N-appy)] (7) behaves very similar to its parent
compound 6. X-ray analysis (Figure SI 1) shows that the pendent picolyl unit points away from
the cobalt centre while leaving the sp2 hybridization on the bridging nitrogen. It is interesting
that the replacement of a bridging carbonyl unit by an amine results in such a drastic decrease
in both TOF and TON of the catalyst. Clearly, this underlines the importance of fine tuning
electronics for both, rates and stabilities.
The structurally similar complexes [CoBr(appy)]Br (2a), [CoBr(C-appy)]Br (3) and [Co(hpy)]Br2
(4) exhibited very good performances in TOFs and TONs >10 nmol H2/s and >10'000 H2/Co,
respectively. The lower performance of 3 vs. 2a coincides with our hypothesis; a distortion
towards a trigonal prismatic geometry leads to higher activities and stabilities. Still, TOFs and
TONs for 4 are similar to 3, although the model would imply a better performance than for 2a.
Catalyst 4 is hexa-coordinated and the availability of a free site for proton reduction is not as
evident as in 2a or 3 (Figure SI 2, left). Since H2 evolution was observed, structural
reorganization of the original complex must take place. Likewise, an entire bpy unit is de-
coordinated upon reduction which allows further reaction steps towards H2 production.
In summary, after performing photocatalytic benchmark with a set of 11 cobalt-based
polypyridyl proton reduction catalysts in water at pH 5, two limiting scenarios are identified; at
low concentrations (1 µM), significant differences emerge between the catalysts in terms of
stability and more than 20 kTONCo are only possible for 2a and 5. Both, the amine bridged
bipyridyl catalysts 6 and 7together with the tetradentate HECs 1a and 1b show constant but
low TONsCo of 500 - 1000 at all concentrations, implying limitation by stability as compared to
the other seven catalysts. In parallel, rates are slow for 1a, 6 and 7 while 1b is fast. Cyclic,
pyrphyrin based HECs 8 and 9 are very robust in terms of TONs. Although slow, they compare
well with the other methylene-bis-bipyridyl based structures. All 5 catalysts featuring the
9
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
methylene-bis-bipyridyl unit (2a, 2b, 3, 4 and 5) are among the most active catalysts, both in
terms of stability (660 to 950 µmol H2 at 20 µM, corresponding to 66 – 95 % conversion of the
electron donor), and TONsCo ≥ 10’000. In terms of rate, 2a and 5 outperform the other three
HECs by roughly a factor of 2.
Mechanistic Investigation: Electrochemical analysis of cobalt based HECs in water is
challenging for several reasons: i) mercury electrodes are needed, as typical cobalt polypyridyl
catalysts display overpotentials of 400 - 700 mV;[32] ii) labile coordination sites are fundamental
for catalysis, but can lead to mixtures of active compounds in aqueous electrolytes; iii) proton
reduction is intrinsically an irreversible process and identification of half-wave potentials
requires stringent search in the scan rate domain; iv) reduction can lead to lower charge
density or neutral compounds, which may deposit on the electrode;[21, 22] v) most importantly,
proton reduction is per se pH dependent, and experimental variation in the pH domain is basic
for understanding the complex pH dependencies of the hydrogen evolution reaction (HER). Ott
and coworkers showed that examination of a HEC in a Britton Robinson buffer (BRB, equimolar
mixture of H3PO3, AcOH, B(OH)3) at various pH values gave valuable insights into the
properties of a water reduction catalyst.[43] Thummel and coworkers investigated a series of
tetradentate, planar cobalt polypyridyl catalysts in buffered aqueous solution, and found two
pH independent reductions, followed by a catalytic current enhancement.[21, 22] Similar studies
have been performed by Savéant et al. with an O2 reduction catalyst[103] as well as by Llobet
et al. with ruthenium based WOC.[104]
A detailed mechanistic picture for H2 generation in water was obtained previously for catalyst
1a by transient absorption spectroscopy.[49] This study employed the RePS / TEOA
(triethanolamine) system for photo-generation of reducing equivalents in buffered solutions at
pH 8 to 8.5. Fast generation of CoI was followed by a concerted proton coupled electron
transfer (PCET) step (CoI / CoII-H couple, no accumulation of CoIII-H), with a slow second
protonation (pKa(CoII-H)), resulting in H2 formation. The last protonation step appears to be
rate limiting (see black solid lines in Scheme 4).
CoII
CoI
Co"0"
III-H
II-H
Co
Co
I-H
Co
Co
Co
III-H2
II-H2
I-H
Co
2
Scheme 4: Plausible reduction (horizontal) and protonation (vertical) steps for H2 formation by CoII HER catalysts,
excluding bimolecular pathways. Concerted proton coupled electron transfer (PCET) steps run diagonally from top
left to bottom right; all reduction and protonation reactions are formally drawn as metal based, but can well be ligand
based in some cases; protonated cobalt species are formally drawn as hydrides, but the polarity of the cobalt
hydride bond might vary. Black arrows: proposed H2 evolving pathway for 1a at pH 8;[49] grey: proposed pathway
at pH < 7.
To rationalise the mechanism of H2 formation with polypyridyl catalysts in water, we performed
electrochemical analysis from pH 2 to 11 in a in buffered solution (Britton Robinson buffer).[105]
Stability of the catalysts at selected pH was independently assessed by UV/Vis titration (see
Figure SI 25 - Figure SI 29). Correlation of catalytic current enhancement with H2 formation
10
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
was quantified by in-line GC/TCD analysis (linear sweep voltammetry (LSV), 0.1 mV/s, see
Figure SI 22 to Figure SI 24). A hanging mercury drop electrode (HMDE) was selected as
working electrode since mercury displays the highest reported overpotential for HER[28] and a
new mercury drop was available for each scan, avoiding artefacts from deposition on
subsequent measurements. Reference experiments on the blank Hg electrode and on
[CoII(OH2)6]2+ solutions were crucial to substantiate the role of the catalysts from background.
A Pt counter electrode was used together with a Ag/AgCl (3 M KCl) reference electrode. All
experiments were independently referenced by measuring a [MV]Cl2 solution of known
concentration (E(MV2+/+) = -440 mV vs. NHE;[106, 107] D(300 K) = 7.1 × 10-6 cm2 s-1).[108]
The electrochemical response of an unbuffered solution of 1a displays only one reversible
reduction assigned to the CoII/I couple at -0.91 V vs NHE (D(300 K) = 4.5 × 10-6 cm2 s-1; Figure
SI 12). The shape of the signal implies adsorption of the reduced catalyst on the electrode
surface. Scan- and concentration rate dependencies are consistent with a reversible process,
indicating fast dissolution upon re-oxidation.[21, 22] In contrast, electrochemistry of 1a in buffered
solution displays a catalytic current enhancement at -0.93 V vs NHE, and a pH dependent
peak on the anodic shoulder of the catalytic signal (51 mV/pH, Figure 4, right, pH < 7). The
second peak shows an increasing current density, proportional to the pH (Figure SI 10), and
irreversibility at pH < 7, indicative for catalysis. To correlate with current enhancement, H2
formation was independently confirmed by in-line GC/TCD experiments (Figure SI 23). High
rates of H2 formation at -0.93 V were observed at pH 2 and 4 with a significant decrease when
going to pH 7. This current density behaviour coincides with the assignment of the second
reduction as a catalytic process. Based on this evidence and consistent with previous
reports,[49] we assign the first shoulder to a PCET from CoII to CoIII-H (1 electron / 1 proton;
magenta in Figure 4, left). The second, catalytic peak is assigned to a reduction of CoIII-H to
CoII-H up to pH = 7, and a reduction of CoII to CoI from pH 7.5 to 9.4 (green / orange in Figure
4, left). The pKa of coordinated water was independently determined to be 9.4 (Figure SI 25),
which causes the cathodic shift in reduction potential above this pH (CoII-OH to CoI). The
deprotonated form CoII-OH, is hardly soluble, causing the current response to drop above pH
8. Below pH 2, decomplexation is observed, as evidenced by LC-MS (Figure SI 26). These
data evidence i) a pKa(CoI/CoIII-H) of ~7.25, and a pKa(CoII-H/CoII-H2) of 3 or smaller (since no
pH dependency of the 2nd reduction is observed down to pH 3), ii) the potentials for reduction
of CoIII-H and CoII are very similar (~-0.93 V), and iii) protonation of CoII-H is slow and limits
catalysis in all cases. Scheme 4 summarizes these data: below pH 7 the grey arrows and
above 7.5 the black arrows are operative, in line with the Pourbaix diagram drawn in Figure 4,
left.
Figure 4: Left: Pourbaix diagram of 1a, right: voltage current response for [CoBr(tpy)]Br (1a) as a function of pH;
(1 mM in 40 mM BRB) referenced to NHE with [MV]Cl2. CE: Pt, RE: Ag/AgCl in 3 M KCl, WE: HMDE, 100mV/s (see
11
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
also Figure SI 9 - Figure SI 12).
This results match very well those previously suggested by Rodenberg et al. at pH 8 to 8.5:[49]
In this report, photochemical reduction of CoII resulted in CoI, which accumulated transiently,
before transformation to CoII-H in a concerted PCET took place. Protonation of CoII-H is
irreversible due to fast H2 release from CoII-H2. Still, the reaction is slow (2.2 × 107 M-1s-1) in
the weakly alkaline pH range (see Scheme 4, black). Furthermore, this mechanism accounts
for the results obtained in photocatalytic runs under continuous light irradiation; rates of H2
formation for 1a increase from pH 5.5 to 4.0 (Figure 3), consistent with the above experiment.
In parallel, the stability decreased when going from pH 4 to 5.5, since slow H2 formation at
higher pH results in accumulation of reduced PS-, leading ultimately to its unproductive
decomposition.[50, 95-97]
The concept can be further brought in relationship to the tetradentate HEC 1b which bears two
electron donating methoxy substituents in the para position on the bipyridine; the redox
potentials shift cathodically by 80 mV (E(CoII/I) = -0.94 (1a) vs -1.02 V (1b) vs NHE, (see Figure
SI 13 and Figure SI 14). The cobalt based orbitals are destabilized by the negative resonance
effects of the π-donating methoxy substituents in para position (σp = -0.27),[109-111] as predicted
in a computational study on a series of cobaloximes catalysts.[89] In contrast to 1a, the CoII/CoIII-
H couple of 1b eludes a quantitative analysis as it is only visible as a weak shoulder on the
catalytic peak at -1.05 V vs NHE. Comparing 1a and 1b a similar, but cathodically shifted
electrochemical response indicates a similar pKa(CoI/CoIII-H). This is unexpected since an
increase in the pKa(CoI/CoIII-H) for 1b due to negative resonance effects would be predicted
as computationally found for cobaloximes (four methoxy vs H groups cause a positive shift of
the pKa(CoI/CoIII-H) of ~10). However, the pKa of coordinated water is very similar (9.4 for 1a
and 9.1 for 1b, Figure SI 25 and Figure SI 28), indicative for a stabilisation by the methoxy
substituents. Likely, different molecular orbitals are involved in reduction (dxz, dyz) and CoI
protonation (dz2), the former being destabilised through negative resonance, the latter
stabilised through inductive effects of the methoxy group (σm = 0.12).[109] As for 1a, the catalytic
peak potential was not dependent on the pH (i.e. a CoIII-H/CoII-H2 couple) down to pH 2.3. The
current enhancement towards lower pH was identical for both HECs, and depended on the
scan rate and the proton concentration (see Figure SI 10 and Figure SI 11). In the
photochemical assay, reduced RuPS (Epc(RuPS2+/+) = -1.2 V vs NHE, see Figure SI 15) is
sufficiently strong to account for this increase in potential. Thus, at present the two rates for
protonation of CoII-H, resp. the pKa(CoII-H/CoII-H2) cannot be compared. Nonetheless, an
increased catalytic rate was found for 1b (17 nmol H2/s) as compared to 1a (5 nmol H2/s), at
pH 5. Prevalence of resonance effects for protonation of CoII-H could thus explain why 1b is
indeed faster as 1a.
For an in depth mechanistic understanding and a benchmarking of mechanistically comparable
catalysts, a similar comparison can be made for 2a and 2b. The electrochemistry of
pentadentate, bis-bipyridyl based catalyst 2a and 2b reveals only a reduced pH dependency
in the precatalytic couples as compared to their tetradentate analogues (Figure SI 16 and
Figure SI 17). 2a displays a non-catalytic couple at -0.61 V vs NHE without pH dependency up
to pH 9.0. A cathodic shift of 56 mV/pH at pH 10 and higher, as required by the pKa(CoII-
OH2/CoII-OH) of 10.1 (Figure SI 27) is then observed. This first reduction is followed by a
second, less pronounced couple at -0.92 V vs NHE, again not showing any pH dependency.
This couple is hidden by the onset of catalysis at pH < 4. No peak or plateau was found on the
catalytic wave at 0.1 V/s up to Ipc/Ipc0 = 20. Instead, the catalytic onset shows a pH dependency
12
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
of 34 mV/pH (at Ipc/Ipc0 = 5), indicative for a two electron / one proton process in the rate limiting
step. HEC 2b is different to 2a as it displays only one reversible, pre-catalytic couple at high
pH (-0.90 V vs NHE), which is also covered by the catalytic onset at pH < 5 (Figure SI 17). The
catalytic onset shows a pH dependency of 75 mV/pH (at Ipc/Ipc0 = 5), indicating a proton coupled
electron transfer in the rate limiting step. For 2a, the first reduction is assigned to a cobalt
(CoII/I), and the second to a ligand based process (formally CoI/0), in agreement to literature
reports.[82] Assuming the same assignments for 2b, the four methoxy substituents shift the CoII/I
couple cathodically by 290 mV. This is significantly more than the effect of the two methoxy
substituents in 1b (80 mV) resulting in an even stronger destabilisation of the electron
accepting dxz or dyz orbitals. Importantly, in this pH range, no further precatalytic reductions
are needed to initiate catalysis. In the photochemical assay, 2b is two times slower than 2a but
performs equally well in terms of stability. As compared to the strong negative pH – rate
corelation observed for the tetradentate catalyst 1a, a weak positive corelation is found for 2a
in the photochemical assay, in line with electrochemistry.
Long, Chang, Castellano et al. reported on the effects of electron accepting -CF3 or donating -
NMe2 groups on tetra- and penta-dentate pyridyl cobalt catalysts.[77, 78, 82] In the case of a
pentapyridyl catalyst (Co(Me2Py5)) with no bipyridyl subunit, -CF3 shifts the first reduction and
the catalytic onset anodically and cathodically for -NMe2.[78] Introduction of a -CF3 group on the
pyridine subunit in another pentadentate ligand, namely ethylene bridged pyridyl-bis-bipyridyl
(Co(MePyBpy2), similar to 2a and 2b) displays an anodic shift for the CoII/I reduction but
unexpectedly, also for the two following, bipyridine based reductions. As opposed to the purely
pyridine based catalysts, the catalytic onset (occurring on the bipyridine based reductions) is
observed at a similar potential, but with a slightly reduced intensity.[82] In both cases,
photocatalysis is not improved by -CF3 substitution in terms of initial rates or stabilities.[82]
Interestingly, however, in the case of the tetradentate MePy2Bpy (similar to 1a and 1b) a
distinct effect of -CF3 substitution on the pyridine unit in photocatalysis is apparent: both rate
and stability decreased significantly as compared to the unsubstituted catalyst.[77] At pH 4, a
shoulder on the catalytic wave is clearly visible for the -CF3 substituted catalyst, assigned to a
CoII/I transition.[77] Considering the pH dependent nature of this shoulder observed for 1a, it is
tempting to speculate that the same would hold true for the catalysts described in this work,
i.e. a CoII/CoIII-H followed by a CoIII/II-H couple to initiate catalysis. It is, moreover, interesting
to note that the substitution, even though diminishing the activity, slightly shifts the pH optima
of both the bipyridine containing substituted catalyst to higher values (pH 4 to 4.5).[77, 82]
Obviously, the electron accepting substituent has a distinct influence on the pKa of one of the
catalytically active species. This dependency deserves further analysis, as highlighted here,
by electrochemistry in water as a function of the pH.
Conclusion
The assessment or qualification of the performance of different HECs as attempted in this
report with the benchmarking of eight novel and three known cobalt polypyridyl catalysts in
photocatalytic assays may be based on different criteria. Absolute TOFs or TONs under
conditions ideal for the particular catalyst is one, comparing these numbers under equal
catalytic parameters another. Both viewpoints may overlap with respect to a conclusion for
improving performance and we have tried in this report to correlate physical and electronic
structure of these eleven HECs with activity and stability, while considering their sometimes
different mechanisms. The acyclic, methylene bridged bis-bipyridyl framework emerges as a
superior polypyridyl ligand framework, both in terms of catalytic rate and catalyst stability.
Substituents on the methylene bridge alter catalysis substantially, with pentadentate pyridyl 2a
13
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
and unsubstituted 5 on top. The introduction of an α-picolyl (3) or a third bipyridyl (4), decrease
the activity and stability, whereas methoxy substituents on the bis-bipyridyl (2b) decrease
activity, while maintaining stability. Cyclic structure 8 and 9 give stable, yet slow catalysts,
highlighting the need for structural flexibility. Amine bridged bis-bipyridyl scaffolds (as for 6 and
7) give slow catalysts with stability limited to 103 turnovers in cobalt, as opposed to 104 for the
carbon bridged structures under our conditions. The tetradentate methylene-bis-pyridyl-
bipyridyl scaffold L1a/b gives highly pH sensitive cobalt catalysts, which are clearly inferior at
pH 5 in terms of stability, even though methoxy substituents increase the maximal rate of
catalysis significantly.
A detailed mechanistic picture is obtained for the tetradentate methylene-bis-pyridyl-bipyridyl
catalyst 1a by a combination of electrochemistry, spectroscopy and hydrogen measurements.
At low pH the CoII resting state undergoes a concerted PCET to CoIII-H, which is easily reduced
to CoII-H before slowly releasing H2. A pKa for CoIII-H of ~7.25 results, above which a pH
independent reduction to CoI occurs, yet followed by a PCET to give CoII-H. H2 formation thus
only occurs at significantly reduced rates the higher the pH, and decomposition of reduced PS
prevails, limiting the overall stability in the photochemical assay. Introduction of two methoxy
substituents in the para position of the pyridyl in 1a gave 1b. All reductions were found
cathodically shifted as compared to 1a, but surprisingly the pKa for CoIII-H did not rise. We
interpret this by a mixture of inductive and resonance effects, the latter prevailing in the
destabilization of orbitals relevant for reduction, the former in stabilizing the orbital relevant for
protonation. Nonetheless, a higher rate of H2 formation was found for 1b as compared to 1a in
the photochemical assays, possibly due to an increase of the second, more relevant rate of
protonation for CoII-H.
Furthermore, the developed design strategy, coupled with our synthetic scheme, allows for
future rational design of new polypyridyl scaffolds, to further corroborate the elaborated trends.
Taken all together, the systematic evaluation of HECs especially with respect to the mode of
action of 1a, further development towards new classes of polypyridyl catalysts with significantly
reduced overpotential through decoupling of reduction and protonation steps becomes
possible.
Acknowledgments
Financial support from the Swiss National Science Foundation SNSF Sinergia project
CRSII2_160801/2 and from the University of Zürich Research Priority Project (URPP) Light to
Chemical Energy Conversion “LightChEC” is gratefully acknowledged. We thank Prof. Tony
Linden for helpful crystallographic discussions.
Supplementary Information
Information about the Instrumentation, syntheses, crystallographic information, detailed
photocatalytic data, electrochemical data, DLS analysis and LSV with in-line GC/TCD.
Literature
[1]
[2]
S. Styring, Faraday Discuss. 2012, 155, 357-376.
T. R. Cook, D. K. Dogutan, S. Y. Reece, Y. Surendranath, T. S. Teets, D. G. Nocera,
Chem. Rev. 2010, 110, 6474-6502.
[3]
Y. Tachibana, L. Vayssieres, J. R. Durrant, Nat. Photonics 2012, 6, 511-518.
14
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
[4]
[5]
G. Centi, S. Perathoner, Chemsuschem 2010, 3, 195-208.
E. S. Andreiadis, M. Chavarot-Kerlidou, M. Fontecave, V. Artero, Photochem.
Photobiol. 2011, 87, 946-964.
[6]
[7]
T. S. Teets, D. G. Nocera, Chem. Commun. 2011, 47, 9268-9274.
S. Berardi, S. Drouet, L. Francas, C. Gimbert-Surin, M. Guttentag, C. Richmond, T.
Stolla, A. Llobet, Chem. Soc. Rev. 2014, 43, 7501-7519.
W. T. Eckenhoff, R. Eisenberg, Dalton Trans. 2012, 41, 13004-13021.
V. Artero, M. Chavarot-Kerlidou, M. Fontecave, Angew. Chem. Int. Edit. 2011, 50,
7238-7266.
[8]
[9]
[10]
J. D. Blakemore, R. H. Crabtree, G. W. Brudvig, Chem. Rev. 2015, 115, 12974–
13005.
[11]
[12]
N. Sutin, C. Creutz, E. Fujita, Comment. Inorg. Chem. 1997, 19, 67-92.
Y. Halpin, M. T. Pryce, S. Rau, D. Dini, J. G. Vos, Dalton Trans. 2013, 42, 16243-
16254.
[13]
[14]
S. Losse, J. G. Vos, S. Rau, Coord. Chem. Rev. 2010, 254, 2492-2504.
P. Zhang, M. Wang, F. Gloaguen, L. Chen, F. Quentel, L. Sun, Chem. Commun.
2013, 49, 9455-9457.
[15]
[16]
[17]
Z. Han, R. Eisenberg, Acc. Chem. Res. 2014, 47, 2537-2544.
S. C. Marinescu, J. R. Winkler, H. B. Gray, PNAS 2012, 109, 15127-15131.
J. L. Dempsey, B. S. Brunschwig, J. R. Winkler, H. B. Gray, Acc. Chem. Res. 2009,
42, 1995-2004.
[18]
W. T. Eckenhoff, W. R. McNamara, P. Du, R. Eisenberg, Bba-Bioenergetics 2013,
1827, 958-973.
[19]
[20]
P. A. Jacques, V. Artero, J. Pecaut, M. Fontecave, PNAS 2009, 106, 20627-20632.
K. Ladomenou, M. Natali, E. Iengo, G. Charalampidis, F. Scandola, A. G.
Coutsolelos, Coord. Chem. Rev. 2015, 304-305, 38-54.
L. P. Tong, A. Kopecky, R. F. Zong, K. J. Gagnon, M. S. G. Ahlquist, R. P. Thummel,
Inorg. Chem. 2015, 54, 7873-7884.
[21]
[22]
[23]
L. P. Tong, R. F. Zong, R. P. Thummel, J. Am. Chem. Soc. 2014, 136, 4881-4884.
M. L. Helm, M. P. Stewart, R. M. Bullock, M. R. DuBois, D. L. DuBois, Science 2011,
333, 863-866.
[24]
[25]
[26]
A. Zarkadoulas, M. J. Field, V. Artero, C. A. Mitsopoulou, Chemcatchem 2017, 9,
2308-2317.
A. Zarkadoulas, M. J. Field, C. Papatriantafyllopoulou, J. Fize, V. Artero, C. A.
Mitsopoulou, Inorg. Chem. 2016, 55, 432-444.
W. T. Eckenhoff, W. W. Brennessel, R. Eisenberg, Inorg. Chem. 2014, 53, 9860-
9869.
[27]
[28]
F. Gloaguen, T. B. Rauchfuss, Chem. Soc. Rev. 2009, 38, 100-108.
E. S. Rountree, B. D. McCarthy, T. T. Eisenhart, J. L. Dempsey, Inorg. Chem. 2014,
53, 9983-10002.
[29]
[30]
[31]
[32]
C. Costentin, S. Drouet, M. Robert, J. M. Saveant, J. Am. Chem. Soc. 2012, 134,
11235-11242.
J. R. McKone, S. C. Marinescu, B. S. Brunschwig, J. R. Winkler, H. B. Gray, Chem.
Sci. 2014, 5.
A. Zarkadoulas, E. Koutsouri, C. Kefalidi, C. A. Mitsopoulou, Coordin. Chem. Rev
2015, 304, 55-72.
N. Queyriaux, R. T. Jane, J. Massin, V. Artero, M. Chavarot-Kerlidou, Coord. Chem.
Rev. 2015, 304, 3-19.
[33]
[34]
B. Probst, C. Kolano, P. Hamm, R. Alberto, Inorg. Chem. 2009, 48, 1836-1843.
B. Probst, A. Rodenberg, M. Guttentag, P. Hamm, R. Alberto, Inorg. Chem. 2010, 49,
6453-6460.
[35]
[36]
M. Heberle, S. Tschierlei, N. Rockstroh, M. Ringenberg, W. Frey, H. Junge, M. Beller,
S. Lochbrunner, M. Karnahl, Chem-Eur. J. 2017, 23, 312-319.
H. Junge, N. Rockstroh, S. Fischer, A. Bruckner, R. Ludwig, S. Lochbrunner, O.
Kuhn, M. Beller, Inorganics 2017, 5.
15
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
[37]
[38]
[39]
[40]
J. Windisch, M. Orazietti, P. Hamm, R. Alberto, B. Probst, Chemsuschem 2016, 9,
1719-1726.
B. Probst, M. Guttentag, A. Rodenberg, P. Hamm, R. Alberto, Inorg. Chem. 2011, 50,
3404-3412.
M. Guttentag, A. Rodenberg, R. Kopelent, B. Probst, C. Buchwalder, M. Brandstätter,
P. Hamm, R. Alberto, Eur. J. Inorg. Chem. 2012, 2012, 59-64.
D. Moonshiram, C. Gimbert-Surinach, A. Guda, A. Picon, C. S. Lehrnann, X. Y.
Zhang, G. Doumy, A. M. March, J. Benet-Buchholz, A. Soldatov, A. Llobet, S. H.
Southworth, J. Am. Chem. Soc. 2016, 138, 10586-10596.
A. Rodenberg, M. Orazietti, M. Mosberger, C. Bachmann, B. Probst, R. Alberto, P.
Hamm, Chemphyschem 2016, 17, 1321-1328.
[41]
[42]
[43]
L. Mintrop, J. Windisch, C. Gotzmann, R. Alberto, B. Probst, P. Kurz, J. Phys. Chem.
B 2015, 119, 13698-13706.
W. M. Singh, M. Mirmohades, R. T. Jane, T. A. White, L. Hammarstrom, A. Thapper,
R. Lomoth, S. Ott, Chem. Commun. 2013, 49, 8638-8640.
C. Creutz, N. Sutin, B. S. Brunschwig, J. Am. Chem. Soc. 1979, 101, 1297-1298.
W. K. Lo, C. E. Castillo, R. Gueret, J. Fortage, M. Rebarz, M. Sliwa, F. Thomas, C. J.
McAdam, G. B. Jameson, D. A. McMorran, J. D. Crowley, M. N. Collomb, A. G.
Blackman, Inorg. Chem. 2016, 55, 4564-4581.
[44]
[45]
[46]
[47]
M. Natali, Acs Catal 2017, 7, 1330-1339.
M. Natali, R. Argazzi, C. Chiorboli, E. Iengo, F. Scandola, Chem-Eur. J. 2013, 19,
9261-9271.
[48]
[49]
[50]
[51]
D. Streich, Y. Astuti, M. Orlandi, L. Schwartz, R. Lomoth, L. Hammarstrom, S. Ott,
Chem-Eur. J. 2010, 16, 60-63.
A. Rodenberg, M. Orazietti, B. Probst, C. Bachmann, R. Alberto, K. K. Baldridge, P.
Hamm, Inorg. Chem. 2015, 54, 646-657.
C. V. Krishnan, C. Creutz, D. Mahajan, H. A. Schwarz, N. Sutin, Isr. J. Chem. 1982,
22, 98-106.
J. L. Dempsey, J. R. Winkler, H. B. Gray, J. Am. Chem. Soc. 2010, 132, 16774–
16776.
[52]
[53]
[54]
[55]
[56]
C. Costentin, M. Robert, J. M. Saveant, A. Tatin, PNAS 2015, 112, 6882-6886.
J. M. Saveant, Chemelectrochem 2016, 3, 1967-1977.
C. Costentin, J.-M. Savéant, Chemelectrochem 2014, 1, 1226-1236.
N. Elgrishi, M. B. Chambers, M. Fontecave, Chem. Sci. 2015, 6, 2522-2531.
N. Elgrishi, B. D. McCarthy, E. S. Rountree, J. L. Dempsey, Acs Catal 2016, 6, 3644-
3659.
[57]
[58]
M. D. Sampson, C. P. Kubiak, Inorg. Chem. 2015, 54, 6674-6676.
D. Brazzolotto, M. Gennari, N. Queyriaux, T. R. Simmons, J. Pecaut, S. Demeshko,
F. Meyer, M. Orio, V. Artero, C. Duboc, Nat. Chem. 2016, 8, 1054-1060.
C. Costentin, G. Passard, M. Robert, J. M. Saveant, J. Am. Chem. Soc. 2014, 136,
11821-11829.
[59]
[60]
P. F. Huo, C. Uyeda, J. D. Goodpaster, J. C. Peters, T. F. Miller, Acs Catal 2016, 6,
6114-6123.
[61]
[62]
C. C. L. McCrory, N. K. Szymczak, J. C. Peters, Electrocat. 2016, 7, 87-96.
A. E. King, Y. Surendranath, N. A. Piro, J. P. Bigi, J. R. Long, C. J. Chang, Chem.
Sci. 2013, 4, 1578-1587.
[63]
N. Kaeffer, A. Morozan, J. Fize, E. Martinez, L. Guetaz, V. Artero, Acs Catal 2016, 6,
3727-3737.
[64]
[65]
[66]
D. Xue, Q. X. Peng, S. Z. Zhan, Inorg. Chem. Commun. 2017, 82, 11-15.
A. Dutta, J. A. Roberts, W. J. Shaw, Angew. Chem. Int. Ed. 2014, 126, 6605 –6609.
A. Dutta, S. Lense, J. A. S. Roberts, M. L. Helm, W. J. Shaw, Eur. J. Inorg. Chem.
2015, 2015, 5218-5225.
[67]
[68]
A. Dutta, B. Ginovska, S. Raugei, J. A. Roberts, W. J. Shaw, Dalton Trans. 2016.
W. M. Singh, T. Baine, S. Kudo, S. Tian, X. A. N. Ma, H. Zhou, N. J. DeYonker, T. C.
Pham, J. C. Bollinger, D. L. Baker, B. Yan, C. E. Webster, X. Zhao, Angew. Chem.
Int. Ed. 2012, 51, 5941–5944.
16
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
[69]
[70]
[71]
M. Vennampalli, G. C. Liang, L. Katta, C. E. Webster, X. Zhao, Inorg. Chem. 2014,
53, 10094-10100.
N. Priyadarshani, A. Dutta, B. Ginovska, G. W. Buchko, M. O'Hagan, S. Raugei, W. J.
Shaw, Acs Catal 2016, 6, 6037-6049.
S. Lense, M.-H. Ho, S. Chen, A. Jain, S. Raugei, J. C. Linehan, J. A. S. Roberts, A.
M. Appel, W. J. Shaw, Organomet. 2012, 31, 6719-6731.
J. L. Dempsey, J. R. Winkler, H. B. Gray, J. Am. Chem. Soc. 2010, 132, 1060-1065.
A. Fihri, V. Artero, M. Razavet, C. Baffert, W. Leibl, M. Fontecave, Angew. Chem.-Int.
Ed. 2008, 47, 564-567.
[72]
[73]
[74]
[75]
[76]
[77]
V. Artero, J.-M. Saveant, Energ. Environ. Sci. 2014, 7, 3808-3814.
X. Hu, B. S. Brunschwig, J. C. Peters, J. Am. Chem. Soc. 2007, 129, 8988-8998.
C. C. McCrory, C. Uyeda, J. C. Peters, J. Am. Chem. Soc. 2012, 134, 3164-3170.
R. S. Khnayzer, V. S. Thoi, M. Nippe, A. E. King, J. W. Jurss, K. A. El Roz, J. R.
Long, C. J. Chang, F. N. Castellano, Energy Environ. Sci. 2014, 7, 1477.
Y. J. Sun, J. P. Bigi, N. A. Piro, M. L. Tang, J. R. Long, C. J. Chang, J. Am. Chem.
Soc. 2011, 133, 9212-9215.
[78]
[79]
[80]
H. I. Karunadasa, C. J. Chang, J. R. Long, Nature 2010, 464, 1329-1333.
R. Matheu, S. Neudeck, F. Meyer, X. Sala, A. Llobet, Chemsuschem 2016, 9, 3361-
3369.
[81]
[82]
[83]
[84]
[85]
[86]
[87]
[88]
J. P. Bigi, T. E. Hanna, W. H. Harman, A. Chang, C. J. Chang, Chem. Commun.
2010, 46, 958-960.
M. Nippe, R. S. Khnayzer, J. A. Panetier, D. Z. Zee, B. S. Olaiya, M. Head-Gordon, C.
J. Chang, F. N. Castellano, J. R. Long, Chem. Sci. 2013, 4, 3934.
D. Z. Zee, T. Chantarojsiri, J. R. Long, C. J. Chang, Acc. Chem. Res. 2015, 48, 2027-
2036.
E. Joliat, S. Schnidrig, B. Probst, C. Bachmann, B. Spingler, K. K. Baldridge, F. von
Rohr, A. Schilling, R. Alberto, Dalton Trans. 2016, 45, 1737-1745.
C. Bachmann, B. Probst, M. Oberholzer, T. Fox, R. Alberto, Chem. Sci. 2016, 7, 436-
445.
C. Bachmann, B. Probst, M. Guttentag, R. Alberto, Chem. Commun. 2014, 50, 6737-
6739.
M. Guttentag, A. Rodenberg, C. Bachmann, A. Senn, P. Hamm, R. Alberto, Dalton
Trans. 2013, 42, 334-337.
C. Bachmann, M. Guttentag, B. Spingler, R. Alberto, Inorg. Chem. 2013, 52,
6055−6061.
[89]
[90]
[91]
B. H. Solis, S. Hammes-Schiffer, J. Am. Chem. Soc. 2011, 133, 19036-19039.
U. Neumann, F. Voegtle, Chem. Ber. 1989, 122, 589-591.
J. C. Knight, A. J. Amoroso, P. G. Edwards, R. Prabaharan, N. Singh, Dalton Trans.
2010, 39, 8925-8936.
[92]
J. C. Knight, S. Alvarez, A. J. Amoroso, P. G. Edwards, N. Singh, Dalton Trans. 2010,
39, 3870-3883.
[93]
[94]
[95]
[96]
B. Spingler, S. Schnidrig, T. Todorova, F. Wild, Cryst. Eng. Comm. 2012, 14, 751-757
C. V. Krishnan, N. Sutin, J. Am. Chem. Soc. 1981, 103, 2141-2142.
K. Kalyanasundaram, Coordin. Chem. Rev. 1982, 46, 159-244.
Q. G. Mulazzani, S. Emmi, P. G. Fuochi, M. Z. Hoffman, M. Venturi, J. Am. Chem.
Soc. 1978, 100, 981-983.
[97]
[98]
[99]
Q. G. Mulazzani, M. Dangelantonio, N. Camaioni, M. Venturi, J. Chem. Soc. Faraday
T. 1991, 87, 2179-2185.
R. Gueret, C. E. Castillo, M. Rebarz, F. Thomas, A. A. Hargrove, J. Pecaut, M. Sliwa,
J. Fortage, M. N. Collomb, J. Photoch. Photobio. B 2015, 152, 82-94.
D. Shevchenko, M. F. Anderlund, A. Thapper, S. Styring, Energ. Environ. Sci. 2011,
4, 1284-1287.
[100] A. Krezel, R. Latajka, G. D. Bujacz, W. Bal, Inorg Chem 2003, 42, 1994-2003.
[101] C. Di Giovanni, C. Gimbert-Surinach, M. Nippe, J. Benet-Buchholz, J. R. Long, X.
Sala, A. Llobet, Chem. Eur. J. 2016, 22, 361-369.
17
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701511
ChemSusChem
[102] S. Mandal, S. Shikano, Y. Yamada, Y. M. Lee, W. Nam, A. Llobet, S. Fukuzumi, J.
Am. Chem. Soc. 2013, 135, 15294-152947.
[103] C. Costentin, H. Dridi, J. M. Saveant, J. Am. Chem. Soc. 2015, 137, 13535-13544.
[104] A. Llobet, Molecular Water Oxidation Catalysis, Wiley, United Kingdom, 2014.
[105] H. T. S. Britton, R. A. Robinson, J. Chem. Soc. 1931, 1456-1462.
[106] A. A. Krasnovsky, G. P. Brin, Dokl. Akad. Nauk Sssr+ 1965, 163, 761-&.
[107] M. Kirch, J. M. Lehn, J. P. Sauvage, Helv Chim Acta 1979, 62, 1345-1384.
[108] A. J. Elliot, D. R. Mccracken, G. V. Buxton, N. D. Wood, J. Chem. Soc. Faraday T.
1990, 86, 1539-1547.
[109] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165-195.
[110] H. Basch, H. B. Gray, Inorg. Chem. 1967, 6, 365-&.
[111] H. B. Gray, C. J. Ballhausen, J. Am. Chem. Soc. 1963, 85, 260-&.
18
This article is protected by copyright. All rights reserved.