Inorganic Chemistry
Article
A cobalt corrole with one C F and two p-aminophenyl
shifts the first corrole reduction to higher potentials, which is
in agreement with the preference of the reduced Co(II) corrole
toward pentacoordination, and consistent with the behavior
observed by cyclic voltammetry, EPR, and DFT calculations.
6
5
substituents (Co(BAPC)Py ) was also shown to perform water
2
oxidation when oxidized twice, with a kobs of 0.2 s− at 1.4 V vs
1
NHE, measured in MeCN with 3% (1.7 M) H O (∼0.7 V vs
2
+
24
FcH /FcH). We summarize our electrocatalytic findings
As dissociation of pyridine is a requirement for binding of H O
2
together with reported data for C F containing Co corroles in
and achieving maximum catalytic performance, three strategies
appear feasible. First, preparing the cobalt corroles 1 and 2
with an apical ligand other than pyridine; one possibility would
6
5
Table 4.
Our results clearly show that in the case of 1Py and 2Py
2
2
48,49
the third corrole oxidation is catalytically active for H O
be DMSO,
which has been shown to favor pentacoordi-
2
oxidation, with only a marginal increase in current at the
second oxidation process. The kobs values obtained with 1Py2
nation. Second, immobilizing the corroles in glassy carbon,
5
0
22
graphite, carbon nanotubes, or FTO electrodes and
perform heterogeneous water oxidation in aqueous media.
Third, performing homogeneous water oxidation studies in
aqueous media, for which a water-soluble corrole must be
and 2Py are larger than those from previously reported cobalt
2
corroles with C F substituents, even with the “hangman”
6
5
carboxylate, and almost 1 order of magnitude larger than kobs
values reported for Co(tpfc) and Co(BAPC)Py , in similar
synthesized. A derivative of 1Py where the methyl ester
2
2
conditions.
groups have been hydrolyzed to the corresponding carboxylic
acid groups would seem a suitable target in that respect. We
are currently working on some of these leads.
Additionally, the open-shell nature of our catalytically active
species offers the prospect of novel spectroscopic insights into
mechanistic details. The mechanism of water oxidation
catalyzed by cobalt corroles has been proposed to involve
These three strategies require careful characterization of the
species involved in electrocatalysis, whether they are corroles 1
and 2 with different apical ligands, adsorbed on electrode
surfaces, or a water-soluble analogue of 1. Furthermore, it is
important to compare any new catalytic and mechanistic
studies with previously carefully studied systems. The electro-
chemical, spectroscopic, and computational results obtained in
this work provide a solid understanding of the physicochemical
properties of cobalt corroles in several oxidations states, which
will facilitate the analysis of the behavior of related cobalt
corroles during water oxidation catalysis. Particularly, this work
has shown that for cobalt corroles with moderately electron-
withdrawing groups, the third oxidation is active toward water
oxidation, thus opening new alternatives in the mechanistic
analysis of this class of catalysts.
•
+
•−
water nucleophilic attack (WNA) on a [Co(III)L (O )]
1
4,47
species,
(Figure 9, lower panel, left). As can be observed in
Figure 9 (top panel), the doubly and triply oxidized
+
/0
[
1Py(O)] corroles differ in the charge and spin state of
the corrole ligand, but not in the electronic configuration of the
cobalt-oxyl moiety (essentially Co(III)-O ). Therefore, it is
possible that in both oxidation states the same mechanism for
•
−
H O attack is operative. Doubly oxidized cobalt corrole species
2
possess a total spin of either S = 0 or S = 1, and therefore will
be EPR-silent or difficult to observe. On the other hand, our
triply oxidized active species would be a spin S = 1/2 system,
as would be the immediate product of the attack by H O
2
(Figure S9, lower panel, right). Therefore, ideally, some of
In summary, our experimental findings show that using
moderate electron withdrawing substituents on cobalt corroles
has a profound effect on the catalytic efficiency for water
oxidation and point to a triply oxidized cobalt corrole
intermediate as active species. In order to confirm or refute
this hypothesis, further studies should be performed in which
these species may be detected by EPR; this is not currently
possible with Co(tpfc) or β-fluorinated Co “hangman” corrole,
for which a third oxidation would likely be very hard to
achieve.
The potentials at which we obtained the maximum kobs
values for 1Py and 2Py were higher by about 0.3−0.6 V,
2
2
(
electro)chemical oxidations coupled to spectroscopic techni-
compared to the potentials at which k was reported for
obs
ques such as EPR, and electronic structure calculations, could
play an important role. In this aspect, our careful character-
ization of the EPR signals associated with several possible
species involved in cobalt corrole electrochemistry provide a
solid base for identification of certain intermediates in water
oxidation catalysis.
Co(tpfc), Co “hangman” corrole and Co(BAPC)Py . This is
2
reasonable considering that the latter cobalt corroles displayed
electrocatalytic water oxidation activity upon the second
oxidation, while 1Py and 2Py displayed maximum activity
2
2
upon the third oxidation.
This difference, however, should not be seen as a definitive
limitation, as we have observed that the potential of the third
oxidation is highly sensitive to the substitution of neutral
pyridine by anionic apical ligands, such as hydroxide. Figure
S40 shows the change in onset potential for the third oxidation
of 2Py with increasing H O concentration, which shifts from
4. CONCLUSION
In this work, we have synthesized two cobalt corroles with
carboxylic ester and nitro- substituents in their meso-phenyl
groups, determined the structure of one of them by single-
crystal X-ray diffraction, and studied their (electro)chemical
and spectroscopic features by UV−vis and EPR spectroscopy,
cyclic voltammetry and spectroelectrochemical methods, as
well as density functional theory electronic structure
calculations. The synthesized metallocorroles were shown to
catalyze water oxidation upon the third corrole oxidation, in
contrast with previously reported C F -containing cobalt
2
2
1
.27 V with no H O to 1.19 V with 1.44 M H O. To
2
2
qualitatively observe the effect of removing one apical pyridine
ligand on the electrocatalytic water oxidation, we performed
cyclic voltammetry experiments of 1Py in MeCN (0.1 M
2
Bu NPF ) with 2.13 M H O before and after adding 0.5 equiv
4
6
2
of H SO (Figure S41). It can be observed that upon addition
2
4
of acid there is a large cathodic shift (300 mV) of the third
oxidation, and a smaller cathodic shift (130 mV) of the second
6
5
corroles, for which the second oxidation has been shown to
−
1
oxidation of 1Py . These changes are consistent with a
be active. The obtained TOF value for 1Py was 1.86 s at
2
2
−
1
dissociation of pyridine leading to fast coordination of H O,
1.29 V with 1.30 M H O, whereas for 2Py it was 1.67 s at
2 2
2
which upon deprotonation lowers the redox potentials of the
second and third oxidation. The small amount of acid also
1.38 V with 1.44 M H O. These values are larger than for
previously reported cobalt corroles with pentafluorophenyl
2
J
Inorg. Chem. XXXX, XXX, XXX−XXX