8734 Inorganic Chemistry, Vol. 48, No. 18, 2009
Cape et al.
these extremes, from which it is evident that mechanisms
involving water exchange at the cis-aqua coordination
site are fundamentally incapable of reproducing the
experimental data. It is also clear from these results that
the absence of appreciable formation of 36O2 cannot be
ascribed to a rapid exchange of water under catalytic
conditions that masks coupling and reductive elimination
of coordinated oxo atoms by intramolecular or bimole-
cular pathways (Scheme 1). Specifically, it is inconceiva-
ble that a rapid loss of only one coordinated water,
leading to an inappreciable formation of 36O2, could
occur without a loss of the second, which would necessa-
rily lead to an inappreciable formation of 34O2 as well, in
contrast with the experimental results (Figures 2A and 3).
The reaction of Ce4þ with [(bpy)2Ru(OOH)ORu-
(OH)(bpy)2]4þ or a similar peroxo intermediate formed
by reaction of the [Ru(dO)2]O core in {5,5} might also
lead to the formation of 32O2 if the reaction proceeded by
coupling of a ceric-coordinated water ligand with the
terminal peroxo O atom. Subsequent cleavage of the
peroxo O-O bond, rather than the Ru-O bond, would
lead to a net formation of O2 from two solvent molecules
with the 18O label remaining in the Ru coordination
sphere. However, 32O2/34O2 ratios measured for the
photochemically driven reactions at pH 7 were indentical
within experimental uncertainty (Figure 4). For these
lack coordinated water have supported just such a me-
chanism.34 In this proposed mechanism, oxidation to
RuIV is followed by H2O addition, after which the se-
ven-coordinate complexes undergo further two-electron
oxidation with a loss of protons to give L6RuVIdO
species that react with H2O to form peroxo-bound inter-
mediates. Electronic rearrangement then leads to the
release of O2 and regeneration of the catalyst in its
original (RuII) oxidation state. The turnover rate con-
stants for these catalysts calculated from the reported
data are kcat e 4 ꢀ10-2 s-1 at 20 °C, in 0.1 M triflic acid,
which is an order of magnitude greater than the overall
value for {5,5} measured in 0.5 M triflic acid (kcat ≈ 5 ꢀ
10-3 s-1).5 In these studies with mononuclear complexes,
initial rates within homologous series increased with
increasing electron donation from bipyridine substitu-
ents, which is opposite to the fractional order of con-
tribution from the two-solvent pathway reported here for
the dinuclear catalyst (Figure 4). The order observed for
the μ-oxo dimers is consistent with a pathway involving
“covalent hydration” of the bipyridine ligands, a reaction
which is promoted in other nitrogen heterocycles by
electron withdrawal from heterocyclic rings and retarded
by steric constraints at potential binding sites imposed by
the substituents.15 Correspondingly, both factors should
disfavor this pathway in the methylated derivatives but
have opposing effects in the carboxylated derivative,
which is consistent with the observed pattern (Figure 4).
One additional consideration bearing on these mechan-
isms is the absence of water exchange during catalytic
turnover of the dimer. Specifically, it is difficult to
imagine a mechanism where expansion of the ruthenium
coordination sphere by H2O would not also lead to
measurable exchange with the coordinated aqua ligand
in these catalysts, particularly when {5,5} decomposition
is rate-limiting.4,5 Finally, we note that recent theoretical
modeling studies of the “Tanaka catalyst”,35 a dimeric
ruthenium complex containing cis-oriented aqua ligands,
suggest that O2 formation occurs via redox cycling of the
electroactive benzoquinone ligands without altering the
oxidation states of bound RuII ions.36 This electrocatalyst
may therefore represent an extreme example of noninno-
cent ligand participation in water oxidation.
3þ
latter reactions, which made use of Ru(bpy)3 analogs
and SO4•- as oxidants, it is difficult to imagine this sort of
coupling occurring at all, much less with the same selec-
tivity as Ce4þ. Consequently, reaction models that are
independent of the oxidant identity are inherently more
appealing.
Comparisons with Other Proposed Mechanisms. The
failure of these alternative mechanisms to account for
the data strongly suggests that a bona fide pathway does
exist for O-O bond formation from two solvent mole-
cules. Whatever its nature, this pathway is apparently
quite general for this class of catalysts, since it is observed
for several bipyridine analogs and in media with widely
varying compositions (Figure 4). We have adapted the
extensive literature on covalent hydration15 and pseudo-
base formation16 by nitrogen heterocycles and concepts
3þ
developed from studies on M(bpy)3 decomposition in
basic solutions12-14 and other reactions of coordination
complexes32 to suggest a pathway involving noninnocent
participation of the bipyridine ligands (Figure 1). As
noted in the Introduction, we have amassed data that
are consistent with this mechanism; none of this is defi-
nitive, however. For example, the radical signals and NIR
bands that are diagnostic of OH• addition to the ring
could be associated with “dead end” species in equilibrium
with the active form of the catalyst.17
An alternative mechanism involving expansion of the
Ru primary coordination sphere by the addition of
H2O,33 which then reacts with a second water molecule,
is difficult to exclude. Recent combined experimental and
DFT theoretical studies examining water oxidation cat-
alyzed by mononuclear polypyridyl RuII complexes that
The pathway for water oxidation catalyzed by another
dinuclear Ru complex ion,37 cis,cis-[RuII(trp)(OH2)]2(μ-
bpp)3þ (where trp=2,20:6,200-terpyridine and bpp=2,6-
bis(pyridyl-pyrazolate)) has recently been determined by
18O-isotopic labeling studies.38 This ion differs from the
“blue dimer” in having a pyrazolate bridging ligand in
place of an oxo atom and having stable redox states
between {2,2} and {4,4}. Both experimental data and
DFT calculations39 indicate that the highest oxidation
state has a [Ru(dO)]2L core and is the only oxidation
state capable of evolving O2. Selective 18O-labeling of the
(34) Tseng, H.-W.; Zong, R.; Muckerman, J. T.; Thummel, R. Inorg.
Chem. 2008, 47, 11763–11773.
(35) Wada, T.; Tsuge, K.; Tanaka, K. Inorg. Chem. 2001, 40, 329–337.
(36) Muckerman, J. T.; Polyanski, D. E.; Wada, T.; Tanaka, K.; Fujita, E.
Inorg. Chem. 2008, 47, 1787–1802.
::
ꢀ
(32) Sagues, J. A. A.; Gillard, R. D.; Lancashire, R. J.; Williams, P. A.
(37) Sens, C.; Romero, I.; Rodriguez, M.; Llobet, A.; Parella, T.;
Benet-Buchholz, J. J. Am. Chem. Soc. 2004, 126, 7798–7799.
(38) Romain, S.; Bozoglian, F.; Sala, X.; Llobet, A. J. Am. Chem. Soc.
2009, 131, 2768–2769.
J. Chem. Soc., Dalton Trans. 1979, 193–198. Gillard, R. D. Coord. Chem. Rev.
1975, 16, 67–94.
(33) Serpone, N.; Ponterini, G.; Jamieson, M. A.; Bolletta, F.; Maestra,
M. Coord. Chem. Rev. 1983, 50, 209–302.
(39) Yang, X.; Baik, M.-H. J. Am. Chem. Soc. 2008, 130, 16231–16240.