Water oxidation by a mononuclear ruthenium catalyst: Characterization of the intermediates

Article abstract of DOI:10.1021/ja203249e

A detailed characterization of intermediates in water oxidation catalyzed by a mononuclear Ru polypyridyl complex [Ru^II-OH₂] ²⁺ (Ru = Ru complex with one 4-t-butyl-2,6-di-(1′,8′- naphthyrid-2′-yl)-pyridine ligand and two 4-picoline ligands) has been carried out using electrochemistry, UV-vis and resonance Raman spectroscopy, pulse radiolysis, stopped flow, and electrospray ionization mass spectrometry (ESI-MS) with H₂¹⁸O labeling experiments and theoretical calculations. The results reveal a number of intriguing properties of intermediates such as [Ru^IV=O]²⁺ and [Ru ^IV-OO]²⁺. At pH > 2.9, two consecutive proton-coupled one-electron steps take place at the potential of the [Ru^III-OH] ²⁺/[Ru^II-OH₂]²⁺ couple, which is equal to or higher than the potential of the [Ru^IV=O] ²⁺/[Ru^III-OH]²⁺ couple (i.e., the observation of a two-electron oxidation in cyclic voltammetry). At pH 1, the rate constant of the first one-electron oxidation by Ce(IV) is k₁ = 2 × 10⁴ M^-1 s^-1. While pH-independent oxidation of [Ru^IV=O]²⁺ takes place at 1420 mV vs NHE, bulk electrolysis of [Ru^II-OH₂]²⁺ at 1260 mV vs NHE at pH 1 (0.1 M triflic acid) and 1150 mV at pH 6 (10 mM sodium phosphate) yielded a red colored solution with a Coulomb count corresponding to a net four-electron oxidation. ESI-MS with labeling experiments clearly indicates that this species has an O-O bond. This species required an additional oxidation to liberate an oxygen molecule, and without any additional oxidant it completely decomposed slowly to form [Ru^II-OOH]⁺ over 2 weeks. While there remains some conflicting evidence, we have assigned this species as ¹[Ru^IV-·²-OO]²⁺ based on our electrochemical, spectroscopic, and theoretical observations alongside a previously reported analysis by T. J. Meyer group (J. Am. Chem. Soc. 2010, 132, 1545-1557).

Full text of DOI:10.1021/ja203249e

ARTICLE
pubs.acs.org/JACS
Water Oxidation by a Mononuclear Ruthenium Catalyst:
Characterization of the Intermediates
Dmitry E. Polyansky,*,† James T. Muckerman,† Jonathan Rochford,†,§ Ruifa Zong,‡ Randolph P. Thummel,‡
and Etsuko Fujita*,†
†Chemistry Department, Brookhaven National Laboratory, Upton, New York 11973-5000, United States
‡Department of Chemistry, University of Houston, Houston Texas 77204-5003, United States
S Supporting Information
b
ABSTRACT: A detailed characterization of intermediates in
water oxidation catalyzed by a mononuclear Ru polypyridyl
complex [RuIIꢀOH2]2+ (Ru = Ru complex with one 4-t-butyl-
2,6-di-(10,80-naphthyrid-20-yl)-pyridine ligand and two 4-pico-
line ligands) has been carried out using electrochemistry,
UVꢀvis and resonance Raman spectroscopy, pulse radiolysis,
stopped flow, and electrospray ionization mass spectrometry
(ESI-MS) with H218O labeling experiments and theoretical
calculations. The results reveal a number of intriguing proper-
ties of intermediates such as [RuIVdO]2+ and [RuIVꢀOO]2+. At pH > 2.9, two consecutive proton-coupled one-electron steps take
place at the potential of the [RuIIIꢀOH]2+/[RuIIꢀOH2]2+ couple, which is equal to or higher than the potential of the [RuIVdO]2+/
[RuIIIꢀOH]2+ couple (i.e., the observation of a two-electron oxidation in cyclic voltammetry). At pH 1, the rate constant of the first
one-electron oxidation by Ce(IV) is k1 = 2 ꢁ 104 Mꢀ1 sꢀ1. While pH-independent oxidation of [RuIVdO]2+ takes place at 1420 mV
vs NHE, bulk electrolysis of [RuIIꢀOH2]2+ at 1260 mV vs NHE at pH 1 (0.1 M triflic acid) and 1150 mV at pH 6 (10 mM sodium
phosphate) yielded a red colored solution with a Coulomb count corresponding to a net four-electron oxidation. ESI-MS with labeling
experiments clearly indicates that this species has an OꢀO bond. This species required an additional oxidation to liberate an oxygen
molecule, and without any additional oxidant it completely decomposed slowly to form [RuIIꢀOOH]+ over 2 weeks. While there
remains some conflicting evidence, we have assigned this species as 1[RuIVꢀη2-OO]2+ based on our electrochemical, spectroscopic,
and theoretical observations alongside a previously reported analysis by T. J. Meyer’s group (J. Am. Chem. Soc. 2010, 132, 1545ꢀ1557).
’ INTRODUCTION
16-electron RuIV species that subsequently undergoes attack by water
at the metal center to expand the coordination shell to seven.29 Sun
and co-workers have structurally characterized a similar 7-coordinate
intermediate as a dimeric species for a system in which the equatorial
ligand is 6,60-dicarboxy-2,20-bipyridine.25,46 The other class of mono-
nuclear Ru catalyst has water bound to the metal in the resting state of
the catalyst. Initial oxidation of the catalyst by a sacrificial oxidant such
as Ce(IV) results in the loss of two electrons accompanied by the loss
of two protons to produce a Ru(IV) oxo species. Proton-coupled
electron transfer (PCET) reactions for the oxidation of [Ru(tpy)
(bpy)(OH2)]2+ to [Ru(tpy)(bpy)(O)]2+ were investigated as early
as 1984,48 but their relevance to catalytic activity for water oxidation
was only recently recognized.32 Meyer30,34,49ꢀ51 and Berlinguette35,36
have elegantly investigated the kinetics and mechanisms of water
oxidation with their [Ru(NNN)(NN)(OH2)]2+ catalysts (NNN =
tpy or tpy-like molecule such as 2,6-bis(benzimidazol-20-yl)pyridine,
NN = bpm, bpy, 4,40-(HOOC)2bpy, and 4,40-(MeO)2bpy).
The major pathway seems to proceed through a [RuIVdO]2+
species produced by proton-coupled electron transfer (PCET)
Artificial photosynthesis holds the promise of providing solar
fuels from water powered by sunlight.1ꢀ12 Considerable progress
has been made in the development of oxidation catalysts that
might serve as one component of such a solar conversion
system.13ꢀ46 Many of these catalysts are based on ruthenium poly-
pyridyl complexes, and most of these complexes can be roughly
categorized as dinuclear15,17ꢀ26 or mononuclear28ꢀ34 catalysts.
Initially, it was thought that two adjacent metal centers would be
needed to enable the eventual formation of the critical OꢀO bond.
More recently, it has been discovered that dioxygen can be produced
by the nucleophilic interaction of a water molecule with a single
RudO center such as [Ru(tpy)(bpy)(O)]3+ and [Ru(tpy)-
(bpm)(O)]3+ (tpy = 2,20:60,200-terpyridine, bpy = 2,20-bipyridine,
bpm = 2,20-bipyrimidine), and this finding has generated some
interesting mechanistic interpretations.
In one type of mononuclear Ru catalyst, a tetradentate ligand
occupies the equatorial plane of the molecule, and two mono-
dentate pyridines occupy the axial sites.25,28,45ꢀ47 The pyridines
do not exchange with a water (solvent) molecule, prompting
Thummel and co-workers to propose a mechanism that involves
two initial one-electron oxidations to produce a 6-coordinate,
Received: April 8, 2011
Published: August 04, 2011
r
2011 American Chemical Society
14649
dx.doi.org/10.1021/ja203249e J. Am. Chem. Soc. 2011, 133, 14649–14665
|

Journal of the American Chemical Society
ARTICLE
Chart 1. Structure of [RuIIꢀOH2]2+
’ EXPERIMENTAL SECTION
Materials. The complex [RuIIꢀOH2]2+ was prepared as described
previously.28 All chemicals used were 99% grade or higher and used without
further purification, except for the following. Trifluoromethanesulfonic acid
was ReagentPlus grade (Aldrich), distilled under vacuum and immediately
used to prepare a 1.0 M aqueous stock solution. The solution was stored
under refrigeration (ca. 10 °C). Ceric ammonium nitrate was dried under
vacuum and used to prepare acidic (triflic acid) solutions. Ceric solutions
were titrated with a known concentration of sodium oxalate to determine
accurate concentrations of the ceric ion. Aqueous solutions were prepared
with distilled water that had been passed through a Millipore ultrapurifica-
tion system. Blanket gases (N2O, Ar) were UHP grade (99.999%).
Spectroscopic Measurements. UVꢀvis spectra were measured
on a Hewlett-Packard 8452A diode-array spectrophotometer or Varian
Cary 500 dual-beam spectrometer. NMR spectra were measured on a
Bruker UltraShield 400 MHz spectrometer. Gas analysis was performed
with a QMS 100 residual gas analyzer (Stanford Research). An Applied
Photophysics stopped-flow system configured for two-syringe mixing
was used to carry out stopped flow experiments. Spectral changes were
detected with a PDA detector in the 300ꢀ800 nm range. Single
wavelength kinetic traces were recorded using a PMT detector. In a
typical experiment, one syringe was charged with an excess of Ce(IV)
solution in 0.2 M triflic acid, and the other syringe contained a solution
of the ruthenium complex in water. The spectral change was recorded
after equal amounts of both solutions were mixed in a 1 cm stopped-flow
cell. The dead time of the mixing setup was 2 ms.
Resonance Raman spectroscopy. Resonance Raman spectra
were measured using focused output (1 mm) of the second harmonic of a
YAG:Nd3+ laser (532 nm; Continuum, Powerlite 7010, 10 ns, 2 Hz, 20 mJ/
pulse). The front-scattered light was collected by a 100 mm achromatic lens
and collimated to a 200 mm achromatic lens. The second lens was
f-matched with the SpectraPro 300i spectrograph (Acton) equipped with
a liquid nitrogen cooled Spec-10 CCD (Acton). A long pass filter
(RazorEdge LP03-532 from Semrock) was placed in front of the spectro-
graph entrance slits to eliminate stray light from the laser excitation. The
spectrograph and CCD control as well as spectral acquisition was archived
with WinSpec software (Acton). The spectra were recorded at CCD
temperature of ꢀ100 °C and corrected for background noise and cosmic
ray hits. The spectrograph was calibrated using known Raman shift values
for the acetonitrile/toluene mixture. The calibration was performed before
and after each measurement using an external standard, and the deviation
between two measurements was found to be always less than 1 cmꢀ1. The
sample for Raman measurements was prepared by electrolyzing ca.
0.3ꢀ0.4 mM aqueous solutions of [RuIIꢀOH2]2+ and subsequent con-
centration of these solutions to ca. 1.5ꢀ2 mM under vacuum. The resulting
concentrated solutions were placed into a syringe which was connected to a
PEEK capillary tube. The end of the capillary was mounted in the focal plane
of the 100 mm collection lens, and a drop of liquid was carefully expelled
from the end of the capillary. The capillary was aligned using an XꢀY
translation stage in such way that the focused laser beam was going through
the hanging drop of liquid. A typical spectrum was recorded using 10 s
accumulation time and 16 averages, followed by replacement of the drop
with the fresh sample. The data postprocessing included the averaging of at
least 8 spectra followed by a baseline correction. The use of the hanging
drop technique allowed recording Raman spectra unobstructed by intense
Raman lines originated from a cell material. UVꢀvis spectra taken before
and after Raman experiments showed no significant decomposition of the
sample due to exposure to the laser radiation.
reactions, which is further oxidized to [RuVdO]3+, a key inter-
mediate that reacts with water to produce an OꢀO bond. Both
groups proposed possible competing reaction pathways that involve
[RuIVꢀOO]2+, [RuVꢀOO]3+, and [RuIIꢀO2H2]2+ intermediate
species, and Berlinguette35,36 proposed additional pathways via
disproportionation of [RuIVdO]2+ and the oxidative addition of
[CeIV(NO3)5]ꢀ to [RuIVdO]2+ to form [RuIVꢀOO]2+. While the
RuVdO/RuIVdO potentials of these Ru centers (1.77ꢀ1.89 V) are
close to the thermodynamic limit of oxidation by Ce(IV) (the CeIV/
CeIII potential is 1.61 V in 1.0 M HNO3), the addition of 1 equiv of
Ce(IV) to [RuIVdO]2+ is proposed to generate [RuVdO]3+,
which, in turn, undergoes OꢀO bond formation. However, spectro-
scopic, electronic, and geometric characterization of these RuꢀOO,
RuꢀOOH, and RuꢀO2H2 species remains unclear.
In the current study, we examine the complex [RuIIꢀOH2]2+
(Ru = Ru(NPM)(pic)2, NPM = 4-t-butyl-2,6-di-(10,80-naphthyrid-
20-yl)-pyridine, pic = 4-picoline, Chart 1), which is one of the first
reported Ru-based mononuclear water oxidation catalysts.28 Unlike
the [Ru(tpy)(NN)(H2O)]2+ (where NN represents 2,20-bipyri-
dine and its derivatives) species mentioned above, in this catalyst the
water molecule occupies a fourth binding site in the equatorial plane.
The X-ray single-crystal structure shows that the coordinated water
molecule is H-bonded to a noncoordinated nitrogen on one of the
two 1,8-naphthyridyl rings appended to the central pyridine.28 The
complex has two axial picolines and in that respect resembles
the first type of catalyst having a tetradentate equatorial ligand.
The presence of an internal basic site may make this complex well
suited for proton-coupled electron transfer events.
Here we present experimental results on the one- and two-
electron oxidized species, the pathways of their formation, their
electronic structures, and their reactivities and discuss the mechan-
ism of water oxidation catalyzed by [RuIIꢀOH2]2+ based on
collective observations from electrochemistry, UVꢀvis and reso-
nance Raman spectroscopy, pulse radiolysis, stopped flow, ESI-MS,
H218O labeling experiments, and theoretical calculations. The most
intriguing result is the electrochemical formation of [RuIVꢀOO]2+
observed with a Coulomb count corresponding to a net four-electron
oxidation when 1260 mV vs NHE at pH 1 or 1150 mV at pH 6 was
applied to a solution containing [RuIIꢀOH2]2+, despite the fact that
the RuVdO/RuIVdO potential is 1420 mV. (NB: We use the
formal oxidation state of “IV” for the OꢀO bonded species in
accordance with the convention employed in various previous
publications on water oxidation with Ru complexes; however, it
could be a species between [RuIVꢀOO]2+ and [RuIIIꢀ(OOꢀ•)]2+)
The red colored species produced by this electrolysis requires 1
equiv of oxidant to produce O2. Without the oxidant, the species
slowly decomposes over two weeks to form [RuIIꢀOOH]+ that
is identified by ESI-MS and other spectroscopic techniques.
Electrochemical Measurements. All potentials are reported vs
NHE. Electrochemical measurements of redox reactions of [RuIIꢀOH2]2+
were conducted with a BAS 100b electrochemical analyzer from Bioana-
lytical Systems. Cyclic voltammograms and square wave voltammograms
in aqueous solutions were measured using 0.5 mM solutions of
14650
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
[RuIIꢀOH2]2+ in water containing 20% 2,2,2-trifluoroethanol, 0.1 M
sodium trifluoromethanesulfonate, and 10 mM sodium phosphate (in the
2ꢀ10 pH range) or 0.1 M trifluoromethanesulfonic acid (at pH 1). In
contrast to acetonitrile, trifluoroethanol does not exchange with water in
the ruthenium coordination sphere but still enhances the solubility of
[RuIIꢀOH2]2+. A glassy carbon disk was used as a working electrode, a
platinum wire as a counter electrode, and Ag/AgCl (sat. NaCl) as a
reference electrode. For measurements in acetonitrile, 0.1 M [(t-Bu)4N]-
[PF6] was used as an electrolyte, and a nonaqueous Ag/AgNO3 electrode
was used as a reference (potentials converted to NHE using an external
Fc+/Fc standard(640 mVvsNHE52). The scanrate in all CVexperiments
was 100 mV sꢀ1 unless noted otherwise. In bulk electrolysis experiments,
glassy carbon was replaced with a platinum gauze electrode, and the
platinum counter electrode was isolated with a Vycor frit.
Pulse Radiolysis. Pulse radiolysis studies were carried out using the
BNL 2 MeV van de Graaff accelerator using electron pulses (pulse width
of 40ꢀ500 ns) that led to irradiation doses of 10ꢀ1000 rad (ca. 0.5ꢀ5
μM primary radicals) generated in solution. A thiocyanate solution (0.01
M KSCN, 0.026 M N2O) was used for dosimetry taking G((SCN)2ꢀ) =
6.13 (G = number of species formed per 100 eV of energy absorbed by
the solution) and ε472nm = (7590 ( 230). The optical path of the cell was
2 cm. All measurements were carried out in aqueous solutions contain-
ing 10 mM NaHCO3/Na2CO3 at pH 10 saturated with N2O at 25 °C.
Under these conditions, the conversion of the primary radicals to the
carbonate radical was complete by the first microsecond. Quoted rate
constants have an error of ca. 15%. All rates measured in the pulse
radiolysis studies are averages of at least three measurements. The pH of
the solution was adjusted by addition of sodium hydroxide or
trifluoromethanesulfonic acid.
Oxygen Measurements. Gases produced during catalytic water
oxidation were analyzed with a QMS 300 Gas Analyzer (Stanford
Research Systems). The headspace of the reaction vessel was connected
to a capillary, leading to the inlet of the gas analyzer. The reaction vessel
was connected to the argon supply line to maintain constant pressure
inside the vessel. Deaerated ceric solutions were injected into the
deaerated catalyst solution through a septum. The analyzer was set up
to monitor m/z signals which corresponded to oxygen, nitrogen, and
carbon dioxide. In experiments with isotopically labeled water, an 85%
enrichment of 18O was used.
Electronic Structure Calculations. All calculations were per-
formed with the Gaussian 09 program package,53 and all DFT calculations
were carried out using the B3LYP hybrid functional.54ꢀ58 We employed
the B3LYP/ECP28MWB(1f,0 g)[Ru];59,60 6-311G(d,p)[H,C,N,O]61,62//
B3LYP/ECP28MWB(1f,0g)[Ru]; 6-31G(d)[O,N]; 6-31G[H,C] met-
hod.63ꢀ68 The ECP28MWB(1f,0g) is a triple-ζ plus polarization and
diffuse functions basis, (8s7p6d1f)/[6s5p3d1f], for Ru that is common to
the small and large basis sets. The 6-31G(d)[O,N]; 6-31G[H,C] small basis
for the nonmetal atoms is a double-ζ plus polarization basis for O and N and
double-ζ for C and H.64 The large basis for the nonmetal atoms is a triple-ζ
plus polarization basis for H, C, N, and O. For the case of [RuIIꢀOH2]2+,
the small basis has 580 functions, and the large basis has 1105 functions.
For all species considered in the proposed catalytic water oxidation
cycle, a single explicit solvent water molecule was included in the
calculation to provide hydrogen bonding to the water species bound
to the metal center. The effect of the bulk solvent was included in all
geometry optimization and vibrational frequency calculations of species
with standard states in solution through the use of the C-PCM
method69ꢀ71 using UAKS radii. The absolute standard Gibbs free
energy in aqueous solution (or in the gas phase for H2 and O2), along
with appropriate values of the absolute free energy of the solvated proton
and gas-phase electron, were used to calculate standard free energy
changes of reaction and standard reduction potentials (see “Details of
Electrochemistry in Aqueous Solution” in Supporting Information).
The reduction potentials were referenced to a value for the absolute
Figure 1. Gas evolution analysis during chemical (Ce(IV) in triflic acid,
pH 1) oxidation of [RuIIꢀOH2]2+ (50 μM) in H218O. The amount of
Ce(IV) relative to the catalyst was 80 equivalents.
potential of the normal hydrogen electrode (NHE) obtained using the
experimental absolute free energies of H2 and an electron in the gas
phase together with the experimental value of the absolute free energy of
the solvated proton in water. Representative thermodynamic cycles are
provided in Schemes S1ꢀS3 (Supporting Information). Additional TD-
B3LYP calculations were carried out for selected species to aid in their
identification from UVꢀvis spectra in electrochemical and pulse radi-
olysis experiments.
In cases in which there was a question of whether the ground electronic
state of a species is a singlet or a triplet, three kinds of additional
calculations were carried out. Each of these additional methods goes
beyond standard densityfunctionaltheory intreatingmulticonfigurational
effects. The first of these was a broken symmetry (BS) B3LYP calculation
to see if there was a singlet state with antiferromagnetically coupled
unpaired electrons with lower energy than the closed-shell singlet. The
second was a CASSCF calculation of both the singlet and triplet species.
The CASSCFcalculation explored the possibility of there being important
valence configurations, e.g., GVB-CI electron pairs, not accounted for by
the single-configuration B3LYP method. The third was a multireference
MP2 calculation on the [RuIVdO]2+ species based on the CASSCF wave
function that includes the effect of dynamic correlation in addition to the
static valence correlation in the reference function.
’ RESULTS
Evidence for Catalytic Water Oxidation Mediated by
[RuIIꢀOH2]2+. Catalytic performance of water oxidation catalysts
is typically evaluated by analyzing catalytic current during the
electrochemically driven oxidation reaction or by utilizing Ce-
(IV) salts such as ceric ammonium nitrate as a sacrificial oxidant.
While ceric salts are very potent oxidation reagents (the E° of
CeIV/CeIII is 1.3ꢀ1.9 V vs NHE),72 the potential of the CeIV/
CeIII couple is strongly dependent on the nature of the acid and
its concentration.72ꢀ75
As was previously reported, the presence of catalytic amounts
of [RuIIꢀOH2]2+ in an acidic solution of Ce(IV) leads to oxygen
formation with TN = 26028,29 as determined by monitoring the
formation of O2 gas. In the current study, the analysis of the
gaseous products was extended to include not only oxygen and its
isotopes but also other gases (Figure 1). The absence of measur-
able amounts of carbon dioxide indicates that decomposition of
[RuIIꢀOH2]2+ involving oxidation of the organic ligands does not
take place under the reaction conditions (ca. 20 TN). CO2 evolution
14651
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
Figure 2. pH-dependent electrochemistry of [RuIIꢀOH2]2+: Pourbaix diagram (left). The dotted line corresponds to the calculated potential for
[RuIVdO]2+ + H2O ꢀ eꢀ f [RuIIIꢀOOH]2+ + H+, and the red stars indicate the conditions of bulk electrolysis experiments. Square wave
voltammograms recorded in the pH range 0.5ꢀ9.5 (right).
has been previously observed as a result of ruthenium polypyridyl
complex oxidation and was attributed to ligand decomposition.13
Electrochemistry of [RuIIꢀOH2]2+. During electrochemical
oxidation of [RuIIꢀOH2]2+ in aqueous acidic solution (pH 1),
the catalytic current started to develop above ca. 1500 mV and
increased as the concentration of [RuIIꢀOH2]2+ increased as
shown in Figure S1 (Supporting Information).
the section “Characterization of One- and Two-Electron Oxidized
Intermediates”. The disproportionation of the one-electron-oxidized
species in the pH range above 2.9 indicates that the potential of the
[RuIVdO]2+/[RuIIIꢀOH]2+ couple is equal to or lower than that of
the [RuIIIꢀOH]2+/[RuIIꢀOH2]2+ couple, and the process corre-
sponding to ꢀ59 mV/pH in the region of pH 2.9ꢀ12 of the Pourbaix
diagram corresponds to two-electron oxidation.
The situation is different for the oxidation of [RuIIꢀOH2]2+
on the lower side of the pKa of [RuIIIꢀOH2]3+. The absence of a
dependence of the potential on pH indicates that the first step is
not proton coupled and thus corresponds to the formation of
[RuIIIꢀOH2]3+ (eq 3); however, we did not observe a redox
couple with a slope of ꢀ118 mV/pH corresponding to the
formation of [RuIVdO]2+. This differs from the case of
[RuII(tpy)(bpy)(OH2)]2+ and [RuII(tpy)(bpm)(OH2)]2+ in which
the first oxidation of [RuII(tpy)(bpy)(OH2)]2+ (or [RuII(tpy)
(bpm)(OH2)]2+) toform[RuIII(tpy)(bpy)(OH2)]3+ (or[RuIII(tpy)
(bpm)(OH2)]3+) is followed by a pH-dependent redox couple
with the slope of ꢀ118 mV/pH corresponding to the formation
of [RuIV(tpy)(bpy)(O)]2+ (or [RuIV(tpy)(bpm)(O)]2+). The
nonobservation of the corresponding process in the case of
cis-[Ru(CNC)(n-Bu-CN)(OH2)]2+ was attributed previously76
to slow electrode kinetics.
2þ
2þ
½RuII ꢀ OH2ꢂ ꢀ eꢀ ꢀ Hþ f ½RuIII ꢀ OHꢂ
ð1Þ
ð2Þ
2þ
2þ
½RuIII ꢀ OHꢂ ꢀ eꢀ ꢀ Hþ f ½RuIVdOꢂ
2þ
3þ
½RuII ꢀ OH2ꢂ ꢀ eꢀ f ½RuIII ꢀ OH2ꢂ
ð3Þ
3þ
2þ
½RuIII ꢀ OH2ꢂ ꢀ eꢀ ꢀ 2Hþ f ½RuIVdOꢂ
ð4aÞ
ð4bÞ
ð5Þ
3þ
3þ
½RuIII ꢀ OH2ꢂ ꢀ 2eꢀ ꢀ 2Hþ f ½RuVdOꢂ
2þ
3þ
½RuIVdOꢂ ꢀ eꢀ f ½RuVdOꢂ
The pH dependence of the electrochemical oxidation of
[RuIIꢀOH2]2+ in aqueous solution is summarized in Figure 2
and eqs 1ꢀ5. Ruꢀaqua polypyridyl complexes such as
[Ru(tpy)(bpy)(OH2)]2+ are known to lose protons and elec-
trons and easily reach higher oxidation states.48 However, the
Pourbaix diagram of [RuIIꢀOH2]2+ (Figure 2) is rather simple
compared to that of [Ru(tpy)(bpy)(OH2)]2+.48 The pH-depen-
dent part of the Pourbaix diagram has a slope of ca. ꢀ59 mV/pH
in the region of pH 2.9ꢀ12. The formation of [RuIIIꢀOH]2+ by
a coupled one-proton and one-electron process is normally
invoked for such a phenomenon; however, a coupled two-proton
and two-electron process to form [RuIVdO]2+ is also possible, as
found for cis-[Ru(CNC)(n-Bu-CN)(OH2)]2+ (CNC = 2,6-bis-
(n-butylimidazol-20-ylidene)pyridine, n-Bu-CN = 2-(n-butylimidazol-
20-ylidene)pyridine).76 The UVꢀvis spectrum of [RuIIꢀOH2]2+ did
not change upon acidꢀbase titration in the range of pH 1ꢀ13.5,
indicating that there is no deprotonation of the coordinated water or
protonation of the noncoordinating N atom of the naphthyridine part
of the ligand, which is consistent with the Pourbaix diagram. The
Pourbaix diagram shows that the pKa of the one-electron-oxidized
species, [RuIIIꢀOH2]3+, is around 2.9. We will discuss in detail the
disproportionation of [RuIIIꢀOH]2+ observed by pulse radiolysis in
A second oxidation process (eq 5) at 1420 mV in the Pourbaix
diagram was found to be pH-independent over a wide range of
proton concentrations (pH 0.9ꢀ10). It should be noted that this
[RuVdO]3+/[RuIVdO]2+ potential is much lower than those of the
[RuII(NNN)(NN)(OH2)]2+ species (e.g., [RuII(tpy)(bpy)(OH2)]2+
1800 mV,35,36 [RuII(tpy)(bpm)(OH2)]2+ 1650 mV).34 It seems
that the formation of [RuIVdO]2+ at low pH is followed by the
production of [RuVdO]3+ at relatively low applied potential. This may
lead to a high catalytic activity for water oxidation at low overpotential.
Finally, in the pH region below ca. 0.9, the signal around 1420
mV starts overlapping with the catalytic current, making it
difficult to determine the peak position, but it appears to have
a ꢀ59 mV/pH slope, which corresponds to two-electron oxida-
tion of [RuIIIꢀOH2]3+ coupled to the loss of two protons
leading to the formation of [RuVdO]3+ (eq 4b).
Theoretical Investigation of Possible Intermediates. To
identify possible intermediates, we calculated geometries, energetics,
and spectroscopic properties of [RuIIꢀOH2]2+, [RuIIIꢀOH]2+,
1[RuIVdO]2+, 3[RuIVdO]2+, [RuVdO]3+, [RuIIIꢀOOH]2+,
3[RuIVꢀη1-OO]2+, 1[RuIVꢀη2-OO]2+, 1[RuIVꢀcyc-OON]2+
14652
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
Table 1. Calculated Standard Free Energies of the Species Derived from [RuIIꢀOH2]2+ Involved in the Catalytic Water Oxidation
Using the Large Basisa
system
reservoir
2H2O
2H2O, eꢀ
rel. total G* (eV vs NHE)
ΔG* (eV)b
1[RuIIꢀOH2]2+
0.000
0.961
1.231
2.532
3.121
4.196
4.397
5.551
6.125
6.500
4.760
5.991
0.000
0.412
0.000
ꢀ0.024
2[RuIIIꢀOH2]3+
2[RuIIIꢀOH]2+
0.961
1.231
1.571
2.159
1.664
1.865
1.355
1.929
2.304
ꢀ0.792
0.440
2H2O, H+, eꢀ
2H2O, 2H+, 2eꢀ
2H2O, 2H+, 2eꢀ
2H2O, 2H+, 3eꢀ
H2O, 3H+, 3eꢀ
H2O, 4H+, 4eꢀ
H2O, 4H+, 4eꢀ
H2O, 4H+, 4eꢀ
O2, 4H+, 4eꢀ
O2, 5H+, 5eꢀ
3[RuIVdO]2+
1[RuIVdO]2+
2[RuVdO]3+
2[RuIIIꢀOOH]2+
3[RuIVꢀη1-OO]2+
1[RuIVꢀη2-OO]2+
1[RuIVꢀcyc-OON]2+ c
1[RuIIꢀOH2]2+
2[RuIIIꢀOH]2+
1[RuIIꢀHOOH]2+ c
1[RuIIꢀOOH2]2+ c,d
1[RuII(NPMH+)(pic)2ꢀOH2]3+ c
1[RuIIꢀOH2]2+
H+
Net system reaction: 1[RuIIꢀOH2]2+ f 1[RuIIꢀOH2]2+, ΔG* = 0 or
2[RuIIIꢀOH]2+ f 2[RuIIIꢀOH]2+, ΔG* = 0
Net reservoir reaction: 2H2O f O2 + 4H+ + 4eꢀ
ΔG* = ꢀ4FEo(O2 + 4H+ + 4eꢀ f 2H2O) [should be 4.92 eV, calcd 4.76 eV]
a The preceding superscript denotes spin multiplicity. b The free-energy difference between a listed species and the lowest-energy one-electron-reduced
species. c These species are not likely involved in the water-oxidation reaction but are listed here for energy comparison purposes. d According to our
DFT calculations, one of the protons of this species is transferred to form [RuII(NPMH+)(pic)2ꢀOOH]2+ (see Supporting Information).
(i.e., a [RuIVꢀη1-OO]2+ species with a strong interaction with
the noncoordinating N atom of the naphthyridine ligand),
[RuIIꢀHOOH]2+, and [RuIIꢀOOH2]2+ (interestingly, one of the
protons of this species is transferred to form [RuII(NPMH+)-
(pic)2ꢀOOH]2+). The calculated standard free energies are listed
in Table 1. The superscript number at the beginning of each species
designation represents the spin multiplicity of that species. The
important geometric and spectroscopic parameters are summarized
in Table S1 (Supporting Information). We will discuss these results
when we describe the experimental data. All UVꢀvis “stick spectra”
from TD-DFT calculations were convoluted with a Gaussian
broadening function with a width of 0.18 eV.
Characterization of One- and Two-Electron Oxidized Inter-
mediates. Ce(IV) is frequently used as a sacrificial oxidant for
catalyzed water oxidation. The reaction between [RuIIꢀOH2]2+
and Ce(IV) was examined at pH 1 in the presence of triflic acid. As seen
in Figure 3a, titration of dilute acidic solutions of [RuIIꢀOH2]2+ (8 ꢁ
10ꢀ7 M, 0.1 M triflic acid) with Ce(IV) shows gradual removal of the
starting material and formation of the oxidized species upon consump-
tion of ca. 2.7 equiv of Ce(IV). The presence of several clear isosbestic
points indicates that the final product is produced without significant
accumulation of the intermediate one-electron oxidized species. Addi-
tional Ce(IV) does not lead to any change in the spectrum. The
formation of the oxidized species is reversible in dilute acidic solutions
(<10ꢀ6 M), and [RuIIꢀOH2]2+ can be almost completely recovered
upon the addition of ferrous ammonium sulfate as a reducing reagent.
The small amount of the starting complex corresponding to the
incomplete recovery of the (green) spectrum vs the (blue) spectrum
in Figure 3b after back-titration with ferrous ammonium sulfate might
be due to a small amount of precipitation of [RuIIꢀOH2]2+ over time
or the chemical reactivity of the oxidized species.
Analysis of the Pourbaix diagram (Figure 2) shows that at pH 1
[RuIVdO]2+ may exist in equilibrium with [RuVdO]3+. We
estimate the ratio of [RuVdO]3+/[RuIVdO]2+ at pH 1 to be
about 1:1 owing to only a few millivolts separation between the two
species. This may be the reason we needed to add more than 2 equiv
of Ce(IV) to obtain the final spectrum (blue) in Figure 3a.
The rate of the reaction between [RuIIꢀOH2]2+ and an excess of
Ce(IV) was monitored by following the spectral change in the 2ꢀ
100 ms time window using a stopped-flow apparatus. The disap-
pearance of the absorption features characteristic of [RuIIꢀOH2]2+
was observed, but no intermediate one-electron-oxidized species was
observed (Figure S2, Supporting Information). From the rate of
disappearance of the MLCT band of [RuIIꢀOH2]2+ at ca. 610 nm, a
rate constant of (2.0 ( 0.3) ꢁ 104 Mꢀ1 sꢀ1 was obtained. It is not
unusual for Ru(II) polypyridyl complexes to react rather slowly
with CeIV, and rate constants as low as 2 ꢁ 102 Mꢀ1 sꢀ1 (at pH 1)
have been reported.34 The one-electron-oxidized species
[RuIIIꢀOH2]3+ was not observed spectroscopically in our experi-
ments at pH 1, similar to findings reported previously.34
Observation of One- and Two-Electron Oxidized Intermedi-
ates by Pulse Radiolysis. Ce(IV) oxidation at pH 1 may be
complicated owing to the recent finding of its noninnocent nature as
a one-electron oxidant.36,77 We therefore studied the formation and
reactivity of the one- and two-electron-oxidized species to elucidate
these properties in the high pH range using the pulse radiolysis
technique. Pulse radiolysis allows the production of various redox
active species in solution and provides a means for following
chemical reactivity using time-resolved UVꢀvis spectroscopy. The
carbonate radical was chosen as the sacrificial one-electron oxidant.
The redox potential of the carbonate radical (1.59 V vs NHE)78 is
similar to the potential of Ce(IV) with HNO3 (1.61 V vs NHE at
14653
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
Figure 3. (a) Titration of [RuIIꢀOH2]2+ (8 ꢁ 10ꢀ7 M, 10 cm path cell) with Ce(IV) and (b) back-titration of 2-electron-oxidized species with ferrous
ammonium sulfate in water at pH 1.
Figure 4. Experimental (left) and calculated (right) UVꢀvis spectra of [RuIIꢀOH2]2+ (black), one-electron oxidized species (red), and two-electron
oxidized species (blue). The calculated spectrum for the two-electron oxidized species is that of 3[RuIVdO]2+.
pH 0).72 The carbonate radical can be produced efficiently only at
pH above 9 since the pKa of the bicarbonate ion is around 10.
Creating oxidative conditions during radiolysis in acidic solutions,
below ca. pH 3, where the yield of the hydrogen atom becomes
significant, proves problematic. Because the hydrogen atom exhibits
diverse reactivity, its production in this particular study was avoided.
The reaction between the carbonate radical and [RuIIꢀOH2]2+
at pH 10 yielded the one-electron-oxidized species (eq 6) with a
distinct absorption spectrum (Figure 4, left) which agrees with
the spectrum of [RuIIIꢀOH]2+ obtained using TD-DFT-
(B3LYP) calculations (Figure 4, right). The one-electron-oxi-
dized species is produced with a rate constant of (2.8 ( 0.5) ꢁ
109 Mꢀ1 sꢀ1 (with ionic strength 0.06 M).
The disappearance of the absorption at 450 nm (attributed
to [RuIIIꢀOH]2+) is concomitant with partial recovery of the
610 nm band and is described by a second-order rate law (Figure
S3ꢀS4, Supporting Information). This reaction was assigned to
disproportionation of the one-electron-oxidized species to produce
1 equiv of the starting material and 1 equiv of the two-electron-
oxidized species (eq 6a). These observations suggest that the RuIII/II
couple is equal to or more positive than the RuIV/III couple. As will be
discussed later, theoretical results on these potentials are consistent
with this interpretation (see the “Theoretical Analysis of the Electro-
chemical Oxidation of [RuIIꢀOH2]2+” section below).
The reaction of [RuIIꢀOH2]2+ with an excess of the carbonate
radical was also studied in an attempt to produce a two-electron-
oxidized species (eqs 6 and 7). This reaction can be described in
terms of two stepwise one-electron oxidations. In addition, the
carbonate radical is known to react with water according to eq 8,
which is second order in the carbonate radical. The data
measured for reactions between [RuIIꢀOH2]2+ and an excess
of the carbonate radical were fitted by a kinetic model based on
eqs 6ꢀ 8 using numerical integration.79 The rate constant
corresponding to the first oxidation was fixed at the value, k6 =
2þ
2þ
CO3 þ ½RuII ꢀ OH2ꢂ f HCO3 þ ½RuIII ꢀ OHꢂ
ð6Þ
ð6aÞ
ð7Þ
•ꢀ
ꢀ
2þ
2þ
2þ
2½RuIII ꢀ OHꢂ f ½RuII ꢀ OH2ꢂ þ ½RuIVdOꢂ
2þ
2þ
CO3 þ ½RuIII ꢀ OHꢂ f HCO3 þ ½RuIVdOꢂ
•ꢀ
ꢀ
2H2O þ 2CO3 f 2CO3 þ H2O2 þ 2Hþ
ð8Þ
•ꢀ
2ꢀ
(2.8 ( 0.5) ꢁ 109 Mꢀ1
s
ꢀ1, that was obtained independently
The one-electron-oxidized species, [RuIIIꢀOH]2+, produced in
the pulse radiolysis experiment is not stable and slowly disappears
on the time scale of ca. 1 min (kdisp = (6.5 ( 0.5) ꢁ 103 Mꢀ1 sꢀ1).
from experiments under pseudo-first-order conditions (with an
excess of [RuIIꢀOH2]2+). The traces obtained from these fits
agree well with experimental data. For example, the behavior of
14654
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
the 450 nm trace is typical of an intermediate species B in an
A-to-B-to-C kinetic model and is well described by the proposed
mechanism (Figure S5, Supporting Information). Further kinetic
analysis based on the proposed model indicates that nearly
complete conversion of [RuIIꢀOH2]2+ into the two-electron-
oxidized species should be achieved using a minimum of 2.5
equiv of the carbonate radical relative to [RuIIꢀOH2]2+. The rate
constants obtained from the fits with the kinetic model at λ =
450 nm are: k6= (2.8 ( 0.5) ꢁ 109 Mꢀ1 sꢀ1; k7 = (2.6 ( 0.3) ꢁ
107 Mꢀ1 sꢀ1;k8 = (7.9 ( 0.9) ꢁ 106 Mꢀ1 sꢀ1. The deviation in rate
constants was within 15% for all wavelengths. Our value of the
CO3•ꢀ decay (k8) was in good agreement with reported values.80,81
Interestingly, the spectra of the two-electron-oxidized species
produced at pH 1 by Ce(IV) and at pH 10 by the carbonate
radical match fairly well (see Figure S6, Supporting Information),
indicating that the same species can be formed over a wide pH
range. Since the Pourbaix diagram of [RuIIꢀOH2]2+ (Figure 2)
reveals that several species might be in equilibrium (i.e.,
[RuIIIꢀOH2]3+, [RuIVdO]2+, and [RuVdO]3+) at pH 1, and
the final spectrum of the two-electron-oxidized species is ob-
tained by an addition of 2.7 equiv of Ce(IV), the spectra of
[RuIVdO]2+ and [RuVdO]3+ might be very similar. We should
point out again that additional Ce(IV) beyond 2.7 equiv does not
change the spectrum. In fact, Berlinguette et al. have reported
that the spectra of Ru (IV) and Ru(V) species of Ru(trp)(bpy)-
(O) are very similar.35,36
Figure 5. UVꢀvis absorption spectrum of [RuIIꢀOH2]2+ (black) and
species formed after bulk electrolysis of [RuIIꢀOH2]2+ in water at pH 6
at 1150 mV vs NHE (red). Inset: Calculated spectra of [RuIIꢀOH2]2+,
3[RuIVꢀη1-OO]2+ (blue), and 1[RuIVꢀη2-OO]2+ (green).
potential (i.e., until the final current was less than 1% of the initial
current) yielded a red colored solution with a Coulomb count
corresponding to the net four-electron oxidation of [RuIIꢀOH2]2+,
despite applying potentials much less than that of the [RuVdO]3+/
[RuIVdO]2+ couple (1420 mV). The results were similar for both
high and low pH experiments.
Theoretical characterization of [RuIVdO]2+ is described in
the Supporting Information. It should be noted that a weak band
is experimentally observed around 670 nm (Figure 3), which
probably corresponds to the 747 nm band predicted theoretically
for the triplet state (Figure S7, Supporting Information).
The species produced during the bulk electrolysis experiment
has a distinct absorption spectrum with the metal-to-ligand
charge transfer (MLCT) band (with some contribution from
tpy intraligand transitions) blue-shifted ca. 80 nm compared
to [RuIIꢀOH2]2+ (Figure 5). The calculated spectrum of
3[RuIVꢀη1-OO]2+ seems to match well; however, we cannot
Characterization of the Two-Electron-Oxidized Species in
CH3CN. While dilute (<10ꢀ5 M) acidic solutions of the two-
electron-oxidized species can be characterized by UVꢀvis spec-
troscopy and ESI-MS spectrometry immediately after their
formation, more concentrated samples are not stable in aqueous
media. To characterize the two-electron-oxidized species in the
absence of water, a sample was prepared by rapid mixing of
aqueous [RuIIꢀOH2]2+ with 2 equiv of Ce(IV) in triflic acid (pH 1)
followed by rapid precipitation with NH4PF6. The resulting solid
precipitate was vacuum-dried and dissolved in dry acetonitrile
and characterized by ESI-MS and UVꢀvis spectroscopy. The
UVꢀvis and MS spectra were consistent with [RuIVdO]2+ but
not [RuIIꢀNCCH3]2+. We also prepared [RuII(tpy)(bpy)(O)]2+
by bulk electrolysis in 0.1 M deuterated triflic acid in D2O at 1370
mV vs NHE. This species is stable even at high concentration. The
1H NMR spectra of [RuII(tpy)(bpy)(O)]2+ in D2O and
[RuIVdO]2+ in CD3CN are shown in Figures S13 and S14
(Supporting Information), respectively. Both spectra are charac-
teristic of a paramagnetic species, with several resonances in the
negative parts per million region that are typical of ruthenium oxo
complexes,82ꢀ87 consistent with our theoretical results with the
assignment as 3[RuIVdO]2+ but not 3[RuIIIꢀO•ꢀ]2+. However, a
low-spin RudO complex has recently been reported.88 The cyclic
voltammogram of [RuIVdO]2+ in CH3CN shows a reversible redox
couple with E1/2 = 1380 mV (Figure S15, Supporting Information).
Products of the Electrochemical Four-Electron Oxidation of
[RuIIꢀOH2]2+. Aqueous solutions of [RuIIꢀOH2]2+ were electro-
lyzed at 1260 mV vs NHE at pH 1 (0.1 M triflic acid) and 1150 mV at
pH 6 (10 mM sodium phosphate). Both solutions contained 20% by
volume of trifluoroethanol to maintain solubility of the complex at a
concentration of about 1 mM. Exhaustive electrolysis at constant
3
exclude the possibility of a mixture of [RuIVꢀη1-OO]2+ and
1[RuIVꢀη2-OO]2+. The ESI-MS analysis of the solution ob-
tained after electrolysis indicates the presence of a species with
m/z = 355 and 710, which correspond to [RuIVꢀ(16O16O)]2+
(Mꢀ1 = 710). The isotope pattern of this proposed structure
matches well the observed pattern (Figure S16, Supporting
Information). Finally, electrolysis of [RuIIꢀOH2]2+ in H218O
(>95% 18O) yields a species with m/z = 357 and 714 correspond-
ing to the structure [RuIVꢀ(18O18O)]2+ (Mꢀ1 = 714; Figure
S17, Supporting Information), confirming that the source of O
atoms is water. It should be noted that similar bulk electrolysis
experiments were conducted with dilute solutions (ca. 75 μM) of
[RuIIꢀOH2]2+ but in the absence of trifluoroethanol. Although
an accurate Coulomb count was not possible due to the low
concentration of [RuIIꢀOH2]2+, the UVꢀvis and mass spectra
of the electrolysis product were identical to those obtained in
experiments with trifluoroethanol as a cosolvent.
The resonance Raman spectra of [RuIIꢀOH2]2+ and the
product obtained by bulk electrolysis of [RuIIꢀOH2]2+, pre-
sumably [RuIVꢀOO]2+, are shown in Figure S18 (Supporting
Information). The peak positions in these two spectra are similar,
except for two new bands at 547 and 930 cmꢀ1 (and changes in
intensities of a few extra bands in the 1250ꢀ1600 cmꢀ1 region)
observed in the spectrum of the oxidized species. The difference
in relative intensities of the Raman bands of the starting material
and the oxidized species is most likely due to differing resonance
enhancement of the vibrational transitions in these two
16
species. The resonance Raman spectra of [RuIVꢀ O16O]2+ and
14655
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
Figure 6. Resonance Raman spectra of [RuIIꢀOH2]2+ after bulk
electrolysis at 1150 mV in H216O (blue trace) and H218O (red trace).
[RuIVꢀ O18O]2+ shown in Figure 6 are almost identical except
18
Figure 7. Bottom: spectral evolution of [RuIVdO]2+ (25 μM) at pH 1
(triflic acid) that indicates formation of an equal amount of
[RuIIꢀOH2]2+ and [RuIVꢀOO]2+. Red curve is the spectrum of
[RuIVdO]2+, and the orange curve is the final spectrum. The spectral
interval is 2 min (the experimental time scale is ca. 120 min). Top:
spectra of [RuIIꢀOH2]2+ (black) and [RuIVꢀOO]2+ obtained by bulk
electrolysis of [RuIIꢀOH2]2+ (blue).
for the red shift of the low-energy vibration (547 f 535 cmꢀ1
)
and the shifts of several bands in the 760ꢀ960 cmꢀ1 region (the
top right spectra in Figure 6). Due to the overlapping vibrational
modes of the ligands, it is not easy to observe the OꢀO and
RudO stretching frequencies. The Raman bands at 930 and
800 cmꢀ1 seem to shift to 920 and 760 cmꢀ1, respectively, by
changing the sample preparation in H216O to H218O. The NMR
spectrum of [RuIVꢀOO]2+ in D2O is indicative of the presence
of a paramagnetic center together with sharp signals, which may
suggest an existence of 1[RuIVꢀη2-OO]2+ in a mixture with the
triplet species (Figure S19, Supporting Information).
observed most likely due to its low initial concentration and its fast
reaction with a water molecule. The m/z 347.8 was observed when
[RuIVdO]2+ was precipitated with NH4PF6 immediately after
preparation by addition of 2 equiv of Ce(IV) to [RuIIꢀOH2]2+
in acidic aqueous solution (pH 1, triflic acid), dried under vacuum,
and transferred into dry acetonitrile (Figure S23, Supporting
Information, bottom panel).
Reactivity of the Two-Electron-Oxidized Species at pH 1.
As presented earlier, very dilute acidic solutions of the two-
electron-oxidized species (8 ꢁ 10ꢀ7 M) prepared by rapid mixing of
[RuIIꢀOH2]2+ with 2 equiv of Ce(IV) appear to be stable on the
time scale of several minutes in the absence of Ce(IV). While such
samples may possibly be an equilibrium mixture of [RuIVdO]2+
and other species, here we simply denote them as [RuIVdO]2+. In
higher concentrations (above ca. 10ꢀ5 M), the formation of approx-
imately equal concentrations of [RuIVꢀOO]2+ and [RuIIꢀOH2]2+
was observed (Figure 7) in 2 h. These products were identified by
UVꢀvis spectroscopy and ESI-MS analysis. The absorption spectra
of [RuIVꢀOO]2+ and [RuIIIꢀOOH]2+ were predicted by TD-
DFT(B3LYP) calculations, and that of 3[RuIVꢀη1-OO]2+ matches
better in terms of the spectral shape and oscillator strength (Figure
S20, Supporting Information). We should note, however, that
[RuIIIꢀOOH]2+ cannot be totally excluded as the assignment of
the oxidized species since the difference in m/z is only 0.5 units and
the calculated spectrum of 1[RuIVꢀη2-OO]2+ may not be accurate
due to the open-shell singlet character missing from its calculated
electronic structure. The kinetics of the formation of these products
appear to be complex but certainly depend on the initial concentra-
tion of [RuIIꢀOH2]2+ (Figure S21, Supporting Information). ESI-
MS of the reaction mixture injected immediately after preparation
by an addition of 2 equiv of Ce(IV) to [RuIIꢀOH2]2+ at pH 1
The final species observed in experiments with 2, 3, and 4
(Figures S23ꢀS25, Supporting Information) equiv of Ce(IV)
have m/z 356.4. We cannot identify whether these are products
of ionization reactions, but [RuIIIꢀOOH]2+ (356.4) and
[RuIIꢀHOOH]2+ (356.9) are possible products.
Electrochemical Products Generated via [RuIVdO]2+. The
bulk electrolysis of [RuIIꢀOH2]2+ discussed in the “Products of
the Electrochemical Four-Electron Oxidation of [RuIIꢀOH2]2+
”
section indicates that our expected product [RuIVdO]2+ changed
to [RuIVꢀOO]2+ by an additional net two-electron oxidation at
154 mV above E1/2(RuIVdO/RuIIꢀOH2) and 270 mV below
E1/2(RuVdO/RuIVdO) at pH 6. The square wave voltammetry
of the bulk electrolysis product with net four-electron oxidation
at 1150 mV shows the most intense peak around 700 mV, which
is likely the redox reaction shown as eq 9, together with weak
peaks around 1030 and 1170 mV (Figure 8). We carried out
further cyclic and square wave voltammograms of [RuIIꢀOH2]2+
at pH 6 to understand this phenomenon and to detect the
product(s). The comparison of the anodic and cathodic voltam-
metric scans during electrochemical studies of [RuIIꢀOH2]2+
revealed the formation of a new species when the potential is
scanned from ca. 1500 to 250 mV at pH 6 (Figure 8).
shows m/z356.8 for [RuIVdO OH2]2+ together with m/z348.3
3 3 3
probably for [RuIIꢀOH2]2+ and m/z 355.1 for [RuIVꢀOO]2+
(Figure S22, Supporting Information, top panel). The intensity of
signals in the range of m/z 354ꢀ359 cannot be explained by the
isotope pattern of a single species and changes in complicated ways
with time. The m/z signal corresponding to [RuVdO]3+ was not
The Pourbaix diagram in Figure 9 obtained from the cathodic
scans from 1240 mV at various pH shows new pH-dependent
features (e.g., the green line), indicating that the electrochemical
product is a conjugate acid. Increasing the quiet time at the
beginning of each cathodic scan at 1240 mV led to an increase in
14656
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
Figure 9. Pourbaix diagram of the species produced by electrochemical
oxidation of [RuIIꢀOH2]2+ in water after holding the potential at 1240
mV for 2 s and scanning in the cathodic direction. The magenta dotted
line corresponds to the calculated potential for [RuIVdO]2+ + H2O ꢀ
eꢀ f [RuIIIꢀOOH]2+ + H+ and was not observed.
to the η1 mode for the formation of [RuIIIꢀOOH]2+ may not take
place on the time scale of the square-wave voltammogram. The
broad signals observed in the NMR might be due to a small
amount of coexisting geometric isomer, 3[RuIVꢀη1-OO]2+, and/
2
or contamination by [RuIIIꢀOOH]2+, together with the sharp
1
signals of [RuIVꢀη2-OO]2+. We will discuss the nature of this
species in the Discussion section.
No oxygen production from [RuIVꢀOO]2+ was found under our
experimental conditions in the absence of an added oxidant. The
decomposition of [RuIVꢀOO]2+ in water (pH 6) can be observed
after several hours by following the changes in the UVꢀvis absorp-
tion spectrum, and complete conversion was achieved after ca. two
weeks (Figure S26, Supporting Information). The UVꢀvis spec-
trum of the decomposition product does not correspond to
[RuIIꢀOH2]2+ but closely resembles it. The ESI-MS analysis of
the product indicates the presence of a species corresponding to the
formula [RuIIꢀOOH]+ (Figure S27, Supporting Information). The
1H NMR spectrum (in D2O) confirms the assignment to diamag-
netic [RuIIꢀOOH]+ based on the number of protons contained in
organic ligands and sharp signals (Figure S28, Supporting Informa-
tion, bottom) and is different from that of [RuIIꢀOH2]2+ reported
previously (Figure S28, Supporting Information, top).28 The NMR
of the product(s) indicates [RuIIꢀOOH]+ as an almost pure
product and supports the reaction shown in eq 10.
Theoretical Analysis of the Electrochemical Oxidation of
[RuIIꢀOH2]2+. Calculations of the absolute standard free energy
of species postulated to be involved in the catalytic water
oxidation cycle were carried out to construct the free-energy
profile of the catalytic reaction and relevant features of Pourbaix
diagrams for comparison with the experimental results. The
results for the species we have identified as catalytic intermediates
are summarized in Table 1. The “system” species in the second and
last rows of the table are identical, i.e., [RuIIIꢀOH]2+. The proposed
catalytic step from [RuIVdO]2+ to [RuIIIꢀOOH]2+ may occur
directly or with [RuVdO]3+ as an intermediate, as discussed below.
Also, a molecule of water solvent could possibly displace triplet
O2 from either singlet or triplet [RuIVꢀOO]2+ species (the
singlet would require spinꢀorbit coupling and would likely be
an activated process) in an exothermic reaction to regenerate
singlet [RuIIꢀOH2]2+. Finally, O2 evolution could result from
Figure 8. Top panel: Cyclic and square wave voltammograms of
[RuIIꢀOH2]2+ in water (pH 6). Bottom panel: Cyclic and square wave
voltammograms of the bulk electrolysis product (pH 6) at 1150 mV vs NHE.
The inset shows square wave anodic (top) and cathodic (bottom) scans.
the electrochemical products. The current at the pH-dependent
potential shown by the red line in Figure 9 appears almost
negligible inthe first few scans, but the intensitygraduallyincreases
with repeated cathodic scans. This suggests that the initial species,
which is likely [RuIIIꢀOOH]2+, is produced at an applied
potential above E1/2(RuIVdO/RuIIꢀOH2) and is reduced to
[RuIIꢀOOH]+ (and/or its protonated species at low pH) at
potentials shown by the green line (eqs 10 and 11).
It should be noted that the applied potential of 1240 mV, i.e.,
180 mV less positive than that for E1/2(RuV/RuIV), seems
sufficient to carry out the reaction shown as the net process in
eq9 duringthe quiettime beforethecathodic scans, and eventually
the reactions shown as eqs 10 and 11 seem to take place.
2þ
2þ
½RuIVdOꢂ þ H2O ꢀ eꢀ f ½RuIII ꢀ OOHꢂ þ Hþ
ð9Þ
ð10Þ
ð11Þ
½RuIII ꢀ OOHꢂ þ eꢀ f ½RuII ꢀ OOHꢂþ
2þ
2þ
2þ
½RuIII ꢀ OOHꢂ þ eꢀ þ Hþ f ½RuII ꢀ HOOHꢂ
Interestingly, the bulk electrolysis product exhibits currents at the
potential corresponding to the red line without any current at the
lower potential of the green line. This discrepancy seems to
suggest that the species prepared by bulk electrolysis is possibly
1[RuIVꢀη2-OO]2+. The geometric conversion from the η2 mode
14657
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
the oxidation of singlet or triplet [RuIVꢀOO]2+ to produce
2[RuIIIꢀOH]2+ as evidenced in an attempted geometry optimi-
zation of “2[RuVꢀη1-OO]3+”, the 1eꢀ oxidized species of
3
3[RuIVꢀη1-OO]2+, in which the O2 moiety is spontaneously
expelled from the metal center (RuꢀO distances increasing from
2.03 and 2.12 Å to 4.47 and 4.62 Å) and while the explicit
molecule of H2O solvent hovers closer tothemetalcenter(RuꢀO
distance decreasing from 4.85 to 4.54 Å).
The species listed in the “reservoir” column of the table are
those that must eventually be transferred to a system species from
the solution (i.e., the two water molecules that will eventually be
oxidized to O2) and those that are removed from the system to
the solution or to the gas phase (i.e., protons, electrons, and
eventually O2) along the catalytic pathway. Consideration of the
sum of the free energies of the system and reservoir species for
each intermediate allows the construction of a meaningful
standard free energy profile along the reaction pathway, as shown
in the “rel. total G* (eV vs NHE)” column of Table 1. It is clear
that the net reaction around the catalytic cycle in the system is
Figure 10. Calculated Pourbaix diagram for electrochemical oxidation
of [RuIIꢀOH2]2+ in water.
2
2þ
2
2þ
½RuIII ꢀ ðOHÞ H2Oꢂ f ½RuIII ꢀ ðOHÞ H2Oꢂ , ΔG ¼ 0
ꢃ
applied potential than for the [RuIVdO H2O]2+ + H2O f
3
3
3
[RuIIIꢀ(OOH) H2O]2+ + H+ step at low pH but that the
3
while the net reaction in the reservoir is
subsequent addition of H2O and removal of a proton would be
slightly endothermic. At pH above ca. 3.5, the calculations
2H2O f O2 þ 4Hþ þ 4eꢀ, ΔG° ¼ 4FE°ðORRÞ
where ORR stands for the oxygen reduction reaction.
predict that the potential for the direct [RuIVdO H2O]2+
+
3
H2O f [RuIIIꢀ(OOH) H2O]2+ + H+ step would drop below
3
that for [RuIVdO H2O]2+ f [RuVdO H2O]3+ (Figure 10).
The column with the heading “rel. total G* (eV vs NHE)” in
Table 1 references the electron transfer steps to the normal
hydrogen electrode (NHE), which is evaluated here (using
the methodogy described in the Supporting Information) to be
4.157 eV in aqueous solution. The calculated value of the
standard free energy change for the ORR, again using the same
methodology, is 4.760 eV and is completely independent of
the catalyst calculations. It results only from the calculated values
of the free energies of O2(g) and H2O (liquid), the values of
the standard free energies of the solvated proton and gas-phase
electron used, and the value of the absolute potential of the
normal hydrogen electrode (NHE). The error (ꢀ0.156 eV, or
ꢀ0.039 eV per electron) in the calculated standard free energy
change of the ORR (4.916 eV) is most likely due to DFT doing a
poor job on the electronic energy of the ground state of O2(g).
This estimated error in the last step of the cycle may account for
the underestimated value of the standard free energy change,
which contrasts with the experimental observations.
The last column of Table 1, under the heading “ΔG* (eV)”,
lists the standard free-energy changes for the various steps in the
proposed catalytic mechanism. The standard free-energy differ-
ence listed for a given species is relative to the corresponding
one-electron-reduced species with the lowest free energy. We do
not imply that the “assignment” of the species with lowest
calculated free energy is necessarily correct (see discussion
regarding [RuIVdO]2+ above) but that there is a species of that
chemical formula that has at least as low an energy as that listed.
All the electron-transfer steps are proton coupled (PCET) except
for [RuIVdO H2O]2+ f [RuVdO H2O]3+, which is a simple
3
3
The Pourbaix diagram constructed from the energies of DFT
optimized structures of various redox and acidꢀbase states of
species [RuIIꢀOH2 H2O]2+ is shown in Figure 10. A similar,
3
though somewhat simpler, diagram based only on the results of our
calculations with the small basis set, which exhibits the same two-
electron, two-proton nature of the oxidation of [RuIIꢀOH2 H2O]2+
3
to [RuIVdO H2O]2+ as in our experiments, is presented in
3
Figure S29 (Supporting Information). The Pourbaix diagram in
Figure 10 indicates sequential 1eꢀ/1H+ oxidation steps from
[RuIIꢀOH2]2+ to [RuIVdO]2+ separated by 70 mV. In addition
to these features, there is a 1eꢀ oxidation of [RuIIꢀOH2]2+ to
[RuIIIꢀOH2]3+, a 1eꢀ/2H+ oxidation of [RuIIIꢀOH2]3+ to
[RuIVdO]2+, a 1eꢀ oxidation of [RuIVdO]2+ to [RuVdO]3+,
and a 2eꢀ/2H+ oxidation of [RuIIIꢀOH2]3+ to [RuVdO]3+.
The [RuIIIꢀOOH]2+ intermediate is reached either by the 1eꢀ/
1H+ oxidation of [RuIVdO]2+ involving its reaction with a water
molecule coupled to the loss of a proton or the attack of H2O on
[RuVdO]3+ and the elimination of a proton in a subsequent
thermal (nonelectrochemical) reaction. These two oxidations are
indicated by dashed lines in Figure 10 in the pH region where the
PCET process is predicted to be favored. This thermal step is
calculated to be slightly endothermic, and the preceding 1eꢀ
oxidation of [RuIVdO]2+ to [RuVdO]3+ becomes more and
more positive relative to that for the production of [RuIVdO]2+
with increasing pH. In contrast, the electrochemical conversion
of [RuIVdO]2+ to [RuIIIꢀOOH]2+ is predicted to be possible at
all pH values in the diagram and to be favored over the formation
of [RuVdO]3+ at pH above 3.5 (although the calculated position
of this PCET potential may be too high). We have also examined
the pH dependence of the onset of catalytic current, as shown in
Figure S30 (Supporting Information). The onset of background-
subtracted catalytic current shows no dependence on pH at low
values of pH, then a somewhat complicated behavior as a function
of pH between pH 7 and 8.5, and finally an ca. ꢀ59 mV/pH
dependence at pH > 8.5. Such behavior is consistent with a switchover
3
3
electron transfer step which requires a subsequent proton
transfer step to reach [RuIIIꢀ(OOH) H2O]2+. Therefore, all
3
the free-energy changes except for the [RuIVdO H2O]2+
f
3
[RuVdO H2O]3+ and [RuIIꢀOH2]2+ f [RuIIIꢀOH2]3+ steps
3
will be pH dependent, with ΔGox = ΔGox* ꢀ 0.059 pH relative
3
to NHE. The computed results suggest that the step
[RuIVdO H2O]2+ f [RuVdO H2O]3+ would occur at a lower
3
3
14658
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
Table 2. Standard Reduction Potentials and pKa Values of
Intermediate Species Derived from the Experimental
Pourbaix Diagrams
couple
type
1eꢀ
2eꢀ/2H+
1eꢀ/2H+
2eꢀ/2H+
1eꢀ
1eꢀ/1H+
1eꢀ/1H+
1eꢀ
E° (V vs NHE)
[RuIIIꢀOH2]3+/[RuIIꢀOH2]2+
[RuIVdO]2+/[RuIIꢀOH2]2+
[RuIVdO]2+/[RuIIIꢀOH2]3+
[RuVdO]3+/[RuIIIꢀOH2]3+
[RuVdO]3+/[RuIVdO]2+
[RuIIIꢀOOH]2+/[RuIVdO]2+ a
[RuIVꢀOO]2+/[RuIIIꢀOOH]2+
[RuIIIꢀOOH]2+/[RuIIꢀOOH]+
[RuIIIꢀOOH]2+/[RuIIꢀHOOH]2+
[RuIIIꢀOH2]3+
1.180
1.351
1.523
1.473
1.420
1.621
0.950
0.295
0.505
2.9
1eꢀ/1H+
pKa
[RuIIꢀOOH2]3+
pKa
3.5
a Includes theoretical value of 0.201 eV for the standard free energy of
the reaction [RuVdO]3+ + H2O f [RuIIIꢀOOH]2+ + H+.
from the RuVdO pathway to a direct pathway involving RuIVdO.
The calculated structures of all key intermediates in the proposed
catalytic cycles are shown in Figure S31 (Supporting Information).
Those of [RuIIꢀHOOH]2+, [RuII(NPMH+)(pic)2ꢀOOH]2+,
and [RuII(NPMH+)(pic)2ꢀOH2]3+ are shown in Figure S32
(Supporting Information).
Using the experimental standard reduction potentials and pKa
values derived from the Pourbaix diagrams in Figures 2 and 9 and
listed in Table 2 and the calculated energetics listed in Table 1,
one can construct the experimental and theoretical LatimerꢀFrost
diagrams, as shown in Figure 11 (under standard conditions) and
Figure S33 (Supporting Information) (at pH 7). The calculated
ΔG* of the thermal reaction [RuVdO]3+ + H2O f
[RuIIIꢀOOH]2+ + H+ was used in the experimental diagrams.
In these diagrams, the cumulative free energy change relative to
the resting state of the catalyst complex, [RuIIꢀOH2]2+, for
producing the various intermediates is plotted as a function of the
number of electrons, n, removed in the oxidation process. Any
intermediate that lies above the line joining its neighboring
points is unstable with respect to disproportionation. The largest
vertical distance from the line joining the points corresponding
to n = 0 and n = 4 or n = 1 and n = 5, corresponding to the
thermodynamic 4-electron oxidation potential, to an intermedi-
ate represents an intrinsic thermodynamic overpotential of the
catalyst for a 0 f 4 cycle or a 1 f 5 cycle, respectively. It is clear
from both panels of Figure 11 that in the present case this is
associated with the [RuIIIꢀOOH]2+ intermediate and lower
(only 257 mV in the experimental plot) for the 1 f 5 cycle than
for the 0 f 4 cycle. We see no evidence in our experiments or
our calculations of the blue lines in Figure 11 that would
Figure 11. Experimental (top) and theoretical (bottom) Latimerꢀ
Frost diagrams for the oxidation of [RuIIꢀOH2]2+ along the proposed
catalytic pathway under standard conditions (i.e., pH 0).
basis predict that the potential for the second step is 70 mV larger
than that of the first step (Figure 10). While the difference
obtained in the two calculations is certainly within the calcula-
tional error, this feature of the results has implications regarding
the topology of the Pourbaix diagram.
The oxidation state assignments in intermediates suggested by
the calculations are in a few cases unconventional for mono-
nuclear ruthenium oxidation catalyst species. If the assignments
were based primarily on the computed spin density distribution
in open-shell species and, of course, the relative energetics of
singlet and triplet intermediates, they would differ from the
conventional ones. The first of the these assignments along the
proposed catalytic pathway would be the designation of inter-
3
culminate a 0 f 4 cycle either by direct elimination of O2
mediate [RuIVdO H2O]2+ as triplet [RuIVdO]2+ arising from
through the PCET oxidation of [RuIIIꢀOOH]2+ or by the
decomposition of [RuIVꢀOO]2+ following the PCET oxida-
tion of [RuIIIꢀOOH]2+.
3
an intermediate spin triplet state at the metal center but having
3
some [RuIIIꢀ(O•ꢀ)]2+ character as discussed in “Theoretical
Characterization of [RuIVdO]2+ and [RuIVꢀOO]2+” in the
Another interesting aspect of the calculations is that those
with the small basis set predict that the sequential PCET
steps [RuIIꢀOH2 H2O]2+ f [RuIIIꢀOH H2O]2+ and
2
Supporting Information. The intermediate [RuVdO H2O]3+
3
is assigned conventionally because it would arise from removing
an electron from the metal side in either panel of Scheme S4
(Supporting Information), leading to the unpaired electron in
the π* orbital being shared more or less equally between the
metal center and oxygen atom and consistent with the computed
spin density distribution shown in Figure S10 (Supporting Information).
3
3
[RuIIIꢀOH H2O]2+ f [RuIVdO H2O]2+ should occur as a
3
3
single two-electron, two-proton step with standard free-energy
change 1.06 eV (1.06 V standard reduction potential) as shown
in Figure S29 (Supporting Information). On the other hand, the
results incorporating a single-point calculation with the large
14659
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
Scheme 1. Energetics of the Net Reaction 2[RuIV=O]2+
+
equal amounts of [RuIVꢀOO]2+ and [RuIIꢀOH2]2+ can be ration-
H2O f [RuꢀOH2]2+ + [RuIVꢀOO]2+
alized in terms of Scheme 1 for the net reaction (2[RuIVdO]2+
+
H2O f [RuIVꢀOO]2+ + [RuIIꢀOH2]2+). The energetics of the
elementary steps leading to the net reaction are taken from Table 2
and show that the overall process is exothermic. In fact, the main
route for the formation of [RuIV(NNN)(NN)(OO)]2+ has been
previously considered via ([RuVdO]3+ +H2Of[RuIIIꢀOOH]2+ +
H+ and [RuIIIꢀOOH]2+ f [RuIVꢀOO]2+ + H+ + eꢀ).30,34,36
Identity of [RuIVꢀOO]2+. Resonance Raman and ESI-MS
measurements were made on samples prepared by four-electron
oxidation of [RuIIꢀOH2]2+ at pH 6. The [RuIIꢀOH2]2+ con-
tent should be at trace amounts, if present at all. The ESI-MS
analysis of the solution used for resonance Raman measurements
indicated the presence of a species with m/z = 355 (M/2) and
16
710 (Mꢀ1), which correspond to [RuIVꢀ O16O]2+. The sample
Finally, the present DFT calculations predict that it is the triplet
[RuIVꢀη1-OO]2+ rather than the singlet [RuIVꢀη2-OO]2+ species
that results from the [RuIIIꢀ(OOH) H2O]2+ f [RuIVꢀ(OO)
prepared in H218O gives m/z = 357 and 714 corresponding to
18
the structure [RuIVꢀ O18O]2+. It should be pointed out that
Wasylenko et al. observed an ESI-MS signal at m/z 523.15 and
assigned it as [Ru(tpy)(bpy)(OO)]+.36 This is a +1 charged
species andnot the+2 chargedspecies astheyandothersproposed
in the mechanism of water oxidation (eqs 12 and 13).
3
3
H2O]2+ step, as shown in the lower half of Table S2 and in Scheme
S5 (Supporting Information). The singlet [RuIVꢀη2-OO]2+ species
(i.e., the conformer with the OꢀO axis approximately in the
equatorial plane) is computed to lie 13.2 kcal molꢀ1 higher in
standard free energy than the triplet [RuIVꢀη1-OO]2+ species, and,
3þ
2þ
½RuVdOꢂ þ H2O f ½RuIII ꢀ OOHꢂ þ Hþ
ð12Þ
3
like the [RuIVdO]2+ species, arises from an intermediate spin
triplet Ru(IV) metal center. The spin density distribution of the
triplet state (Figure S10, Supporting Information) indicates con-
siderable spin on the two oxygen atoms (and very little spin on the
Ru center), which can be accounted for by the two singly occupied
molecular orbitals having much more peroxo π* character than
Ru dπ character. The localization of the unpaired spins on the
peroxo π* orbitals imparts strong [RuIIꢀη1-3(OO)]2+ character to
this species, consistent with the computed spin density shown in
Figure S10 (Supporting Information). Table S2 (Supporting Infor-
mation) shows that the singlet (η2) state is stabilized by GVB-CI
involving the RuꢀO π and (7%) π* orbitals, similar to the singlet
state of [RuIVdO]2+ in trading an intermediate-spin metal center,
albeit one with extensive mixing with the O22ꢀ π* orbitals, for a low-
spin metal center and a vacant π* orbital available for GVB-CI. Here
again, the singlet state is predicted to be stabilized enough by valence
configuration interaction to be the ground electronic state. Similar to
the case of the triplet η1 species, the donation of two electrons to
orbitals strongly localized on the metal center should impart strong
[RuIIꢀη2-1(OO)]2+ character to singlet [RuIVꢀη2-OO]2+.
2þ
2þ
½RuIII ꢀ OOHꢂ ꢀ eꢀ f ½RuIV ꢀ OOꢂ þ Hþ
ð13Þ
In fact, when an aqueous solution containing [RuIVꢀ O16O]2+
16
decomposed slowly with the λmax shift from 512 to 580 nm, it
16
produces [RuIVꢀ O16OH]+ with m/z 712 for [RuIIꢀOOH]+
andm/z 356.5 for [RuIIꢀOOH]+ + H+. Therefore, webelievethat
the species Wasylenko et al. observed to have an ESI-MS signal at
m/z 523.15 is likely [Ru(tpy)(bpy)(OOH)]+ (MW 523.54).
While mass spectrometry and Coulomb count experiments
suggest the formation of a species corresponding to [RuIVꢀOO]2+,
these experimental data do not provide any insight into the bonding
between oxygen atoms within the complex. On the other hand, the
resonance Raman spectrum of [RuIVꢀOO]2+ reveals a new distinct
vibration at 547 cmꢀ1, which red-shifts only 12 cmꢀ1 upon 16O/18O
substitution. While the direction of the shift is consistent with the
presence of a heavier atom (18O), its magnitude appears to be
smaller than is predicted from the simple diatomic oscillator model
(27 cmꢀ1). The assignment of this vibration to a metal-oxo stretch
(RuIVdO) is unlikely since that is usually observed around
800 cmꢀ1 with isotopic shifts in the range of 33ꢀ45 cmꢀ1
(Table S3, Supporting Information).89ꢀ92 However, the vibrational
frequencies of the metalꢀoxygen bond in metal superoxo or hydro-
peroxo compounds are reported to be in the 500 cmꢀ1 region with
small isotopic shifts, e.g., 12ꢀ21 cmꢀ1, upon 16O/18O substitution
(Table S4, Supporting Information).34,90,93ꢀ111 For example, in a
study reported by Bakac et al.,97 the vibrational frequency of the
CrꢀO bond was observed at 503 cmꢀ1 by measuring the resonance
Raman spectrum of the η1-superoxochromium(III) complex, and
interestingly, its isotopic shift was only 12 cmꢀ1 upon 16O/18O
substitution. It was proposed that significant coupling of the CrꢀO
stretch with other vibrational modes (e.g., the CrꢀOꢀO bend) leads
to the discrepancy between experiment and the calculation based on a
diatomic oscillator model. The frequency corresponding to the OꢀO
vibration in superoxo, peroxo, and hydroperoxo metal complexes is
usually observed in the region between 830 and 1200 cmꢀ1 showing
isotopic shifts between 17 and 68 cmꢀ1 (Table S4, Supporting
Information).
The RuꢀO and OꢀO vibrations are calculated to be 470 cmꢀ1
(RuꢀO2 symmetric stretch), 609 cmꢀ1 (RuꢀO2 antisymmetric
stretch), and 1105 cmꢀ1 (OꢀO stretch) for 1[RuIVꢀη2-OO]2+,
as opposed to 348 cmꢀ1 (RuꢀO stretch) and 1352 cmꢀ1 (OꢀO
stretch) for the 3[RuIVꢀη1-OO]2+ species (see Table S1, Supporting
Information). Since the calculated OꢀO stretches do not match well
with experimental data, we investigated 1[RuIVꢀcyc-OON]2+ where
the OꢀO interacts with a noncoordinating nitrogen of the NPM
ligand. However, this species is energetically 21.88 kcal molꢀ1 higher
3
than [RuIVꢀη1-OO]2+ (and 8.65 kcal molꢀ1 higher than
1[RuIVꢀη2-OO]2+). The calculated spectroscopic properties of
this species (absorption max at 500 and 698 nm) are also in
disagreement with experimental data.
’ DISCUSSION
Reactivity of [RuIVdO]2+ in Water (pH 1). The experimental
observation that aqueous [RuIV=O]2+ solutions in concentrations
higher than ca. 10ꢀ5 M spontaneously react to form approximately
14660
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
To our knowledge, there are no single-crystal X-ray diffraction
studies on ruthenium complexes with the η1-OO moiety. We
have summarized the OꢀO vibrational frequency of structurally
characterized ruthenium complexes with the η2-OO moiety in
Table S4 (Supporting Information). The OꢀO stretching
frequency of ruthenium complexes with the η2-OO moiety is
in the range of 850ꢀ920 cmꢀ1 regardless of the formal oxidation
state of ruthenium. While some spectral changes are observed for
η2 configuration. The species [RuIVꢀη2-OOH]2+ does not seem
to reduce to [RuIVꢀη2-OOH]+ or [RuIVꢀη2-HOOH]2+ at
potentials greater than 100 mV. At an applied potential at 1240
mV, [RuIIIꢀOOH]2+ is produced, which is reduced to
[RuIIꢀOOH]+ and [RuIIꢀHOOH]2+ at the potential of the
green line in Figure 9. Once the produced [RuIIIꢀOOH]2+ is
further oxidized at 1240 mV, [RuIVꢀη1-OO]2+ likely forms and
may eventually convert to [RuIVꢀη2-OO]2+. While a pendent
base on the NPM ligand could possibly be involved in the above
reactions via a hydrogen bonding interaction with the OOH
moiety, we do not have any clear evidence of it, and our DFT
calculations indicate otherwise.
ν
16 in the high-frequency region (750ꢀ950 cmꢀ1) upon
16Oꢀ O
isotopic substitution for the [RuIVꢀOO]2+ species,112 the
definitive identification of the band corresponding to the OꢀO
vibration can be complicated by the strong Raman activity of the
ligand vibrational modes and a relatively low Raman intensity of
the OꢀO vibration. It should be noted that the RuꢀO and OꢀO
vibrations are calculated to be 470 cmꢀ1 (RuꢀO2 symmetric
stretch), 609 cmꢀ1 (RuꢀO2 antisymmetric stretch), and
1105 cmꢀ1 (OꢀO stretch) for singlet [RuIVꢀη2-OO]2+, as
opposed to 221 cmꢀ1 (RuꢀO stretch) and 1350 cmꢀ1 (OꢀO
stretch) for the triplet [RuIVꢀη1-OO]2+ species. These vibra-
tional modes are mixed with other modes in most cases. It is clear
that the calculated values for the OꢀO stretch are not good
matches (i.e., no observation of a vibrational frequency shift by
isotopic labeling experiments in the region of 1000ꢀ1500 cmꢀ1),
and it is difficult to predict even the geometry of the MꢀO2
moiety. While Concepcion et al. assigned 1015 cmꢀ1 as the OꢀO
stretch based on DFT calculations for [RuIV(Mebimpy)(bpy)
(η2-OO)]2+ (Mebimpy = 2,6-bis(benzimidazol-2-yl)pyridine), their
assignment was not confirmed by labeling experiments.34 As seen in
Table S4 (Supporting Information), experimental data on MꢀOꢀ
OꢀM species indicate that OꢀO stretch frequencies change over
quite a wide range (800ꢀ1200 cmꢀ1) due to different bond
strengths, electron distributions, geometries, etc. Also, any hydrogen
bonding of the bound OꢀO (or OOH) with solvent water molecules
and/or interaction with a noncoordinating nitrogen of the NPM
ligand may decrease the OꢀO or MꢀO stretching frequencies
1
In conclusion, we have assigned this species as [RuIVꢀη2-
OO]2+ based on electrochemical, spectroscopic, and theoretical
plausibility alongside a previously published conclusion by the
T. J. Meyer’s group34 and an X-ray structural database for
RuꢀOO species, although there remains some conflicting evi-
1
dence, especially the higher energy of [RuIVꢀη2-OO]2+ com-
pared to 3[RuIVꢀη1-OO]2+ predicted by DFT and BS-DFT, but
not by CASSCF calculations.
Oxygen Release from [RuIVꢀOO]2+. Despite the fact that the
formation of the OꢀO bond is achieved at relatively low potentials
via reaction of an [RuVdO]3+ or [RuIVdO]2+ complex, the oxygen
molecule remains bound to the Ru center in [RuIVꢀOO]2+ and
requires an additional oxidation to be liberated into solution. This
unusual stability of [RuIVꢀOO]2+ is evident from the fact that, if
left in solution, [RuIVꢀOO]2+ was slowly reduced to form
[RuIIꢀOOH]+. However, the addition of Ce(IV) to [RuꢀOO]2+
results in oxygen evolution. On the other hand, the oxygen evolution
from an analogue of [RuIVꢀOO]2+ reported by Meyer’s group34
occurs thermally, but very slowly (k=7.5ꢁ 10ꢀ4 sꢀ1). Thisthermal
process could involve conversion to the triplet [RuIVꢀη1-OO]2+
state as a first step, followed by displacement of triplet O2 by a water
molecule to form singlet [RuIIꢀOH2]2+, in a manner analogous to
that indicated by the blue lines in both the experimental and
theoretical LatimerꢀFrost diagrams of Figure 11. This reaction
was found to be accelerated by several orders of magnitude upon
additional oxidation of [RuIVꢀOO]2+. The reactivity of the singlet
[RuIVꢀOO]2+ complex is similar to that in one branch of a previ-
ously proposed mechanism of its analogue studied by Meyer’s
group,34 i.e., sequential or concurrent oxidation and addition of H2O
with loss of a proton and elimination of triplet O2. The additional
oxidation of [RuIVꢀOO]2+ apparently lowers the kinetic barrier for O2
release and returns the catalyst not to its original [RuIIꢀOH2]2+ state
but rather to the one-electron-oxidized [RuIIIꢀOH]2+ intermediate.
Identity of [RuIVdO]2+. There have been extensive studies on
RuIVdO species that display the triplet spin state82ꢀ87 but are
not [RuIIIꢀO•ꢀ]2+. NMR spectra of [RuIV(tpy)(bpy)(O)]2+ in
D2O and [RuIVdO]2+ in CD3CN indicate that these complexes
exhibit typical triplet spectra (Figures S13 and S14, Supporting
Information). It is of interest that [RuIV(tpy)(bpy)(O)]2+ is
stable in water (pH 1), but [RuIVdO]2+ decomposes to
[RuIVꢀOO]2+ and [RuIIꢀOH2]2+ in 2 h. While the role of a
pendent base of the NPM ligand is not clear in this reaction, the
RuV/RuIV potential is almost 400 mV less positive than that of
[RuIV(tpy)(bpy)(O)]2+, indicating that the [RuVdO]3+ species
might easily form.
1
(e.g., see [RuIVꢀcyc-OON]2+ in Table 1). The low vibrational
frequency (547 cmꢀ1) and small isotopic shift (12 cmꢀ1) for RuꢀO
of [RuIVꢀOO]2+ are reasonable for those of η1-OO or η2-OO
complexes. The small isotopic shift observed in the RuꢀO stretching
frequencies for both [RuIVꢀη1-OO]2+ and [RuIVꢀη2-OO]2+
is, however, consistent with our DFT calculations, which indicate
that the vibrational modes are mixed with various other modes and, by
analogy to the above-mentioned case of CrꢀOO, with small isotopic
shifts.
Now we discuss the electrochemistry of [RuIVꢀOO]2+. A
sample was prepared by bulk electrolysis (four-electron oxidation
of [RuIIꢀOH2]2+) at pH 6, and the electrochemical data shown
in the bottom panel of Figure 8 were taken after performing
resonance Raman and ESI-MS measurements. Currents were
observed at potentials corresponding to the red and blue lines,
but not the green line, in Figure 9 during both anodic and
cathodic scans. However, the cathodic scan starting at 1240 mV
of [RuIIꢀOH2]2+ led to the formation of a new species that has
reduction potentials at the blue and green lines in Figure 9.
Eventually, after repeated scans, another species that has a
reduction potential at the red line started to be observed. These
results suggest the existence of geometric isomers for
[RuIVꢀOO]2+. We tentatively assigned the electrochemically
prepared sample used for the resonance Raman measurement to
be [RuIVꢀη2-OO]2+. When the [RuIVꢀη2-OO]2+ species is
reduced, protonation of one of the O atoms may take place while
retaining (at least during the voltammetric time scale) the
Mechanism of Formation of [RuIVꢀOO]2+. The proposed
cycle for the water oxidation reaction catalyzed by [RuIIꢀOH2]2+
is shown in Figure 12. The main focus of the discussion will be on
the crucial step of OꢀO bond formation and the identity of the
[RuIVꢀOO]2+ species.
14661
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
keff = kOꢀO ꢁ fi, so that the 1/e time for conversion under the
experimental conditions at pH “i” should be τi = 1/(kOꢀO ꢁ fi).
Under the conditions of the underpotential bulk electrolysis at pH
1 and 6, f1 = 2.0 ꢁ 10ꢀ3 and f6 = 2.7 ꢁ 10ꢀ5. This gives τ1 = 15 h
and τ6 = 1100 h, both negligibly slow compared to the experi-
mental time scale (20ꢀ40 min). However, if kOꢀO were 2 orders
of magnitude faster than that estimated by Concepcion et al., τ1 =
8.7 min and τ6 = 11 h. The shorter value of τ1 appears to be of the
same order as the experimental time scale for underpotential
electrolysis; however, τ6 still remains too large. While this experi-
mental observation and the background-subtracted onset of catalytic
current of [RuIIꢀOH2]2+ at various pH (Figure S30, Supporting
Information) do not offer definitive proof of the “direct route”, it
demonstrates that [RuIIIꢀOOH]2+ can be formed in large part by
some pathway that does not involve [RuVdO]3+. Any reactions
involving direct interactions with the ceric ammonium salt, as
proposed for other systems,36 can be dismissed in our case since
no cerium was used in the electrolysis experiments.
Figure 12. Proposed catalytic cycle of water oxidation reaction cata-
lyzed by [RuIIꢀOH2]2+.
The proposed mechanism assumes that the formation of the
OꢀO bond may take place either by nucleophilic attack of a
water molecule on [RuVdO]3+ (eq 12), as proposed for other
mononuclear ruthenium catalysts reported previously,34,50 or
through the reaction of [RuIVdO]2+ with a water molecule and
concomitant removal of an electron and a proton (eq 9, the so-
called “direct pathway”).
There is as of yet no clear understanding of the exact mechanism
for the direct pathway. An early theoretical study by Wang et al.113
demonstrated that formation of an OꢀO bond at a RuIV center is
unfavorable; however, later experimental studies have proposed
that [RuIV(tpy)(bpy)(O)]2+ can react with a water molecule to
yield [RuII(tpy)(bpy)(H2O2)]2+.36 We do not have strong experi-
mental evidence for such a reaction except that: (1) two molecules
of [RuIVdO]2+ convert completely to [RuIVꢀOO]2+ and
[RuIIꢀOH2]2+ in less than 2 h (i.e., the reactions via [RuVdO]3+
seem too slow as mentioned above); (2) the observation of m/z356.9
The thermodynamics of both oxidation reactions of
[RuIVdO]2+ were predicted theoretically (Figure 10 and
Table 1), and it was found that the formation of [RuVdO]3+
(eq 5) at standard conditions (pH 0) requires a lower applied
potential compared to the direct pathway. The subsequent
addition of a water molecule to [RuVdO]3+ (eq 12) is, however,
a thermodynamically slightly uphill reaction requiring an addi-
tional 4.64 kcal molꢀ1 under standard conditions (Table 1).
Given that the concentration of the water reactant in that thermal
reaction is 55.4 M, the reaction should proceed fairly well. At
standard conditions, the net change in free energy is the same for
the direct pathway (eq 9) and the formation of [RuVdO]3+
followed by the addition of water (eq 5 + eq 12). However, in the
high pH region the direct pathway becomes more energetically
favorable compared to the formation of [RuVdO]3+ since the
former is a proton-coupled reaction and the latter is not
(Figure 12). An analysis of the experimental LatimerꢀFrost
diagram at pH 7 indicates that generation of [RuVdO]3+ from
[RuIVdO]2+ is more endothermic, compared to the “direct
pathway” to produce [RuIIIꢀOOH]2+ (Figure S33, Supporting
Information). The possible “branching” of the mechanism at
RuIV and RuV oxo species was suggested by Berlinguette36 based
on a kinetic analysis of spectroscopically observed intermediates
and the overall catalytic rate of O2 production. These multiple
pathways become easily accessible since powerful oxidants such
as Ce(IV) are used to drive these reactions.
For example, consider the bulk electrolysis experiments at pH 1
with an applied potential of 1260 mV, and at pH 6 with an applied
potential of 1150 mV, i.e., 160 and 270 mV below the 1420 mV of
the RuVdO/RuIVdO couple, respectively. From the Nernst
equation for the one-electron reduction of [RuVdO]3+, we obtain
[RuVdO3+]/[RuIVdO2+] = exp((Ei ꢀ E°)/RT) ꢄ fi under the
experimental conditions at pH “i”. The subsequent reaction of
nucleophilic attack by a water molecule on [RuVdO]3+ (eq 12)
for a similar mononuclear ruthenium complex was estimated by
Concepcion et al.34 to have a pseudo first-order rate constant,
kOꢀO, of 9.6 ꢁ 10ꢀ3 sꢀ1. Assuming a rapid pre-equilibrium and
neglecting the back reaction of [RuIIIꢀOOH]2+, both assump-
tions favoring the production of [RuIIIꢀOOH]2+ through the
[RuVdO]3+ route, the effective pseudo first-order rate constant
for the conversion of [RuIVdO]2+ to [RuIIIꢀOOH]2+ becomes
that corresponds to [RuIIꢀHOOH]2+ (or [RuIVdO H2]2+);
3 3 3
and (3) the onset of background-subtracted catalytic current in
CV scans shows a change in pH dependence from low values of
pH to higher values, with an ca. ꢀ59 mV/pH dependence at
pH > 8.5, consistent with a switchover from the [RuVdO]3+
pathway to a direct pathway involving [RuIVdO]2+. We assume
that an equilibrium concentration of a [RuIVdO H2O]2+
3 3 3
complex can exist in solution with a water molecule in close
proximity to the Ru center, which can be further oxidized and
deprotonated to form an OꢀO bonded species.
Finally, the proposed [RuIVꢀη1-OOH]2+ (Figure 12) was
not observed, probably due to facile proton-coupled oxidation
to form [RuIVꢀη1-OO]2+ and possibly further conversion to
[RuIVꢀη2-OO]2+. However, the puzzle that the lowest-en-
ergy state found from DFT calculations corresponds to a
[RuIVꢀη1-OO]2+ complex still remains. CASSCF calculations
favor the [RuIVꢀη2-OO]2+ state, but because their energies do
not reflect the dynamic correlation included in the DFT calcula-
tions, they are only semiquantative. Do state-of-the-art calcula-
tions have limitations in predicting the energetics, spin densities,
and vibrational properties of [RuIVꢀOO]2+ species?
’ CONCLUSIONS
This study describes the characterization of intermediates in
water oxidation catalyzed by a mononuclear ruthenium complex
containing a dinaphthyridyl pyridine ligand having two basic sites
not directly bound to the metal center. Electrochemistry,
UVꢀvis and resonance Raman spectroscopy, pulse radiolysis,
stopped flow, and ESI-MS with H218O labeling experiments, and
theoretical calculations were collectively used to identify the
electronic, spectroscopic, and structural properties of intermedi-
ates. The results reveal a number of intriguing properties of
intermediates such as [RuIVdO]2+ and [RuIVꢀOO]2+.
14662
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
The key findings are the following:
Present Addresses
§Department of Chemistry, University of Massachusetts, Boston,
MA 02125.
(a) The potential of the [RuIIIꢀOH]2+/[RuIIꢀOH2]2+ couple
is equal to or higher than the potential of the [RuIVdO]2+/
[RuIIꢀOH]2+ couple. The rate constant of the first oxidation
by Ce(IV) is (2.0 ( 0.3) ꢁ 104 Mꢀ1 sꢀ1 (at pH 1).
(b) At pH 1 both [RuIVdO]2+ and [RuVdO]3+ exist in
equilibrium in a ratio of ca. 1:1.
’ ACKNOWLEDGMENT
We thank Dr. Norman Sutin and Dr. Carol Creutz for helpful
discussions. The work at Brookhaven National Laboratory
(BNL) is funded under contract DE-AC02-98CH10886, and
the work at Houston is funded under contract DE-FG02-
07ER15888 with the U.S. Department of Energy and supported
by its Division of Chemical Sciences, Geosciences, & Bios-
ciences, Office of Basic Energy Sciences. The BNL authors also
thank the U.S. Department of Energy for funding under the BES
Hydrogen Fuel Initiative. RZ and RPT also thank the Robert A.
Welch Foundation (E-621).
(c) The formation of [RuIIIꢀOOH]2+ can proceed via forma-
tion of [RuVdO]3+ followed by nucleophilic attack by a
water molecule at pH < 1; however, this pathway cannot
account for the product formation at pH 6. An alternative
pathway was proposed for the reaction of [RuIVdO]2+ with
a water molecule accompanied by the concomitant removal
of an electron and a proton (“direct pathway”). The direct
pathway (or some other as of yet unidentified pathway)
becomes predominant at higher pH in underpotential
bulk electrolysis experiments and in the onset of catalytic
current in background-subtracted CV scans as a function
of pH.
’ REFERENCES
(1) Eisenberg, R.; Nocera, D. G. Inorg. Chem. 2005, 44, 6799–6801.
(2) Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci. U.S.A. 2006,
103, 15729–15735.
(3) Alstrum-Acevedo, J. H.; Brennaman, M. K.; Meyer, T. J. Inorg.
Chem. 2005, 44, 6802–6827.
(d) While there remains some conflicting evidence, the RuꢀOO
1
species is most likely [RuIVꢀη2-OO]2+ based on our
electrochemical, spectroscopic, and theoretical data alongside
a previously published conclusion by T. J. Meyer’s group.34
(e) [RuIVꢀOO]2+ slowly transforms to [RuIIꢀOOH]+ with-
out releasing oxygen, but in the presence of additional
Ce(IV) oxidant, [RuIVꢀOO]2+ yields O2.
(4) Brimblecombe, R.; Dismukes, G. C.; Swiegers, G. F.; Spiccia, L.
Dalton Trans. 2009, 9374–9384.
(5) Huynh, M. H.; Meyer, T. J. Chem. Rev. 2007, 107, 5004–5064.
(6) Eisenberg, R.; Gray, H. B. Inorg. Chem. 2008, 47, 1697–1699.
(7) Balzani, V.; Credi, A.; Venturi, M. ChemSusChem 2008, 1, 26–58.
(8) Youngblood, W. J.; Lee, S. H. A.; Maeda, K.; Mallouk, T. E. Acc.
Chem. Res. 2009, 42, 1966–1973.
(9) Cook, T. R.; Dogutan, D. K.; Reece, S. Y.; Surendranath, Y.;
Teets, T. S.; Nocera, D. G. Chem. Rev. 2010, 110, 6474–6502.
(10) Walter, M. G.; Warren, E. L.; McKone, J. R.; Boettcher, S. W.;
Mi, Q. X.; Santori, E. A.; Lewis, N. S. Chem. Rev. 2010, 110, 6446–6473.
(11) Dau, H.; Zaharieva, I. Acc. Chem. Res. 2009, 42, 1861–1870.
(12) Dau, H.; Limberg, C.; Reier, T.; Risch, M.; Roggan, S.; Strasser,
P. ChemCatChem 2010, 2, 724–761.
’ ASSOCIATED CONTENT
Supporting Information. CV of [RuIIꢀOH2]2+ (pH 1);
S
b
spectral changes of [RuIIꢀOH2]2+ upon Ce(IV) titration;
[Ce(IV)] dependence of disappearance of [RuIIꢀOH2]2+; kinetic
data for the reaction of [RuIIꢀOH2]2+ with the carbonate radical;
calculated UVꢀvis spectra for 1[RuIVdO]2+, 3[RuIVdO]2+,
[RuIIꢀOH2]2+, [RuIVꢀη1-OO]2+, and [RuIVꢀη2-OO]2+; active
space orbitals and orbital populations for 1[RuIVdO]2+,
3[RuIVdO]2+, 3[RuIVꢀη1-OO]2+, and 1[RuIVꢀη2ꢀOO]2+;
(13) Ghosh, P. K.; Brunschwig, B. S.; Chou, M.; Creutz, C.; Sutin, N.
J. Am. Chem. Soc. 1984, 106, 4772–4783.
3
2
computed spin densities for [RuIVdO]2+, [RuVdO]3+, and
3[RuIVꢀOO]2+; observed spectra for [RuIVdO]2+ at pH 1 and
10; NMR spectrum of the solution containing [RuIVꢀOO]2+,
[RuIVdO]2+, and [RuIV(tpy)(bpy)O]2+; CV of [RuIVdO]2+ in
CH3CN; ESI-MS of [RuIVꢀOO]2+ prepared in H216O and
H218O; resonance Raman spectra of [RuIIꢀOH2]2+ and
[RuIVꢀOO]2+; ESI-MS of the reaction mixture after addition of
Ce(IV); ESI-MS and NMR spectra of [RuIIꢀOOH]+; calculated
Pourbaix diagram; CV scans showing onset of background-
corrected catalytic current vs pH; calculated structures of key inter-
mediates; experimental LatimerꢀFrost diagram at pH 7; theoretical
characterization of [RuIVdO]2+ and [RuIVꢀOO]2+; details of
calculated electrochemistry in aqueous solution; several thermo-
dynamic cycles; schematic frontier orbital diagrams; table of calcu-
lated energies of [RuIVdO H2O]2+ and [RuIVꢀ(OO) H2O]2+;
(14) Ruttinger, W.; Dismukes, G. C. Chem. Rev. 1997, 97, 1–24.
(15) Gersten, S. W.; Samuels, G. J.; Meyer, T. J. J. Am. Chem. Soc.
1982, 104, 4029–4030.
(16) Yagi, M.; Kaneko, M. Chem. Rev. 2001, 101, 21–35.
(17) Wada, T.; Tsuge, K.; Tanaka, K. Inorg. Chem. 2001, 40,
329–337.
(18) Sens, C.; Romero, I.; Rodriguez, M.; Llobet, A.; Parella, T.;
Benet-Buchholz, J. J. Am. Chem. Soc. 2004, 126, 7798–7799.
(19) Hurst, J. K. Coord. Chem. Rev. 2005, 249, 313–328.
(20) Muckerman, J. T.; Polyansky, D. E.; Wada, T.; Tanaka, K.;
Fujita, E. Inorg. Chem. 2008, 47, 1787–1802.
(21) Cape, J. L.; Siems, W. F.; Hurst, J. K. Inorg. Chem. 2009, 48,
8729–8735.
(22) Bozoglian, F.; Romain, S.; Ertem, M. Z.; Todorova, T. K.; Sens,
C.; Mola, J.; Rodriguez, M.; Romero, I.; Benet-Buchholz, J.; Fontrodona,
X.; Cramer, C. J.; Gagliardi, L.; Llobet, A. J. Am. Chem. Soc. 2009, 131,
15176–15187.
(23) Xu, Y. H.; Akermark, T.; Gyollai, V.; Zou, D. P.; Eriksson, L.;
Duan, L. L.; Zhang, R.; Akermark, B.; Sun, L. C. Inorg. Chem. 2009, 48,
2717–2719.
(24) Concepcion, J. J.; Jurss, J. W.; Brennaman, M. K.; Hoertz, P. G.;
Patrocinio, A. O. T.; Iha, N. Y. M.; Templeton, J. L.; Meyer, T. J. Acc.
Chem. Res. 2009, 42, 1954–1965.
3
3
three tables for vibrational frequencies and their isotopic shifts of
MdO, MꢀO, and OꢀO bonds; calculated Cartesian coordinates
of key intermediates in water oxidation catalysis by [RuIIꢀ(OH2)
(H2O)]2+; and the complete ref 53. This material is available free of
charge via the Internet at http://pubs.acs.org.
(25) Duan, L. L.; Fischer, A.; Xu, Y. H.; Sun, L. C. J. Am. Chem. Soc.
2009, 131, 10397–10399.
(26) Jurss, J. W.; Concepcion, J. C.; Norris, M. R.; Templeton, J. L.;
Meyer, T. J. Inorg. Chem. 2010, 49, 3980–3982.
’ AUTHOR INFORMATION
Corresponding Author
dep@bnl.gov; fujita@bnl.gov
14663
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
(27) Wasylenko, D. J.; Ganesamoorthy, C.; Koivisto, B. D.; Berlinguette,
C. P. Eur. J. Inorg. Chem. 2010, 3135–3142.
(61) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72,
5639–5648.
(28) Zong, R.; Thummel, R. P. J. Am. Chem. Soc. 2005, 127,
12802–12803.
(29) Tseng, H. W.; Zong, R.; Muckerman, J. T.; Thummel, R. Inorg.
(62) Raghavachari, K.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem.
Phys. 1980, 72, 650–654.
(63) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54,
724.
Chem. 2008, 47, 11763–11773.
(30) Concepcion, J. J.; Jurss, J. W.; Templeton, J. L.; Meyer, T. J.
J. Am. Chem. Soc. 2008, 130, 16462–16463.
(31) McDaniel, N. D.; Coughlin, F. J.; Tinker, L. L.; Bernhard, S.
(64) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972,
56, 2257–2261.
(65) Hariharan, P. C.; Pople, J. A. Theor. Chem. Acc. 1973, 28, 213–222.
(66) Hariharan, P. C.; Pople, J. A. Mol. Phys. 1974, 27, 209–214.
(67) Gordon, M. S. Chem. Phys. Lett. 1980, 76, 163–168.
(68) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.;
DeFrees, D. J.; Pople, J. A.; Gordon, M. S. J. Chem. Phys. 1982,
77, 3654–3665.
(69) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995–2001.
(70) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem.
2003, 24, 669–681.
(71) Klamt, A.; Sch€u€urmann, G. J. Chem. Soc., Perkin Trans. 2
1993, 799–805.
(72) Wadsworth, E.; Duke, F. R.; Goetz, C. A. Anal. Chem. 1957,
29, 1824–1825.
J. Am. Chem. Soc. 2008, 130, 210–217.
(32) Masaoka, S.; Sakai, K. Chem. Lett. 2009, 38, 182–183.
(33) Hull, J. F.; Balcells, D.; Blakemore, J. D.; Incarvito, C. D.;
Eisenstein, O.; Brudvig, G. W.; Crabtree, R. H. J. Am. Chem. Soc. 2009,
131, 8730–8731.
(34) Concepcion, J. J.; Tsai, M. K.; Muckerman, J. T.; Meyer, T. J.
J. Am. Chem. Soc. 2010, 132, 1545–1557.
(35) Wasylenko, D. J.; Ganesamoorthy, C.; Koivisto, B. D.; Henderson,
M. A.; Berlinguette, C. P. Inorg. Chem. 2010, 49, 2202–2209.
(36) Wasylenko, D. J.; Ganesamoorthy, C.; Henderson, M. A.;
Koivisto, B. D.; Berlinguette, C. P. J. Am. Chem. Soc. 2010, 132,
16094–16106.
(37) Kanan, M. W.; Nocera, D. G. Science 2008, 321, 1072–1075.
(38) Romain, S.; Vigara, L.; Llobet, A. Acc. Chem. Res. 2009, 42,
1944–1953.
(73) Bishop, E.; Cofre, P. Analyst 1981, 106, 316–322.
(74) Bondareva, T. N.; Stromberg, A. G. Zh. Obshch. Khim. 1955,
25, 666–671.
(39) Sartorel, A.; Miro, P.; Salvadori, E.; Romain, S.; Carraro, M.;
Scorrano, G.; Di Valentin, M.; Llobet, A.; Bo, C.; Bonchio, M. J. Am.
Chem. Soc. 2009, 131, 16051–16053.
(40) Geletii, Y. V.; Botar, B.; Koegerler, P.; Hillesheim, D. A.;
Musaev, D. G.; Hill, C. L. Angew. Chem., Int. Ed. 2008, 47, 3896–3899.
(41) Geletii, Y. V.; Huang, Z. Q.; Hou, Y.; Musaev, D. G.; Lian, T. Q.;
Hill, C. L. J. Am. Chem. Soc. 2009, 131, 7522–7523.
(42) Sala, X.; Romero, I.; Rodriguez, M.; Escriche, L.; Llobet, A.
Angew. Chem., Int. Ed. 2009, 48, 2842–2852.
(43) Romero, I.; Rodriguez, M.; Sens, C.; Mola, J.; Kollipara, M. R.;
Francas, L.; Mas-Marza, E.; Escriche, L.; Llobet, A. Inorg. Chem. 2008,
47, 1824–1834.
(44) Duan, L.; Xu, Y.; Zhang, P.; Wang, M.; Sun, L. Inorg. Chem.
2009, 49, 209–215.
(45) Duan, L.; Xu, Y. H.; Gorlov, M.; Tong, L. P.; Andersson, S.;
Sun, L. C. Chem.—Eur. J. 2010, 16, 4659–4668.
(46) Nyhlen, J.; Duan, L. L.; Akermark, B.; Sun, L. C.; Privalov, T.
Angew. Chem., Int. Ed. 2010, 49, 1773–1777.
(47) Zhang, G.; Zong, R.; Tseng, H.-W.; Thummel, R. P. Inorg.
Chem. 2008, 47, 990–998.
(75) Maverick, A. W.; Yao, Q. Inorg. Chem. 1993, 32, 5626–5628.
(76) Masllorens, E.; Rodriguez, M.; Romero, I.; Roglans, A.; Parella,
T.; Benet-Buchholz, J.; Poyatos, M.; Llobet, A. J. Am. Chem. Soc. 2006,
128, 5306–5307.
(77) Yoshida, M.; Masaoka, S.; Abe, J.; Sakai, K. Chem.--Asian J.
2010, 5, 2369–2378.
(78) Huie, R. E.; Clifton, C. L.; Neta, P. Int. J. Radiat. Appl. Instrum.
Part C 1991, 38, 477–481.
(79) The numerical integration feature is a part of the PRWin
software package developed at BNL for data aquisition and analysis of
pulse radiolysis data.
(80) Chen, S.-N.; Cope, V. W.; Hoffman, M. Z. J. Phys. Chem. 1973,
77, 1111–1116.
(81) Weeks, J. L.; Rabani, J. J. Phys. Chem. 1966, 70, 2100–2106.
(82) Moyer, B. A.; Meyer, T. J. Inorg. Chem. 1981, 20, 436–444.
(83) Dobson, J. C.; Seok, W. K.; Meyer, T. J. Inorg. Chem. 1986, 25,
1513–1514.
(84) Roecker, L.; Meyer, J. J. Am. Chem. Soc. 1987, 109, 746–754.
(85) Che, C.-M.; Tang, W.-T.; Wong, W.-T.; Lai, T.-F. J. Am. Chem.
Soc. 1989, 111, 9048–9056.
(48) Takeuchi, K. J.; Thompson, M. S.; Pipes, D. W.; Meyer, T. J.
Inorg. Chem. 1984, 23, 1845–1851.
(86) Che, C.-M.; Yam, V. W. W. J. Am. Chem. Soc. 1987, 109,
1262–1263.
(49) Concepcion, J. J.; Jurss, J. W.; Norris, M. R.; Chen, Z. F.;
Templeton, J. L.; Meyer, T. J. Inorg. Chem. 2010, 49, 1277–1279.
(50) Chen, Z. F.; Concepcion, J. J.; Hu, X. Q.; Yang, W. T.; Hoertz,
P. G.; Meyer, T. J. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 7225–7229.
(51) Chen, Z. F.; Concepcion, J. J.; Hull, J. F.; Hoertz, P. G.; Meyer,
T. J. Dalton Trans 2010, 39, 6950–6952.
(87) Benet-Buchholz, J.; Comba, P.; Llobet, A.; Roeser, S.; Vadivelu,
P.; Wadepohl, H.; Wiesner, S. Dalton Trans. 2009, 5910–5923.
(88) Kojima, T.; Hirai, Y.; Ishizuka, T.; Shiota, Y.; Yoshizawa, K.;
Ikemura, K.; Ogura, T.; Fukuzumi, S. Angew. Chem., Int. Ed. 2010, 49,
8449–8453.
(89) Yamada, H.; Koike, T.; Hurst, J. K. J. Am. Chem. Soc. 2001, 123,
(52) Zanello, P. Inorganic Electrochemistry. Theory, Practice and
Application; The Royal Society of Chemistry: Cambridge, 2003.
(53) Gaussian09, revision B.01; Gaussian, Inc.: Wallingford, CT,
2009 (see SI for complete reference).
12775–12780.
(90) Paeng, I. R.; Nakamoto, K. J. Am. Chem. Soc. 1990, 112,
3289–3297.
(91) Fujii, H.; Kurahashi, T.; Tosha, T.; Yoshimura, T.; Kitagawa, T.
(54) Becke, A. D. Phys. Rev. A 1988, 38, 3098–3100.
(55) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37,
785–789.
J. Inorg. Biochem. 2006, 100, 533–541.
(92) Almond, M. J.; Crayston, J. A.; Downs, A. J.; Poliakoff, M.;
Turner, J. J. Inorg. Chem. 1986, 25, 19–25.
(56) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58,
1200–1211.
(93) Bajdor, K.; Nakamoto, K.; Kanatomi, H.; Murase, I. Inorg. Chim.
Acta 1984, 82, 207–210.
(57) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(58) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J.
J. Phys. Chem. 1994, 98, 11623–11627.
(59) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H. Preuss.
Theor. Chim. Acta 1990, 77, 123–141.
(94) Diemente, D.; Hoffman, B. M.; Basolo, F. J. Chem. Soc., Chem.
Commun. 1970, 467–468.
(95) Nakamoto, K.; Nonaka, Y.; Ishiguro, T.; Urban, M. W.; Suzuki,
M.; Kozuka, M.; Nishida, Y.; Kida, S. J. Am. Chem. Soc. 1982, 104,
3386–3391.
(60) Martin, J. M. L.; Sundermann, A. J. Chem. Phys. 2001, 114,
3408–3420.
(96) Shan, X. P.; Que, L. Proc. Natl. Acad. Sci. U.S.A. 2005, 102,
5340–5345.
14664
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665

Journal of the American Chemical Society
ARTICLE
(97) Bakac, A.; Scott, S. L.; Espenson, J. H.; Rodgers, K. R. J. Am.
Chem. Soc. 1995, 117, 6483–6488.
(98) Spiro, T. G. In Iron Porphyrins; Lever, A. B. P., Gray, H. B., Eds.;
Addison-Wesley: Reading, MA, 1983; Vol. 2, pp 89ꢀ159.
(99) Wagner, W. D.; Paeng, I. R.; Nakamoto, K. J. Am. Chem. Soc.
1988, 110, 5565–5567.
(100) Nakamura, A.; Tatsuno, Y.; Yamamoto, M.; Otsuka, S. J. Am.
Chem. Soc. 1971, 93, 6052–6058.
(101) Horn, R. W.; Weissberger, E.; Collman, J. P. Inorg. Chem.
1970, 9, 2367–2371.
(102) Kryatov, S. V.; Chavez, F. A.; Reynolds, A. M.; Rybak-
Akimova, E. V.; Que, L.; Tolman, W. B. Inorg. Chem. 2004,
43, 2141–2150.
(103) Bajdor, K.; Oshio, H.; Nakamoto, K. J. Am. Chem. Soc. 1984,
106, 7273–7274.
(104) Brunold, T. C.; Solomon, E. I. J. Am. Chem. Soc. 1999, 121,
8277–8287.
(105) Mizoguchi, T. J.; Lippard, S. J. J. Am. Chem. Soc. 1998, 120,
11022–11023.
(106) He, C.; Barrios, A. M.; Lee, D.; Kuzelka, J.; Davydov, R. M.;
Lippard, S. J. J. Am. Chem. Soc. 2000, 122, 12683–12690.
(107) MacMurdo, V. L.; Zheng, H.; Que, L. Inorg. Chem. 2000,
39, 2254–2255.
(108) Momenteau, M.; Reed, C. A. Chem. Rev. 1994, 94, 659–698.
(109) Suzuki, M.; Ishiguro, T.; Kozuka, M.; Nakamoto, K. Inorg.
Chem. 1981, 20, 1993–1996.
(110) Nakamoto, K.; Suzuki, M.; Ishiguro, T.; Kozuka, M.; Nishida,
Y.; Kida, S. Inorg. Chem. 1980, 19, 2822–2824.
(111) Nakamoto, K.; Watanabe, T.; Ama, T.; Urban, M. W. J. Am.
Chem. Soc. 1982, 104, 3744–3745.
(112) The shifts of 40ꢀ60 cmꢀ1 can be identified for several peak
combinations in the 760ꢀ1050 cmꢀ1 range.
(113) Wang, L. P.; Wu, Q.; Van Voorhis, T. Inorg. Chem. 2010, 49,
4543–4553.
14665
dx.doi.org/10.1021/ja203249e |J. Am. Chem. Soc. 2011, 133, 14649–14665