Half-sandwich iridium complexes for homogeneous water-oxidation catalysis

Article abstract of DOI:10.1021/ja104775j

Iridium half-sandwich complexes of the types Cp*Ir(N-C)X, [Cp*Ir(N-N)X]X, and [CpIr(N-N)X]X are catalyst precursors for the homogeneous oxidation of water to dioxygen. Kinetic studies with cerium(IV) ammonium nitrate as primary oxidant show that oxygen evolution is rapid and continues over many hours. In addition, [Cp*Ir(H₂O) ₃]SO₄ and [(Cp*Ir)₂(μ-OH) ₃]OH can show even higher turnover frequencies (up to 20 min ^-1 at pH 0.89). Aqueous electrochemical studies on the cationic complexes having chelate ligands show catalytic oxidation at pH > 7; conversely, at low pH, there are no oxidation waves up to 1.5 V vs NHE for the complexes. H₂¹⁸O isotope incorporation studies demonstrate that water is the source of oxygen atoms during cerium(IV)-driven catalysis. DFT calculations and kinetic experiments, including kinetic-isotope-effect studies, suggest a mechanism for homogeneous iridium-catalyzed water oxidation and contribute to the determination of the rate-determining step. The kinetic experiments also help distinguish the active homogeneous catalyst from heterogeneous nanoparticulate iridium dioxide.

Full text of DOI:10.1021/ja104775j

Published on Web 10/21/2010
Half-Sandwich Iridium Complexes for Homogeneous
Water-Oxidation Catalysis
†
†
‡,§
†
James D. Blakemore, Nathan D. Schley, David Balcells, Jonathan F. Hull,
#
†
,§
,†
Gerard W. Olack, Christopher D. Incarvito, Odile Eisenstein,* Gary W. Brudvig,*
,
†
and Robert H. Crabtree*
Department of Chemistry, Yale UniVersity, P.O. Box 208107, New HaVen,
Connecticut 06520, United States, Departament de Qu ´ı mica, UniVersitat Aut o` noma de
Barcelona, 08193 Bellaterra, Barcelona, Catalonia, Spain, Institut Charles Gerhardt, UniVersite´
Montpellier 2, CNRS 5253, cc 15001, Place Eug e` ne Bataillon 34095, Montpellier, France, and
Department of Geology and Geophysics, Yale UniVersity, P.O. Box 208109, New HaVen,
Connecticut 06520, United States
Received June 1, 2010; E-mail: robert.crabtree@yale.edu; gary.brudvig@yale.edu;
odile.eisenstein@univ-montp2.fr
Abstract: Iridium half-sandwich complexes of the types Cp*Ir(N-C)X, [Cp*Ir(N-N)X]X, and [CpIr(N-N)X]X
are catalyst precursors for the homogeneous oxidation of water to dioxygen. Kinetic studies with cerium(IV)
ammonium nitrate as primary oxidant show that oxygen evolution is rapid and continues over many hours.
In addition, [Cp*Ir(H
2
2
O)
0 min at pH 0.89). Aqueous electrochemical studies on the cationic complexes having chelate ligands
show catalytic oxidation at pH > 7; conversely, at low pH, there are no oxidation waves up to 1.5 V vs
3 4 2 3
]SO and [(Cp*Ir) (µ-OH) ]OH can show even higher turnover frequencies (up to
-
1
18
NHE for the complexes. H
2
O isotope incorporation studies demonstrate that water is the source of oxygen
atoms during cerium(IV)-driven catalysis. DFT calculations and kinetic experiments, including kinetic-isotope-
effect studies, suggest a mechanism for homogeneous iridium-catalyzed water oxidation and contribute to
the determination of the rate-determining step. The kinetic experiments also help distinguish the active
homogeneous catalyst from heterogeneous nanoparticulate iridium dioxide.
Introduction
Beyond the OEC, other synthetic catalysts are known for
water oxidation. Manganese model systems for the oxygen-
4
The oxidation of water to dioxygen is carried out by green
plants, algae, and cyanobacteria in order to obtain the reducing
equivalents required for sustaining life processes. Consequently,
evolving complex have been described and are most effectively
driven by two-electron oxo-donating primary oxidants such as
5
Oxone (potassium peroxymonosulfate). In contrast, ruthenium
6
the mechanism of water oxidation has attracted considerable
water oxidation catalysts have been shown to operate with one-
1
attention. The oxygen-evolving complex (OEC), a Mn
4
Ca oxo-
electron primary oxidants such as cerium(IV) ammonium
7
bridged cluster in the enzyme photosystem II, catalyzes water
nitrate. These one-electron oxidants may better mimic the
2
oxidation. This catalysis is thought to involve formation of a
multiple, sequential, single-electron transfer events which occur
in natural photosynthesis, and studies on water-oxidation
catalysts driven by these one-electron oxidants show promise
for adaptation to light-driven systems (i.e., artificial photosyn-
thesis). However, improving our understanding of water-
oxidation catalysis is a key hurdle to realizing these alternative
high-valent manganese oxo species, which undergoes nucleo-
philic attack by water, forming an O-O bond. A schematic
version of the proposed pathway is given in eq 1.
3
MdO]n+ + OH f M
(n-2)+
+ O + 2H + 2e
+
-
(1)
[
8
2
2
energy schemes.
In 2008, Bernhard and co-workers reported the first mono-
9
nuclear iridium water oxidation catalyst. By treating a chloride-
†
‡
§
#
bridged iridium(III) dimer containing cyclometalated phenylpy-
ridine ligands with silver triflate, they were able to generate
Department of Chemistry, Yale University.
Universitat Aut o` noma de Barcelona.
Institut Charles Gerhardt.
Department of Geology and Geophysics, Yale University.
(
(
1) (a) McEvoy, J. P.; Brudvig, G. W. Chem. ReV. 2006, 106, 4455–
(4) Cady, C. W.; Crabtree, R. H.; Brudvig, G. W. Coord. Chem. ReV.
2008, 252, 444–455.
4
483. (b) Dau, H.; Limberg, C.; Reier, T.; Risch, M.; Roggan, S.;
Strasser, P. ChemCatChem 2010, 2, 724–761.
2) Ferreira, K. N.; Iverson, T. M.; Maghloaoui, K.; Barber, J.; Iwata, S.
Science 2001, 303, 1831–1838.
(5) (a) Limburg, J.; Vrettos, J. S.; Liable-Sands, L. M.; Rheingold, A. L.;
Crabtree, R. H.; Brudvig, G. W. Science 1999, 283, 1524–1527. (b)
Limburg, J.; Vrettos, J. S.; Chen., H.; de Paula, J. C.; Crabtree, R. H.;
Brudvig, G. W. J. Am. Chem. Soc. 2001, 123, 423–430. (c) Tagore,
R.; Crabtree, R. H.; Brudvig, G. W. Inorg. Chem. 2008, 47, 1815–
1823. (d) Kurz, P.; Berggren, G.; Anderlund, M. F.; Styring, S. Dalton
Trans. 2007, 4258–4261.
(3) (a) Brudvig, G. W.; Crabtree, R. H. Proc. Natl. Acad. Sci. U.S.A. 1986,
8
1
3, 4586–4588. (b) Brudvig, G. W. Phil. Trans. R. Soc. B 2008, 363,
211–1219. (c) Brudvig, G. W.; Thorp, H. H.; Crabtree, R. H. Acc.
Chem. Res. 1991, 24, 311–316.
10.1021/ja104775j  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 16017–16029 9 16017

A R T I C L E S
Blakemore et al.
Figure 1. Iridium(III) catalysts for homogeneous water oxidation.
cationic bis-solvento iridium(III) catalyst precursors that oxidize
water to dioxygen in the presence of cerium(IV) ammonium
nitrate.
clear first-order kinetics. In contrast, the chelate-bound precur-
sors 1, 3, and 8 follow clear first-order kinetics over the entire
range studied with cerium(IV) ammonium nitrate. Furthermore,
moving to N-N chelate ligands yields cationic complexes 3-6
that are fully soluble in water, even in the iridium(III) state.
This makes these systems more amenable to detailed kinetic
analysis and aqueous electrochemical studies.
The presence of potentially oxidizable methyl groups on the
Cp* ligand encouraged us to examine the behavior of the
corresponding cyclopentadienyl (Cp) complexes. We now report
synthesis and characterization of Cp complexes 7 and 8, which
have both Cp and bipyridine in the inner coordination sphere
of iridium(III). Interestingly, these are the first iridium(III)
complexes reported with both Cp and chelate ligands, due to
the synthetic challenges associated with preparation of the Cp-
containing precursors. These complexes are generally less active
for water oxidation than their Cp* analogues and show lower
activities per iridium at higher catalyst loadings, which suggests
a multinuclear deactivation pathway or change in the rate-
determining step. However, long time-course oxygen-evolution
measurements suggest that the Cp complexes are less susceptible
to deactivation.
Since numerous iridium catalysts were available to us from
work in other areas, we decided to test a selection for water-
oxidizing chemistry. We focused on a catalyst precursor
containing the pentamethylcyclopentadienyl ligand (Cp*), which
we hoped would provide an electron rich environment, likely
to stabilize the necessary high-valent intermediates required for
water oxidation. In a preliminary communication, we reported
Cp*Ir complexes 1 and 2 (Figure 1) as highly active catalyst
10
precursors for water oxidation. These Cp*Ir(N-C)X catalysts
have turnover frequencies far above those reported for many
other systems, on the order of ten turnovers per minute.
Furthermore, the catalysts are mononuclear, are easy to make,
and continue to catalyze oxygen evolution over hours.
We now report full details on 1 and 2 and also compare their
behavior with related complexes 3-12 (Figure 1). In particular,
the tris-aqua complex 11 and the dimer 12 can show a greater
rate of oxygen evolution than that of 1-10 but do not follow
(
6) (a) Romain, S.; Vigara, L.; Llobet, A. Acc. Chem. ReV. 2009, 42, 1944–
1
953. (b) Concepcion, J. J.; Jurss, J. W.; Brennaman, M. K.; Hoertz,
A key concern in organometallic systems of the type
presented here is the possibility of ligand oxidation and even
degradation; in such a case, heterogeneous materials such as
P. G.; Patrocinio, A. O. T.; Iha, N. Y. M.; Templeton, J. L.; Meyer,
T. J. Acc. Chem. ReV. 2009, 42, 1954–1965. (c) Romero, I.; Rodr ´ı guez,
M.; Sens, C.; Mola, J.; Kollipara, M. R.; Franc a` s, L.; Mas-Marza, E.;
Escriche, L.; Llobet, A. Inorg. Chem. 2008, 47, 1824–1834. (d) Sens,
C.; Romero, I.; Rodriguez, M.; Llobet, A.; Parella, P.; Benet-Buchholz,
J. J. Am. Chem. Soc. 2004, 126, 7798–7799. (e) Gilbert, J. A.;
Eggleston, D. S.; Murphy, W. R.; Geselowitz, D. A.; Gersten, S. W.;
Hodgson, D. J.; Meyer, T. J. J. Am. Chem. Soc. 1985, 107, 3855–
11
IrO could be the true catalyst. This potential heterogeneous-
2
homogeneous ambiguity is a problem in which we have long
12
been interested. We carefully probed this question and find
that the evidence favors a homogeneous origin for the catalysis
3
864. (f) Concepcion, J. J.; Jurss, J. W.; Templeton, J. L.; Meyer,
at short time scales. In particular, k
rate data show distinct behavior for authentic IrO
, although this difference in isotope effect cannot rigorously
/k
data based on initial
H
D
T. J. J. Am. Chem. Soc. 2008, 130, 16462–16463. (g) Concepcion,
J. J.; Jurss, J. W.; Templeton, J. L.; Meyer, T. J. Proc. Natl. Acad.
Sci. U.S.A. 2008, 105, 17632–17635. (h) Sala, X.; Romero, I.;
Rodriguez, M.; Escriche, L.; Llobet, A. Angew. Chem., Int. Ed. 2009,
2
versus catalyst
3
rule out the formation of unusual form(s) of iridium oxide from
the various organometallic iridium precursors.
4
8, 2842–2852. (i) Wasylenko, D. J.; Ganesamoorthy, C.; Koivisto,
B. D.; Henderson, M. A.; Berlinguette, C. P. Inorg. Chem. 2010, 49,
202–2209. (j) Duan, L. L.; Fischer, A.; Xu, Y. H.; Sun, L. C. J. Am.
Water oxidation has been studied computationally for several
2
1
3
14
Chem. Soc. 2009, 131, 10397–10398. (k) Hurst, J. K.; Cape, J. L.;
Clark, A. E.; Das, S.; Qin, C. Y. Inorg. Chem. 2008, 47, 1753–1764.
catalytic systems, including mononuclear, dinuclear and
15
polyoxometalate Ru complexes and the oxygen-evolving
complex of Photosystem II. The reaction mechanisms pro-
(
l) Concepcion, J. C.; Jurss, J. W.; Norris, M. R.; Chen, Z.; Templeton,
16
J. L.; Meyer, T. J. Inorg. Chem. 2010, 49, 1277–1279.
(
(
7) Sridharan, V.; Men e´ ndez, J. C. Chem. ReV. 2010, 110, 3805–3849.
8) (a) Youngblood, W. J.; Lee, S.-H. A.; Kobayashi, Y.; Hernandez-
Pagan, E. A.; Hoertz, P. G.; Moore, T. A.; Moore, A. L.; Gust, D.;
Mallouk, T. E. J. Am. Chem. Soc. 2009, 131, 926–927. (b) Gust, D.;
Moore, T. A.; Moore, A. L. Acc. Chem. ReV. 2009, 42, 1890–1898.
posed to date, which involve the participation of a high-valent
metal oxo center, can be classified in two major categories: (1)
O
2
formation by intramolecular coupling of two oxo ligands
and (2) O formation by intermolecular addition of H O to an
oxo ligand.
2
2
(
c) Hetterscheid, D. G. H.; van der Vlugt, J. I.; de Bruin, B.; Reek,
J. N. H. Angew. Chem., Int. Ed. 2009, 48, 8178–8181. (d) Hurst, J. K.
Science 2010, 328, 315–316.
(
9) McDaniel, N. D.; Coughlin, F. J.; Tinker, L. L.; Bernhard, S. J. Am.
Chem. Soc. 2008, 130, 210–217.
(11) (a) Harriman, A.; Thomas, J. M.; Millward, G. R. N. J. Chem. 1987,
11, 757–762. (b) Nakagawa, T.; Bjorge, N. S.; Murray, R. W. J. Am.
Chem. Soc. 2009, 131, 15578–15579. (c) Nakagawa, T.; Beasley,
C. A.; Murray, R. W. J. Phys. Chem. C 2009, 113, 12958–12961.
(12) Anton, D. R.; Crabtree, R. H. Organometallics 1983, 2, 855–859.
(
10) 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,
8
730–8731.
1
6018 J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010

Half-Sandwich Iridium Complexes
A R T I C L E S
environment; Table 1 shows the catalytic rates for 1-12 and
some simple precursor materials. The rate of oxygen evolution
is greater for 3 than for any of the analogous complexes 4-6.
Specifically, the tmeda complex 6 shows a slower rate than the
bipyridine complex 3. Kinetic studies suggest that tmeda is
quickly protonated and lost from the iridium center under the
catalytic conditions, giving essentially the same precursor as
the tris-aqua complex 11. Both 6 and 11 show a non-first-order
rate dependence on catalyst concentration (vide infra), and at 5
µM catalyst, the rates of oxygen evolution from 6 and 11 are
nearly identical (see Supporting Information).
The presence of potentially oxidizable, benzylic C-H bonds
on the Cp* ligand could provide a site for ligand modification
under the catalytic conditions. Oxidation of the Cp* ligand of
[
Cp*IrCl]
2
has been observed upon treatment with iodine(III)
reagents to give acetoxylated or hydroxylated products, depend-
17
Figure 2. Reaction mechanism postulated for iridium-catalyzed water
oxidation.
ing on the reaction conditions. To investigate whether a similar
ligand oxidation was taking place and was involved in either
catalyst decomposition or catalyst modification, we replaced the
Cp* ligand with Cp. In addition to removing the oxidizable
methyl groups, this change also affects the steric effect and
A plausible reaction mechanism for iridium-catalyzed water
oxidation, taking into account the mononuclear nature of our
catalysts, is shown in Figure 2. With a neutral chloro catalyst,
18
donor power of the ligand and exchange rates of the ancillary
-
such as 1 or 2, the Cl ligand is replaced by a molecule of H
2
O
19
ligands.
yielding a cationic Ir(III) aqua complex. This intermediate is
oxidized by cerium(IV) to an Ir(V) oxo complex, which acts as
active species in the formation of the O-O bond by reacting
with water. The oxidation of the resulting peroxo intermediate
In contrast to the wealth of literature on Cp* iridium
complexes, Cp complexes of iridium are much less common.
This difference can be ascribed to the lack of an easily accessible
and versatile starting material for their synthesis comparable to
yields the final O
coordination of H
step is characterized by means of DFT calculations.
2
product, and the catalyst is recovered by
O. In this paper, the key O-O bond formation
[
2 2
Cp*IrCl ] , a readily prepared, air-stable material. The analo-
2
gous CpIr dihalide materials are polymeric and not directly
accessible from hydrated iridium trichloride. Instead, for the
synthesis of 7 and 8 (Figure 4), we employed a procedure
Results and Discussion
2
0
developed by Heinekey and co-workers involving direct
Water Oxidation Catalyzed by 1-12. Complexes 3-12 were
synthesized and characterized (see Experimental Section and
Supporting Information), and no unusual features were encoun-
oxidation of CpIr(C with I to give [CpIrI . Employing
2
H
4
)
2
2
2 x
]
2
CpIr(C
2
H
4
)
2
instead of the more easily prepared CpIr(η -
cyclooctene)
2
allowed us to obtain exceptionally pure material
1
0
tered. In relation to our previously reported catalysts 1 and 2,
2
-6 replace the LX-type cyclometalated ligands with the L -
by sublimation of the bis-ethylene complex.
3
Looking at the Cp catalysts, we find that 7, the analogue of
Cp* complex 3, has a slower initial turnover rate. The rates of
oxygen evolution from nitrate 7 and hexafluorophosphate 8,
type bipyridine, phenanthroline, 2,2′-bipyrimidine, and tetram-
ethylethylenediamine (tmeda) ligands. These compounds were
isolated as their water-soluble chloride salts, having both one
inner- and one outer-sphere chloride. The crystal structures of
the new compounds 6 and 7 (Figure 3) confirm the expected
atom connectivity.
All the complexes in Figure 1 catalyze water oxidation with
cerium(IV) as primary oxidant. The initial, maximal rate of
oxygen evolution is affected by changes in coordination
21
show essentially similar rates of oxygen evolution. The nitrate
salt 7 has excellent solubility in water, and this complex was
(
15) (a) Kuznetsov, A. E.; Geletii, Y. V.; Hill, C. L.; Morokuma, K.;
Musaev, D. G. J. Am. Chem. Soc. 2009, 131, 6844–6854. (b) Geletii,
Y. V.; Huang, Z.; Hou, Y.; Musaev, D. G.; Lian, T.; Hill, C. L. J. Am.
Chem. Soc. 2009, 131, 7522–7523. (c) Geletii, Y. V.; Besson, C.; Hou,
Y.; Yin, Q.; Musaev, D. G.; Qui n˜ onero, D.; Cao, R.; Hardcastle, K. I.;
Proust, A.; K o¨ gerler, P.; Hill, C. L. J. Am. Chem. Soc. 2009, 131,
17360–17370. (d) Geletii, Y. V.; Botar, B.; K o¨ gerler, P.; Hillesheim,
D. A.; Musaev, D. G.; Hill, C. L. Angew. Chem., Int. Ed. 2008, 47,
3896–3899. (e) 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.
(
13) (a) Kohl, S. W.; Weiner, L.; Schwartsburd, L.; Konstantinovski, L.;
Shimon, L. J. W.; Ben-David, Y.; Iron, M. A.; Milstein, D. Science
2
009, 324, 74–77. (b) Li, J.; Shiota, Y.; Yoshizawa, K. J. Am. Chem.
Soc. 2009, 131, 13584–13585. (c) Yang, X.; Hall, M. B. J. Am. Chem.
Soc. 2010, 132, 120–130. (d) 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.
(16) (a) Lundberg, M.; Blomberg, M. R. A.; Siegbahn, P. E. M. Inorg.
Chem. 2004, 43, 264–274. (b) Siegbahn, P. E. M. Inorg. Chem. 2008,
47, 1779–1786. (c) Siegbahn, P. E. M. Acc. Chem. Res. 2009, 42,
1871–1880. (d) Siegbahn, P. E. M. J. Am. Chem. Soc. 2009, 131,
18238–18239. (e) Sproviero, E. M.; Gasc o´ n, J. A.; McEvoy, J. P.;
Brudvig, G. W.; Batista, V. S. J. Am. Chem. Soc. 2008, 130, 6728–
6730.
2
010, 107, 7225–7229. (e) Wang, L.-P.; Wu, Q.; Voorhis, T. V. Inorg.
Chem. 2010, 49, 4543–4553.
(
14) (a) Bartolotti, L. J.; Pedersen, L. G.; Meyer, T. J. Int. J. Quantum
Chem. 2001, 83, 143–149. (b) Yang, X.; Baik, M.-H. J. Am. Chem.
Soc. 2004, 126, 13222–13223. (c) Yang, X.; Baik, M.-H. J. Am. Chem.
Soc. 2006, 128, 7476–7485. (d) Yang, X.; Baik, M.-H. J. Am.
Chem. Soc. 2008, 130, 16231–16240. (e) Romain, S.; Bozoglian, F.;
Sala, X.; Llobet, A. J. Am. Chem. Soc. 2009, 131, 2768–2769. (f)
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,
(17) Park-Gehrke, L. S.; Freudenthal, J.; Kaminsky, W.; DiPasquale, A. G.;
Mayer, J. M. Dalton Trans. 2009, 1972–1983.
(18) Fendrick, C. M.; Schertz, L. D.; Mintz, E. A.; Marks, T. J. Inorg.
Synth. 1992, 29, 193–238.
(19) (a) Wang, D.; Angelici, R. J. Inorg. Chem. 1996, 35, 1321–1331. (b)
Hughes, R. P.; Lindner, D. C.; Liable-Sands, L. M.; Rheingold, A. L.
Organometallics 2001, 20, 363–366.
1
31, 15176–15187. (g) Batista, E. R.; Martin, R. L. J. Am. Chem.
Soc. 2007, 129, 7224–7225. (h) Concepcion, J. J.; Tsai, M.-K.;
Muckerman, J. T.; Meyer, T. J. J. Am. Chem. Soc. 2010, 132, 1545–
(20) Heinekey, D. M.; Millar, J. M.; Koetzle, T. F.; Payne, N. G.; Zilm,
K. W. J. Am. Chem. Soc. 1990, 112, 909–919.
1
557.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010 16019

A R T I C L E S
Blakemore et al.
Figure 3. Crystal structures of 6 (left) and 7 (right) with ellipsoids shown at 50% probability. Counteranions have been omitted for clarity.
Table 1. Rates of Iridium-Catalyzed Oxygen Evolution
complexes 9 and 10 were studied and proved to be excellent
water oxidation catalysts with similar activity to each other.
These two ligands are likely lost under the catalytic conditions
initial rate
TOF (t.o.
b
min )
(µmol L- min
1
-1
a
)
-1
compound
formula
2+
3
4
5
6
7
8
9
1
1
1
1
Cp*Ir(bipy)Cl
2
72 ( 3
19 ( 2
42 ( 2
31 ( 3
46 ( 3
38 ( 2
18 ( 1
17 ( 1
27.2 ( 0.5
99 ( 9
49 ( 4
101 ( 5
14.4 ( 0.7
3.9 ( 0.4
8.4 ( 0.4
6.3 ( 0.6
9.3 ( 0.6
7.6 ( 0.3
3.5 ( 0.2
3.2 ( 0.1
5.5 ( 0.1
10.4 ( 0.9
10.0 ( 0.9
10.1 ( 0.5
0.43 ( 0.05
0.08 ( 0.01
to yield the CpIr(H
2
O)
3
fragment, analogous to 11.
c
Cp*Ir(bipyrim)Cl
Cp*Ir(phen)Cl
Cp*Ir(tmeda)Cl
[CpIr(bipy)I]NO
[CpIr(bipy)I]PF
2
To test the possibility that all the ligands on the complexes
2
1
-12 were removed by oxidation, we looked at iridium
2
trichloride. When injected into cerium(IV) solutions, IrCl
shows a turnover frequency of 0.4 min , or 3% of that for 3.
Surprisingly, potassium hexachloroiridate, K IrCl , a common
precursor to oxide materials of iridium shows a near-zero rate
of oxygen evolution. This is consistent with results obtained
previously.
3
2
·nH O
3
c
,
d
-1
6
CpIr(PPh
CpIr(PPh
3
)COI
)CO
2
2
6
d
c
c
0
1
3
11
Cp*Ir(H
[(Cp*Ir)
Cp*Ir(ppy)Cl
[Cp*IrCl
IrCl ·nH
IrCl
2 3
O) SO
4
2
2
(µ-OH)
3
]OH
2
2
d
d
c
[
Cp*IrCl
IrCl ·nH
IrCl
2
]
2
]
2 2
For the bipyridine complex 3, the initial rate of oxygen
evolution is first-order in iridium (vide infra). In the first minute
of catalysis with 3, a sigmoidal response is encountered, in
which a lag phase is followed by essentially constant oxygen
evolution. The oxygen evolution rate reaches a maximum value
within the first minute (Figure 5), in line with the behavior
3
2
O
3
2
O
2.1 ( 0.2
0.39 ( 0.05
K
2
6
K
2
6
a
Measured by Clark-electrode oxygen assay; triplicate data sets were
collected for each rate measurement. Errors (presented as (1σ) were
calculated from variance in the three measurements. Data were analyzed
for the first 30 s of oxygen evolution. These initial rates are also the
1
0
-
1
previously reported for catalysts 1 and 2. However, within
several minutes, the rate does begin to decrease. This behavior
is also seen in the case of catalyst 1 (see Supporting Informa-
tion). We cannot completely exclude the possibility that the
decrease in rate is the result of oxidation of the catalyst,
including both the chelate ligand or the Cp* methyl groups.
Since the Clark-electrode oxygen assay shows essentially no
instrumental lag time, any observed lag phase is real. The lag
time in Figure 5, between nominal injection of catalyst at zero
time and the onset of oxygen evolution, may be a result of slow
chloride dissociation from 3, required for substrate water
binding. Alternatively, since oxidation appears to be rate-limiting
in these studies (vide infra), the lag phase may be related to
slow oxidation of the metal complex to the required high-valent
intermediate. ESI mass spectra (see Supporting Information)
collected on solutions of catalyst 3 and cerium(IV) ammonium
nitrate show populations of the Cp*Ir(bipy) fragment with
nitrate, chloride, or hydroxide bound in the open site, suggesting
an analogous solution-phase equilibrium of these species.
Regardless, in the presence of a large excess of cerium(IV),
the robustness of catalyst 3 seems to be limited.
maximal rates. Conditions: 5.0((0.1) µmol L catalyst; 78 mM Ce(IV);
b
pH 0.89; vessel temperature maintained at 25 °C. Turnovers (t.o.) are
-
1
c
2
mol O (mol Ir) . Calculated from initial rate data. Compound is
isolated as a hydrate and is hygroscopic. Uncertainty in waters of
hydration causes additional uncertainty in turnover numbers. Catalyst
d
injected as a solution in acetonitrile; improved analytical methods have
led us to modify our previously reported rates for complex 1 and we
now prefer the values given in this table; the presence of the chloride
counterion was found not to accelerate the oxygen-evolution rate. The
previously reported Cp*Ir(ppy)OTf analogue of 1 gives an identical
-
1
-1
rate to 1: 47 ( 5 µmol L min
.
Figure 4. Synthesis of the CpIr(III) catalysts 7 and 8.
18
also used for O incorporation studies confirming incorporation
of oxygen atoms from water into the product oxygen (vide infra).
In the expectation that triphenylphosphine and carbon mon-
oxide would both be removed by irreversible oxidation,
(21) The salt of 8 with hexafluorophosphate counterion is hygroscopic,
suggesting a source of the difference in turnover frequency when
compared with 7 at the same nominal catalyst loading.
(
22) Ginzburg, S. I.; Yuzko, M. I. Russ. J. Inorg. Chem. 1965, 10, 444–
448.
1
6020 J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010

Half-Sandwich Iridium Complexes
A R T I C L E S
Figure 7. Dependence of oxygen-evolution rate on the concentration of
catalysts 1, 3, and 8.
Figure 5. Oxygen-evolution traces for complex 3 as measured by Clark-
electrode oxygen assay.
Figure 8. Dependence of oxygen-evolution rate on the concentration of
catalysts 1, 3, and 8.
suggesting that iodide is quickly lost from the metal center under
catalytic conditions.
Figure 6. Comparison of oxygen evolution traces for a catalyst concentra-
tion of 5 µM.
Although the catalysts 1, 3, and 8 have different rates of
oxygen evolution (Table 1), a plot of the rate dependence with
catalyst concentration shows that they closely resemble one
another at low catalyst loadings (see Figure 7). Studies on 1
The oxygen evolution traces differ among the various
catalysts (Figure 6). Within three minutes, complex 3 has
significantly dropped in rate, as noted above. The analogous
Cp complex 7 shows a different trace, with no apparent
deactivation up to 10 min after injection, although there is clearly
lag phase-like behavior since the trace is noticeably upwardly
convex. Thus, substitution of Cp* with Cp affects not only the
turnover rate but also the profile of oxygen evolution and lag
phase. This difference in rate profile is reflected in long
timecourse studies as well (vide infra). Complexes 10 and 11
show essentially no lag phase, but 11 has a higher sustained
rate of oxygen evolution. Finally, iridium trichloride hydrate
shows insignificant oxygen evolution compared to the organo-
metallic complexes. The variation in the nature of the lag phase
between different complexes suggests that each complex
behaves somewhat differently. For the Cp complex 7, mass
spectra show evidence of hydroxide- and nitrate-bound species
in solutions with cerium(IV) ammonium nitrate. Treatment of
this solution with excess ammonium hexafluorophosphate
followed by extraction with dichloromethane gave [CpIr(bipy)-
1
0
showed that oxygen evolution is first-order in iridium.
However, we were surprised to find that the rate of oxygen
evolution for the phenylpyridine (1) and bipyridine (3) systems
track each other at low catalyst loadings. Similarly, the Cp
complex 8 tracks with both 1 and 3 at low catalyst loadings. A
conventional log-log plot gives the order of reaction in catalyst.
For 1, 3, and 8, the slopes of these lines are: 0.98 ( 0.01, 0.91
( 0.02, and 0.94 ( 0.03 (see Supporting Information for the
double log plots). For each of these catalysts, the rate of oxygen
evolution is essentially first-order with respect to iridium,
suggesting that bimolecular steps do not contribute to catalysis
with 1, 3, or 8. However, at higher catalyst loadings (upward
from ca. 5 µM), catalyst 8 undergoes a deactivation process
and the rate of catalysis tends toward zero-order behavior (see
Figure 8 and Supporting Information).
The water-oxidation behavior of complex 11, [Cp*Ir(H
2 3
O) ]-
SO , is surprising. Under the conditions of our Clark-electrode
4
(
O-NO
2
)]PF
6
which we were able to crystallographically
oxygen assay, the turnover frequency clearly increases at higher
catalyst loadings. A log-log plot gives a slope of 1.31 ( 0.06,
suggesting that dimerization may contribute to catalysis (see
characterize (see Supporting Information). However, no iodide-
bound complexes are detected in the mass spectra of 7,
J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010 16021

A R T I C L E S
Blakemore et al.
Table 2. Catalyst Turnovers Measured with Headspace Oxygen
a
Assay
turnovers
in first
hour
turnovers
in 8 h
turnovers
per hour
compound
formula
1
3
7
Cp*Ir(ppy)Cl
Cp*Ir(bipy)Cl
155
93
214
128
349
320
738
281
44
40
92
35
2
[CpIr(bipy)I]NO
[Cp*Ir(H O) ]SO
4
3
1
1
2
3
a
Conditions: 5.0 µmol L-1 catalyst; 78 mM Ce(IV); pH 0.89; vessel
(mol
temperature maintained at 21 °C; turnovers are defined as mol O
2
-
1
Ir) . Turnovers per hour are the average over the 8-h run. Error is (5%
for the first hour, and (9% after 8 h, reflecting increased uncertainty at
long reaction times (triplicate measurements).
2
3
This gives an overall turnover frequency of 40 per hour. The
complexes 1 and 11 give similar initial turnover numbers and
overall frequencies. On the basis of these results and the
appearance of the oxygen-evolution traces (see Supporting
Information), the catalysts are deactivated within the first two
hours of catalysis. This is also observable in the initial rate data,
as most of the Cp*-bearing compounds show decreases in rate
within minutes of injection. We cannot exclude the possibility
that the decrease in rate is the result of oxidation of the catalyst,
such as of the chelate ligand or of the Cp* methyl groups. If
the chelate ligand alone were oxidized and lost, we would expect
to obtain activity like that of the tris-aqua species; this is not
observed (Table 2). If the Cp* were also oxidized and/or lost,
catalytically active Ir-O particulate species could be formed.
The first few minutes of reaction are kinetically well-behaved
and are ascribed to homogeneous catalysis, but the origin of
the longer term activity cannot be so confidently ascribed. With
all this in mind, it is intriguing to note that Cp complex 7 gives
a much higher overall turnover number but a lower initial rate.
This agrees with initial rate data collected by Clark electrode
and points to increased oxidative stability of the catalyst. Indeed,
compared to 1, 3, or 11, Cp-catalyst 7 shows more than double
the turnover frequency over 8 h.
Figure 9. Dependence of oxygen-evolution rate on concentrations of 3
and 11.
Figure 9 and Supporting Information). Additionally, the log-log
plot clearly shows nonlinear behavior, suggesting equilibrium
speciation processes affect the rate of oxygen evolution. For
the distinct regime at higher catalyst loadings, the order in
iridium is 1.64 ( 0.03. Due to the saturation of the aqueous
solution with O
upper limit for the Clark assay, which only records dissolved
, is around 30 µM for the tris-aqua complex. Even with this
limitation, at this catalyst loading, a turnover frequency of 20
2
and the consequent formation of bubbles, the
O
2
-
1
min is seen.
With the complex 11 in hand, we next moved to characterize
the behavior of the related dimer 12. On a per-iridium basis,
the performance of complexes 11 and 12 is similar at 5 µM
catalyst. However, at higher loadings of the dimer 12, the rate
of oxygen evolution is greater than first-order in iridium, with
a log-log plot slope of 1.53 ( 0.05 (see Supporting Informa-
tion), again suggesting that the dimeric structure of 12 contrib-
utes to oxygen evolution. The non-first-order behavior at low
loading of catalyst suggests a monomer-dimer equilibrium that
affects the rate of catalysis. Turnover frequencies for the dimer
Water was confirmed as the source of the oxygen gas
produced by using measurements of 18O-incorporation. Stable
isotope ratio mass spectrometry (SIR-MS, see Experimental
Section) is exquisitely sensitive to changes in isotopic abundance
in gases, and examines parts per thousand (ppt) changes in
isotope distribution of near natural-abundance systems. Con-
sequently, only a small amount of isotope label is required. In
-1
1
2 are up to 25 turnovers min on a per-iridium basis, or 50
-1
min on a dimer basis.
In our original report on water-oxidation catalysis with Cp*Ir
complexes, we noted that [Cp*IrCl ] was competent for oxygen
2 2
evolution. We have confirmed this result here, and find high
turnover numbers with this system. However, this dimeric
material shows a similar concentration-dependent rate plot when
compared to 12 (see Supporting Information). As with 12, this
suggests that a dimeric structure contributes to catalysis.
Chloride oxidation to hypchlorite or chlorine gas is a
potentially competing reaction in these water-oxidizing systems.
Comparing the rates of the compounds Cp*Ir(ppy)Cl (1) and
Cp*Ir(ppy)OTf, we see no difference either in the observed
initial rate of water oxidation or in the appearance of the
subsequent oxygen-evolution data. Similarly, with dimers
2
the first experiment, purified natural-abundance H O was used
to prepare solutions of cerium(IV) oxidant and catalyst 7. Upon
injection of catalyst, oxygen evolution commenced and reached
a steady rate, giving a total 18O/ O ratio of 0.200%. In the
16
18
second experiment, a small amount of 95% H
2
O in natural-
abundance H O was used to prepare solutions of cerium(IV)
2
24
18
16
and 7, with the result that the O/ O ratio was then 0.217%.
18
16
The expected change in the O/ O ratio was 0.023% based on
18
the added O label. This is in good agreement with the
measured 0.017% change, which corresponds to a change of
8
.5 ppt, and confirms that water is the source of the oxygen
[
2 2
Cp*IrCl ] and 12, there is no significant change in initial rate
gas produced.
or catalytic behavior. This suggests that the presence of chloride
does not significantly impact the water oxidation reaction in
these systems. This also applies to the Cp-containing systems,
since upon chloride addition to solutions of chloride-free catalyst
Among homogeneous water-oxidizing systems, ruthenium
6,13
catalysts are among the best characterized.
In this context,
analogues of 3 were synthesized with ruthenium, as well as
rhodium. No oxygen evolution activity was detected using
7
, no significant change in oxygen evolution is observed.
Solutions containing catalysts and excess oxidant continue
to evolve oxygen over hours (see Table 2). For example, a
solution with catalyst 3 gives a total of 320 turnovers in 8 h.
(
(
23) There was sufficient cerium(IV) for more than 3500 turnovers.
18
24) A 35 µ` L amount of 95% H
abundance H O.
2
O was added to 12.5 mL of natural-
2
1
6022 J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010

Half-Sandwich Iridium Complexes
A R T I C L E S
cerium(IV) as primary oxidant for Cp*RuCl(2,2′-dipyridyl),
RuCl(2-phenylpyridine-κC,N)(p-cymene), and [Cp*RhCl(2,2′-
dipyridyl)]Cl. We anticipate that the lack of water-oxidizing
behavior is due to kinetically slow oxidation of the metal
complexes, which limits access to the high-valent intermediates
required for oxygen evolution.
A surprising feature of the experimental work is the near
independence of the results on the nature of the chelating ligand
employed. We, therefore, considered the possibility that the
chelating ligand was removed under the oxidizing conditions
2
+
and that the true catalyst is always solvated [Cp*Ir] . Kinetic
studies show that while the reaction is first-order in catalyst for
compounds containing chelate ligands, 1 and 3, the order of
the reaction is greater than one for the solvated [Cp*Ir]2
fragment, 11, which suggests loss of chelating ligand is not
occurring. This is consistent with ESI+ MS data, which show
retention of both the Cp*/Cp and chelate ligands for catalysts
+
3
and 8 upon treatment with excess cerium(IV). Finally, the
oxygen-evolution rates of compounds 3, 4, 5, and 6 vary (above
the level of experimental error) with the identity of the chelate
ligand, suggesting that the choice of ligand, slightly but
significantly, affects the rate of catalysis.
Comparison of our work with the prior experimental studies
4
-6
cited above
indicates that the present catalysts may have
certain advantages. These catalysts provide good activity and
are suitable for electrochemical or spectroelectrochemical stud-
ies. They have an easily modified ligand environment, a factor
that may assist in their attachment to other components of more
elaborate architectures, such as a solar water-splitting cell.
However, the kinetic behavior of the complexes, at least with
the harsh conditions required for use of cerium(IV) as primary
oxidant, suggest that long-term robustness may be limited with
these catalysts. Generally, though, as a third row metal, the
metal-ligand bonds are also expected to be much more robust,
kinetically and thermodynamically, than in the case of base
Figure 10. pH-dependent cyclic voltammograms of complex 3 in aqueous
solution (scan rate: 20 mV/s; electrode surface area: 0.09 cm ).
2
5
metal catalysts such as our own terpyridine-Mn system or than
equimolar solutions of cerium(IV) and cerium(III). Since we
prepare solutions of cerium(IV) which can be conservatively
expected to initially contain less than 1% cerium(III), the
oxidizing power of the cerium(IV) solutions used in the chemical
oxidant-driven catalysis is as high as 1.7 V. Thus, at low pH
values, the potential required for oxygen evolution is likely
above 1.5 V.
6
the second row systems based on Ru. Building on Bernhard’s
9
work, they also open up the way to the use of organometallic
precursors for water-oxidation catalysis. On the other hand,
iridium is a precious metal, so high value and research
applications seem the most appropriate.
Aqueous Electrochemistry. Cerium(IV) studies are limited
to the strongly acidic pH values needed to stabilize the oxidant.
However, electrochemistry allows us to examine the oxidation
behavior of these complexes over a much wider pH range.
Although electrochemistry with complex 1 is possible in
acetonitrile (see Supporting Information), the complex is not
soluble in aqueous solutions in the iridium(III) oxidation state.
The cationic iridium(III) complexes 3-6, 7, and 11 are soluble
in water, allowing us to now examine electrochemically driven
oxidations in aqueous solution.
In the range of pH 7-11.5, irreversible oxidation waves
appear between 1.15 and 1.35 V vs NHE. At higher pH values
(
ca. 11), the lower potential wave at ca. 1.15 V grows in
intensity, eventually becoming the dominant, catalytic oxidation
at pH ) 12. The current at higher pH values exceeds the current
at the lower pH values, consistent with catalysis. From the
appearance of the voltammograms, the catalysis is not occurring
26
under purely kinetic conditions, as described by Sav e´ ant. This
may be a consequence of a fast rate of reaction compared to
diffusion at the chosen scan rate of 20 mV/s. Regardless, the
appearance of multiple oxidation waves suggests that different
reactions take place as the pH is changed. This may be related
to a pH-dependent speciation process, resulting in bound water
deprotonation at higher pH values, or pH-dependent reactions
of oxidized species. Such a speciation process would be expected
to affect the catalytic behavior of the metal complex, as is seen
in the voltammograms in Figure 10.
Cyclic voltammograms collected for complex 3 at various
pH values are shown in Figure 10. A prominent catalytic wave
at ca. pH ) 12, assigned to catalytic water oxidation, is
associated with the complex. Below pH ) 7, there is no
oxidation up to 1.5 V vs NHE, suggesting that catalysis with
cerium(IV) occurs at oxidizing potentials above 1.5 V. This is
consistent with the literature redox potential for cerium(IV), ca.
2
5
1
.61 V. Additionally, the Nernst equation dictates that the
oxidizing potential of a concentrated solution of cerium(IV) will
be greater than this literature value, which is calculated for
Alternatively, the 2 equiv of chloride present in solution could
be oxidized to chlorine or hypochlorite. However, oxidative
(
25) Schilt, A. A. Anal. Chem. 1963, 35, 1599–1602.
(26) Sav e´ ant, J.-M. Chem. ReV. 2008, 108, 2348–2378.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010 16023

A R T I C L E S
Blakemore et al.
Table 3. Summary of KIE Data
KIE at 8 mM
cerium(IV)
KIE at 243 mM
cerium(IV)
compound
3
IrO
0.65
1.60
1.15
-
2
nanoparticles
formation of the O-O bond of the final O
2
product (vide infra).
For complex 11, the rate of oxygen evolution is zero-order in
oxidant (see Supporting Information).
In order to distinguish between homogeneous and heteroge-
neous catalysis, we undertook a study of the H/D kinetic isotope
H D
effect by determining the k /k ratio for both our organometallic
complex 3 and IrO nanoparticles (Table 3) using Clark-
2
electrode oxygen-evolution data. Since the rate-determining step
appears to change based on the oxidant concentration for 3, the
kinetic isotope effect (KIE) was measured in the two distinct
regimes. For complex 3, at 8 mM cerium(IV), the measured
KIE is 0.65. This is a large and relatively unusual inverse isotope
effect, which corresponds to a near doubling of the rate of water
2 2
oxidation when D O is substituted for H O.
This inverse KIE is consistent with rate-determining oxidation
from Ir(IV) to Ir(V) occurring by way of a transition state
containing a cerium(IV) bound to the iridium(IV) center via a
bridging hydroxyl group. Upon formation of this transition state,
2
the hybridization at the oxygen center will become sp , causing
an increase in the O-H bond strength. The O-D bond will
strengthen more than the O-H bond, resulting in the measured
inverse KIE. Although there is no precedent for an inverse KIE
in water oxidation, there is precedent for inverse isotope effects
during the formation or cleavage of C-H bonds at metal
2
8
centers. Often these inverse isotope effects involve pre-
equilibrium conditions, which give rise to the observed variation
in rate. In the case of the inverse isotope effect measured for
complex 3, pre-equilibrium coordination of Ce(IV) to the
hydroxyl group of Ir(IV) is consistent with the near first-order
dependence on oxidant found in our experiments (vide supra).
In contrast, at 8 mM cerium(IV), iridium dioxide nanopar-
ticles give a very different KIE of 1.6. This normal isotope effect
is consistent with rate-determining O-O bond formation
occurring by nucleophilic attack of water to an iridium oxo
species. Furthermore, oxygen evolution rapidly decays in the
case of iridium dioxide nanoparticles, already attaining a near-
zero rate within 15 min from the time of injection into the
Figure 11. Oxygen-evolution rate dependence on the concentration of
cerium(IV).
background scans run with equivalent concentrations of chloride
without the iridium complex do not show significant currents.
Furthermore, upon multiple oxidizing scans in cyclic voltam-
metry, the oxidation waves seen for complex 3 do not decrease
in intensity. This is consistent with a catalytic process, rather
than stoichiometric oxidation of chloride. Finally, since chloride
oxidation is a pH-independent process, whereas water oxidation
is pH dependent, the increasingly intense oxidations observed
for 3 at higher pH values are consistent with water oxidation.
pH dependent electrochemistry for complex 4 is very similar
to that of 3 and is provided in the Supporting Information. The
electrochemical behavior of 3 and 4 is reversible upon acidifica-
tion. Complexes 11 and 12 show complex and unique electro-
cerium(IV) solution. IrO
poor performance for oxygen evolution with cerium(IV).
2
has been shown previously to have
1
1a
With complex 3 at 243 mM cerium(IV), the measured KIE
is 1.2. This value is also consistent with rate-determining O-O
bond formation. The value for this KIE is similar to other
measured kinetic isotope effects observed for O-O bond
formation and is best considered a secondary isotope effect.
4+
For example, cis,cis-[(bipy)
2
Ru(H
2
O)]
2
O
and [(terpy)Mn-
2
7
chemical behavior, which is discussed elsewhere.
3+
(
H
2
O)O]
2
have been found to have KIE values for oxygen
5
b,29
Rate Dependence on the Concentration of Cerium(IV) and
Kinetic Isotope Effects. A plot of the cerium(IV) concentration-
dependence of oxygen evolution for complex 3 (Figure 11)
shows two regimes. At lower cerium(IV) concentration, the
kinetics are less than first-order in oxidant, with a log-log plot
slope of 0.73 ( 0.01 (see Supporting Information). At higher
loadings, the kinetics become zero-order in cerium(IV). This
change in kinetic behavior suggests that the rate-determining
step (rds) has changed. With a low oxidant loading, the rds may
be the oxidation of the catalyst to a high-valent metal-oxo active
species, whereas at higher oxidant loading, the rds may be the
evolution of 1.7 and 1.64, respectively.
2
For IrO at 243 mM
cerium(IV), no catalysis was observed. Upon injection of the
nanoparticles into the oxidant, a short and rapid burst of oxygen
release was detected, consistent with stoichiometric release of
oxygen from the nanoparticles.
(
(
27) Blakemore, J. D.; Schley, N. D.; Olack, G. W.; Incarvito, C. D.;
Brudvig, G. W.; Crabtree, R. H. Chem. Sci. 2010, DOI: 10.1039/
c0sc00418a.
28) (a) Janak, K. E. In ComprehensiVe Organometallic Chemistry III;
Crabtree, R. H.; Mingos, M. P., Eds.; Elsevier: Oxford, 2007. (b)
Parkin, G. Acc. Chem. Res. 2009, 42, 315–325.
1
6024 J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010

Half-Sandwich Iridium Complexes
A R T I C L E S
From the above information, it is clear that the homogeneous
catalyst 3 has a distinct behavior from the previously studied
nanoparticulate IrO material. The differences in KIE data show
2
that oxidation to the key intermediate required for O-O bond
formation occurs through different intermediates in complex 3
2
versus IrO . The rapid decay in activity at 8 mM cerium(IV)
and the lack of catalytic activity at higher oxidant concentrations
demonstrate that the catalyst from organometallic complex 3 is
2
distinct from the nanoparticulate IrO material, when water
oxidation is driven with cerium(IV). Furthermore, examination
by transmission electron microscopy (TEM) of solutions
containing cerium(IV) and catalyst 3 do not show any evidence
2
for formation of nanoparticulate (IrO ) materials.
Mallouk et al. have previously shown that iridium dioxide
nanoparticles have a KIE value of 1.0 for oxygen evolution,
3
+
30
3
when oxidized with [Ru(bipy) ] . The kinetic data for the
nanoparticles suggested that the rate-determining step was
oxidation from Ir(III) to Ir(IV). Since these results are different
than those obtained here with cerium(IV), it appears that the
nature of the primary oxidant used to drive water oxidation is
closely tied to the kinetics of oxygen evolution. Additionally,
the identity of the catalytically competent oxidation state for
O-O bond formation in various iridium oxide systems has been
debated. Harriman et al. have implicated an iridium(V) species
as the catalytically competent intermediate required for oxygen
IrdO orbitals of CpIr(O)(ppy)+ seen down
π π
-p
Figure 12. Schematic d
the IrdO axis.
3
1
evolution in iridium oxide colloids, while Murray et al. have
implicated an iridium(VI) species as the catalytically competent
intermediate in electrochemically driven systems with small
1
1b,c
nanoparticles.
DFT Calculations on the Formation of the O-O Bond. The
formation of the O-O bond is one of the key steps in the
1
3,15
oxidation of water to dioxygen.
the DFT(B3LYP) level (see Computational Details) for catalyst
(Figure 1), which is modeled by replacing the Cp* ligand by
Cp. Following the reaction mechanism represented in Figure
, our model catalyst is oxidized to an Ir(V) oxo complex,
This step is studied at
1
2
+
CpIrO(ppy) , which promotes O-O bond formation by reacting
with water. In view of the high catalytic activities observed in
our experiments, this step should involve a low energy pathway.
The challenging O-O bond formation step requires the
coupling of two O atoms having opposite reactivity, since the
Figure 13. O-O bond formation in pathway A. Energies are given in
-
1
kcal mol
.
O of H
oxo has to act as electrophile. Because of the poor nucleophilic
character of H O, a highly electrophilic metal-oxo active species
is needed for efficient catalysis. The ability of CpIrO(ppy) to
promote the formation of an O-O bond with H O stems from
its electronic structure. As noted in our previous work, the
set of orbitals associated with the IrdO π bond (Figure 12) is
filled with six electrons. Four of them are in the two bonding
2
O has to act as nucleophile, whereas the O of the metal-
3
2
of H
2
O to the oxo ligand. A key point of this pathway is that
O to some proton acceptor,
2
+
one proton has to transfer from H
2
in order to benefit from the enhanced nucleophilicity of the
incipient hydroxide ion. The nature of the proton acceptor is
explored in our calculations. Three different mechanisms,
labeled as A, B, and C (Figures 13-15), are characterized, each
one involving a different proton acceptor. The global stabilizing
effect of water on all extrema is evaluated by single-point
calculations using the CPCM method (see Computational
Details). The counterion is not explicitly introduced in the
modeling since the cations and anions most likely form solvent
2
1
0
d
π
+p
antibonding π*
π* orbital, which is preferentially located on Ir rather than on
π
orbitals, π
N
C
and π , whereas the other two are in the
N
orbital. The presence of the doubly occupied
N
O, increases the electrophilicity of O and weakens the IrdO
bond. The electrophilicity of the oxo ligand is also promoted
6
separated ion pairs in water. Since the final product is a d Ir(III)
low spin complex, the calculations are carried out for the singlet
electronic state.
by the presence of the empty π*
C
orbital, which lies low in
energy at the LUMO level.
3
3
+
The proposed active species, CpIr(O)(ppy) , is mononuclear
and contains only one oxo ligand. Therefore, a reasonable
mechanism for O-O bond formation is the intermolecular attack
(
31) Nahor, G. S.; Hapiot, P.; Neta, P.; Harriman, A. J. Phys. Chem. 1991,
5, 616–621.
32) Alternatively, O-O bond formation may involve OH instead of H
9
-
(
2
O,
2
because OH is much more nucleophilic than H O. Nevertheless, at
-
(
(
29) Yamada, H.; Siems, W. F.; Koike, T.; Hurst, J. K. J. Am. Chem. Soc.
004, 126, 9786–9795.
30) Morris, N. D.; Suzuki, M.; Mallouk, T. E. J. Phys. Chem. A 2004,
08, 9115–9119.
2
the strongly acidic pH at which the reaction is performed, the hydroxide
concentration is more than 10 orders of magnitude lower than that of
H O.
2
1
J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010 16025

A R T I C L E S
Blakemore et al.
involves a significant energy barrier when the proton of H
transferred to the N of the ppy ligand.
In pathway B (Figure 14), the oxo ligand of CpIrO(ppy)
2
O is
+
not only generates the new O-O bond, but it also accepts the
proton released by H O, i.e., the oxo inserts into the O-H bond.
2
In the transition state, B-TS, the formation of the O-O bond,
d(O· · ·O) ) 1.99 Å, is concerted with the proton transfer from
H O to oxo, d((H)O· · ·H) ) 1.53 Å and d(O· · ·H) ) 1.00 Å.
2
The relaxation of B-TS toward reactant and product yields
intermediates B-IN1, which is analogous to A-IN1, and B-IN2,
respectively. In B-IN2, the dihydroperoxide product is formed,
1
d(HO-OH) ) 1.46 Å, and κ -bound to Ir, as illustrated by the
two different Ir-OH distances of 2.24 Å and 3.11 Å. The
energies associated with the transition state and product of
-
1
-1
pathway B, 21.6 kcal mol and 1.3 kcal mol , respectively,
are lower than those associated with pathway A. This suggests
that O-O bond formation is more favorable when the proton
2
released by H O is transferred to the oxo ligand, rather than to
the N of the ppy ligand.
Pathway B may be further accelerated by including an
additional water molecule to assist the proton transfer. In the
real system, the catalyst is solvated by H O, which may,
2
Figure 14. O-O bond formation in pathway B. Energies are given in kcal
-
1
mol
.
therefore, act both as reactant, by promoting the formation of
the O-O bond, and as cocatalyst, by assisting the water-to-
oxo proton transfer. This possibility is explored by introducing
2
one additional molecule of H O in the more favorable pathway,
B, which is named pathway C (Figure 15). The stabilizing
influence of water was also explored for pathway A but this
pathway D is found to be energetically less favorable than
pathway C and is not presented further (see Supporting
3
4
Information for further details).
In the prereaction complex, C-IN1, the catalyst interacts with
a water dimer through a H-bond, d(H· · ·O(Ir)) ) 2.06 Å. In
the cyclic transition state, C-TS, H
whereas H O(2) assists the proton transfer. C-TS involves the
concerted formation and cleavage of three bonds: (1) O-O bond
formation by H O(1), d(O· · ·O) ) 1.50 Å, (2) proton transfer
from H O(1) to H O(2), d(O· · ·H) ) 1.40 Å and d(H· · ·O) )
.08 Å, and (3) proton transfer from H O(2) to the oxo ligand,
d(O· · ·H) ) 1.00 Å and d(H· · ·O) ) 1.86 Å. This is a proton
transfer chain in which one proton is transferred through H O(2)
2
O(1) attacks the oxo ligand,
2
2
2
2
1
2
2
and involves a planar five-membered ring (Ird)O· · ·O(1)· · ·H· · ·
O(2)· · ·H· · ·O(dIr), dihedral angle Ω(O· · ·O(1)· · ·H· · ·O(2))
) 2.2°. The relaxation of C-TS toward product yields a
dihydroperoxo complex, C-IN2, which is analogous to B-IN2.
Figure 15. O-O bond formation in pathway C. Energies are given in kcal
-
1
mol
.
-
1
Pathway C is exothermic by -11.1 kcal mol and involves a
In pathway A (Figure 13), the N of the ppy ligand acts as
proton acceptor. The reaction starts with the formation of an
intermediate, A-IN1, in which the oxo ligand is H-bonded to a
-
1
low energy transition state, which lies 8.0 kcal mol above
reactants. These results show that O-O bond formation by
+
CpIrO(ppy) involves a low energy pathway when the two
molecule of H O, d(H· · ·O) ) 1.89 Å. This complex lies 10.6
2
kcal mol-1 below the starting reactants. In the following step,
following requirements are fulfilled: (1) the oxo ligand, rather
than to the N of the ppy ligand, acts as proton acceptor and (2)
the proton transfer is assisted by an additional molecule of water.
The kinetic isotope effect (KIE) has been also calculated using
A-IN1 is transformed into the hydroperoxo complex A-IN2, in
which the O-O bond has been formed, d(O-OH) ) 1.48 Å,
and the N of the ppy ligand has been protonated, d(N-H) )
all frequencies. The k
H
/k
D
values computed for pathways A, B,
1
.05 Å. The protonation of the N of ppy causes the cleavage of
the Ir-N bond; 2.07 Å in A-IN1, 3.43 Å in A-IN2. In the
transition state, A-TS, the O of H O approaches the oxo ligand,
+
(
33) The triplet state of CpIrO(ppy) , which is generated by promoting
2
one electron to the π* orbital, has radical oxyl character, with a local
C
spin density on oxygen of 1.14. Some theoretical studies have
d(O· · ·O) ) 1.85 Å, and transfers one proton to the N of the
ppy ligand, d(O· · ·H) ) 1.21 Å and d(H· · ·N) ) 1.27 Å. The
three atoms involved in the proton transfer are very close to
the expected linear arrangement, R(O· · ·H· · ·N) ) 174.0°.
suggested that oxyl character promotes water oxidation in photosystem+
II (ref 16a). Nevertheless, we find that water oxidation by CpIrO(ppy)
becomes much more endothermic in the triplet oxyl state. For example,
the ∆E associated with A-IN1 f A-IN2 increases from 26.7 kcal
-
1
-1
-1
mol , in the singlet, to 52.4 kcal mol , in the triplet. This is due to
the reduction of the metal center from Ir(V) to Ir(III), which, in the
octahedral environment of the catalyst, has strong preference for a
low spin diamagnetic configuration.
Pathway A is endothermic by 16.1 kcal mol and involves a
transition state lying 25.4 kcal mol- above reactants. These
results show that O-O bond formation is endothermic and
1
1
6026 J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010

Half-Sandwich Iridium Complexes
A R T I C L E S
3
5
and C are 4.1, 2.2, and 1.6, respectively. The decrease of k
H
/
reaction is significantly lowered when the proton is transferred
to the oxo ligand with the assistance of one additional molecule
of water.
k
D
from A to B and from B to C, shows that the amount of
O-H cleavage in the transition state relative to reactant is
smaller for the proton transfer to the oxo ligand (B) than to the
N of the ppy ligand (A) and is further reduced when the proton
O (C). Interestingly, the A > B > C
trend in k /k parallels that of the transition state energies, which
suggests that O-H cleavage has a critical influence on the
energy barrier. The KIE calculated for pathway C, 1.6, which
is normal and secondary, is close to the experimental value
obtained for 3 at high cerium(IV) concentration, 1.15 (Table
Experimental Section
General. All solvents were of commercial grade and dried over
activated alumina prior to use. Chloride-free tetra-n-butylammonium
nitrate was prepared by shaking an aqueous solution of sodium
nitrate and tetrabutylammonium hydrogen sulfate in a separatory
transfer is assisted by H
2
H
D
37
funnel and extracting the product with dichloromethane. All other
reagents were obtained from major commercial suppliers and used
3
8
39
withoutfurtherpurification.[Cp*IrCl
2
]
2
, CpIr(C
2
H
4
)
2
, [Cp*RhCl(2,2′-
40
41
dipyridyl)]Cl, Cp*RuCl(2,2′-dipyridyl), RuCl(2-phenylpyridine-
3
), when O-O bond formation would be the rate-determining
4
2
43
44
κC,N)(p-cymene), CpIr(PPh
3
)(CO), [CpIr(PPh
3
)(CO)I]I, [(Cp*Ir)
and colloidal IrO nanoparticles
ca. 20 nm diameter) were prepared according to literature
procedures. Complexes of the type [Cp*Ir(N-N)Cl]Cl, with 2,2′-
2
(µ-
step. These results support pathway C as a plausible mechanism
for the formation of the O-O bond. In the experimental
conditions, the proton transfer may be further assisted by more
45
46
OH)
3
]OH, [Cp*Ir(H
2
O)
3
]SO
4
,
2
47
(
48
35
49,50
molecules of H
2
O, which would reduce even further both the
dipyridyl, 1,10-phenanthroline, and 2,2′-bipyrimidine
ligands
were prepared by published procedures. NMR spectra were recorded
at room temperature on 400 or 500 MHz Bruker spectrometers and
referenced to the residual solvent peak (δ in ppm and J in Hz).
Elemental analyses were performed by Atlantic Microlab, Inc.
calculated energy barrier and KIE.
The calculations show that the energy barrier for the formation
of the O-O bond is significantly influenced by the nature of
the proton acceptor. The intramolecular proton transfer to the
nitrogen of the coordinated ppy ligand is less favorable than to
an outer-sphere water molecule. In contrast, a theoretical study
by Yang and Hall on Ru-catalyzed water oxidation, suggests
5
-
[(η C Me )Ir(N,N,N′,N′-Tetramethylethylenediamine)Cl]Cl.
5
5
[Cp*IrCl ] (0.101 g, 0.127 mmol) was dissolved in 7 mL of
2
2
dichloromethane under nitrogen. To this solution was added neat
N,N,N′,N′-tetramethylethylenediamine (0.05 mL, 0.3 mmol). Within
minutes, the color of the solution had changed from deep orange
to yellow. After stirring 2.5 h at room temperature, the reaction
solution was added to 40 mL of diethyl ether and the hygroscopic
yellow precipitate collected by vacuum filtration in air. Yield 0.089
1
3c
that the O-H bond is cleaved intramolecularly. However,
the participation of outer-sphere water has been invoked in
1
3b
related chemistry and is involved in water oxidation by Ru
1
5a
1
polyoxometallates.
Participation of the solvent in proton
g (68%). H NMR (500 MHz, MeOD-d ) δ 3.29 (s, 6H), 3.10-3.01
4
13
transfer reactions involving transition metal complexes is well
(m, 2H), 2.96 (s, 6H), 2.83-2.72 (m, 2H), 1.57 (s, 15H). C NMR
126 MHz, MeOD-d
Calcd for Ir Cl
3
6
(
4
) δ 89.47, 65.36, 58.55, 54.96, 9.51. Anal.
·2H O. C, 34.90; H, 6.41; N, 5.09; O, 5.81.
documented and also seems to be of great importance in the
present case.
1
C
16
H
31
N
2
2
2
Found: C, 34.67; H, 6.34; N, 4.93; O, 6.03. Crystals suitable for
X-ray diffraction study were obtained by diffusion of diethyl ether
into a saturated dichloromethane solution.
In summary, our calculations show that the metal oxo active
+
species, modeled by CpIrO(ppy) , is capable of forming an
5
39,51
O-O bond by intermolecular attack of H
This reaction is accompanied by a proton transfer from H
the oxo ligand, which is assisted by an additional molecule of
O.
2
O to the oxo ligand.
[(η -C
5
H
5
)IrI
2
]
x
.
A 50 mL flame-dried Schlenk flask was
4 2
H ) (0.215 g, 0.686 mmol) in a nitrogen-
charged with CpIr(C
2
2
O to
(
(
36) (a) Filippov, O. A.; Filin, A. M.; Tsupreva, V. N.; Belkova, N. V.;
Lled o´ s, A.; Ujaque, G.; Epstein, L. M.; Shubina, E. S. Inorg. Chem.
006, 45, 3086–3096. (b) Filippov, O. A.; Tsupreva, V. N.; Golubin-
skaya, L. M.; Krylova, A. I.; Bregadze, V. I.; Lled o´ s, A.; Epstein,
L. M.; Shubina, E. S. Inorg. Chem. 2009, 48, 3667–3678.
37) Lund, H.; Hammerich, O. Organic electrochemistry, 4th ed.; M.
Dekker: New York, 2001.
H
2
2
Conclusions
Iridium half-sandwich complexes Cp*Ir(N-C)X, [Cp*Ir-
N-N)X]X, and [CpIr(N-N)X]X are highly active catalyst
(
(38) Ball, R. G.; Graham, W. A. G.; Heinekey, D. M.; Hoyano, J. K.;
McMaster, A. D.; Mattson, B. M.; Michel, S. T. Inorg. Chem. 1990,
precursors for the homogeneous oxidation of water to dioxygen.
Kinetic studies with cerium(IV) ammonium nitrate as primary
oxidant show that oxygen evolution is rapid and continues over
2
9, 2023–2025.
(
(
(
39) Heinekey, D. M.; Millar, J. M.; Koetzle, T. F.; Payne, N. G.; Zilm,
K. W. J. Am. Chem. Soc. 1990, 112, 909–919.
40) Kolle, U.; Grutzel, M. Angew. Chem., Int. Ed. Engl. 1987, 26, 567–
many hours. In addition, [Cp*Ir(H
2 3 4 2
O) ]SO and [(Cp*Ir) (µ-
5
70.
OH) ]OH show even higher turnover frequencies. Aqueous
3
41) (a) Fagan, P. J.; Ward, M. D.; Calabrese, J. C. J. Am. Chem. Soc.
1989, 111, 1698–1719. (b) Koelle, U.; Kossakowski, J. J. Organomet.
Chem. 1989, 362, 383–398.
electrochemical studies on the water-soluble cationic complexes
show a catalytic oxidation wave at a potential of 1.2-1.4 V at
pH > 7.0 but, at lower pH, show no oxidation waves below 1.5
(
42) (a) Djukic, J.-P.; Berger, A.; Duquenne, M.; Pfeffer, M.; de Cian, A.;
Kyritsakas-Gruber, N. Organometallics 2004, 23, 5757–5767. (b)
Bennett, M. A.; Smith, A. K. J. Chem. Soc., Dalton Trans. 1974, 2,
18
2
V vs NHE. H O isotope incorporation studies show that water
2
33–241.
is the source of oxygen atoms during cerium(IV)-driven
catalysis. DFT calculations show that in the key O-O bond
formation step, water attacks the oxo ligand of the active species
by losing one proton. The energy profile associated with this
(
43) Blais, M. S.; Rausch, M. D. Organometallics 1994, 13, 3557–3563.
(44) Oliver, A. J.; Graham, W. A. G. Inorg. Chem. 1970, 9, 2653–2657.
(
(
(
45) Nutton, A.; Bailey, P. M.; Maitlis, P. M. J. Chem. Soc., Dalton Trans.
981, 1997–2002.
46) Ogo, S.; Makihara, N.; Watanabe, Y. Organometallics 1999, 18, 5470–
1
5
474.
47) Hara, M.; Waraksa, C. C.; Lean, J. T.; Lewis, B. A.; Mallouk, T. E.
(
34) We also find pathway D, in which a molecule of H
2
O assists the proton
J. Phys. Chem. A 2000, 104, 5275–5280.
transfer to the ppy ligand of pathway A. Nevertheless, the transition
(48) Ziessel, R. J. Chem. Soc., Chem. Commun. 1988, 16–17.
(49) Govindaswamy, P.; Canivet, J.; Therrien, B.; S u¨ ss-Fink, G.; Stepnicka,
P.; Ludvik, J. J. Organomet. Chem. 2007, 692, 3664–3675.
(50) Gabrielsson, A.; van Leeuwen, P.; Kaim, W. Chem. Commun. 2006,
4926–4927.
state associated with pathway D, D-TS (see Supporting Information),
-
1
lies 14.4 kcal mol above that of pathway C, C-TS.
‡
(
35) The kinetic isotope effect was computed as k
H
/k
D
) exp[(∆G -
D
‡
∆
H
G )/RT], with T ) 298.15 K.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010 16027

A R T I C L E S
Blakemore et al.
5
4
2
filled glovebox and 30 mL of dry, degassed dichloromethane was
added, followed by iodine (0.185 g, 0.729 mmol). The reaction
was stirred at room temperature for 1 h, during which time a fine
purple-brown precipitate formed in solution. The precipitate was
collected by filtration in air and washed with two 5 mL portions of
dichloromethane. Yield 0.272 g (78%). As previously reported, the
polymeric material is insoluble in water and most organic solvents.
least-squares refinement on F were applied until convergence
of unweighted and weighted factors of R ) ∑|F | - |F |/∑|F |; R
) ]} . Crystal data and experimental
o
c
o
w
2
2 2
2 2 1/2
) {∑[w(F
o
- F
c
) ]/∑[w(F
o
details are included in the Supporting Information.
Kinetic Studies. Measurements of initial oxygen evolution rate
were made with a YSI Clark-type oxygen electrode. Prior to
beginning each set of experiments, the gas-permeable membrane
was replaced to ensure a high-quality response. The electrode,
secured in a Teflon tube, was inserted into a tight-fitting water-
jacketed glass vessel. The system was kept at a constant temperature
of 25 °C. In a typical experiment, a freshly prepared Ce(IV) solution
in Milli-Q water (5 mL, 78 mM; pH 0.87) was allowed to equilibrate
with stirring for 7 min. Data collection over 5-7 min showed
equilibration of the system, and when a steady baseline was
achieved, 10 µL of catalyst solution was injected. Oxygen evolution
commenced immediately and typically exceeded the maximum
value for the system within 10 min of injection.
For studies on oxygen evolution at longer time scales, a
headspace fluorescence-probe oxygen assay was used. The apparatus
(Ocean Optics, Dunedin, FL) consists of a probe (FOXY-R) and a
custom glass cell with a threaded fitting for the probe and an airtight
septum for purging. A fluorescent dye is embedded in a sol-gel
matrix in the probe tip. Upon exposure to oxygen, the quenching
of the dye’s fluorescence is recorded by a fluorimeter connected to
the probe via a fiber-optic cable. A Stern-Volmer plot generated
from standard oxygen concentrations in the cell is then used to
convert the raw fluorescence lifetime data into headspace oxygen
concentration. In a typical experiment, a freshly prepared Ce(IV)
solution in Milli-Q water (5 mL, 78 mM; pH 0.87) was allowed to
equilibrate with stirring after being purged with nitrogen gas for
20 min. Data collection over 20 min showed equilibration of the
system, and when a steady baseline had been achieved, 10 µL of
catalyst solution was injected. Depending on the experimental
conditions, oxygen evolution was apparent within 5 min.
6
In DMSO-d , the solid dissolves to give an orange solution with a
1
singlet in the H NMR at 6.10 ppm. Over time, the signal at 6.10
decreases in intensity with a new singlet appearing at 6.61 ppm.
Bright red crystals corresponding to the sulfur-bound DMSO adduct
5
1
η -cyclopentadieneyl-κ -dimethylsulfoxo-S-iridium(III) iodide grew
from a concentrated DMSO-d solution which was allowed to stand
open in air for several days (see the Supporting Information).
6
5
5
[
(η -C
5
H
5
)IrI(2,2′-dipyridyl)]PF
6
5 5 2 x
. [(η -C H )IrI ] (0.162 g,
0
.317 mmol), 2,2′-dipyridyl (0.0549 g, 0.352 mmol), and potassium
hexafluorophosphate (0.0881 g, 0.479 mmol) were combined in an
oven-dried 50 mL Schlenk flask under nitrogen, and 30 mL of dry,
degassed acetonitrile was added. The suspension was heated to 50
°
C under nitrogen, gradually lightening in color until it was an
entirely homogeneous solution after approximately 4 h. After being
heated for 16 h, the now-yellow solution was cooled to room
temperature and reduced in volume to 1-2 mL on a rotary
evaporator. Addition of 50 mL of deionized water gave a yellow
precipitate which was collected by filtration in air. The solid was
washed first with water, then diethyl ether and finally dried in vacuo.
1
Yield 0.158 g (73%). H NMR (400 MHz, DMSO-d
6
) δ 9.54 (d, J
)
2
1
5.0 Hz, 2H), 8.80 (d, J ) 8.0 Hz, 2H), 8.31 (m, 2H), 7.74 (m,
H), 6.33 (s, 5H). 13C NMR (126 MHz, DMSO-d
) δ 157.07,
55.03, 140.49, 128.16, 124.83, 79.66. Anal. Calcd for
6
Ir
1
I
1
C
15
H
13
N
2
P
1
F
6
. C, 26.29; H, 1.91; N, 4.09. Found: C, 26.50; H,
1
.79; N, 4.24.
5
5
[
(η -C
5
H
5
)IrI(2,2′-dipyridyl)]NO
3
. [(η -C
5
H
5
)IrI
2
]
x
(0.1036 g,
0
.203 mmol), 2,2′-dipyridyl (0.0365 g, 0.234 mmol), and tetra-n-
butylammonium nitrate (0.0936 g, 0.307 mmol) were combined in
an oven-dried 50 mL Schlenk flask under nitrogen, and 20 mL of
dry, degassed acetonitrile was added. The reaction was stirred at
Cerium(IV) ammonium nitrate (CAN) of the highest grade
99.99%) was purchased from Sigma-Aldrich and kept in a
desiccator prior to use. In the case of comparative rate experiments,
the pH of the solutions was adjusted by using concentrated nitric
acid solutions. Deuterated cerium(IV) ammonium nitrate was
(
5
0 °C for 13 h before being cooled to room temperature and
evaporated to dryness on a rotary evaporator. The resulting solid
was dissolved in 40 mL of deionized water, and the aqueous
solution was extracted with three 30 mL portions of dichlo-
romethane. The combined organic portions were discarded, and the
aqueous solution once again evaporated to dryness. The remaining
solid was washed once with a 40 mL portion of a 2:1 mixture of
acetone with diethyl ether to remove any residual tetra-n-butylam-
2
prepared by dissolution of CAN in excess 99% D O (Cambridge
Isotope Laboratories, Cambridge, MA) followed by removal of
solvent under reduced pressure and drying overnight. Solutions for
KIE at lower cerium(IV) loading were prepared by first acidifying
solvent with HNO
Laboratories).
3 3 3
or DNO (99% DNO from Cambridge Isotope
monium salts, giving 0.075 g of the product as a yellow powder
18O Incorporation and Stable Isotope Ratio Mass Spectrometry
SIR-MS). Analyses were performed using a Thermo-Finnigan
GasBench II connected to a Thermo-Finnigan Delta XP stable
isotope ratio mass spectrometer operating in continuous-flow mode.
Both the GasBench and Delta
software (version 3.0). Using the GasBench II, the headspace of a
vial is sampled via a nested sampling needle. Helium flows (ca.
1
(
61% yield). H NMR (500 MHz, DMSO-d
6
) δ 9.54 (d, J ) 5.7
(
Hz, 2H), 8.80 (d, J ) 8.2 Hz, 2H), 8.31 (s, 2H), 7.74 (s, 2H), 6.32
plus
5H). 1 C NMR (126 MHz, DMSO-d
3
(
1
6
) δ 157.07, 155.00, 140.47,
28.14, 124.83, 79.65. Anal. Calcd for Ir . C, 29.91;
1 1 15 13 3 3
I C H N O
plus
XP were controlled by Isodat
H, 2.18; N, 6.98. Found: C, 29.69; H, 2.05; N, 6.89. Crystals suitable
for X-ray diffraction study were obtained by diffusion of diethyl
ether into a saturated acetonitrile solution.
-1
0
.5 µL min ) into the vial through an upper aperture in the needle,
X-ray Diffraction Studies of Compounds 6, 8, CpIrI
2
(DMSO-
and the resultant mixture flows back through a lower aperture and
into an inner capillary. The gas mixture is then directed into a
Nafion (DuPont) drying tube and out through a 10-µL sampling
loop in a Valco switching valve. Once the sampling loop is flushed
with headspace gas, it is switched to inject the sample in the loop
onto an integral gas chromatograph with a Varian CP7551 PLOT
fused silica 25 m × 0.32 mm column (coating: Poraplot Q) at 32
S), and CpIr(O-NO )(bipy)PF . Crystal samples were mounted in
2
6
a polyimide MiTeGen loop with immersion oil. All measurements
were made on a Rigaku SCXMini diffractometer with filtered
Mo-KR radiation at a temperature of 223 K. Three omega scans
consisting of 180 data frames each were collected with a frame
width of 1.0° for compounds 6 and 8. Two omega scans consisting
of 180 data frames each were collected for CpIrI
2
(DMSO-S) and
°
2 2
C to separate O from CO . The separated gases are then carried
CpIr(O-NO )(bipy)PF . The data frames were processed and scaled
2
6
through a second drying tube to an open split where a glass capillary
continually draws in helium and sample gases to the mass
spectrometer for analysis. The dilution factor of the open split can
5
2
using the Rigaku CrystalClear. The data were corrected for
Lorentz and polarization effects. The structure was solved by direct
5
3
methods and expanded using Fourier techniques. The non-
hydrogen atoms were refined anisotropically and hydrogen atoms
were treated as idealized contributions. The final cycle of full-matrix
(52) CrystalClear and CrystalStructure, Rigaku/MSC, The Woodlands,
Texas, 2005.
(
53) SHELXS (direct methods): Sheldrick, G. M., 1990. SHELXL (refine-
(
51) Yamazaki, H. Bull. Chem. Soc. Jpn. 1971, 44, 582.
6028 J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010
ment): Sheldrick, G. M.; Schneider, T. R., 1997.
1

Half-Sandwich Iridium Complexes
A R T I C L E S
be adjusted to decrease the signal intensity, if necessary. In the
mass spectrometer, the samples are ionized (electron ionization)
The vibrational data were used to relax the geometry of each
transition state toward reactants and products, in order to confirm
its nature. Free energies were calculated within the harmonic
approximation for vibrational frequencies. The full set of frequencies
was used to calculate KIE. The effect of the solvent, water, was
and analyzed in three cups set to O
3
1
2
peaks with masses 32, 33, and
4. The relative nominal amplifications for the three cups are 1,
00, and 300. Under the conditions we used, the sample vials are
6
1
only slightly pressurized because the sampling gas mixture from
the vial either was directly vented (with the Valco valve set to
modeled by single-point calculations with the CPCM method.
The influence of the counteranion was not implemented since the
cations and the anions should form separated ion pairs in water.
The zero-point, thermal, entropy, and solvent corrections did not
alter the energy profiles to any great extent. These corrections are
given in the Supporting Information. The spin densities were
“
inject” mode) or was vented after going through the sampling loop
(
with the Valco valve set to “load” mode). This technique has been
described and used previously by our group.
5
5
Electrochemistry. Electrochemical measurements were made on
an EG&G Princeton Applied Research Model 273A potentiostat/
galvanostat using a standard three-electrode configuration. A basal
62
obtained from NPA (Natural Population Analysis) calculations.
2
plane graphite electrode (surface area: 0.09 cm ) was used as the
Acknowledgment. This material is based upon work supported
as part of the Argonne-Northwestern Solar Energy Research
(ANSER) Center, an Energy Frontier Research Center funded by
the U.S. Department of Energy, Office of Science, Office of Basic
Energy Sciences under Award Number DE-PS02-08ER15944
working electrode to reduce background water oxidation. The
electrode consisted of a brass cylinder, sheathed in a Teflon tube.
At the tip of the brass, a two-part silver conducting epoxy (Alfa
Aesar) was used to firmly attach the basal plane carbon electrode
surface to the brass. Finally, the tip was sealed with the organic
solvent-resistant, electrically insulating, two-part epoxy Tra-bond
(
G.W.B., R.H.C., and J.D.B.). O.E. thanks the Minist e` re de
l’Enseignement Sup e´ rieur et de la Recherche and the CNRS for
funding. D.B. thanks Sanofi-Aventis for a postdoctoral fellowship
(2007-2009) and the Spanish MICINN for his current Juan de la
Cierva position. J.D.B. thanks Katherine Shinopoulos for helpful
discussions.
2
151 (Emerson and Cuming, Canton, MA). Immediately prior to
experiments, the working electrode was polished with 1 µm alumina
paste, washed with copious amounts of water, and allowed to dry
completely. Then the surface of the working electrode was
resurfaced with tape to restore the gray, basal surface.
For aqueous electrochemical experiments, a platinum wire was
used as the counter electrode, and a saturated calomel electrode
Supporting Information Available: NMR spectra for new
(
SCE) was used as the reference electrode (NHE vs SCE: +241
mV). Experiments were carried out in unbuffered solutions contain-
ing 0.1 M KNO (Johnson Matthey, electronics grade) as the
5
compounds 6, 7, and 8. Time-dependent NMR spectra for [(η -
5
C
C
5
H
5
)IrI
2
]
x
. Crystal structures and parameters for 6, 7 (η -
3
5
5
H
5
)IrI
2
(dimethyl sulfoxide-S), and (η -C
5 5 2
H )Ir(O-NO )-
supporting electrolyte.
For nonaqueous electrochemical studies, a platinum wire was
6
(bipy)PF . Kinetic data, log-log plots, oxygen-evolution traces,
used as the counter electrode, and Ag/0.1 M AgNO
the reference electrode. Ferrocene was used as an external standard
3
was used as
and ESI+ MS data. Cyclic voltammograms of 1, 3, and 6.
Optimized geometries of all stationary points reported in the
article, with the associated potential and CPCM energies and
the zero-point, thermal, and entropy energy corrections. Com-
plete list of authors for ref 58. This material is available free of
charge via the Internet at http://pubs.acs.org.
+
to calibrate the reference electrode versus NHE (Fc/Fc : +0.690
56
V vs NHE in MeCN). Experiments were carried out in still-dried
acetonitrile solution containing 0.1 M tetrabutylammonium per-
chlorate (Fluka, electrochemical grade) as the supporting electrolyte.
Computational Details. Calculations were performed at the
5
7
58
DFT(B3LYP) level with Gaussian03. Geometry optimizations
5
9
JA104775J
were carried out with the 6-31G** basis set for O, N, C, and H
and the Stuttgart-Bonn scalar relativistic ECP with associated basis
6
0
set for Ir. These basis sets were used to compute the energies
given in the text. All geometries were fully optimized without any
symmetry or geometry constrains. The nature of all stationary points
was confirmed by the analytical calculation of their frequencies.
(
58) Frisch, M. J. et al. GAUSSIAN 03, reVision D.01; Gaussian, Inc.:
Wallingford, CT, 2004.
(59) (a) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Phys. Chem. 1972, 56,
257–2261. (b) Ehlers, A. W.; B o¨ hme, M.; Dapprich, S.; Gobbi, A.;
2
H o¨ llwarth, A.; Jonas, V.; K o¨ hler, K. F.; Stegmann, R.; Veldkamp,
A.; Frenking, G. Chem. Phys. Lett. 1993, 208, 111–114. (c) H o¨ llwarth,
A.; B o¨ hme, H.; Dapprich, S.; Ehlers, A. W.; Gobbi, A.; Jonas, V.;
K o¨ hler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G. Chem. Phys.
Lett. 1993, 203, 237–240. (d) Hariharan, P. C.; Pople, J. A. Theor.
Chim. Acta 1973, 28, 213–222.
2
2 2
(
(
54) Least squares function minimized: ∑w(F
o
c
- F ) .
55) Chen, H.; Tagore, R.; Olack, G.; Vrettos, J. S.; Weng, T.-C.; Penner-
Hahn, J.; Crabtree, R. H.; Brudvig, G. W. Inorg. Chem. 2007, 46,
3
4–43.
(
(
56) Barrette, W. C., Jr.; Johnson, H. W., Jr.; Sawyer, D. T. Anal. Chem.
984, 56, 1560.
57) (a) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. ReV. B 1988, 37, 785–
89. (b) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (c)
Stephens, P.; Devlin, F.; Chabalowski, C.; Frisch, M. J. Phys. Chem.
994, 98, 11623–11627.
(60) (a) Andrae, D.; H a¨ ussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor.
Chim. Acta 1990, 77, 123–141. (b) Bergner, A.; Dolg, M.; K u¨ chle,
W.; Stoll, H.; Preuss, H. Mol. Phys. 1993, 30, 1431–1441.
(61) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995–2001.
(62) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899–
926.
1
7
1
J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010 16029