Visible light-driven hydrogen production from aqueous protons catalyzed by molecular cobaloxime catalysts

Article abstract of DOI:10.1021/ic900389z

A series of cobaloxime complexes-([Co(dmgH)₂pyCI] (1), [Co(dmgH)₂(4-COOMe-py)CI] (2), [Co(dmgH)₂(4-Me ₂N-py)CI] (3), [Co(dmgH)(dmgH₂)CI₂] (4), [Co(dmgH)₂(py)₂](PF

Full text of DOI:10.1021/ic900389z

4
952 Inorg. Chem. 2009, 48, 4952–4962
DOI:10.1021/ic900389z
Visible Light-Driven Hydrogen Production from Aqueous Protons Catalyzed by
Molecular Cobaloxime Catalysts
Pingwu Du, Jacob Schneider, Genggeng Luo, William W. Brennessel, and Richard Eisenberg*
Department of Chemistry, University of Rochester, Rochester, New York 14627
Received February 25, 2009
A series of cobaloxime complexes;([Co(dmgH) pyCl] (1), [Co(dmgH) (4-COOMe-py)Cl] (2), [Co(dmgH) (4-Me N-
2
2
2
2
py)Cl] (3), [Co(dmgH)(dmgH )Cl ] (4), [Co(dmgH) (py) ](PF ) (5), [Co(dmgH) (P(n-Bu) )Cl] (6), and [Co
3
2
2
2
2
6
2
(
dmgBF ) (OH ) ] (7), where dmgH = dimethylglyoximate monoanion, dmgH = dimethylglyoxime, dmgBF =
2 2 2 2 2 2
(difluoroboryl)dimethylglyoximate anion, and py = pyridine;were synthesized and studied as molecular catalysts
for the photogeneration of hydrogen from systems containing a Pt terpyridyl acetylide chromophore and triethano-
lamine (TEOA) as a sacrificial donor in aqueous acetonitrile. All cobaloxime complexes 1-7 are able to quench the
0
luminescence of the Pt(II) chromophore [Pt(ttpy)(CtCPh)]ClO (C1) (ttpy = 4 -p-tolyterpyridine). The most effective
electron acceptor for hydrogen evolution is found to be complex 2, which provides the fastest luminescence quenching
rate constant for C1 of 1.7 ꢀ 10 M
4
9
-1 -1
s . The rate of hydrogen evolution depends on many factors, including the
stability of the catalysts, the driving force for proton reduction, the relative and absolute concentrations of system
components (TEOA, Co molecular catalyst, and sensitizer), and the ratio of MeCN/water in the reaction medium. For
example, when the concentration of TEOA increases, the rate of H photogeneration is faster and the induction period
2
is shorter. Colloidal cobalt experiments and mercury tests were run to verify that the system is homogeneous and that
catalysis does not occur from in situ generated colloidal particles during photolysis. The most effective system
-
5
examined to date consists of the chromophore C1 (1.1 ꢀ 10 M), TEOA (0.27 M), and catalyst complex 1
-
4
(
2.0 ꢀ 10 M) in a MeCN/water mixture (24:1 v/v, total 25 mL); this system has produced ∼2150 turnovers of H after
2
only 10 h of photolysis with λ > 410 nm.
Introduction
Colloidal platinum is a well-known heterogeneous catalyst
for hydrogen production. Numerous successful systems
The development of renewable energy from the sun is a
major challenge in addressing the problem of abundant and
environmentally benign energy for sustainable development
to produce H have been constructed utilizing colloidal pla-
2
tinum nanoparticles in combination with different chrom-
2
+ 3,9-11
12-15
1-8
ophores, such as Ru(bpy)3
,
Zn(II) porphyrins, 1
globally.
A promising energy conversion pathway con-
6,17
platinum(II)
terpyridyl
[Ir(C N)2(N N)]PF
acetylide
complexes,
sidered by many researchers is the development of efficient
catalytic systems that can convert the energy of solar photons
into stored chemical potential by splitting water into hydro-
gen and oxygen. Hydrogen with its high specific enthalpy of
combustion, its use in fuel cells, and its benign combustion
product (water) is the ideal fuel to be produced using solar
energy, and it will serve to reduce mankind’s dependence on
fossil fuels and subsequent emissions of the greenhouse gases.
∧
∧
18-20
cyclometalated
complexes,
6
(
9) Kalyanasundaram, K.; Kiwi, J.; Gratzel, M. Helv. Chim. Acta 1978,
1, 2720–2730.
(10) Kirch, M.; Lehn, J.-M.; Sauvage, J.-P. Helv. Chim. Acta 1979, 62,
1345–1384.
6
(
(
11) Lehn, J.-M.; Sauvage, J.-P. Nouv. J. Chim. 1977, 1, 449–451.
12) Adar, E.; Degani, Y.; Goren, Z.; Willner, I. J. Am. Chem. Soc. 1986,
1
08, 4696–4700.
*
To whom correspondence should be addressed. E-mail: eisenberg@
(13) Okura, I. Coord. Chem. Rev. 1985, 68, 53.
(
14) Okura, I.; Kaji, N.; Aono, S.; Kita, T. Inorg. Chem. 1985, 24, 451–
453.
(15) Okura, I.; Kim-Thuan, N. J. Mol. Catal. 1979, 6, 227–230.
16) Du, P.; Schneider, J.; Jarosz, P.; Eisenberg, R. J. Am. Chem. Soc.
chem.rochester.edu.
(
(
1) Bard, A. J.; Fox, M. A. Acc. Chem. Res. 1995, 28, 141–145.
2) Chakraborty, S.; Wadas, T. J.; Hester, H.; Schmehl, R.; Eisenberg, R.
(
Inorg. Chem. 2005, 44, 6865–6878.
:
:
(
(
(
3) Gratzel, M. Acc. Chem. Res. 1981, 14, 376–384.
4) Gratzel, M. Inorg. Chem. 2005, 44, 6841–6851.
5) Alstrum-Acevedo, J. H.; Brennaman, M. K.; Meyer, T. J. Inorg. Chem.
2006, 128, 7726–7727.
(17) Du, P.; Schneider, J.; Jarosz, P.; Zhang, J.; Brennessel, W. W.;
Eisenberg, R. J. Phys. Chem. B 2007, 111, 6887–6894.
::
2
005, 44, 6802–6827.
(18) Goldsmith, J. I.; Hudson, W. R.; Lowry, M. S.; Anderson, T. H.;
Bernhard, S. J. Am. Chem. Soc. 2005, 127, 7502–7510.
(19) Grzybowski, J. J.; Urbach, F. L. Inorg. Chem. 1980, 19, 2604–2608.
(20) Lowry, M. S.; Goldsmith, J. I.; Slinker, J. D.; Rohl, R.; Pascal, J. R. A.;
Malliaras, G. G.; Bernhard, S. Chem. Mater. 2005, 17, 5712–5719.
(
(
6) Esswein, A. J.; Nocera, D. G. Chem. Rev. 2007, 107, 4022–4047.
7) Lewis, N. S.; Nocera, A. G. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
1
5729–15735.
8) Eisenberg, R.; Nocera, D. G. Inorg. Chem. 2005, 44, 6799–6801.
(
pubs.acs.org/IC
Published on Web 4/27/2009
© 2009 American Chemical Society

Article
Inorg. Chem., Vol. 48, No. 11, 2009
4953
2
1-23
24
halogenated fluorescein dyes,
methyl acridine orange,
hydrogen-forming reaction and allow comparisons between
25
and conjugated polymers. However, due to the high price
of platinum metal, many researchers are exploring earth-
abundant, less expensive catalysts such as cobalt, nickel,
to replace the colloidal plati-
num catalyst. Early studies reported the successful use of
as a hydrogen-generating catalyst to replace
colloidal platinum in systems using Ru(bpy)
recently, [Ir(ppy) (bpy)] photosensitizers.
different H -generating systems.
2
Experimental Section
26-31
and iron complexes
Chemicals. Cobalt chloride hexahydrate, cobalt acetate
tetrahydrate, dimethylglyoxime (dmgH
pyridine (4-Me N-py), methyl isonicotinate (4-COOMe-py),
2
), 4-(dimethylamino)
2+
Co(bpy)3
2
2+
and, more
Within the
and triethanolamine (TEOA) are commercially available (Al-
drich) and were used as received. HPLC-grade acetonitrile
3
18,32
2+
2
(
MeCN) was purchased from Fisher and used without further
purification. The cobalt complexes [Co(dmgH) pyCl] (1), [Co
past decade, cobaloximes (cobaloximes are bis(dialkyl or
diarylglyoximate)cobalt complexes) have been reported for
hydrogen production based initially on electrocatalytic
3
7
2
3
8
39
(
dmgH)(dmgH )Cl ] (4), [Co(dmgH) (py) ](PF ) (5), [Co
2 2 2 2 6
4
0
41
28-31
2 2 2 2 2 3
(dmgBF ) (OH ) ] (7), [Co(dmgH) (P(n-Bu) )Cl] (6), and
the platinum chromophores C1-C7 (Scheme 3)
synthesized as reported in the literature.
studies
investigations.
and their respective co-workers reported that Co(dmgBF2)2
dmgBF = difluoroboryldimethylglyoximate) can electro-
and extended more recently to photochemical
16,47
were
33-36
In the earlier studies, Artero and Peters
[
sample of [Co(dmgH)(dmgH )Cl ] was suspended in 50 mL of
Co(dmgH) (4-COOMe-py)Cl] (2). A 500 mg (1.48 mmol)
2
(
2
2 2
catalytically produce hydrogen at an impressively low over-
potential of 50 mV in acetonitrile in the presence of
methanol. One equivalent of triethylamine was then added to
the flask, after which the complex dissolved within 5 min,
changing from a green suspension to a clear brown solution.
Methyl isonicotinate, 203 mg (1.48 mmol), was then added and
the solution stirred for 1 h following formation of a brown
precipitate (10-20 min). The suspension was filtered and the
precipitate washed with water (10 mL), ethanol (10 mL), and
CF COOH and also generate hydrogen using other acids
3
+
(
such as tosic acid, p-cyanoanilinium, and Et NH ) as
3
the proton source. In 2008, Artero and co-workers emp-
2+
loyed related cobaloxime complexes with Ru(bpy) , [Ir
3
2
+
(
ppy) (bpy)] , and ReBr(CO) (phen) to produce hydrogen
2
3
diethyl ether (10 mL) to give [Co(dmgH)
2
(4-COOMe-py)Cl] (2)
33,34
photochemically,
but in these systems, the proton source
1
600 mg, 93%). H NMR (400 MHz, CDCl ): 8.40 (d, 2H, py),
(
7.70 (d, 2H, py), 3.87 (s, 3H, -COOMe), 2.35 (s, 12H, dmg ).
3
was [Et NH]BF or [Et NH]Cl and not water. In fact, these
2-
3
4
3
systems encountered serious problems with low efficiency for
hydrogen generation when water was added to the system.
In 2008, we extended our initial report of the photochemi-
Anal. calcd for C H N ClCoO : C, 39.02; H, 4.58; N, 15.17.
6
Found: C, 38.88; H, 4.64; N, 15.02.
1
Characterization. H NMR spectra were recorded on a
1
5
21
5
cal generation of H from aqueous protons and a sacrificial
Bruker Avance-400 spectrometer (400.1 MHz). Absorption
spectra were recorded using a Hitachi U2000 scanning spectro-
photometer (200-1100 nm). Luminescence spectra were ob-
tained using a Spex Fluoromax-P fluorimeter corrected for the
spectral sensitivity of the photomultiplier tube and the spectral
output of the lamp with monochromators positioned for a 2 nm
band-pass. Solution samples were degassed by three freeze-
pump-thaw cycles.
2
electron source in a system containing the Pt terpyridyl
acetylide chromophore C1 and a colloidal Pt catalyst to
one having the same chromophore but with the Co dmg
3
5
complex 1 as the molecular catalyst for H formation. In
2
this paper, we present a much more detailed study of this
system and expand our investigations to include other Co
catalysts, different Pt sensitizers, variation of the solvent, and
the effect of sacrificial donor and concentrations. The results
lead to conclusions about different steps in the overall
Electrochemistry. Cyclic voltammetry experiments were
conducted on an EG&G PAR 263A potentiostat/galvanostat
using a three-electrode single-compartment cell including a
glassy carbon working electrode, a Pt wire auxiliary electrode,
and a Ag wire reference electrode. For all measurements,
samples were degassed by bubbling argon through the solution
for 10 min. Tetrabutylammonium hexafluorophosphate (Flu-
ka) was used as the supporting electrolyte (0.1 M), and ferrocene
was employed as an internal redox reference. All redox poten-
(
21) Bi, Z.; Qian, Y.; Huang, J.; Xiao, Z.; Yu, J. J. Photochem. Photobiol.
A: Chem. 1994, 77, 37–40.
(
(
22) Hashimoto, K.; Kawai, T.; Sakata, T. Chem. Lett. 1983, 5, 709–712.
23) Hashimoto, K.; Kawai, T.; Sakata, T. Nouv. J. Chem. 1984, 8, 693–
7
00.
24) Islam, S. D. M.; Yoshikawa, Y.; Fujitsuka, M.; Ito, O.; Komatsu, S.;
Usui, Y. J. Photochem. Photobiol. A: Chem. 2000, 134, 155–161.
25) Jiang, D. L.; Choi, C. K.; Honda, K.; Li, W. S.; Yuzawa, T.; Aida, T.
J. Am. Chem. Soc. 2004, 126, 12084–12089.
+
tials were measured relative to the ferrocenium/ferrocene (Fc /
(
4
2
Fc) couple (0.40 V vs SCE) used as an internal standard and
then adjusted to NHE (SCE vs NHE = 0.24 V). All scans were
(
-
1
.
done at 100 mV s
(
(
26) Na, Y.; Pan, J.; Wang, M.; Sun, L. Inorg. Chem. 2007, 46, 3813–3815.
27) Na, Y.; Wang, M.; Pan, J.; Zhang, P.; Akermark, B.; Sun, L. Inorg.
˚
Quenching Experiments. All quenching experiments
were done by adding calibrated amounts of quencher into the
chomophore solution at room temperature. The Pt chromo-
Chem. 2008, 2805–2810.
28) Razavet, M.; Artero, V.; Fontecave, M. Inorg. Chem. 2005, 44, 4786–
795.
(
-
5
4
1
8
phore solution was held fixed at 2.0 ꢀ 10 M in MeCN/water
(
29) Baffert, C.; Artero, V.; Fontecave, M. Inorg. Chem. 2007, 46, 1817–
(
3:2 v/v). The solutions were degassed through three
824.
(
988–8998.
freeze-pump-thaw cycles. Steady-state luminescence spectra
were then collected for each sample. The relative emission
intensity of the chromophore in the absence and presence of
30) Hu, X.; Brunschwig, B. S.; Peters, J. C. J. Am. Chem. Soc. 2007, 129,
(
31) Hu, X. L.; Cossairt, B. M.; Brunschwig, B. S.; Lewis, N. S.; Peters, J.
C. Chem.Commun. 2005, 4723–4725.
32) Brown, G. M.; Brunschwig, B. S.; Creutz, C.; Matsubara, T.; Sutin,
N. J. Am. Chem. Soc. 1979, 101, 1298–1300.
33) Fihri, A.; Artero, V.; Razavet, M.; Baffert, C.; Leibl, W.; Fontecave,
M. Angew. Chem., Int. Ed. 2008, 47, 564–567.
34) Fihri, A.; Artero, V.; Pereira, A.; Fontecave, M. Dalton Trans. 2008,
567–5569.
35) Du, P. W.; Knowles, K.; Eisenberg, R. J. Am. Chem. Soc. 2008, 130,
2576–12577.
36) Probst, B.; Kolano, C.; Hamm, P.; Alberto, R. Inorg. Chem. 2009, 48,
836–1843.
(
(
(37) Schrauzer, G. N. Inorg. Synth. 1968, 11, 61.
(38) Trogler, W. C.; Stewart, R. C.; Epps, L. A.; Marzilli, L. G. Inorg.
Chem. 1974, 13, 1564–1570.
(39) Gerli, A.; Marzilli, L. G. Inorg. Chem. 1992, 31, 1152–1160.
(40) Bakac, A.; Espenson, J. H. J. Am. Chem. Soc. 1984, 106, 5197–5202.
(41) Toscane, P. J.; Swider, T. F.; Marzilli, L. G.; Bresciani-Pahor, N.;
Randaccio, L. Inorg. Chem. 1983, 22, 3416–3421.
(
5
1
1
(
(
(42) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877–910.

4954 Inorg. Chem., Vol. 48, No. 11, 2009
Du et al.
the quencher was used to calculate the quenching rate constant
by the Stern-Volmer equation.
4
3
Details of Hydrogen Production. For photoinduced hy-
drogen evolution, each sample was made in a 50 mL round-
bottom flask with a volume of 25 mL in MeCN/water (3:2 v/v).
-
5
Typically, the sample contained 1.1 ꢀ 10 M Pt chromophore,
-
2
-4
1
.6 ꢀ 10 M TEOA, and 2.0 ꢀ 10 M cobaloximes. The flask
was sealed with a septum and degassed by bubbling nitrogen for
5 min under atmospheric pressure at room temperature. After
1
that, 5 mL of nitrogen was removed from the flask and 5 mL of
methane (760 Torr) was added to the flask to serve as the
internal standard for gas chromatography (GC) measurements.
The samples were irradiated by a 200W Mercury Xeon lamp
with a filter that cut off light with λ < 410 nm. The quantities
of hydrogen were measured by GC using a Shimadzu
˚
GC-17A chromatograph with a Molecular Sieve 5A column
(
30 m ꢀ 0.53 mm).
X-Ray Structural Determination of Complexes 2 and 3.
Figure 1. Normalized absorption spectra of cobaloxime complexes 1-7
Brown crystals of 2 and red crystals of 3 suitable for single-
crystal X-ray diffraction were maintained at 100.0(1) K, on a
Bruker SMART platform diffractometer equipped with an
APEX II CCD detector. The X-ray source, powered at 50 kV
-5
in MeCN/water (3:2 v/v) at 4.0 ꢀ 10 M.
Scheme 1
˚
and 30 mA, provided Mo KR radiation (λ = 0.71073 A, graphite
monochromator). Space groups were determined based on
systematic absences, intensity statistics, and space group fre-
quencies. Direct methods were used to solve the structures.
4
4
Full-matrix least-squares/difference Fourier cycles were per-
formed, which located the remaining non-hydrogen atoms.
All non-hydrogen atoms were refined with anisotropic displace-
ment parameters. All hydrogen atoms were given riding
models and refined with relative isotropic displacement para-
meters.
Scheme 2
Results and Discussion
Synthesis and Electronic Absorption. Cobaloxime com-
plexes 2 and 3 were synthesized by stirring a 1:1 ratio of
Co(dmgH)(dmgH )Cl and the corresponding pyridine
2
2
3
8
derivative in methanol. One of the chloride ligands in
Co(dmgH)(dmgH )Cl was replaced by the correspond-
ing pyridine derivative to form 2 and 3 with yields of over
2
2
8
5% (Scheme 1).
The UV-vis absorption spectra of complexes 1-7 (see
Scheme 2 for complexes) measured at room temperature
in MeCN/water (3:2 v/v) are shown in Figure 1. Com-
plexes 1-6 exhibit a broad high-energy absorption
4
3
between 220 and 300 nm with ε on the order of 10 dm
-1
-1
mol cm , which corresponds to spin-allowed intrali-
gand (π-π*) transitions. Unlike the Co(III) complexes
1
spectrum with a broad low-energy band between
-6, catalyst 7 possesses a very different absorption
3
4
mol
00 and 530 nm centered at 442 nm (ε ∼ 7750 dm
- -1
1
cm ) corresponding to Co(II) absorption, as
3
0,40
assigned previously.
Structural Characterization of Complexes 2 and 3. The
solid-state molecular structures of 2 and 3, as determined
by single-crystal X-ray crystallography, are shown in
Figure 2 with important crystallographic parameters
and bond lengths and angles given in Tables 1 and 2,
respectively. In both 2 and 3, the refined structure reveals
a six-coordinate Co(III) ion that is ligated by two
coplanar dimethylglyoximate ligands and trans chloride
and pyridine ligands. The observed axial Co-N
pyridine
distances of 1.959 A in complex 2 and 1.946 A in complex
˚
˚
3
agree closely with those of in other Co(III) cobalox-
imate complexes, such as [Co(dmgH) pyCl] with a Co-
2
4
5
˚
distance of 1.959 A. The average Co-Nimine
bond distances of the glyoximate ligands in 2 and 3 are
N
pyridine
˚
1.892 and 1.891 A, respectively, which are consistent with
the values of 1.895 A in [Co(dmgH) pyCl], 1.891 A in
4
5
˚
˚
2
(
43) Vincze, L.; Sandor, F.; Pem, J.; Bosnyak, G. J. Photochem. Photobiol.
999, 120, 11–14.
44) SHELXTL, V.6.14; Bruker AXS: Madison, WI, 2000.
1
(45) Geremia, S.; Dreos, R.; Randaccio, L.; Tauzher, G. Inorg. Chim. Acta
1994, 216, 125–129.
(

Article
Inorg. Chem., Vol. 48, No. 11, 2009
4955
Figure 2. Perspective views of complex 2 (left) and complex 3 (right) with atomic numbering scheme. All hydrogen atoms have been omitted for clarity.
Table 1. Crystallographic, Data Collection, and Structure Refinement Parameters for Complexes 2 and 3
form
brown needle (complex 2)
27ClCoN
red prism (complex 3)
24ClCoN
empirical formula
fw
C
507.82
17
H
5
O
7
C
15
H
446.78
6 4
O
T, K
wavelength, A
100.0(1)
0.71073 A
monoclinic
100.0(1)
0.71073 A
monoclinic
P2 /n
1
4
7.886 (2)
22.361(5)
13.229(3)
˚
˚
cryst syst
space group
Z
P2
4
1
/c
˚
a, A
9.9436(7)
22.955(2)
9.6917(6)
˚
b, A
˚
c, A
R, β, γ, deg
V, A
90°, 91.317(1)°, 90°
2211.6(3)
1.525
0.945
90°, 92.98(3)°, 90°
2329.7(8)
1.274
0.880
90 2. 38 2 ꢀ 0.14 ꢀ 0.12
1.79 to 26.00°
-9 e h e +9, -27 e k e +27, -16 e l e +16
26914
4580/0/244
1.073
R1 = 0.0349, wR2 = 0.0976
R1 = 0.0380, wR2 = 0.0992
3
˚
3
F
calcd, Mg/m
-
1
μ, mm
cryst size, mm
F(000)
3
0.24 ꢀ 0.20 ꢀ 0.04
1056
2.05 to 34.97°
2θ range, deg
limiting indicies
-16 e h e +16, -37 e k e +37, -15 e l e +15
35749
no. of reflns collected
no. of data/restraints /params
goodness of fit
R1, wR2 (l > 2σ)
R1, wR2 (all data)
9578/3/303
1.016
R1 = 0.0514, wR2 = 0.1007
R1 = 0.0964, wR2 = 0.1165
a
b
b
P
P
P
P
a
2
2
2
1/2
b
2
o
GOF = [ [w(F
2
o
- F
c
) ]/(m - n)] , where n and p denote the number of data and parameters. R1 = ||F
o
| - |F
c
||/ |F
o
| and wR2 = [ [w(F
-
P
Co(dmgH) (4-CN-py)Cl],
2
2
) ]/ [w(F
2
o
1/2
2
2
o
2
+ (aP) + bP] and P = 1/3 max(0, F
2
2
F
c
) ]] ), where w = 1/[σ (F
o
) + 2/3F
c
.
4
6
42
˚
[
(
and 461.891
A
in [Co
standard (0.40 V vs SCE, SCE vs NHE = 0.24 V). All
cobaloxime complexes in the present study exhibit the
expected number of reductions and oxidations, and the
results are summarized in Table 3. The redox potentials
were also compared with values measured in DMF solu-
2
dmgH) (2,6-dimethylpyrazine)Cl].
2
Electrochemistry of Cobaloximes. Cyclic voltammetry
studies were performed on all of the complexes in MeCN
with 0.1 M tetra-n-butylammonium hexafluoropho-
sphate as the supporting electrolyte under an argon
atmosphere. All of the redox potentials were measured
2
8
tion and found to have slight but significant differences.
For example, complex 1 in MeCN shows one irreversible
reduction assigned to a Co(III)/Co(II) process at -0.43 V
versus NHE, followed by a reversible reduction of Co(II)
to Co(I) at -0.88 V, whereas in DMF the same complex
exhibits the first and second reduction potentials at -0.37
and -0.76 V, corresponding to a slight anodic
shift. Complex 1 also shows an irreversible oxidation
and adjusted to NHE using the ferrocenium/ferrocene
(
ing the potentials using literature values for the internal
+
Fc /Fc) couple as an internal standard and then adjust-
(
46) Concepcion Lopez, S. A.; Xavier Solans, M. F. Inorg. Chem. 1986, 25,
2
962–2969.

4956 Inorg. Chem., Vol. 48, No. 11, 2009
Du et al.
˚
Table 2. Selected Bond Lengths (A) and Angles (deg) for Complexes 2 and 3
within 10 min and continues to increase in intensity up to
3
5
Complex 2
a limiting value after ca. 50 min. When irradiation is
stopped, the bright yellow color persists under an inert
atmosphere but dissipates upon exposure of the solution
to air within 1-2 min. For the other cobaloximes (2-4,
bond lengths
Co(1)-N(2)
bond angles
N(2)-Co(1)-N(1)
1.884(2)
1.889(2)
1.895(2)
1.900(2)
1.959(2)
2.2279(5)
1.347(2)
1.339(2)
1.344(2)
1.345(2)
81.90(7)
177.08(7)
98.97(7)
97.81(7)
179.44(7)
81.30(7)
91.06(7)
89.02(7)
91.73(7)
91.47(7)
Co(1)-N(1)
Co(1)-N(3)
Co(1)-N(4)
Co(1)-N(5)
Co(1)-Cl(1)
N(1)-O(1)
N(2)-O(2)
N(3)-O(3)
N(4)-O(4)
N(2)-Co(1)-N(3)
N(1)-Co(1)-N(3)
N(2)-Co(1)-N(4)
N(1)-Co(1)-N(4)
N(3)-Co(1)-N(4)
N(2)-Co(1)-N(5)
N(1)-Co(1)-N(5)
N(3)-Co(1)-N(5)
N(4)-Co(1)-N(5)
6
), similar quenching behavior in the presence of TEOA
is also observed.
When complex 7, which contains Co(II), serves as the
electron acceptor, quenching is also observed with a
decrease in the photoluminescence of C1, and in the
presence of TEOA, the formation of a persistent blue
color with maxima at 550 and 650 nm (Figure 3) that
3
0
is characteristic of the corresponding Co(I) absorption
Complex 3
is seen. UV-vis spectra show an isosbestic point at 465
nm, indicating no decomposition during the Co(II)
reduction.
bond lengths
bond angles
Co(1)-N(3)
Co(1)-N(4)
Co(1)-N(1)
Co(1)-N(2)
Co(1)-N(5)
Co(1)-Cl(1)
N(1)-O(1)
N(2)-O(2)
N(3)-O(3)
N(4)-O(4)
1.8831(9)
1.8908(9)
1.8945(9)
1.8976(9)
1.9464(9)
2.2364(3)
1.346(1)
1.341(1)
1.331(1)
1.344(1)
N(3)-Co(1)-N(4)
N(3)-Co(1)-N(1)
N(4)-Co(1)-N(1)
N(3)-Co(1)-N(2)
N(4)-Co(1)-N(2)
N(1)-Co(1)-N(2)
N(3)-Co(1)-N(5)
N(4)-Co(1)-N(5)
N(1)-Co(1)-N(5)
N(2)-Co(1)-N(5)
81.68(4)
98.61(4)
179.24(4)
177.96(4)
98.31(4)
81.36(4)
89.54(4)
89.07(4)
91.63(4)
92.50(4)
Hydrogen Production. A series of experiments using
complexes 1-7 for hydrogen production have been car-
ried out employing visible light irradiation (λ > 410 nm)
and the following concentrations and conditions: 1.1 ꢀ
-5
-4
-2
1
1
0
0
M Pt chromophore, 1.6 ꢀ 10 M TEOA, and 2.0 ꢀ
M cobaloxime complex catalyst in MeCN/water (3:2
v/v). Quantitative determination of generated H was
2
conducted by GC analysis with added methane as an
internal standard, as described in the Experimental Sec-
tion. During initial observations using complex 1, the pH
of the reaction solution was varied between 5 and 13, and
process at +1.37 V (vs NHE) in MeCN, attributable to a
Co(III)/Co(IV) process. All of the other cobalt complexes
except 7 have the same trend in reduction potentials
between MeCN and DMF, with the latter being slightly
both the rate and the amount of H generated was seen to
2
maximize at pH ∼8.5. Below pH 5, no H generation was
2
less negative, 0.01-0.15 V. In contrast, for the dmgBF2
observed, and above pH 12, the amounts of H produced
2
complex 7, the reduction wave of the Co(II)/Co(I) couple
were moderate (less than 15% of that obtained at pH 8.5
during the same amount of irradiation).
in MeCN, -0.29 V versus NHE, is slightly more positive
than the corresponding potential in DMF.
Quenching Studies and Photoreduction of Co Complexes
1-7. Complexes 1-7 were all observed to quench the
luminescence of the Pt chromophore C1. In all samples,
quenching constants (k ) were determined using the
q
Stern-Volmer equation (eq 1), where I0 is the
integrated metal-to-ligand charge-transfer emission in-
2
9,30
Table 4 shows the results of these photolytic hydrogen
production experiments. No hydrogen can be detected in
the absence of any cobaloxime complex (run 9), indicat-
ing that the cobalt catalyst is one of the key components
of the homogeneous hydrogen production system. Com-
plex 2 proved to be the most active of the cobaloximes
4
3
examined, with up to 238 turnovers of H obtained after
2
tensity in the absence of a quencher, I is the corresponding
value in the presence of a quencher, τ is the excited-
5
increased hydrogen yields (run 3).
h of irradiation (run 2), while further photolysis led to
0
state lifetime in the absence of a quencher (for C1, τ =
0
For complex 6, in which the axial pyridine of 1 is
replaced by P(n-Bu) , no hydrogen was evolved in 5 h of
1
6
4
and [Q] is the molar concentration of the quencher.
.6 μs ), k is the bimolecular quenching rate constant,
q
3
photolysis (run 7), despite the fact that the redox potential
of the Co(II)/Co(I) couple (-0.74 V vs NHE in MeCN) is
thermodynamically favorable for proton reduction. It has
been reported that 6 undergoes reaction with NaBH to
4
form the stable Co(III) hydride [CoH(dmgH) (P(n-Bu) )]
3
that yields hydrogen only upon thermolysis at 150 °C. It
is therefore possible to ascribe the poor catalytic activity
of 6 for H generation in the present study to the great
2
stability of [CoH(dmgH) (P(n-Bu) )], as compared with
2
analogous hydride complexes [CoH(dmgH) py] having
2
I0=I ¼ 1 þ kqτ½Qꢁ
All of the cobaloxime complexes except 5 give fast
ð1Þ
2
quenching rates close to the diffusion-controlled limit of
s . Among these complexes, 2 yields
49
1
0
-1 -1
∼
1 ꢀ 10
M
9
-1 -1
the fastest quenching rate at 1.7 ꢀ 10 M
s . It is
noteworthy that complex 5 has a quenching rate cons-
tant 2 orders of magnitude slower than the other coba-
loximes. While the reason for this difference was not
investigated, it could relate to the fact that 5 is the only
complex that does not have a chloride ligand for possible
interaction with the Pt chromophore in the quenching
process.
3
pyridine ligands trans to the hydride that are much less
stable.
49
In the context of H evolution, the Co(II) complex 7, [Co
2
(dmgBF ) (OH ) ], which possesses good air and acid
stability, was studied initially by Espenson and Connolly.
2 2 2 2
50
In the photolysis solutions containing C1, TEOA, and
complex 1, the bright yellow color of Co(II) develops
(
48) Cline, E. D.; Adamson, S. E.; Bernhard, S. Inorg. Chem. 2008, 47,
0378–10388.
49) Schrauzer, G. N.; Holland, R. J. J. Am. Chem. Soc. 1971, 93, 1505–
1506.
(50) Connolly, P.; Espenson, J. H. Inorg. Chem. 1986, 25, 2684–2688.
1
(
(
47) Schneider, J.; Du, P.; Jarosz, P.; Lazarides, T.; Wang, X.; Brennessel,
W. W.; Eisenberg, R. Inorg. Chem. 2009, in press.

Article
Inorg. Chem., Vol. 48, No. 11, 2009
4957
a
Table 3. Electrochemical Potentials (V, vs NHE) for Cobaloximes and C1.
b
c
d
e
-1 -1
Compound
Oxidation
Reduction E1/2/v
Reduction E1/2/v
q
k M s
9
9
9
7
9
9
9
[
[
[
[
[
[
[
[
Co(dmgH)
Co(dmgH)
Co(dmgH)
Co(dmgH)(py)
Co(dmgBF
2
2
2
(py)Cl]
(4-COOMe-py)Cl]
1.37
1.39
1.33
1.64
1.23
1.54
1.35
1.56
-0.43
-0.88
-0.84
-0.88
-0.85
-0.29
-0.55
-0.74
-1.16
-0.37
-0.76
1.3 ꢀ 10
1.7 ꢀ 10
1.2 ꢀ 10
4.0 ꢀ 10
1.3 ꢀ 10
1.0 ꢀ 10
1.6 ꢀ 10
-0.53
-0.59
-0.14
(4-Me
](PF
(OH
Cl
(P(n-Bu)
2
N-py)Cl]
-0.44
-0.02
-0.79
-0.76
-0.33
2
6
)
2
)
2
2 2
) ]
Co(dmgH)
Co(dmgH)
2
2
]
-0.24
-0.67
-0.68
2
3
)Cl]
-0.40
-0.59
-0.59
-1.07
+
Pt(ttpy)(phenylacetylide)]
a
4 6
Electrochemical data were obtained in degassed solution with 0.1 M n-Bu NPF (TBAH) as supporting eletrolyte at room temprature; Scan
42 b
+/0
Rate:100 mV/sec; potential vs NHE converted from internal standard Fc
solution with Fc
NHE; Reductions measured in DMF solution and adjusted to vs NHE. Taken from Ref.
using Ref. ; Irreversible oxidation peak estimated vs NHE in MeCN
+/0
+
/0
c
as an internal standard; Reversible reductions measured in MeCN solution with Fc
as an internal standard and adjusted to vs
Measured in MeCN/water (3:2 v/v); The excited state
d
28
e
f
;
+
17
oxidation potential (vs NHE) of [Pt(ttpy)(phenylacetylide)] (C1) has been estimated as -0.92 V (when written conventionally as a reduction).
Table 4. Photocatalytic Hydrogen Production Using Different Cobaloximes as
a
the Catalysts
run
complex
irrad. time/h
TNs
1
2
3
4
5
6
7
8
1
2
2
3
4
5
6
7
5
5
10
5
5
5
5
5
5
5
193
238
378
106
124
125
<1
<1
<1
<1
25
9
b
1
0
1
1
1
1
c
1
1
5
5
5
d
1
2
56
92
e
1
3
a
General conditions: The reaction was run in MeCN/water
-2
(
1
3:2 v/v) at pH = 8.5 in the presence of 1.6 ꢀ 10 M TEOA, 1.1 ꢀ
-5
-4
b
0
M photosensitizer C1, and 2.0 ꢀ 10 M cobaloximes. In DMSO/
c
d
e
water (3:2 v/v). In DMF/water (3:2 v/v). In MeOH/water (3:2 v/v). In
EtOH/water (3:2 v/v).
from a thermodynamic point of view because of differ-
ences in the media of the quoted potentials, the results
strongly suggest that complex 7 is not thermodynami-
cally able to act as a proton reduction catalyst at
2
+
this pH. An analogous system containing Ru(bpy)3
the photosensitizer with [Co(dmgBF ) (MeCN) ] as
as
2
2
2
the catalyst and ascorbic acid as the sacrificial donor at
pH = 2.0 was examined by us, and while it did yield
hydrogen, it did so in only very small amounts. A sli-
ghtly different result was obtained by Artero and co-work-
ers in nonaqueous media with a system composed of
Figure 3. Formation of Co(I) upon visible light irradiation of a solution
-5
-2
containing 1.1 ꢀ 10 M Pt chromophore C1, 1.6 ꢀ 10 M TEOA, and
-5
2
.0 ꢀ 10 M complex 7 ([Co(dmgBF
2 2 2 2
) (OH ) ]).
2
+
The generation of H was the result of the chromous ion
2
Ru(bpy)3 , Co(dmgBF ) , and [Et NH]Cl astheproton
2 2 3
reduction of 7 and thus became catalytic in Co(II).
Within the past decade, the chemistry of this complex
has been revisited for catalysis of H generation, and it
2
source that yielded 20 TNs of H2 formation under
UV irradiation but much less under visible light
irradiation.
3
3
has been found to have excellent activity for electroca-
talytic hydrogen production in nonaqueous media.
In all of the hydrogen production experiments that
yielded H , induction periods were observed. Typically,
2
8,30
2
Accordingly, we examined this complex in the present
photochemical system for hydrogen production (run 8).
Despite its advantages for electrochemical proton re-
hydrogen evolution commenced only after 1-2 h of
irradiation of the sample. The observed induction periods
are consistent with the notion that Co(III) must first be
converted into Co(II) and then Co(I) for the photoge-
neration of H to occur. This hypothesis is supported by
2
UV-vis spectroscopic monitoring of the reaction system
as a function of time.
Solvent variation has a substantial effect on the rate
and amount of hydrogen production. Smaller quantities
of hydrogen were produced in systems having a different
duction in terms of low overpotential for H evolution,
2
3
5
the complex did not function satisfactorily as a catalyst
in the current photochemical system for hydrogen pro-
duction. In this run, the system pH was set to 8.5, at
+
which value the H reduction potential is ca. -0.48 V,
which is negative relative to the Co(II)/Co(I) couple
(
-0.29 V vs NHE, Table 3). While not rigorously correct

4958 Inorg. Chem., Vol. 48, No. 11, 2009
Du et al.
Figure 5. Comparison of hydrogen production using different cobalox-
-
5
-2
imes with 1.1 ꢀ 10 M Pt chromophore C1, 1.6 ꢀ 10 M TEOA, and
-5
2
.0 ꢀ 10 M catalyst at pH = 8.5. (a) Complex 1; (b) complex 2;
(
c) complex 3.
5
2
Figure 4. Comparison of hydrogen production using different ratios of
M
TEOA, and 2.0 ꢀ 10 M complex 1 at pH = 8.5.
alkyl cobaloximes. Interestingly, however, these three
complexes, 1-3, exhibit different activities for hydrogen
production, as shown in Figure 5. Complex 2 with the
electron-withdrawing axial ligand (4-MeOOC-py) is more
active than complex 1 (Table 4, run 1), while 3 with the
-5
-2
MeCN/water with 1.1 ꢀ 10 M Pt chromophore C1, 1.6 ꢀ 10
-5
organic solvent, albeit with the same 3:2 v/v ratio
with water, for example, in DMSO (Table 4, run 10, no
appreciable H was seen within 5 h), in DMF (run 11, 25
more electron-donating pyridine (4-Me N-py; Table 4, run
2
) is less active than 1. With regard to the induction period,
2
4
TNs were obtained within 5 h), in MeOH (run 12, 56 TNs
within 5 h), and in EtOH (run 13, 92 TNs within 5 h). The
solvent dependence for hydrogen evolution from this
system probably results from a number of factors includ-
ing solvent polarity, stabilization of reduction intermedi-
ates, solvent coordination ability for binding to Co, the
the plot (Figure 5) shows that hydrogen generation com-
mences following the same order as the activity: 2 > 1 > 3.
This ordering is also in accord with modest changes in the
quenching rate constants (2 > 1 > 3), consistent with the
importance of electron transfer from the excited chromo-
phore to the Co catalyst despite the absence of a correla-
tion in the Co reduction potentials.
driving force dependence for the ultimate H generation
2
reaction (as discussed below, either the Co(II)/Co(I)
couple or reduction of a Co(III) hydride intermediate),
and catalyst stability.
In a study of the effects of axial pyridine substituents on
H production via electrocatalysis, Artero et al. found that
the largest rates of H production were obtained with
2
2
Further studies on medium effects were done to inves-
tigate the catalytic activity of the photochemical hydro-
gen reaction by changing the ratio of MeCN and water.
The results are shown in Figure 4, demonstrating that the
photochemical reaction has a higher rate of hydrogen
production when the ratio of MeCN/water is increased.
cobaloximes that possessed electron-donating axial pyr-
2
8
idines, in contrast with the present results. This differ-
ence in activity may be a consequence of the different
media and optimal pH ranges employed in the two studies.
Complexes 4 and 5 were also examined as the potential
hydrogen catalysts. Both complexes have the same D
2
h
-
When the MeCN/H O ratio is increased to 24:1 from 3:2,
2
symmetry with trans Cl ligands for 4 and trans pyridine
ligands for 5. These two complexes showed different
reactivity for hydrogen production, as shown in Figure 6.
While complex 4 did produce hydrogen after 1 h of
the rate of H generation essentially doubles, with H TNs
2
2
growing to 390 from 193 after 5 h of irradiation. The
induction time for producing hydrogen is also shortened
at higher MeCN concentrations, decreasing from 80 to 15
irradiation, the plot of H production versus time leveled
2
min when the MeCN/H O ratio changes from 3:2 to 24:1.
2
off after 4 h, indicating instability of the catalyst complex
during photolysis. For complex 5, hydrogen production
exhibited a linear relationship with the irradiation time
even after 6 h of photolysis, but 5 yielded a slower rate of
hydrogen production than 1, possibly because of its slow
oxidative quenching rate constant with photoexcited C1
The effect of the MeCN/H O ratio may be attributable to
2
5
1
differences in the reduction potentials in different media
or a change in solvent dielectric constant as the MeCN/
H O ratio changes (the solvent dielectric constants for
2
4
8
MeCN and H O are 36.6 and 80, respectively).
2
7
-1 -1
There is no significant trend in the potentials for the
Co(II)/Co(I) couple between complexes 1-3 in which
the axial pyridine ligand is modified in the 4- position
(
Table 3, k = 4.0 ꢀ 10 M
s ).
In the initial report of the photochemical production of
q
H using 1 as the hydrogen-generating catalyst, [Pt(ttpy)
2
to possess either an electron-donating group (-NMe ) as
(CtCPh)]ClO (C1) was employed as the light-harvesting
4
2
3
chromophore. In the present study, the related chro-
5
in 3 or an electron-withdrawing group (-COOMe) as in
2, an observation in accord with earlier work on
mophores C2-C4 (Scheme 3) were examined as to their
effectiveness under similar experimental conditions: λ >
(
51) Abe, R.; Sayama, K.; Sugihara, H. J. Sol. Energy Eng. 2005, 127
-5
4
10 nm; Pt(II) chromophore, 1.1 ꢀ 10 M; TEOA, 1.6 ꢀ
4
3
13–416.
-2 -4
10
M; Co complex 1, 2.0 ꢀ 10 M in MeCN/water
(
52) Schrauzer, G. N.; Windgassen, R. J. J. Am. Chem. Soc. 1966, 88,
738–3743.
(3:2 v/v), pH = 8.5. Figure 7A shows the time dependence

Article
Inorg. Chem., Vol. 48, No. 11, 2009
4959
of hydrogen production in this reaction system for the
different Pt chromophores. The plot demonstrates that
the rate of hydrogen generation follows the order C4 >
C2 > C3 > C1, which is dependent on both the absorp-
tivities of the chromophores and their quenching rate
constants with Co complex 1, as has been discussed
previously. The differences between the induction per-
iods in Figure 7 relate to the effectiveness of the reduction
of Co(III) to Co(II) and then to Co(I) before the onset of
likely be the first step toward H production. Oxidative
2
quenching was established in separate experiments invol-
ving Co complex 1 with the cyclometalated chromo-
phores C5-C7 and yielded near diffusion-controlled
9
-1 -1
quenching rate constants k of 3.6 ꢀ 10 M
s
, 2.9 ꢀ
q
9
10 M
-1 -1
9
-1 -1
s s , respectively. In the
presence of both a reductive quencher (TEOA) and
and 2.1 ꢀ 10 M
1
7
/\
/\
complex 1, visible light excitation of the C N N chro-
mophores generated a yellow color characteristic of Co
(II) species, as seen in similar experiments with the Pt(II)
terpyridyl acetylide chromophores and confirmed by
UV-vis spectroscopy.
H generation, as well as to instrument sensitivity at low
2
H levels.
2
Interestingly, H was also observed when Pt(II) terpyr-
2
idyl acetylide complexes were replaced by the cyclometa-
Examination of the influence of the TEOA concentra-
tion on the rate of H production was examined by the
2
runs shown in Figure 8. Only the concentration of TEOA
is changed between these runs, with all other concentra-
tions and conditions kept the same: λ > 410 nm; chro-
/
\
/\
lated Pt(II) acetylide complexes Pt(C N N-Ph-R)
(
(
CtCPh) where R = Me (C5), COOMe (C6), and P(O)
OEt) (C7) (Scheme 3). The experiments were run under
2
the same reaction conditions as used for C1-C4, and the
results are given in Figure 7B. The photochemical activity
of C5 is comparable with that of C1, yielding 170 TNs in 5
h. Within the same photolysis time, systems containing
C6 and C7 generate 90 TNs and 120 TNs, respectively,
which are less efficient than C5. While the luminescence of
C5-C7 is not readily quenched by the sacrificial donor
-
5
-4
mophore C1, 1.1 ꢀ 10 M; Co complex 1, 2.0 ꢀ 10
M
in MeCN/water (3:2 v/v), pH = 8.5. When the TEOA
concentration was 6.7 mM, photolysis led to only 63 TNs,
while in the presence of a TEOA concentration of
134 mM, 500 TNs were achieved with a yield of 21%
based on TEOA in 5 h. With a further increase in TEOA
concentration to 0.27 M, ∼1000 turnovers were obtained
(
TEOA), oxidative quenching of the excited states of C5-
C7 by the Co catalyst complexes indeed occurs, and this is
after 10 h of irradiation. Along with greater yields of H
2
generation as [TEOA] is increased, the induction period
for H generation was also reduced. An explanation for
2
increased turnovers of H and decreased induction peri-
2
ods with greater concentrations of TEOA is most likely
related to faster reductive quenching that also leads to a
more rapid rate of transferring the second electron from
+
•
the decomposing TEOA radical cation.
A further investigation of the effect of [TEOA] was also
conducted in the systems for which the ratio of MeCN/
H O was 24:1. As above, both the rate of hydrogen
2
production and the induction period were positively
affected by increasing TEOA concentrations. The H2
production system gave 390, 432, and 575 turnovers in 5
h as [TEOA] was varied from 16.1 to 47.5 and 134 mM,
respectively. With a TEOA concentration of 0.27 M,
∼
2150 turnovers were obtained after 10 h of irradiation,
giving an average rate of hydrogen production of ∼215
turnovers per hour.
The Mechanism of Cobalt Species in Hydrogen Produc-
tion. The mechanism of the visible-light-driven formation
of hydrogen in the system containing the Pt chromophore
C1 and the cobaloxime catalyst 1 has been briefly dis-
Figure 6. Comparison of hydrogen production using different cobalox-
-5
imes 1, 4, and 5 under the following conditions: 1.1 ꢀ 10 M Pt
-2
-5
chromophore C1, 1.6 ꢀ 10 M TEOA, and 2.0 ꢀ 10 M cobaloxime
3
5
at pH = 8.5.
cussed before. The production of hydrogen involves
-5
-2
Figure 7. Hydrogen production using different Pt(II) chromophores under the following conditions: 1.1ꢀ 10 M Pt chromophore, 1.6 ꢀ 10 M TEOA,
-5
and 2.0 ꢀ 10 M complex 1 at pH = 8.5. (A) Pt(II) terpyridyl acetylide chromophores; (B) cyclometalated Pt(II) acetylide chromophores.

4960 Inorg. Chem., Vol. 48, No. 11, 2009
Du et al.
-5
Figure 8. (A) Comparison of hydrogen production under different concentrations of TEOA (together with 1.1 ꢀ 10 M Pt chromophore C1 and 2.0 ꢀ
-
5
-5
10
M complex 1 at pH = 8.5). (B) Comparison of hydrogen production under a higher ratio of MeCN/water (v/v 24:1; together with 1.1 ꢀ 10 M Pt
-5
chromophore C1 and 2.0 ꢀ 10 M complex 1 at pH = 8.5).
Scheme 3
+
the Co(III) hydride reacts with H to give H2 and
Co(III), which is subsequently reduced. An alternative
single Co mechanism involves reduction of the Co(III)
two irreversible processes;decomposition of the TEOA
sacrificial donor and the release of hydrogen;and two
reversible cycles;a photochemical Pt cycle and a cataly-
tic Co cycle. The photochemical cycle can proceed along
two paths, one based on reductive quenching and
the other on oxidative quenching of the photoexcited
chromophore. In the reductive quenching pathway, the
excited chromophore C1* will be quenched by TEOA to
hydride to yield Co(II)-H that subsequently reacts with
+
H
able proposal for H formation that has been put forth
to form H + Co(II). A kinetically distinguish-
2
2
by others involves the reaction of two Co(III) hydrides to
give H + 2 Co(II).
2
8-31,50
2
-
+
+
form C1 and TEOA , after which TEOA loses a
proton and another electron on its decomposition path
to glycolaldehyde and di(ethanol)amine. Electron trans-
To address the kinetic dependence of the Co catalyst in
the photochemical system studied here, a series of photo-
lyses were run in which the cobalt catalyst concentration
[Co] was varied while all other reaction parameters were
held fixed and the TEOA concentration remained nearly
constant. Complex 1 was used as the hydrogen-evolving
catalyst for these experiments. The results are shown in
Figure 9. For values of [1] less than 0.7 mM, hydrogen
production after 5 h of photolysis exhibits a distinctly
linear dependence on [1]. However, further increases of
the cobalt concentration above 0.7 mM led to an essen-
-
fer from C1 to Co(III) occurs, to regenerate the ground-
state chromophore C1. For the oxidative quenching
process, the Co complex serves as an electron acceptor
to quench excited C1* to give C1 and Co(II). The TEOA
+
+
sacrificial donor will then be oxidized by C1 , leading to
regeneration of the ground state of C1.
The catalytic cobalt cycle begins with reduction of the
catalyst in one-electron steps to Co(I), followed by proto-
nation of the cobaloxime anion to give a Co(III) hydride
intermediate. Spectroscopic study of the system C1 + 1 +
tially constant rate of H generation, consistent with the
2
photochemical steps becoming turnover-limiting at these
values of [Co]. Similar experiments have been reported
recently by the groups of Alberto and Artero for a related
system in which the chromophore employed was ReBr
TEOA shows that, prior to H formation, stepwise reduc-
2
tion to Co(II) and Co(I) occurs (the latter is seen directly at
pH 12, at which only very small amounts of H are
2
3
4,36
generated). During photolysis, in which H is generated,
2
(CO) (bpy).
3
The latter shows a linear dependence on
Co concentration, while the former shows a quadratic
dependence.
3
5
the resting state of the system is Co(II). These obs-
ervations are entirely consistent with the previously
reported work by Espenson and Connolly, Artero and
5
0
For the single-Co mechanisms, the rate constant should
have a linear dependence on [Co], whereas for the two-Co
mechanism, the rate should vary as [Co] . The results in
2
8,29
30,31
co-workers,
and Peters and co-workers.
Subse-
2
quent reaction of the Co(III) hydride formed by pro-
tonation of Co(I) represents a branch point in the mechan-
istic proposals for the generation of H . In one proposal,
Figure 9 clearly show a linear dependence of rate with
[Co] at concentrations below 0.7 mM and thus support a
2

Article
Inorg. Chem., Vol. 48, No. 11, 2009
4961
Figure 9. Hydrogen production under different concentrations of co-
Figure 10. Mercury test for hydrogen production under the following
conditions: 1.1 ꢀ 10 M Pt chromophore C1, 1.6 ꢀ 10 M TEOA, and
-5
balt complex 1 (together with 1.1 ꢀ 10 M Pt chromophore C1 and 1.6 ꢀ
-5
-2
-2
10
M TEOA at pH = 8.5).
-4
2
.0 ꢀ 10 M complex 1 at pH = 8.5, (a) without mercury, (b) with 3 mL
of mercury, and (c) pretreated with mercury (3 mL) overnight, then
filtering off the mercury prior to photolysis.
single-Co mechanism for hydrogen formation in the
present photochemical system. While the two single-Co
mechanisms mentioned above are kinetically indistin-
guishable at this point, we favor the one involving Co
and that it is therefore important to rule out the existence
or formation of colloids in any photochemical system for
making hydrogen. Consequently, experiments were run
53
(
protonation because of favorable charge management
III) hydride reduction to Co(II) hydride followed by
to determine if colloidal cobalt was possibly formed and
acted as the real catalyst in the present photochemical
system. For these experiments, elemental mercury was
-
in this mechanism that results from alternate e and
H
+
additions.
Experimental observations in support of both single Co
used. Mercury is known to poison or inhibit the activity
of metal catalysts including colloids
54,55
and forms an
and two Co mechanisms for hydrogen generation have
been presented in the literature. Artero et al. have sug-
gested the single-Co mechanism in which the Co(III)
56
amalgam with cobalt. Two photolyses were done in
addition to a parallel run of the system described here
-5
-2
+
(1.1 ꢀ 10 M chromophore C1; 1.6 ꢀ 10 M TEOA;
hydride reacts with H to form H in the nonaqueous
2
-4
2
.0 ꢀ 10 M Co(III) complex 1 in MeCN/water (3:2 v/v)
electrochemical system using [Co(dmgBF ) (OH ) ] as
2 2
2
2
at pH = 8.5). In one, elemental mercury was added to the
C1 + 1 + TEOA reaction system (∼3 mL with vigorous
stirring) throughout the course of photolysis. In the
second, the same system was stirred over Hg for 12 h,
after which the mercury was removed and photolysis was
started. The results are shown in Figure 10. The samples
with and without Hg throughout the course of the
photolysis gave nearly identical results, while the sample
pretreated with mercury before photolysis yielded a
slightly lower value, possibly because of minor losses
of the starting solution during filtration. Overall, the
results are consistent with the absence of any effect of
the added mercury and thus support the notion that in
this system the cobaloxime complexes are indeed the
active catalysts.
the catalyst with the weak acids p-cyanoanilinium and
[
+
30
Et NH] as the proton source, while Peters et al. favor
3
the two-Co mechanism for the same catalyst with stron-
ger acids such as HBF , TsOH, CF COOH, and AcOH
4
3
3
6
based on electrochemical modeling. Alberto et al. also
favor the two-Co mechanism for their nonaqueous
photochemical system in which strong acid is used as
the proton source. In fact, both single-Co and two-Co
mechanisms may be operating in all of these systems to
differing respects, depending on conditions.
A distinction for the photochemical system reported
here is that water is an integral part of the system and
photochemical H evolution is most active under weakly
2
basic conditions, maximizing at pH 8.5. The protonation
of Co(I) to form the Co(III) hydride at this pH requires
that the Co(I) species be moderately basic. This require-
ment does not exist for the systems in which strong acid is
used as the proton source. While most of the work using
Conclusion. The cobaloxime complexes 1-7 act as oxi-
dative quenchers of photoexcited Pt chromophore C1.
Quenching studies and Stern-Volmer analysis re-
veal that all of the complexes except for 5 exhibit near
diffusion-controlled rates of quenching; for 5, the quench-
Co molecular catalysts for H generation has emphasized
2
the low overpotential for hydrogen evolution and the
consequent low potential for the reduction of Co(II) to
Co(I), in the present system, the basicity of the Co(I)
species is important and may explain why the difluoro-
borylated complex 7 does not function as a catalyst in the
photochemical system reported here.
We have previously written that purported molecular
catalysts for hydrogen evolution actually served as col-
loidal precursors through photochemical decomposition,
7
-1 -1
ing rate constant is slower at 4.0 ꢀ 10 M s . In a system
composed of a Pt(II) terpyridyl acetylide or cyclometa-
(53) Du, P.; Schneider, J.; Li, F.; Zhao, W.; Patel, U.; Castellano, F. N.;
Eisenberg, R. J. Am. Chem. Soc. 2008, 130, 5056–5058.
(54) Anton, D. R.; Crabtree, R. H. O. Organometallics 1983, 2, 855–859.
(
55) Paul, F.; Patt, J.; Hartwig, J. F. J. Am. Chem. Soc. 1994, 116, 5969–
970.
56) Paklepa, P.; Woroniecki, J.; Wrona, P. K. J. Electroanal. Chem. 2001,
498, 181–191.
5
(

4
962 Inorg. Chem., Vol. 48, No. 11, 2009
Du et al.
+
and subsequent reaction with H to give H with regen-
lated phenyl-bipyridyl acetylide chromophore, TEOA,
and complexes 1-5, the photogeneration of H2 is
observed. All of the chromophores C1-C7 work for
photoinduced hydrogen production. While most of the
studies of hydrogen photogeneration were done with
chromophore C1 and Co catalyst 1, comparative runs
show that chromophore C4 and cobaloxime catalyst 2 are
2
eration of the Co(II) catalyst. Ironically, [Co(dmgBF2)2-
(OH ) ] (7), which is most effective in the electrocatalytic
2
2
+
reduction of H to H , does not work in the present
2
system, in part due to the Co(II)/Co(I) reduction poten-
tial, which is too positive to achieve proton reduction at
pH 7-8.5.
most effective. The rate of H evolution is dependent on
2
solution pH, TEOA concentration, and the ratio of
MeCN/H O with more than 2100 turnovers obtained in
Acknowledgment. We wish to thank the Division of Basic
Sciences, U.S. Department of Energy for financial support
of this research (DE-FG02-90ER14125). P.D. gratefully
acknowledges an Arnold Weissberger Fellowship.
2
1
H O (24:1 v/v).
0 h of irradiation (λ > 410 nm) using C1 and 1 in MeCN/
2
Spectroscopic and mechanistic studies of the system
reveal an induction period to hydrogen evolution and an
initial color change to bright yellow characteristic of Co
Note Added after Print Publication. The print publica-
tion of June 1, 2009 (Vol. 48, issue 11), contained incorrect
data in Scheme 1 and Table 4 on pp 4952–4962. These were
corrected on the web version on August 31, 2009. An
Addition and Correction also appears in the August 31,
(
III) reduction to Co(II). Hydrogen is usually produced
after 1-2 h of photolysis with λ > 410 nm when [TEOA]
-
2
=
(
1.6 ꢀ 10 M. Mechanistically, further reduction to Co
I) occurs, followed by protonation to form a Co(III)
hydride intermediate. Variation of catalyst concentration
2009 (Vol. 48, issue 17), print issue.
reveals a linear relationship with H evolution that is
2
Supporting Information Available: X-ray crystallographic
consistent with a mechanism involving only a single
Co for H generation. We propose that the mechanism
proceeds by reduction of the Co(III) hydride intermediate
data of complexes 2 and 3 in CIF format. This material
is available free of charge via the Internet at http://pubs.
acs.org/.
2