Proton electroreduction catalyzed by cobaloximes: Functional models for hydrogenases

Article abstract of DOI:10.1021/ic050167z

Cobaloximes have been examined as electrocatalysts for proton reduction in nonaqueous solvent in the presence of triethylammonium chloride. [Co ^III(dmgH)₂pyCl], working at moderate potentials (-0.90 V/(Ag/AgCl/3 mol·L^-1 NaCl) and in neutral conditions, is a promising catalyst as compared to other first-row transition metal complexes which generally function at more negative potentials and/or at lower pH. More than 100 turnovers can be achieved during controlled-potential electrolysis without detectable degradation of the catalyst. Cyclic voltammograms simulation is consistent with a heterolytic catalytic mechanism and allowed us to extract related kinetic parameters. Introduction of an electron-donating (electron-withdrawing) substituent in the axial pyridine ligand significantly increases (decreases) the rate constant of the catalytic cycle determining step. This effect linearly correlates with the Hammet coefficients of the introduced substituents. The influence of the equatorial glyoxime ligand was also investigated and the capability of the stabilized BF₂-bridged species [Co(dmgBF₂)₂(OH₂)₂] for electrocatalyzed hydrogen evolution confirmed.

Full text of DOI:10.1021/ic050167z

Inorg. Chem. 2005, 44, 4786−4795
Proton Electroreduction Catalyzed by Cobaloximes: Functional Models
for Hydrogenases
Mathieu Razavet, Vincent Artero,* and Marc Fontecave
Laboratoire de Chimie et Biochimie des Centres Re´dox Biologiques, UMR 5047
CEA/CNRS/UniVersite´ Joseph Fourier, CEA-Grenoble, DRDC/CB, Bat K’,
17 rue des Martyrs, 38054 Grenoble Cedex 09, France
Received February 1, 2005
Cobaloximes have been examined as electrocatalysts for proton reduction in nonaqueous solvent in the presence
of triethylammonium chloride. [CoIII(dmgH)2pyCl], working at moderate potentials ( L-1 NaCl)
0.90 V/(Ag/AgCl/3 mol
−
‚
and in neutral conditions, is a promising catalyst as compared to other first-row transition metal complexes which
generally function at more negative potentials and/or at lower pH. More than 100 turnovers can be achieved during
controlled-potential electrolysis without detectable degradation of the catalyst. Cyclic voltammograms simulation is
consistent with a heterolytic catalytic mechanism and allowed us to extract related kinetic parameters. Introduction
of an electron-donating (electron-withdrawing) substituent in the axial pyridine ligand significantly increases (decreases)
the rate constant of the catalytic cycle determining step. This effect linearly correlates with the Hammet coefficients
of the introduced substituents. The influence of the equatorial glyoxime ligand was also investigated and the capability
of the stabilized BF2-bridged species [Co(dmgBF2)2(OH2)2] for electrocatalyzed hydrogen evolution confirmed.
Introduction
structure resolution of the hydrogenases enzymes which
revealed the composition of their active sites.4 Hydrogenases
are indeed the only molecular catalysts capable of catalyzing
both proton reduction and hydrogen oxidation with efficien-
cies comparable to platinum.5 These reactions proceed at
fascinating binuclear active sites based on iron only ([Fe]-
only H2ase) or both nickel and iron ([FeNi] H2ase) metal
ions coordinating cyanide (CN-) and carbon monoxide (CO)
ligands in a sulfur-rich environment (Figure 1).4 Intense work
was exquisitely and successfully done in the past few years
to synthesize biomimetic models of [Fe]-only H2ase retaining
the sulfur environment, the nature, and number of metal ions
as well as the diatomic carbonyl and cyanide ligands found
at the active site.6-12 The activities, displayed by some of
Because its combustion in fuel cells does not exhaust
greenhouse gas, hydrogen is expected to become a major
energetic vector in the next century.1 However, its production
is currently not sufficient to sustain the world economy and
is mainly achieved through nonecological fossil fuel reform-
ing processes. Alternatively, biomass pyrolysis and water
splitting, in thermolytic, electrolytic, or photolytic pathways,
can afford hydrogen as a carbon-neutral fuel, although this
is not yet cost-effective. Regarding water electrolysis,
whatever the power supplied, two technologies for hydrogen
production are currently economically viable on a small
scale: alkaline electrolyzers using nickel cathodes and PEM
devices using unsustainable noble metals such as platinum
as the electrocatalyst. The latter is the most promising for
various reasons, but it suffers from the costs of both the
membrane and the noble metal, which cannot be reduced at
the moment.2 The development of new catalysts for the
hydrogen evolution reaction (her) based on cheap first-row
transition metals has been a long-term goal for inorganic
chemists.3 This was recently boosted up by the single-crystal
(3) (a) Artero, V.; Fontecave, M. Coord. Chem. ReV., published online
Feb 23, 2005, 10.1016/j.ccr.2005.01.014. (b) Ko¨lle, U. New. J. Chem.
1992, 16, 157.
(4) [Fe]-only hydrogenases: (a) Peters, J. W.; Lanzilotta, W. N.; Lemon,
B. J.; Seefeldt, L. C. Science 1998, 282, 1853. (b) Nicolet, Y.; Piras,
C.; Legrand, P.; Hatchikian, C. E.; Fontecilla-Camps, J. C. Structure
1999, 7, 13. [NiFe] hydrogenases: (c) Volbeda, A.; Charon, M. H.;
Piras, C.; Hatchikian, E. C.; Frey, M.; Fontecilla-Camps, J. C. Nature
1995, 373, 580. (d) Volbeda, A.; Garcin, E.; Piras, C.; de Lacey, A.
L.; Fernandez, V. M.; Hatchikian, E. C.; Frey, M.; Fontecilla-Camps,
J. C. J. Am. Chem. Soc. 1996, 118, 12989. (e) Higuchi, Y.; Yagi, T.;
Yasuoka, N. Structure 1997, 5, 1671. (f) Higuchi, Y.; Ogata, H.; Miki,
K.; Yasuoka, N.; Yagi, T. Structure 1999, 7, 549.
* Author to whom correspondence should be addressed. E-mail:
vartero@cea.fr. Phone: int + 4 38 78 91 06. Fax: int + 4 38 78 91 24.
(1) Towards A Hydrogen Economy (special issue). Science 2004, 305,
13 (Aug).
(5) Jones, A. K.; Sillery, E.; Albracht, S. P. J.; Armstrong, F. A. Chem.
Commun. 2002, 866.
(2) Turner, J. A. Science 2004, 305, 972.
4786 Inorganic Chemistry, Vol. 44, No. 13, 2005
10.1021/ic050167z CCC: $30.25
© 2005 American Chemical Society
Published on Web 06/02/2005

Functional Models for Hydrogenases
Figure 1. Structures of the active sites of [Fe]-only (left) and [NiFe] (right)
hydrogenases. (X are putative ligands.)
these compounds,6c,d,7c,g,i-k,8h,i,9 are noteworthy but still lie
below those obtained with former nonbiomimetic catalysts
such as [Cp*Rh(bipy)(OH2)]2+,13 [Ni(biscyclam)]4+,14 co-
baloximes,15 cobalt porphyrins,16 and related complexes.17
Figure 2. [Co(dmgH)2(py)Cl].
Moreover, biomimics of [Fe]-only H2ase sometimes suffer
from their instability8h since they generally lack chelating
ligands which contribute to the stability of the coordination
sphere during catalysis.3 On the other hand, the fine com-
bination of soft ligands found at the active site probably helps
making the iron center nucleophilic and easily reducible,
stabilizing an open coordination site for substrate binding.
Some of these features can be found in cobaloximes, which
have proved to be interesting electrocatalysts for her in terms
of both cost and working potential.
Cobaloximes (Figure 2) are pseudomacrocyclic bis(dialkyl-
glyoximato)cobalt complexes. They were first developed to
mimic vitamin B12 and related compounds.18 In the reduced
Co(I) state, they are known as powerful nucleophiles. They
can be protonated to yield cobalt(III)-hydride species, which
can further evolve dihydrogen through either protonation of
the hydride moiety or bimolecular reductive elimination.19
However, their catalytic activity for her was only described
twice, first using divalent metal salts as bulk reducing agents
in strongly acidic media15a and second in nonaqueous solvent
in combination with [Ru(bipy)3]2+ as a photosensitizer.15b
We report here a systematic study of proton electroreduction
catalyzed by cobaloximes at glassy carbon electrodes.
(6) (a) Schmidt, M.; Contakes, S. M.; Rauchfuss, T. B. J. Am. Chem.
Soc. 1999, 121, 9736. (b) Gloaguen, F.; Lawrence, J. D.; Schmidt,
M.; Wilson, S. R.; Rauchfuss, T. B. J. Am. Chem. Soc. 2001, 123,
12518. (c) Gloaguen, F.; Lawrence, J. D.; Rauchfuss, T. B. J. Am.
Chem. Soc. 2001, 123, 9476. (d) Gloaguen, F.; Lawrence, J. D.;
Rauchfuss, T. B.; Be´nard, M.; Rohmer, M.-M. Inorg. Chem. 2002,
41, 6573. (e) Lawrence, J. D.; Li, H.; Rauchfuss, T. B.; Be´nard, M.;
Rohmer, M.-M. Angew. Chem., Int. Ed. 2001, 40, 1768. (f) Lawrence,
J. D.; Li, H.; Rauchfuss, T. B. Chem. Commun. 2001, 1482. (g) Li,
H.; Rauchfuss, T. B. J. Am. Chem. Soc. 2002, 124, 726. (h) Lawrence,
J. D.; Rauchfuss, T. B.; Wilson, S. R. Inorg. Chem. 2002, 41, 6193.
(i) Boyle, C. A.; Rauchfuss, T. B.; Wilson, S. R.; Rohmer, M.-M.;
Be´nard, M. J. Am. Chem. Soc. 2004, 126, 15151. (j) Rauchfuss, T. B.
Inorg. Chem. 2004, 43, 14.
(7) (a) Lyon, E. J.; Georgakaki, I. P.; Reibenspies, J. H.; Darensbourg,
M. Y. Angew. Chem., Int. Ed. 1999, 38, 3178. (b) Lyon, E. J.;
Georgakaki, I. P.; Reibenspies, J. H.; Darensbourg, M. Y. J. Am. Chem.
Soc. 2001, 123, 3268. (c) Zhao, X.; Georgakaki, I. P.; Miller, M. L.;
Yarbrough, J. C.; Darensbourg, M. Y. J. Am. Chem. Soc. 2001, 123,
9710. (d) Darensbourg, M. Y.; Lyon, E. J.; Zhao, X.; Georgakaki, I.
P. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3683. (e) Georgakaki, I.
P.; Thomson, L. M.; Lyon, E. J.; Hall, M. B.; Darensbourg, M. Y.
Coord. Chem. ReV. 2003, 238-239, 255. (f) Zhao, X.; Georgakaki, I.
P.; Miller, M. L.; Yarbrough, J. C.; Darensbourg, M. Y. J. Am. Chem.
Soc. 2001, 123, 9710. (g) Zhao, X.; Georgakaki, I. P.; Miller, M. L.;
Mejia-Rodriguez, R.; Chiang, C.-Y.; Darensbourg, M. Y. Inorg. Chem.
2002, 41, 3917. (h) Zhao, X.; Chiang, C.-Y.; Miller, M. L.; Rampersad,
M. V.; Darensbourg, M. Y. J. Am. Chem. Soc. 2004, 125, 518. (i)
Georgakaki, I. P.; Miller, M. L.; Darensbourg, M. Y. Inorg. Chem.
2003, 42, 2489. (j) Chong, D.; Georgakaki, I. P.; Mejia-Rodriguez,
R.; Sanabria-Chinchilla, J.; Soriaga, M. P.; Darensbourg, M. Y. Dalton
Trans. 2003, 4158. (k) Meija-Rodriguez, R.; Chong, D.; Reibenspies,
J. H.; Soriaga, M. P.; Darensbourg, M. Y. J. Am. Chem. Soc. 2004,
126, 12004.
(8) (a) Le Cloirec, A.; Best, S. P.; Borg, S.; Davies, S. C.; Evans, D. J.;
Hughes, D. L.; Pickett, C. J. Chem. Commun. 1999, 2285. (b) Razavet,
M.; Le Cloirec, A.; Davies, S. C.; Hughes, D. L.; Pickett, C. J. J.
Chem. Soc., Dalton Trans. 2001, 3551. (c) Razavet, M.; Davies, S.
C.; Hughes, D. L.; Pickett, C. J. Chem. Commun. 2001, 847. (d)
Razavet, M.; Borg, S. J.; Georges, S. J.; Best, S. P.; Fairhurst, S. A.;
Pickett, C. J. Chem. Commun. 2002, 700. (e) Georges, S. J.; Cui, Z.;
Razavet, M.; Pickett, C. J. Chem.sEur. J. 2002, 8, 4037. (f) Yang,
X.; Razavet, M.; Wang, X.-B.; Pickett, C. J.; Wang, L.-S. J. Phys.
Chem. A 2003, 107, 4612. (g) Evans, D. J.; Pickett, C. J. Chem. Soc.
ReV. 2003, 32, 268. (h) Borg, S. J.; Behrsing, T.; Best, S. P.; Razavet,
M.; Liu, X.; Pickett, C. J. J. Am. Chem. Soc. 2004, 126, 16988. (i)
Tard, C.; Liu, X.; Hughes, D. L.; Pickett, C. J. Chem. Commun. 2005,
133. (j) Zampella, G.; Bruschi, M.; Fantucci, P.; Razavet, M.; Pickett,
C. J.; De Gioia, L. Chem.-Eur. J. 2005, 11, 509.
Experimental Section
Materials. Commercial EPR-grade DMF and electronic-grade
1,2-dichloroethane were degassed by bubbling nitrogen. Tetrahy-
drofuran was distilled from sodium-benzophenone and kept under
argon. Cobalt chloride hexahydrate (Acros), cobalt acetate tetrahy-
drate (Strem chemicals), glyoxime (gH2) (Lancaster), dimethyl-
glyoxime (dmgH2) (Acros), diphenylglyoxime (dpgH2) (Aldrich),
4-(dimethylamino)pyridine (Aldrich), triethylammonium chloride
(Acros), triethylamine (Acros), and tetrafluoroboric acid diethyl
ether complex (Fluka) were used as received. The supporting elec-
trolyte (n-Bu4N)BF4 was prepared from (n-Bu4N)HSO4 (Aldrich)
and NaBF4 (Aldrich) and dried overnight at 80 °C under vacuum.
[Co(dmgH)2(py)Cl],20 [Co(dmgH)2(OH2)2],20 [Co(dmgH)2(py)2]-
(PF6),21 [Co(dmgBF2)2(OH2)2],22 [Co(gH)2(py)Cl],23 [Co(dmgH)2-
(PBu3)Cl],23 [Co(dpgH)2(py)Cl],20 isonicotinoyl chloride hydro-
chloride,24 and (Et3NH)BF425 were prepared as previously described.
(9) Ott, S.; Kritikos, M.; Ackermark, B.; Sun, L.; Lomoth, R. Angew.
Chem., Int. Ed. 2004, 43, 1006.
(10) Nehring, J. L.; Heinekey, D. M. Inorg. Chem. 2003, 42, 4288.
(11) Das, P.; Capon, J.-F.; Gloagen, F.; Pe´tillon, F. Y.; Schollhammer, P.;
Talarmin, J. Inorg. Chem. 2004, 43, 8203.
(12) Song, L.-C.; Yang, Z.-Y.; Bian, H.-Z.; Hu, Q.-M. Organometallics
2004, 23, 3082.
(13) (a) Ko¨lle, U.; Gra¨tzel, M. Angew. Chem., Int. Ed. Engl. 1987, 26,
567. (b) Ko¨lle, U.; Kang, B.-S.; Infelta, P.; Comte, P.; Gra¨tzel, M.
Chem. Ber. 1989, 122, 1869. (c) Cosnier, S.; Deronzier, A.; Vla-
chopoulos, N. J. Chem. Soc., Chem. Commun. 1989, 1259.
(14) Collin, J.-P.; Jouaiti, A.; Sauvage, J.-P. Inorg. Chem. 1988, 27, 1986.
(15) (a) Connolly, P.; Espenson, J. H. Inorg. Chem. 1986, 25, 2684. (b)
Hawecker, J.; Lehn, J.-M.; Ziessel, R. NouV. J. Chim. 1983, 7, 271.
(16) (a) Kellett, R. M.; Spiro, T. G. Inorg. Chem. 1985, 24, 2373. (b) Kellett,
R. M.; Spiro, T. G. Inorg. Chem. 1985, 24, 2378.
(17) Fisher, B.; Eisenberg, R. J. Am. Chem. Soc. 1980, 102, 7363.
(18) Schrauzer, G. N. Acc. Chem. Res. 1968, 1, 97.
(19) Chao, T.-H.; Espenson, J. H. J. Am. Chem. Soc. 1978, 100, 129.
(20) Schrauzer, G. N. Inorg. Synth. 1968, 11, 61.
Inorganic Chemistry, Vol. 44, No. 13, 2005 4787

Razavet et al.
[Co(dmgH)2(py)I] was prepared as [Co(dmgH)2(py)Cl], starting
from a mixture of Co(OAc)2‚4H2O and 2 equiv of NaI instead of
CoCl2‚2H2O.
then modified to include the catalytic steps, and the new param-
eters were refined until one set of value could provide correct
simulation of the voltammograms recorded in the presence of
various amounts (1.5, 3, and 10 equiv) of added acid at 20, 100,
Methods and Instrumentation. 1H NMR spectra were recorded
at room temperature using CDCl3 or acetone-d6 in 5 mm o.d. tubes
with a 300 MHz Bruker AC300 spectrometer equipped with a QNP
probehead. Electronic spectra are recorded on a diode-array HP
845x UV-visible system. Elemental analyses were performed by
the Service Central d’Analyse of the CNRS (Vernaison, France).
All electrochemical measurements were carried under nitrogen in
a thermostated cell at 20 °C. A three-electrode cell consisting of a
glassy carbon (3 mm i.d.) or platinum (2 mm i.d.) disk working
electrode (Radiometer), an auxiliary platinum wire, and an
Ag/AgCl/3 mol‚L-1 NaCl reference electrode, abbreviated as
and 500 mV‚s-1
.
[Co(dmgH)2(4-(Me2N)C5H4N)Cl]. CoCl2‚6H2O (500 mg, 2.15
mmol), dimethylglyoxime (551 mg, 4.70 mmol), and NaOH (86.0
mg, 2.15 mmol) were dissolved in 95% ethanol (20 mL) and heated
to 70 °C. 4-(Dimethylamino)pyridine (263 mg, 2.15 mmol) was
then added and the resulting solution cooled to room temperature.
A stream of air was then passed through the solution for 30 min,
which caused precipitation of a brown solid. The suspension
was stirred for 1 h and filtered. The precipitate was successively
washed with water (5 mL), ethanol (2 × 5 mL), and diethyl ether
(3 × 5 mL). The product was then extracted with acetone. Re-
moval of the solvent from the extracts yielded pure [Co(dmgH)2-
(4-(Me2N)C5H4N)Cl] (465 mg, 1.04 mmol, 48%). Anal. Calcd for
C15H24N6ClCoO4: C, 40.32; H, 5.41; N, 18.81; Cl, 7.94; Co, 13.19.
Found: C, 41.77; H, 5.71; N, 17.74; Cl, 7.51; Co, 12.41. 1H NMR
(300 MHz, CDCl3): δ 18 (broad, OH), 7.62 (d, J ) 6.6 Hz, 2H,
py), 6.28 (d, J ) 6.6 Hz, 2H, py), 2.95 (s, 6H NMe2), 2.38 (s, 12H,
dmg2-).
4-(N-tert-Butylcarboxamido)pyridine. tert-Butylamine (293 µL,
2.80 mmol) was added to a solution of isonicotinoyl chloride
hydrochloride (500 mg, 2.80 mmol) and triethylamine (1.57 mL,
11.2 mmol) in THF (10 mL). The mixture was stirred overnight at
room temperature and filtered. The solvent was removed, and the
product was chromatographied on silica using MeOH/CH2Cl2
(1.2/98.8 (v/v)) as eluant. The product was recovered as a pure
white powder (210 mg, 1.18 mmol, 42%). Spectroscopic data are
similar to those previously described.30
[Co(dmgH)2(4-(t-BuNH(CdO)C5H4N)Cl]. CoCl2‚6H2O (278
mg, 1.17 mmol), dimethylglyoxime (298 mg, 2.57 mmol), and
NaOH (47.0 mg, 1.17 mmol) were dissolved in 95% ethanol (10
mL) and heated to 70 °C. 4-(N-tert-butylcarboxamido)pyridine (208
mg, 1.17 mmol) dissolved in 95% ethanol (3 mL) was then added
and the resulting solution cooled to room temperature. A stream
of air was then passed through the solution for 45 min, which caused
the precipitation of a brown solid. The suspension was stirred for
1 h and filtrated. The precipitate was successively washed with
water (5 mL), ethanol (2 × 5 mL), and diethyl ether (3 × 5 mL).
The product was then extracted with acetone. Removal of the
solvent from the extracts yielded pure [Co(dmgH)2(4-(t-BuNH-
(CdO))C5H4N)Cl] (306 mg, 0.61 mmol, 52%). Anal. Calcd for
C18H28N6ClCoO5: C, 42.99; H, 5.61; N, 16.71; Cl, 7.05; Co, 11.72.
Found: C, 43.33; H, 5.80; N, 17.57; Cl, 6.91; Co, 11.22. 1H NMR
(300 MHz, CDCl3): δ 18 (broad, OH), 8.35 (d, J ) 6.4 Hz, 2H,
py), 7.46 (d, J ) 6.4 Hz, 2H, py), 5.91 (s, 1H, NHt-Bu), 2.36 (s,
12H, dmg2-), 1.38 (s, 9H, CMe3).
(Ag/AgCl), was used. Using this reference electrode, the (Epa
+
Epc)/2 value for the Fc/Fc+ couple is measured in DMF (CH3CN
and 1,2-dichloroethane) at 0.55 V (0.45 and 0.47 V) on both
platinum or glassy carbon electrodes. The Fc+/Fc couple (E° )
0.400 V vs SHE)26 can be used to quote potentials to SHE, when
needed. Cyclic voltamograms were recorded on a EG&G PAR
273A instrument. Solution concentrations were ca. 1.0 mmol‚L-1
for the cobaloxime and 0.1 mol‚L-1 for the supporting electrolyte,
(n-Bu4N)BF4. Electrodes were polished on a MD-Nap polishing
pad with 1 µm monocrystalline diamond DP suspension and DP
lubricant blue (Struers). Additions of Et3NHCl were made by
syringe using a 50 mmol‚L-1 solution in DMF. Bulk electrolysis
and coulometry were carried out on a EG&G PAR 273A instrument
in 1,2-dichloroethane, using a 2 cm2 graphite cathode. The platinum-
grid + carbon-foam counter electrode was placed in a separated
compartment connected with a glass-frit and filled with a 0.3
mol‚L-1 solution of (n-Bu4N)BF4 in degassed 1,2-dichloroethane.
A degassed 1,2-dichloroethane solution (10 mL) containing 0.1
mol‚L-1 (n-Bu4N)BF4 and 0.2 mol‚L-1 (Et3NH)BF4 was first
electrolyzed at -0.90 V/(Ag/AgCl) until the current has dropped
below 1 mA. The catalyst (1.0 mmol dissolved in 500 µL of
degassed dichloroethane) was added, and electrolysis was then
performed at -0.90 V/(Ag/AgCl). Electrolysis cell constant was
determined to be 2 × 10-4 s-1 by performing bulk electrolysis of
ferrocene.27 The faradaic yield was periodically estimated from the
volumetric measure of the amount of evolved gas during the time
needed for a 10 C of charge to pass through the cell. Hydrogen
was tested for purity using a Delsi Nermag DN200 GC chromato-
graph equipped with a thermal conductivity detector (TCD) detector.
Cyclic voltammograms were simulated and best-fitted using
Digisim28 or DigiElch29 software. A first simulation was per-
formed for each catalyst in the absence of added acid. Parameters
were refined until voltammograms recorded at 20, 100, and 500
mV‚s-1 could be simulated using the same set of values. The
resulting parameters were kept unchanged for simulation of
voltammograms recorded in the presence of acid. The model was
Results
(21) Gerli, A.; Marzilli, L. G. Inorg. Chem. 1992, 31, 1152.
(22) Bakac, A.; Espenson, J. H. J. Am. Chem. Soc. 1984, 106, 5197.
(23) Toscane, P. J.; Swider, T. F.; Marzilli, L. G.; Bresciani-Pahor, N.;
Randaccio, L. Inorg. Chem. 1983, 22, 3416.
Electrocatalytic Activity of [Co(dmgH)2(py)Cl]. The
cyclic voltammogram of [CoIII(dmgH)2(py)Cl] recorded in
DMF exhibits one irreversible reduction of Co(III) to
Co(II) (Epc ) -0.61V/(Ag/AgCl)) followed by a reversible
Co(II)/Co(I) process at -0.98 V/(Ag/AgCl). Addition of
increasing amounts, i.e., 1.5, 3.0, and 10 equiv, of a weak
acid, Et3NH+ (pKa ) 10.75), triggers the appearance of a
new irreversible cathodic wave near the Co(II)/Co(I) response
(24) Christensen, J. B. Molecules 2001, 6, 47.
(25) Mohamed, K. S.; Padma, D. K. Indian J. Chem. 1988, 27A, 759.
(26) Koepp, H. M.; Wedt, H.; Strehlow, H. Z. Elektrochem. 1960, 64, 483.
(27) Pletcher, D. A first course in electrode processes; The Electrochemical
Consultancy: Hants, U.K., 1991.
(28) (a) Digisim 3.05 for Windows 95 distributed by the Bioanalytical Corp.
(b) Rudolph, M.; Reddy, D. P.; Feldberg, S. W. Anal. Chem. 1994,
66, 589A.
(29) (a) Rudolph, M. J. Electroanal. Chem. 2003, 543, 23. (c) Rudolph,
M. J. Electroanal. Chem. 2003, 558, 171. (c) Rudolph, M. J.
Electroanal. Chem. 2004, 571, 289.
(30) Bonnet, V.; Mongin, F.; Tre´court,; Que´guiner, G. J. Chem. Soc., Perkin
Trans. 1 2000, 4245.
4788 Inorganic Chemistry, Vol. 44, No. 13, 2005

Functional Models for Hydrogenases
Figure 3. Cyclic voltammograms of [Co(dmgH)2(py)Cl] (1.0 mmol‚L-1) in the presence of various amounts of Et3NHCl recorded in a DMF solution of
n-Bu4NBF4 (0.1 mol‚L-1) on a glassy carbon electrode at 100 mV‚s-1: (a) 0 equiv; (b) 1.5 equiv; (c) 3.0 equiv; (d) 10 equiv (potentials versus Ag/AgCl).
Figure 4. Coulometry for bulk electrolysis at -0.90 V/(Ag/AgCl) of a 1,2-dichloroethane solution of Et3NHBF4 (0.2 mol‚L-1) and n-Bu4NBF4 (0.1
mol‚L-1) at a graphite electrode in the presence of 1.0 mmol‚L-1 Key: [Co(dmgH)2(py)Cl] (bold); [Co(dmgBF2)(OH2)2] (plain); without catalyst (dashed).
Evolution of the current as a function of time during the same experiment in the case of [Co(dmgBF2)(OH2)2] is shown in the inset.
(Figure 3). The behavior of this new wave is typical of a
“total” catalysis, as defined by Save´ant et al.31 and observed
for [Fe(TPP)Cl] (TPP2- ) meso-tetraphenylporphyrinato)
under the same conditions.32 At low acid/catalyst concentra-
tion ratios, the catalytic wave lies at more positive potentials
than that of the Co(II)/Co(I) couple. The latter is still
observed as a shoulder when the voltammogram is recorded
at 100 mV‚s-1 (Figure 3) and is better defined at 20 mV‚s-1
(Figure 5). Increasing the acid/catalyst concentration ratio
raises the height of the new wave and shifts it to more
negative potentials while the Co(II)/Co(I) reversible wave
disappears. This catalytic wave can then be assigned to
proton electroreduction. Bulk electrolysis on a graphite
electrode at -0.90 V/(Ag/AgCl) of a 0.2 mol‚L-1 solution
of Et3NH(BF4) in 1,2-dichloroethane33 in the presence of 1.0
mmol‚L-1 [Co(dmgH)2(py)Cl] produces H2 gas, identified
by gas chromatography, with a 85-100% faradaic yield.
Electrolysis was complete after 2.5 h, corresponding to 100
( 3 turnovers (193 C after subtraction of the blank elec-
trolysis, Figure 4). No degradation of the catalyst is observed
as shown by UV-visible spectroscopy.34,35 The cathodic
current decreases exponentially during electrolysis with a
time constant identical with that of the electrolysis cell.
Proton reduction electrocatalysis is therefore mass transport-
controlled, and the measured turnover frequency (TOF)
depends on the geometry of the electrochemical cell which
can be improved a 100-fold by electrochemical engineering.36
When the electrolysis is performed under the same conditions
but in the absence of catalyst, the volume of evolved H2
does not exceed 1 mL.
Tuning Redox Features. The effect of varying the
electronic properties of the ligand on the catalytic activity
was investigated. The electrochemical features of various
(34) Reaction of [Co(dmgH)2py] with H2 in MeOH yields 2-aminobutan-
3-one oxime as a result of hydrogenation of dimethylglyoxime:
Sima´ndi, L. I.; Szevere´nyi, Z.; Budo´-Za´honyi, E. Inorg. Nucl. Chem.
Lett. 1975, 11, 773.
(35) At this point, proton electroreduction can be started again by adding
of HBF4‚Et2O (270 µL, 2.00 mmol) to the solution containing Et3N
after electrolysis of Et3NHBF4. This was repeated twice, yielding ∼300
TON without observation of catalyst degradation.
(31) (a) Andrieux, C. P.; Blocman, C.; Dumas-Bouchiat, J.-M.; M’Halla,
F.; Save´ant, J.-M. J. Electroanal. Chem. 1980, 113, 19. (b) Save´ant,
J.-M.; Su, K. B. J. Electroanal. Chem. 1984, 171, 341.
(32) Bhugun, I.; Lexa, D.; Save´ant, J.-M. J. Am. Chem. Soc. 1996, 118,
3982.
(33) If electrolysis is carried out in DMF, the solution turns green and the
current drops after one turnover.
(36) Bard, A. J. Anal. Chem. 1963, 35, 1125.
Inorganic Chemistry, Vol. 44, No. 13, 2005 4789

Razavet et al.
C after subtraction of the blank electrolysis, Figure 4). No
degradation of the catalyst is observed by UV-visible
spectroscopy during the reaction.38 As expected from the
cyclic voltametry experiments, the cathodic current is smaller
than that obtained in a similar experiment using [Co(dmgH)2-
(py)Cl] as catalyst. The current initially increases, reaches a
maximum, and then decreases exponentially with a time
constant lower than that of the electrolytic cell (see inset in
Figure 4), in agreement with her being controlled by turnover
frequency and not by mass transport. The induction time
probably corresponds to the formation of the active species
in the bulk. The high current value measured during
electrolysis is possibly due to higher amounts of the active
species accumulating in the bulk as evidenced by the brown
color of the solution after the induction time. This is in
contrast with electrolysis experiments using [Co(dmgH)2pyCl]
for which the active species react immediately nearby the
electrode with only a small part of the introduced complex
actually achieving catalysis at the same time. According to
the proposed mechanism for the reduction of HCl by divalent
metal salts catalyzed by [Co(dmgBF2)2(OH2)2],15 hydrogen
evolution should occur when electrolysis is performed at
-0.55 V/(Ag/AgCl) in DMF or 1,2-dichloroethane. In fact
it does not until the electrolysis potential is lowered to -0.90
V/(Ag/AgCl). This may imply that the reaction with a weak
acid such as Et3NHCl proceeds via another mechanism where
the Co(III)-hydride species formed at -0.55 V/(Ag/AgCl)
should be activated by subsequent reduction at more negative
potential. This system is currently under study.
Figure 5. Cyclic voltammograms of 1.0 mmol‚L-1 [Co(dmgH)2(L)Cl]
(top, L ) py, 20 mV‚s-1; bottom, L ) 4-Me2NC5H4N, 100 mV‚s-1) + 0.1
mol‚L-1 n-Bu4NBF4 in DMF at a glassy carbon electrode in the presence
of Et3NHCl (mmol‚L-1, 1.5 equiv).
Substitution at the axial positions would only be relevant
if the introduced ligand is retained in the reduced Co(II) and
Co(I) states. Whatever the scan rate used in the 20-2000
mV‚s-1 range, the Co(III)/Co(II) system indeed appears
always chemically irreversible, except for [Co(dmgH)2(py)2]-
(PF6) studied in pyridine as solvent: the reduction wave of
the Co(III) species is broad and flattened, and its position
depends on the polishing state of the electrode. The corre-
sponding anodic wave on the reverse scan is shifted by 700
mV, consistent with an ECE mechanism (see eqs 1-3 in
Scheme 1) involving an irreversible loss of an axial ligand
to yield a pentacoordinated Co(II) species.39 The cyclic
voltammograms of [Co(dmgH)2(py)X] (X ) Cl- or I-) and
[Co(dmgH)2(py)2]+ display Co(III)/Co(II) cathodic peaks at
different potentials while the Co(II)/Co(I) reversible features
(-0.98 V/(Ag/AgCl) and the Co(II)/Co(III) anodic peak (0.15
V/(Ag/AgCl) are found at the same potential indicating that
in the reduced state axial halide ligands are more labile than
pyridine. It should be noted that recombination of halide with
the pentacoordinated cobaloxime is not immediate after
cobaloxime derivatives containing different glyoxime ligands
and axial bases are reported in Table 1.
Substitution of hydrogen or phenyl for methyl at the
glyoxime ligands significantly increases the electrochemical
potentials of the Co(II)/Co(I) couple as expected for less
electron-donating ligands but also results in the loss of the
proton reduction electroactivity on the cyclic voltametry time
scale. The same effects are observed when substituting
{BF2}+ for proton bridges between the equatorial ligands:
the potential of the Co(II)/Co(I) couple in the cobalt(II)
species [Co(dmgBF2)2(OH2)2] (-0.55 V/(Ag/AgCl) in DMF)
is the highest in the series and thus the closest to that of
the H+/H2 couple. Yet, the cyclic voltammogram of [Co-
(dmgBF2)2(OH2)2] is not modified upon addition of Et3NHCl.
This cannot exclude the possibility for [Co(dmgBF2)2(OH2)2]
to be active toward Et3NH+ but may reflect that turnover
frequency is low on the cyclic voltametry time scale.37 Bulk
electrolysis on a graphite electrode at -0.90 V/(Ag/AgCl)
of a 0.2 mol‚L-1 solution of Et3NH(BF4) in 1,2-dichloroeth-
ane in the presence of 1.0 mmol‚L-1 [Co(dmgBF2)2(OH2)2]
indeed produces H2 gas, identified by gas chromatography,
with a 75-100% faradaic yield. Electrolysis is almost
complete after 17 h, corresponding to 80 ( 3 turnovers (155
(38) If the electrolysis experiment is stopped before complete consumption
of the acid, the color turns back to yellow and the final UV-visible
spectrum is identical with the initial one. However, if the electrolysis
experiment is continued after total depletion of the acid, the current
drops and a green color develops. This color can be due to an
alternative decomposition pathway of the reactive species, which is
not observed as long as there is enough acid in the solution.
(39) Ngameni et al. previously reported reversible behavior for the
Co(III)/Co(II) couple of [Co(dmgH)2PyI] in CH3CN: Ngameni, E.;
Ngoune, J.; Nassi, A.; Megnamisi Belombe, M.; Roux, R. Electrochim.
Acta 1996, 41, 2571. However, in our hand, this compound show
irreversible chemical behaviour regarding the Co(III)/Co(II) couple.
(37) The observation of an electrocatalytic property using cyclic voltametry
is unfortunately restricted to fairly active species, i.e., with a second-
order rate constant of more than approximately 100 mol‚L-1‚s-1 as
estimated by DuBois and co-workers: Curtis, C. J.; Miedaner, A.;
Ciancanelli, R.; Ellis, W. W.; Noll, B. C.; DuBois, M. R.; DuBois, D.
L. Inorg. Chem. 2003, 42, 216.
4790 Inorganic Chemistry, Vol. 44, No. 13, 2005

Functional Models for Hydrogenases
Table 1. Electrochemical Potentials versus Ag/AgCl (V) Measured in DMF on a Glassy Carbon Electrode for Cobaloximes (Scan Rate: 100 mV‚s-1
)
Epc(CoIII/CoII)
Epa(CoIII/CoII) (reverse scan)
(Epc + Epa)/2(CoII/CoI)
[CoIII(dmgH)2(py)Cl]
-0.59
-0.66
-0.65
-0.44
-0.24
-0.62
0a
0.15
0.06
0.15
0.15
0.15
0.24
0.30
0.40
b
-0.98
-1.01
-0.97
-0.98
-0.98
-0.81
-0.98
-0.55
-0.65
-0.70
[CoIII(dmgH)2(4-Me2NC5H4N)Cl]
[CoIII(dmgH)2(4-t-BuNHCOC5H4N)Cl]
[CoIII(dmgH)2(py)I]
[CoIII(dmgH)2(py)2](PF6)
[CoIII(dmgH)2(PBu3)Cl]
[CoII(dmgH)2(OH2)2]
[CoII(dmgBF2)2(OH2)2]
[CoIII(gH)2(py)Cl]
b
-0.40
[CoIII(dpgH)2(py)Cl]
-0.47c
0.17
a Second scan. b Not observed. c Unresolved in DMF but well defined in CH3CN.
Scheme 1. Mechanism Used for Simulation of the Cyclic
Voltammograms
reoxidation to the Co(III) state (reverse scan) since a new
reduction wave at -0.15 V/(Ag/AgCl), presumably that of
the pentacoordinated cationic [Co(dmgH)2(py)]+ species, can
be observed if a second scan is immediately performed
(see Supporting Information). This [Co(dmgH)2(py)]+/
[Co(dmgH)2(py)] couple appears quasi-reversible with a
peak-to-peak separation of 300 mV (Scheme 1, eq 3).
Substitution of tributylphosphine for pyridine in [Co-
(dmgH)2(py)Cl] increases the potential of the Co(II)/Co(I)
couple to -0.81 V/(Ag/AgCl), but this also results in the
loss of the electrocatalytic activity on the cyclic voltametry
time scale.
Substitution of 4-(dimethylamino)pyridine or 4-(N-tert-
butylamido)pyridine for pyridine in [Co(dmgH)2(py)Cl] does
not result in a significant modification of the potential for
the Co(II)/Co(I) couple as previously reported for alkyl
cobaloximes,40,45 but careful examination of the cyclic
voltammograms recorded at low acid/catalyst ratios reveals
that the more donating the axial ligand, the more active the
catalyst. For 1-1.5 equiv of acid added, and at a scan
rate of 100 mV‚s-1, the catalytic wave is well separated
from the Co(II)/Co(I) wave in the case of [Co(dmgH)2(4-
Me2NC5H4N)Cl] (Figure 5) while these waves are merged
for [Co(dmgH)2(py)Cl] (Figure 3) and could only be resolved
at lower scan rate such as 20 mV‚s-1 (Figure 5). As the
potential shift observed for an EC mechanism is corre-
lated with the reaction rate of the chemical reaction fol-
lowing electron transfer,31,41 this indicates an increased
catalytic activity (TOF) using 4-(dimethylamino)pyridine as
the axial ligand.
Figure 6. Heterolytic and homolytic pathways for proton reduction
catalyzed by a coordination compound.
Mechanism Determination. Proton reduction yields di-
hydrogen following two general pathways depending on
whether the H-H bond is formed via a homolytic or a
heterolytic step. The general mechanistic schemes3 are given
in Figure 6.
Both catalytic cycles include protonation of the metal
center to give a metal-hydride complex. In the homolytic
mechanism, H2 evolution results from a reductive elimination
reaction from two metal-hydride moieties. In the heterolytic
pathway, the metal hydride decomposes by proton attack to
evolve hydrogen, probably via an intermediate dihydrogen-
metal complex. The catalytic cycle is completed by one or
two monoelectronic reduction steps which can occur either
at the unprotonnated metal center or after formation of the
hydride species.
Simulation of the voltammograms obtained for co-
baloximes in the presence of triethylammonium chloride
(41) For an EC mechanism with reversible heterogeneous charge-transfer
E (n ) number of transferred electrons) and rapid chemical reaction
(rate constant k) on the time scale of cyclic voltametry (scan rate, V),
the potential shift is given by the equation Ep ) E1/2 - 0.780(RT/nF)
+ (RT/2nF) ln(kRT/nFV), so that the wave shifts toward positive
potentials by about 30/n mV (at 25 °C) for a 10-fold increase of k
(Bard, A., J.; Faulkner, L. R. Electrochemical Methods; Wiley: New
York, 1980; p 452).
(40) Rillema, D. P.; Endicott, J. F.; Papaconstantinou, E. Inorg. Chem. 1971,
10, 1739.
Inorganic Chemistry, Vol. 44, No. 13, 2005 4791

Razavet et al.
This peak is better defined in the simulations than in the
experimental data. We suppose that this is a consequence of
the modification of the surface state of the electrode upon
electrocatalysis, which is not taken into account in the model.
Indeed, this anodic peak is better resolved when the cyclic
voltammogram is recorded on a hanging mercury electrode
(data not shown). We have chosen height, position, and shape
of the catalytic wave as relevant criteria for the evaluation
of the simulation. Significant extracted data are given in
Table 2.
Charge-transfer parameters (E°, R, ks) and diffusion
coefficients corresponding to eqs 1, 3, and 4 and thermo-
dynamic (Keq2) and kinetic (kf2) parameters for eq 2 were
determined using voltammograms of cobaloximes without
any added acid and kept unchanged for further simulation
in the presence of acid. Heterogeneous charge-transfer
processes appear relatively slow with rate constants ks
ranging from 10-3 to 10-2 cm‚s-1. Equation 2 is a total
reaction with a first-order rate constant ranging from 105 to
109 s-1. Disproportionation reaction 7 was set as diffusion-
controlled. Its equilibrium constant was calculated from the
standard potentials of the corresponding redox couples.
Introduction of this reaction in our model led to a better,
even not optimal (see above), graphical reproduction of the
reverse scan of the electrocatalytic wave.
The value of the product Keq5Keq6 was estimated as 2 ×
109 using the method of Spiro et al.16 and fixed during
simulations. Simulation showed that both reactions 5 and 6
are fully shifted except in the case of the amido derivative
[Co(dmgH)2(4-t-BuNHC(dO)C5H4N)Cl] for which a Keq5
value of 400 was found, in accordance with the electron-
withdrawing property of the amido group.
Reaction 6 was found to be kinetically determinant. The
value of its rate constant kf6 determines the height of the
electrocatalytic wave while the reaction rate constant of eq
5 coupled with the heterogeneous charge-transfer rate
constant ks4 is responsible for the position of the electro-
catalytic wave. Only a lower limit for kf5 at 108 mol-1‚L‚s-1
can be determined since higher values do not modify the
simulated voltammogram. Ammonium ions being more
acidic than water, this value is in accordance with the lower
limit at 104 mol-1‚L‚s-1 determined by Spiro et al. for the
reaction rate of Co(I) porphyrin centers with water.16a
The same simulation and best fitting was undertaken for
[CoII(dmgH)2(OH2)2] with a new model in which eqs 1-3
were deleted and assuming rapid substitution of one water
molecule for the hydride ligand. Significantly lower values
for kf5 (104 mol-1‚L‚s-1) and kf6 (300 mol-1‚L‚s-1) were
obtained.
Figure 7. Cyclic voltammograms of [Co(dmgH)2(py)Cl] (top, 0.970
mmol‚L-1; bottom, 0.833 mmol‚L-1) recorded in DMF solution of n-Bu4-
NBF4 (0.1 mmol‚L-1) at a glassy carbon electrode in the presence of Et3-
NHCl (top, 1.46 mmol‚L-1, 1.5 equiv; bottom, 8.33 mmol‚L-1, 10 equiv).
Scan rate: 100 mV‚s-1. Left: experimental curves. Right: simulated curves.
allowed us to assign a heterolytic nature to the mechanism.
With the homolytic hypothesis, the simulated electrocatalytic
wave indeed displays a peculiar shape with a round toped
peak (see Supporting Information), whereas the classical,
experimentally observed, sharp pointed peak is observed for
the heterolytic model. Furthermore, using a model including
both mechanisms, a very low kinetic constant is obtained
for the homolytic step compared to the reaction rates of the
heterolytic steps, indicating that the contribution of the
homolytic mechanism is minor if any.
The relative order of the two one-electron transfer and two
protonation steps in the heterolytic mechanism can be
immediately determined by a simple examination of the
voltamogram. If the electrocatalytic wave develops at
potentials more negative than the reduction potential of the
catalyst, it implies that the hydride moiety formed by
protonation after reduction of the catalyst should still be
activated by further reduction. On the contrary, an electro-
catalytic wave developing at potential just above a reduction
wave of the catalyst attests for the two one-reduction steps
occurring before the two protonation steps. This is the case
for proton reduction catalyzed by cobaloximes. Such a
mechanism was also ascertained using cyclic voltammetry
by Save´ant et al. for proton electroreduction catalyzed by
iron32 and rhodium42 porphyrin complexes.
Selected parameters gained from simulations are reported
in Table 2.
We thus simulated the voltammograms using the mech-
anism given in Scheme 1 and were able to reproduce
the experimental data (Figure 7). As seen in Figure 7, the
major discrepancies between simulated and experimental
value concern the reverse scan, namely the anodic peak
related to the Co(II)/Co(I) couple at low acid concentrations.
Discussion
In this study, we have reinvestigated the potential of
cobaloximes for the hydrogen evolution reaction (her). Our
results definitively show that [Co(dmgH)2LX] (L ) pyridine
derivatives, X ) halide or labile ligands) are efficient and
robust electrocatalysts:
(42) Grass, V.; Lexa, D.; Save´ant, J.-M. J. Am. Chem. Soc. 1997, 119,
7526.
4792 Inorganic Chemistry, Vol. 44, No. 13, 2005

Functional Models for Hydrogenases
Table 2. Relevant Data Extracted from Simulation of the Voltammograms
catalyst
E°Co(II)/Co(I) (V/(Ag/AgCl))
kf5min (mol-1‚L‚s-1
)
Keq5
k
f6 (mol-1‚L‚s-1
)
kf2 (s-1
)
[CoIII(dmgH)2pyCl]
-0.98
-1.01
-0.97
-0.98
108
108
108
104
5 × 104
5 × 104
4 × 102
1.35 × 104
2.5 × 103
5 × 105
108
109
105
[CoIII(dmgH)2(4-Me2NC5H4N)Cl]
[CoIII(dmgH)2(4-t-BuNHC(dO)C5H4N)Cl]
[CoII(dmgH)2(OH2)2]
103
3 × 102
Table 3. Comparison of Reported Her-Catalyst Properties
catalyst
appl potential
electrode
exptl conditns
TON
ref
poly{pyrrole-[Cp*Rh(bipy)Cl]+} film
[CoPc]/poly(4-vinylpyridine-co-styrene) -0.90 V/(Ag/AgCl) basal-plane
-0.55 V/SCE
glassy carbon
pH 1-4 aq soln
353 (14 h)
13c
0.1 mol‚L-1 pH 1.0 phosphate buffer 2 × 105‚h-1 a 44
pyrolytic graphite
-0.90 V/(Ag/AgCl) graphite
[Co(dmgH)2pyCl]
1,2-C2H4Cl2-0.2 mol‚L-1 Et3NHCl
aq 0.1 mol‚L-1CF3COOH
CH3CN-50 mmol‚L-1 HClO4
pH 2 aq soln
100 (2.5 h)
0.65 (20 min) 16
this work
Co(TMAP),Co(TMPyP, CoTPyP
[Fe2(µ-ADT)(CO)6]b
[Ni(LN4)]+ c
[CpCo{P(OMe)3}2]
[Ru2(C6H3(CF3)2)2DPB]
[HFe2(µ-pdt)(CO)4(CN)(PMe3)]d
[Ni2(biscyclam)]4+
-0.95 V/SCE
-1.48 V/(Fc/Fc+)
-1.1 V/SCE
Hg pool
glassy carbon
glassy carbon
Hg pool
Hg pool
Hg pool
Hg pool
Hg pool
25 (10 min)
12.7 (12 h)
20 (18 h)
1.6 (20 min)
6 (15 min)
100
9
50
51
52
6c
14
32
-1.15 V/SCE
-1.2 V/Ag
pH 5 aq soln
THF + CF3COOH
-1.2 V/(Ag/AgCl)
-1.5 V/SCE
CH3CN-50 mmol‚L-1 H2SO4
0.2 mol‚L-1 pH 7 phosphate buffer
DMF-50 mmol‚L-1 Et3NHCl
[FeTPPCl]
-1.6 V/SCE
24 (1.25 h)
2-
-
a Only TOF is reported for this complex. b ADT2- ) p-Br-C6H4CH2N(CH2S)2
dithiolate.
.
c LN4 ) 2,6-(HN(CH2)3NdC(CH3))2C5H3N-. d pdt2- ) propane-1,3-
(i) They are active toward weak acids such as Et3NH+.
(ii) Their activation overpotentials are low. Proton elec-
troreduction catalyzed by [Co(dmgH)2pyCl] occurs at po-
tentials between -0.91 (1.5 equiv, [Et3NHCl] ) 1.46
mmol‚L-1) and -0.98 V/(Ag/AgCl) (10 equiv, [Et3NHCl]
) 8.30 mmol‚L-1) in DMF at a carbon electrode. Pro-
ton reduction of an Et3NHCl solution occurs at a freshly
polished platinum electrode in DMF at -0.88 V/(Ag/AgCl)
([Et3NHCl] ) 1.46 mmol‚L-1) and -0.95 V/(Ag/AgCl)
([Et3NHCl] ) 8.30 mmol‚L-1). This allows an estimation
of a 30 mV overpotential for her catalyzed by [Co(dmgH)2-
(py)Cl].43
derivative of the biological azadithiolate bridging ligand
[Fe2(µ-ADT)(CO)6] is an active her electrocatalyst working
at moderate potential (-1.48 V/(Fc/Fc+) corresponding to
-1.03 V/(Ag/AgCl)).9 However, [Fe2(µ-ADT)(CO)6]-cata-
lyzed her is studied in 50 mmol‚L-1 perchloric acid solutions
and reduction of protons under these conditions at a freshly
polished platinum electrode occurs in our hands at -0.67
V/(Fc/Fc+), which unfortunately still reveals an important
(810 mV) activation overpotential. Cobalt porphyrins may
be considered as good catalysts for H2 evolution especially
in neutral aqueous solution because of their high working
potential and nucleophilicity, but their utilization is precluded
by their tendency to adsorb on all electrode materials with
subsequent loss of electrocatalytic properties.16 Cobalt(II)
phthalocyanine incorporated in a poly(4-vinylpyridine-co-
styrene) film catalyzes proton electroreduction at -0.90
V/(Ag/AgCl) but in quite acidic conditions (pH 1).44 The
very high reported turnover frequency as compared to other
electrocatalysts studied in the bulk is probably due to the
fact that cobalt(II) phthalocyanine is grafted at the electrode.
Starting from the promising and simple catalyst [Co-
(dmgH)2pyCl], we then tried to improve it regarding both
working potential and stability.
Increasing the electrochemical potential substantially
remains essential in the perspective of an aqueous application,
through immobilization of catalysts at an electrode surface,
where the -0.413 V/SHE value for the apparent potential
of the H+/H2 couple at pH 7 would apply. Some modifica-
tions of the coordination sphere of the cobalt ion reported
here resulted in increased potential of the Co(II)/Co(I) wave
(Table 1) but also in the parallel loss of the electrocatalytic
wave. In fact, species with more positive reduction potentials
are also less nucleophilic and thus exhibit significant catalytic
activity only under strongly acidic conditions. As an example,
330 mV is gained in reduction potential when substituting
(iii) High turnover numbers could be obtained without
any detectable decomposition of the catalyst as shown by
UV-visible spectroscopy.
Moreover, the catalytic mechanism was elucidated using
simulation of cyclic voltammograms. Hydrogen evolution
proceeds via a heterolytic pathway with the key step
consisting in the fairly rapid protonation of a cobalt(III)-
hydride moiety (Table 2) as compared to the similar reaction
of [HFeII(TPP)]- with Et3NH+ (k ) 4 105 mol-1‚L‚s-1).32
Comparison of a given her catalyst with others always
proves difficult because the assay conditions described in
the literature are different. Nevertheless cobaloximes appear
superior to other first-row transition metal her catalysts which
function at more negative potentials and/or at lower pH
(Table 3). The rhodium and therefore unsustainable complex
[Cp*Rh(bipy)Cl]+,13 active at the Rh(I) oxidation state, still
works at a less negative potential. It is interesting to note
that the [Fe]-only hydrogenase model compound bearing a
(43) The apparent pH in nonaqueous media is not well defined, and proton
reduction electrocatalysis is most probably a general rather than a
specific acid catalysis; i.e., the rate of the overall reaction depends
not only on the apparent pH but also on the pKa of the acid used in
the assay. Thus, we believe that the potential of electrocatalyzed proton
reduction should not be compared to the apparent potential for the
H+/H2 couple at the given pH but rather to the electrocatalytic potential
obtained under the same conditions on a platinum electrode, where
the H+/H2 couple is likely to display the most Nerstian behavior.
(44) Zhao, F.; Zhang, J.; Abe, T.; Wo¨hrle, D.; Kanedo, M. J. Mol. Catal.
A 1999, 145, 245.
Inorganic Chemistry, Vol. 44, No. 13, 2005 4793

Razavet et al.
glyoximato for dimethylglyoximato in [CoII(dmgH)2(py)], but
[Co(gH)2(PBu3)]- has been reported to be approximatively
600 times less nucleophilic than [Co(dmgH)2(PBu3)]-.23
The remarkable stability of [Co(dmgBF2)2(OH2)2]45 com-
pared to [Co(dmgH)2pyCl] is worth being considered in the
perspective of higher turnover number applications: in
contrast with other proton-linked cobaloximes which de-
compose in a matter of minutes under acidic conditions, the
half-life of [Co(dmgBF2)2(OH2)2] in 0.05 mol‚L-1 H+ is 5.5
h.45 Replacing protons with {BF2}+ linking groups in
cobaloximes thus dramatically increases their stability toward
hydrolysis.22 Despite the absence of electrocatalytic wave
in cyclic voltametry experiments, bulk electrolysis ex-
periments carried out under the same conditions as for
[Co(dmgH)2pyCl] amazingly revealed promising activity for
her. [Co(dmgBF2)2(OH2)2] therefore appears to be a good
starting point for the design of acid-resistant cobaloxime-
based electrocatalyst, and its solubility in water opens new
and tantalizing fields of investigation, currently under study.
Bulk electrolysis experiments, using [Co(gH)2pyCl] or
[Co(dmgH)2(PBu3)Cl] as catalysts, revealed activities lower
than or comparable to that of [Co(dmgBF2)2(OH2)2]. This is
in accordance with former studies reporting protonation of
[Co(dmgH)2(PBu3)]- to yield [Co(dmgH)2(PBu3)H],46 which
evolves dihydrogen by parallel heterolytic and homolytic
mechanisms.19 With comparable activity but greater instabil-
ity with respect to [Co(dmgBF2)2(OH2)2], these species are
therefore less attractive as electrocatalysts for her.
Although it seems difficult to increase the potential of
cobaloximes without lowering their catalytic activity, the
nucleophilicity of the cobalt(I) center can be significantly
enhanced through substitution at the pyridine ligand and this
without dramatically decreasing the potential of the Co(II)/
Co(I) couple (Table 1). Introduction of the electron-donating
dimethylamino substituent in para position of the axial
pyridine increases the rate constant of the reaction 6 40-
fold. The electron-withdrawing N-tert-butylamido sub-
stituent decreases this rate constant 5-fold. This effect can
be linearly correlated to the Hammett coefficients which
reflect the electronic properties of the substituents (Figure
8).47 The resulting linear free enthalpy of activation corre-
lation (Hammett equation) has a positive slope, in accord-
ance with a nucleophilic attack of the cobalt(III)-hydride
moiety toward protons (Scheme 1, eq 6). However, no effect
of the axial ligand is observed in bulk electrolysis or rotating-
disk experiments using [Co(dmgH)2(L)Cl] (L ) py, 4-t-
BuNHCOC5H4N and 4-Me2NC5H4N) as catalysts probably
because the current response is mass transport-controlled.
Nevertheless, technological applications for H2 production
with optimized mass transport should benefit from the axial
ligand influence on the turnover frequency.
Figure 8. Logarithmic plot and linear correlation of the rate constant of
eq 6 versus Hammet coefficients of the para substituent of the axial pyridine
ligand L in [Co(dmgH)2(L)Cl].
[Ru(bipy)3]2+ and triethanolamine as the reducing source.15b
Degradation (hydrogenation) of the catalyst was observed,
and a 6-fold excess of dmgH was systematically added to
maintain an active complex. Typical experiments were
conducted in DMF and the pH (optimal value 8.9) was
adjusted through addition of acetic acid or dissolution of
carbon dioxide. The system was described as highly active
but insufficiently stable in a long term use. Surprisingly,
[Co(dmgH)2(OH2)2] was the less active cobaloximes of our
study. Although cobaloximes bearing a pyridine as the axial
base are described to be more sensitive to hydrogenation of
the ligand,34 we did not observe such a degradation when
carrying out electrolysis in 1,2-dichloroethane or propylene
carbonate. We hence believe that it should be worth
reinvestigating such systems for hydrogen photoproduction,
either using separated catalysts and photosensitizers48 or
designed bifonctional complexes.49
Conclusion
Cobaloxime complexes are valuable catalysts for proton
electroreduction at moderate working potentials and in
neutral conditions. Readily available from a one-step syn-
thesis starting from classical chemicals of little value, they
can be handled in air in the Co(III) state. Drawbacks could
reside in their instability toward acid (pH < 5) even if highly
acidic conditions are not a specification for her applications.
Their activity and stability can be increased by modification
of the axial or equatorial substituents, respectively. We are
currently investigating the grafting of these species at graphite
or glassy carbon electrode to develop heterogeneous aqueous
applications.
Acknowledgment. The authors thank Eric Saint-Aman
(LEOPR, University Joseph Fourier, Grenoble, France) and
(48) Kirch, M.; Lehn, J.-M.; Sauvage, J.-P. HelV. Chim. Acta 1979, 62,
1345.
(49) (a) Komatsuzaki, N.; Himeda, Y.; Hirose, T.; Sugihara, H.; Kasuga,
K. Bull. Chem. Soc. Jpn. 1999, 72, 725. (b) Balzani, V.; Juris, A.;
Venturi, M. Chem. ReV. 1996, 96, 759.
(50) Efros, L. L.; Thorp, H. H.; Brudvig, G. W.; Crabtree, R. H. Inorg.
Chem. 1992, 31, 1722.
(51) Ko¨lle, U.; Ohst, S. Inorg. Chem. 1986, 25, 2689.
(52) Collman, J. P.; Ha, Y.; Wagenknecht, P. S.; Lopez, M.-A.; Guilard,
R. J. Am. Chem. Soc. 1993, 115, 9080.
In another catalytic application, Lehn et al. used
[Co(dmgH)2(OH2)2] under irradiation in combination with
(45) Lance, A. K.; Goldsby, K. A.; Busch, D. H. Inorg. Chem. 1990, 29,
4537.
(46) Schrauzer, G. N.; Holland, R. J. J. Am. Chem. Soc. 1971, 93, 1505.
(47) σp values were found in the following: Correlation Analysis in
Chemisry: recent adVances; Chapman, N. B., Shorter, J., Eds.; Plenum
Press: New York, 1978; pp 439-540.
4794 Inorganic Chemistry, Vol. 44, No. 13, 2005

Functional Models for Hydrogenases
Supporting Information Available: Experimental data and
simulations, including parameters, using Digisim, and cyclic
voltammograms of the supporting electrolyte (blank experiment),
Et3NHCl, and [Co(dmgH)2pyCl] (2 scans). This material is available
free of charge via the Internet at http://pubs.acs.org.
Martial Billon (DRFMC/SPrAM, CEA-Grenoble, France) for
helpful discussions about voltammograms simulations, Jean-
Marc Moulis and Jacques Meyer (DRDC/BMC, CEA-
Grenoble, France) for the gift of a GC chromatograph, and
Alexandra Ferrignio for performing the first experiments of
this study during an undergraduate training in the laboratory.
IC050167Z
Inorganic Chemistry, Vol. 44, No. 13, 2005 4795