Hydride-containing models for the active site of the nickel-iron hydrogenases

Article abstract of DOI:10.1021/ja105312p

The [NiFe]-hydrogenase model complex NiFe(pdt)(dppe)(CO)₃ (1) (pdt = 1,3-propanedithiolate) has been efficiently synthesized and found to be robust. This neutral complex sustains protonation to give the first nickel-iron hydride [1H]BF₄. One CO ligand in [1H]BF₄ is readily substituted by organophosphorus ligands to afford the substituted derivatives [HNiFe(pdt)(dppe)(PR₃)(CO)₂]BF₄, where PR ₃ = P(OPh)₃ ([2H]BF₄); PPh₃ ([3H]BF₄); PPh₂Py ([4H]BF₄, where Py = 2-pyridyl). Variable temperature NMR measurements show that the neutral and protonated derivatives are dynamic on the NMR time scale, which partially symmetrizes the phosphine complex. The proposed stereodynamics involve twisting of the Ni(dppe) center, not rotation at the Fe(CO)₂(PR₃) center. In MeCN solution, 3, which can be prepared by deprotonation of [3H]BF₄ with NaOMe, is about 10⁴ stronger base than is 1. X-ray crystallographic analysis of [3H]BF₄ revealed a highly unsymmetrical bridging hydride, the Fe-H bond being 0.40 A shorter than the Ni-H distance. Complexes [2H]BF₄, [3H]BF₄, and [4H]BF₄ undergo reductions near -1.46 V vs Fc^0/+. For [2H]BF₄, this reduction process is reversible, and we assign it as a one-electron process. In the presence of trifluoroacetic acid, proton reduction catalysis coincides with this reductive event. The dependence of i _c/i_p on the concentration of the acid indicates that H₂ evolution entails protonation of a reduced hydride. For [2H] ⁺, [3H]⁺, and [4H]⁺, the acid-independent rate constants are 50-75 s^-1. For [2H]⁺ and [3H]⁺, the overpotentials for H₂ evolution are estimated to be 430 mV, whereas the overpotential for the N-protonated pyridinium complex [4H ₂]²⁺ is estimated to be 260 mV. The mechanism of H ₂ evolution is proposed to follow an ECEC sequence, where E and C correspond to one-electron reductions and protonations, respectively. On the basis of their values for its pK_a and redox potentials, the room temperature values of ΔG_H? and ΔG_H- are estimated as respectively as 57 and 79 kcal/mol for [1H]⁺.

Full text of DOI:10.1021/ja105312p

Published on Web 10/06/2010
Hydride-Containing Models for the Active Site of the
Nickel-Iron Hydrogenases
Bryan E. Barton and Thomas B. Rauchfuss*
School of Chemical Sciences, UniVersity of Illinois at Urbana-Champaign,
Urbana, Illinois 61801
Received June 17, 2010; E-mail: rauchfuz@uiuc.edu
Abstract: The [NiFe]-hydrogenase model complex NiFe(pdt)(dppe)(CO)3 (1) (pdt ) 1,3-propanedithiolate)
has been efficiently synthesized and found to be robust. This neutral complex sustains protonation to give
the first nickel-iron hydride [1H]BF4. One CO ligand in [1H]BF4 is readily substituted by organophosphorus
ligands to afford the substituted derivatives [HNiFe(pdt)(dppe)(PR3)(CO)2]BF4, where PR3 ) P(OPh)3
([2H]BF4); PPh3 ([3H]BF4); PPh2Py ([4H]BF4, where Py ) 2-pyridyl). Variable temperature NMR measure-
ments show that the neutral and protonated derivatives are dynamic on the NMR time scale, which partially
symmetrizes the phosphine complex. The proposed stereodynamics involve twisting of the Ni(dppe) center,
not rotation at the Fe(CO)2(PR3) center. In MeCN solution, 3, which can be prepared by deprotonation of
[3H]BF4 with NaOMe, is about 104 stronger base than is 1. X-ray crystallographic analysis of [3H]BF4 revealed
a highly unsymmetrical bridging hydride, the Fe-H bond being 0.40 Å shorter than the Ni-H distance.
Complexes [2H]BF4, [3H]BF4, and [4H]BF4 undergo reductions near -1.46 V vs Fc0/+. For [2H]BF4, this
reduction process is reversible, and we assign it as a one-electron process. In the presence of trifluoroacetic
acid, proton reduction catalysis coincides with this reductive event. The dependence of ic/ip on the
concentration of the acid indicates that H2 evolution entails protonation of a reduced hydride. For [2H]+,
[3H]+, and [4H]+, the acid-independent rate constants are 50-75 s-1. For [2H]+ and [3H]+, the overpotentials
for H2 evolution are estimated to be 430 mV, whereas the overpotential for the N-protonated pyridinium
complex [4H2]2+ is estimated to be 260 mV. The mechanism of H2 evolution is proposed to follow an ECEC
sequence, where E and C correspond to one-electron reductions and protonations, respectively. On the
basis of their values for its pKa and redox potentials, the room temperature values of ∆GH• and ∆GH- are
estimated as respectively as 57 and 79 kcal/mol for [1H]+.
well suited for modeling since these protein are diverse,8
sometimes oxygen-tolerant,9 and are catalytically biased for H2
Introduction
Two families of hydrogenases, the [FeFe]-hydrogenases and
the [NiFe]-hydrogenases,1-4 have stimulated intense work on
the development of bioinspired catalysts for hydrogen proces-
sing.5,6 In the case of the [FeFe]-hydrogenases, the transition
from structural to functional model complexes was rapid owing
to foundational work conducted, albeit without awareness of
the biological connection,7 years before the structural charac-
terization of these proteins. Modeling the [NiFe]-hydrogenases
has proven more challenging, despite the fact that the relevant
proteins had been structurally characterized already in the 1990s.
Relative to the [FeFe]-enzymes, the [NiFe]-hydrogenases are
oxidation,10 which is the more challenging reaction for model
complexes.
Diverse structural models for the [NiFe]-hydrogenases have
been described,11 including many that feature Ni(SR)2Fe cores
complemented by diatomic ligands on Fe.12 Unlike model
complexes for the active site of the [FeFe]-hydrogenases, where
hydrides are numerous, active site models for the [NiFe]-
enzymes have until recently lacked hydride ligands.13,14 Our
contributions to this area of active site modeling have focused
(8) Cammack, R.; Frey, M.; Robson, R. Hydrogen as a Fuel: Learning
from Nature; Taylor & Francis: London, 2001.
(1) De Lacey, A. L.; Ferna´ndez, V. M.; Rousset, M.; Cammack, R. Chem.
ReV. 2007, 107, 4304–4330.
(9) Ludwig, M.; Cracknell, J. A.; Vincent, K. A.; Armstrong, F. A.; Lenz,
O. J. Biol. Chem. 2009, 284, 465–477.
(2) Ogata, H.; Lubitz, W.; Higuchi, Y. Dalton Trans. 2009, 7577–7587.
(3) Fontecilla-Camps, J. C.; Amara, P.; Cavazza, C.; Nicolet, Y.; Volbeda,
A. Nature 2009, 460, 814–822.
(10) Vincent, K. A.; Parkin, A.; Armstrong, F. A. Chem. ReV. 2007, 107,
4366–4413.
(11) Tard, C.; Pickett, C. J. Chem. ReV. 2009, 109, 2245–2274. van der
Vlugt, J. I.; Meyer, F. Met. Ions Life Sci. 2007, 2, 181–239.
(12) Li, Z.; Ohki, Y.; Tatsumi, K. J. Am. Chem. Soc. 2005, 127, 8950–
8951. Ohki, Y.; Yasumura, K.; Kuge, K.; Tanino, S.; Ando, M.; Li,
Z.; Tatsumi, K. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 7652–7657.
Pal, S.; Ohki, Y.; Yoshikawa, T.; Kuge, K.; Tatsumi, K. Chem. Asian
J. 2009, 4, 961–968.
(4) Fontecilla-Camps, J. C.; Volbeda, A.; Cavazza, C.; Nicolet, Y. Chem.
ReV. 2007, 107, 4273–4303.
(5) Heinekey, D. M. J. Organomet. Chem. 2009, 694, 2671–2680.
(6) Gloaguen, F.; Rauchfuss, T. B. Chem. Soc. ReV. 2009, 38, 100–108.
(7) Reihlen, H.; von Friedolsheim, A.; Ostwald, W. Justus Liebigs Ann.
Chem. 1928, 465, 72–96. Hieber, W.; Spacu, P. Z. Anorg. Allgem.
Chem. 1937, 233, 852–864. King, R. B. J. Am. Chem. Soc. 1962, 84,
2460. Fauvel, K.; Mathieu, R.; Poilblanc, R. Inorg. Chem. 1976, 15,
976–978. Winter, A.; Zsolnai, L.; Huttner, G. Z. Naturforsch. 1982,
37b, 1430–1436.
(13) Mealli, C.; Rauchfuss, T. B. Angew. Chem., Int. Ed. 2007, 46, 8942–
8944.
(14) Barton, B. E.; Whaley, C. M.; Rauchfuss, T. B.; Gray, D. L. J. Am.
Chem. Soc. 2009, 132, 6942–6943.
9
10.1021/ja105312p  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 14877–14885 14877

A R T I C L E S
Barton and Rauchfuss
Scheme 1. Proposed Ni-Centered Dynamics for 1 (L ) CO),
2 (L ) P(OPh)3), 3 (L ) PPh3), and 4 (L ) PPh2py)
Figure 1. Structure of the active site of the [NiFe]-hydrogenases.
on developing strategies for installing hydride ligands. For
example, one could attach Ni modules to substitutionally labile
iron hydrides and, complementarily, attach Fe species to Ni
hydrides. We have investigated the former route with partial
success. The species [HFe(CN)2(CO)3]-, which contains a
biomimetic ferrous [HFe(CO)(CN)2]- module, has been pre-
pared efficiently.15 Although [HFe(CN)2(CO)3]- undergoes well-
behaved substitution reactions, it has not yet been condensed
with nickel thiolates. Alternatively, we envisioned reactions of
preformed Ni(SR)2Fe ensembles with the equivalent of H- or
H+. Hydrides can be installed on diferrous dithiolates using
consistent with rotation of the trigonal bipyramidal Ni(dppe)
site (Scheme 1).
Treatment of a CH2Cl2 solution of 1 with HBF4•Et2O or
CF3CO2H resulted in immediate protonation to yield the
respective salt of the cationic hydride. The tetrafluoroborate salt
([1H]BF4) was isolated as a stable red microcrystalline solid
that is soluble in CH2Cl2, THF, MeCN, and MeOH. Its 31P{1H}
NMR spectrum displays a sharp singlet at δ 71. In the 1H NMR
spectrum, the hydride signal appears as a triplet of triplets (J )
6, 0.6 Hz, Figure 2). The 6 Hz coupling is a typical JPH for
nickel phosphine hydrides, for example, [HNi(dppe)2]AlCl4.20
The smaller coupling, confirmed by the {1H-1H} COSY
spectrum, arises from coupling to protons on the dithiolate. The
same cationic hydride can be generated from H2 in the presence
BH4 salts.16 We have generated impure samples of [(dppe)-
-
Ni(µ-pdt)(µ-H)Fe(CO)(dppe)]+ from [(dppe)Ni(µ-pdt)Fe(CO)2-
(dppe)]2+ via this method (dppe ) 1,2-bis(diphenylphosphino)-
ethane, pdt ) 1,3-propanedithiolate).14
The final approach to nickel-iron hydrides calls for proton-
ation of reduced Ni(SR)2Fe species, a route that is well
developed in the modeling of the [FeFe]-hydrogenases.17 Most
structural models feature NiII(SR)2FeII cores, but a few consist
of L2(RS)NiII(SR)Fe0(CO)4 centers, which in principle could be
protonated at iron.18 Despite its reputed instability, the
NiI(SR)2FeI species (dppe)Ni(µ-pdt)Fe(CO)3 (1) was attractive
to us.19 We found that protonation of 1 efficiently affords
[(dppe)Ni(µ-pdt)(µ-H)Fe(CO)3]+ ([1H]+), the first hydride-
containing model for the [NiFe]-hydrogenases. This heterobi-
metallic hydride resembles the active site with respect to the
presence of a L2Ni(H)(SR)2Fe(CO)L2 core (Figure 1), although
it diverges from the biological structure in other ways. The
finding that [1H]+ is an active catalyst for proton reduction
demonstrated that even approximate models could prove
functional, a finding that underscores the importance of hy-
dride ligands in modeling of the hydrogenases.13 In this paper,
we describe progress in modifying the Fe(CO)3 subsite and the
chemical consequences thereof. The resulting derivatives are
amenable to analysis of design features, including stereody-
namics, basicity, and proton relay.
21
of the borane B(C6F5)3 (eq 1).
NiFe(pdt)(dppe)(CO)3 + H2 + B(C6F5)3 f
[HNiFe(pdt)(dppe)(CO)3]HB(C6F5)3 (1)
Monosubstituted complexes [HNiFe(pdt)(dppe)(PR3)(CO)2]-
BF4 were prepared via thermal and photochemical substitution
of [1H]BF4 (eq 2).
[HNiFe(pdt)(dppe)(CO)3]BF4 + L f
[HNiFe(pdt)(dppe)(L)(CO)2]BF4 + CO (2)
[2H]BF4, L ) P(OPh)3
[3H]BF4, L ) PPh3
[4H]BF4, L ) PPh2Py
FT-IR spectra of these adducts feature a pair of comparably
intense νCO bands at about 2025 and 1970 cm-1 (Figure 3). The
positions of these bands indicate the expected sequence of
basicity, that is, P(OPh)3 < PPh2Py < PPh3. The 1H NMR spectra
exhibit doublet of triplets (JPH ) 35, 4 Hz) in the hydride region,
the coupling to the dppe ligands being slightly less than that of
the tricarbonyl hydride, [1H]BF4. In addition, all three mono-
substituted hydrides displayed an additional doublet of triplets
(JPH ≈ 40, 5 Hz) accounting for ∼1% of the sample (see Figure
Results
NiFe(pdt)(dppe)(CO)3 and Hydride Derivatives. Complex 1
has been characterized crystallographically by Schro¨der et al.19
The physical properties of our samples differ from those
previously reported however.19 The 31P{1H} NMR spectrum
consists of a broadened singlet at room temperature that
decoalesces into two doublets at -68 °C. This pattern is
(15) Whaley, C. M.; Rauchfuss, T. B.; Wilson, S. R. Inorg. Chem. 2009,
48, 4462–4469.
(16) van der Vlugt, J. I.; Rauchfuss, T. B.; Whaley, C. M.; Wilson, S. R.
J. Am. Chem. Soc. 2005, 127, 16012–16013.
(17) Koper, M. T. M.; Bouwman, E. Angew. Chem., Int. Ed. 2010, 49,
3723–3725.
(18) Lai, C.-H.; Reibenspies, J. H.; Darensbourg, M. Y. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 2390–2393. Verhagen, J. A. W.; Lutz, M.;
Spek, A. L.; Bouwman, E. Eur. J. Inorg. Chem. 2003, 2003, 3968–
3974.
Figure 2. 1H NMR spectra (500 MHz, CD2Cl2 solution) of [1H]BF4
(bottom, triplet, JPH ) 6 Hz) and of [3H]BF4 (top, doublet of triplets, JPH
) 35 Hz).
(19) Zhu, W.; Marr, A. C.; Wang, Q.; Neese, F.; Spencer, D. J. E.; Blake,
A. J.; Cooke, P. A.; Wilson, C.; Schro¨der, M. Proc. Natl. Acad. Sci.
2005, 102, 18280–18285.
9
14878 J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010

Active Site of the Nickel-Iron Hydrogenases
A R T I C L E S
are fully coalesced at +19 °C, but the SCH2 signals (∆ν )
91 Hz at -30 °C) remain completely distinct. The finding
that more closely spaced SCH2 signals must coalesce at higher
temperatures than the PCH2CH2P signals indicates that the
rocking of the Fe(CO)2PPh3 group is a higher energy process
and rotation of the Ni(dppe) center proceeds with a lower
barrier.
Complex [3H]BF4 was further characterized by X-ray crystal-
lography, which confirmed that PPh3 is cis to the hydride (Figure
5). The Fe(1)-Ni(1) distance of 2.6432(7) Å is only slightly
longer than that of [1H]BF4 (2.6131(14) Å). The Fe-Ni distance
in the D. Vulgaris enzyme for the Ni-C/Ni-R state is 2.55
Å.2 The bridging hydride position for [3H]BF4 is unsymmetrical,
being closer to iron than nickel by 0.40 Å, whereas in [1H]BF4
the hydride ligand is closer to iron by only 0.18 Å. The
unsymmetrical character of the hydride ligand is also indicated
by the diminished value of JPH between the hydride and the
dppe ligand (4 Hz vs 35 Hz for coupling to PPh3). The
Fe(1)-S(1)-Ni(1) angle in [3H]BF4 (71.42(3)°) is almost
identical to that for [1H]BF4 (70.39(6)°). The OC-Fe-CO angle
is 99°, vs the value of 96° calculated from the difference in the
intensities of the two νCO bands.23
Figure 3. FT-IR spectra in the νCO region for CH2Cl2 solutions of the
nickel-iron hydride complexes described in this work (top to bottom):
[1H]BF4; [2H]BF4; [3H]BF4; [4H]BF4. The νCO band for the Ni-R state
occurs at 1936-1948 cm-1, depending on the organism.1
NiFe(pdt)(dppe)(PPh3)(CO)2 (3) can be prepared in analytical
purity by the deprotonation of [3H]BF4 with NaOMe. The νCO
bands of 3 (CH2Cl2: 1971, 1916 cm-1) shift by about 45 cm-1
toward lower energy. The 31P{1H} NMR spectrum (Figure 6)
of a CD2Cl2 solution at room temperature displays a triplet (δ
55) assigned to the PPh3 ligand and two broad signals for the
dppe (δ 77, 45). Upon cooling the sample to -20 °C, the dppe
signals sharpen to the expected AB quartet and the PPh3 signal
appears as a doublet-of-doublets. This dynamic behavior is
analogous to that proposed for 1.
The pyridylphosphine hydride [4H]+ displays more com-
plexity than [2H]+ and [3H]+. Addition of D2O to a
d6-acetone solution of [4H]BF4 in CD2Cl2 solution resulted
in conversion to [4D]BF4 within seconds. Under the same
conditions, the corresponding PPh3 derivative [3H]BF4 was
found to exchange only slowly (t1/2 ) 20 min). Similarly,
[4H]BF4 was rapidly deprotonated by NEt3, whereas depro-
tonation of the analogous PPh3 derivative required hours
under comparable conditions.
Figure 4. 31P{1H} NMR (161 MHz) spectra of CD2Cl2 solutions of [3H]BF4
recorded at various temperatures. The signal at δ 68 is assigned to the
Fe(PPh3) center. The dynamic AB quartet at δ 65 is assigned to the Ni(dppe)
center. Weak signals near δ 59 and 70 arise from trace impurities of
Ni(pdt)(dppe) and [HNiFe(pdt)(dppe)(CO)3]+, respectively.
2). The identity of this second species is unknown, but we
suggest that it is an isomer. Such a species is observed in the
spectra for [2H]+, [3H]+, and [4H]+, but its NMR shift varies
with the identity of the phosphorus ligand.
Variable temperature 31P{1H} NMR spectra provide in-
sights into the dynamics that cannot be readily analyzed for
the more symmetrical [1H]BF4. The signal assigned to dppe
(δ 65) was broad at room temperature but decoalesced at
-60 °C into two doublets. The singlet for PPh3 remained
unchanged throughout this experiment (Figure 4). This
DNMR pattern is consistent with both a rocking of the
Fe(PPh3)(CO)2 subunit between two equivalent structures
(Scheme 2) and rotation of the Ni(dppe) center. We have
previously shown that related diiron dithiolates, for example,
compounds such as Fe2(pdt)(CO)4(diphosphine), are stereo-
chemically nonrigid.22 Variable temperature 13C NMR spectra
show that the PCH2CH2P signals (∆ν ) 180 Hz at -60 °C)
fast: [4H]BF4 + D2O f [4D]BF4 + HOD
slow: [3H]BF4 + D2O f [3D]BF4 + HOD
The IR spectrum of [4H]BF4 is more complex than that for
[2H]BF4 and is indicative of a mixture of three equilibrating
species. In THF solution, bands are observed not only for the
expected cationic hydride [4H]+, but also for neutral 4 and the
N-protonated hydride [4H2]2+. At room temperature, the relative
Scheme 2. Representation of the Fe- and Ni-Centered Dynamic Processes Proposed for [2H]BF4
9
J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010 14879

A R T I C L E S
Barton and Rauchfuss
Figure 5. Structure of [3H]BF4. Selected distances (Å): Fe(1)-Ni(1), 2.6432(7); Fe(1)-H(1), 1.49(3); Ni(1)-H(1), 1.89(3). Selected Bond angles (deg):
S(2)-Fe(1)-S(1), 83.27(4); S(2)-Fe(1)-P(1), 93.58(4); C(1)-Fe(1)-C(2), 99.93(17).
Table 1. Selected Electrochemical Properties of Nickel-Iron
Hydridesa
complex
E1/2 (V) vs Fc+/0
ipa/ipc (at 0.1 V/s)
[1H]+
[2H]+
[3H]+
[4H]+
[4H2]2+
-1.29
-1.44
-1.49
-1.49
-1.28
0.26
0.93
0.06
0.00
-
a Data were collected on 1 mM freshly prepared MeCN solution of
nickel-iron hydride and 0.1 M [NBu4]PF6 as electrolyte.
Figure 6. 31P{1H} NMR spectrum (161 MHz, CD2Cl2, 20 °C) of 3. Signals
at δ 77 and 45 are assigned to dppe, the signal at δ 55 is assigned to PPh3.
and [3H]BF4, undergo at least quasi-reversible reductions at
about -1.5 V (Table 1). The corresponding reduction for
[4H]BF4 was irreversible. For complex [2H]BF4, this couple
was partially reversible with ipa/ipc of 0.93. The one-electron
nature of the reduction of [2H]+ is indicated by the separation
of the peak potentials of ∆Ep ≈ 65 mV. Additionally, the scan
rate dependence for the [2H]+/0 and [2]0/+ couples are very similar.
We have independently established the one-electron nature of
the oxidation of 2 to give the mixed-valence NiIIFeI complex
(Figure 7).24 The oxidation of CH2Cl2 solutions of [2H]BF4,
[3H]BF4, and [4H]BF4 occur near 0.5 V vs Fc0/+ and are
irreversible.
amounts of the organometallic components in the solution were
estimated by simulation of the IR spectrum, giving Kprot (eq 3).
Kprot ) ([4])([4H22+])/([4H]+)2 ≈ 0.03
(3)
Addition of excess CF3CO2H to a CH2Cl2 solution of [4H]+
gave a new hydride with a concomitant increase of νCO by 10
cm-1. This shift is consistent with protonation of the pyridine
ligand giving [4H2]2+ (eq 4). The 1H NMR spectrum of [4H2]2+
features a doublet of triplets centered at δ ) -3.11 (JPH ) 40,
3.5 Hz). The 31P{1H} NMR spectrum resembles that for
[3H]BF4. Consistent with the weak donor ability of PPh2PyH+,
[4H2]2+ is not stable for prolonged periods and degrades to [1H]+
and free phosphine.
Upon addition of CF3CO2H (pKaMeCN ) 12.65,25 E0 ) -0.89
V)26 to CH2Cl2 solutions of [2H]+ and [3H]+, cyclic voltam-
mograms displayed increased cathodic current coinciding with
the hydride reduction event indicative of proton reduction
catalysis (Figure 9). At the relevant potentials (-1.5 V) proton
reduction by the glassy carbon working electrode was confirmed
to be negligible.26 In the case of [4H]+, addition of CF3CO2H
resulted in the appearance of a new catalytic current about 200
mV milder than the [4H]+/0 couple (Figure 10). We attribute
this new feature to catalysis by the N-protonated pyridine
[4H2]2+ complex, consistent with the previous spectroscopic
results. Because [4H2]2+ degrades into [1H]+ at high concentra-
tions of CF3CO2H (10-100 equiv), the voltammograms display
an additional catalytic wave for [1H]+.
[HNiFe(pdt)(dppe)(PPh2Py)(CO)2]+ + CF3CO2H f
-
[HNiFe(pdt)(dppe)(PPh2PyH)(CO)2]2+ + CF3CO2
(4)
4H 2+
[
]
2
Redox Properties and Catalytic Hydrogen Evolution. Cyclic
voltammetry indicates that the complexes [1H]BF4, [2H]BF4,
(20) Schunn, R. A. Inorg. Chem. 1970, 9, 394–395.
(21) Stephan, D. W. Dalton Trans. 2009, 312, 9–3136. Heiden, Z. M.;
Zampella, G.; De Gioia, L.; Rauchfuss, T. B. Angew. Chem., Int. Ed.
2008, 47, 9756–9759.
(22) Justice, A. K.; Zampella, G.; De Gioia, L.; Rauchfuss, T. B.; van der
Vlugt, J. I.; Wilson, S. R. Inorg. Chem. 2007, 46, 1655–1664. Justice,
A. K. Ph.D. Thesis, University of Illinois at Urbana-Champaign, 2008.
(23) Braterman, P. S. Metal Carbonyl Spectra; Academic Press: London,
1975.
(24) Schilter, D.; Rauchfuss, T. B. Unpublished results.
(25) Izutsu, K. Acid-Base Dissociation Constants in Dipolar Aprotic
SolVents; Blackwell Scientific Publications: Oxford, U.K., 1990.
(26) Felton, G. A. N.; Glass, R. S.; Lichtenberger, D. L.; Evans, D. H.
Inorg. Chem. 2006, 45, 9181–9184.
9
14880 J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010

Active Site of the Nickel-Iron Hydrogenases
A R T I C L E S
Figure 7. Cyclic voltammograms of a 1.85 mM MeCN solution of [2H]+ (left) and a 1.68 mM 9:1 MeCN/CH2Cl2 solution (0.1 M [NBu4]PF6) of 2 (right)
at various scan rates, denoted in mV/s.
Figure 8. Scan rate dependence of ipc for the couples [2H]+/0 and 20/+ in
CH2Cl2 solution (1.8 mM complex, 0.1 M [NBu4]PF6).
Figure 10. Cyclic voltammograms of a CH2Cl2 solution (0.74 mM) of
[4H]BF4 with increasing equiv of CF3CO2H (denoted on right). Conditions:
see Figure 9.
Plots of ic/ip (ic ) catalytic current, ip ) peak current in the
absence of acid) vs [CF3CO2H] are linear up to about ic/ip )
16-20 (Figure 11). This linear dependence indicates that H2
evolution follows protonation of the reduced hydride, that is,
is a measure of the acid-dependent rate-determining step. For
the hydrides investigated, this step is the protonation of
the rate is second order in [H+].27 The initial slope of this plot
Figure 9. Cyclic voltammograms of [2H]BF4 (left) and [3H]BF4 (right) with increasing equiv of CF3CO2H (denoted on right). Overpotentials were estimated
by the standard potential for hydrogen evolution from CF3CO2H in MeCN solution. Conditions: ∼0.5 mM in CH2Cl2 (see experimental), 0.1 M [NBu4]PF6, scan
rate 0.1 V/s, glassy carbon working electrode (d ) 3.0 mm); Ag wire pseudoreference with internal Fc standard at 0 V; Pt counter electrode.
9
J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010 14881

A R T I C L E S
Barton and Rauchfuss
Table 2. Thermodynamic Data (E’s in V vs Fc0/+, ∆G’s in kcal/mol
at 298 K for MeCN Solutions, NiFe ) NiFe(pdt)(dppe)L(CO)2) (T )
298 K)
E(NiFe0/+
)
E(NiFe+/2+
)
E(HNiFe+/0
)
∆GH•
∆GH
-
pKa
[1H]+
[3H]+
10.7
14.9
-0.543
-0.722
-0.124
-0.191
-1.29
-1.49
57
58
79
79
not have an established pKa scale, others have found that it
behaves similarly to MeCN solutions.28 A dilute solution of
[1H]+ in PhCN with one equiv of aniline ([PhNH3]BF4 pKaMeCN
) 10.7) afforded a 1:1 equilibrium mixture of [1H]+ and 1,
indicating a pKa of 10.7 for [1H]+. Similarly, four equiv of
4-methoxypyridine ([4-methoxypyridinium)]BF4 pKaMeCN
)
Figure 11. Influence of [acid]-catalyst ratio on catalytic current (ic/ip)
for [1H]BF4 (black circles), [2H]BF4 (red squares), [3H]BF4 (blue triangles),
and [4H]BF4 (green diamonds). Conditions: see Figure 9.
14.23)30 and [3H]+ provided an 2:1 equilibrium ratio of 3 and
[3H]+, as determined by 31P NMR spectroscopy. Although the
organometallic species decomposes over the course of several
hours, we assign the pKa as 14.9 ( 0.1 as the 3:[3H]+ ratio
remained unchanged throughout the decomposition. For com-
parison with these results, the µ-hydrido diiron complex
[HFe2(pdt)(CO)4(PMe3)2]BF4 has pKaMeCN of 12.31
Ni(I)Fe(I) precursors to form hydrides. Catalysis by the sterically
more accessible tricarbonyl 1 displays a steeper initial slope
than the substituted derivatives 2 and 3 even though the latter
are more basic. Relative to 2 and 3, the greater steric acces-
sibility of 1 toward protonation is consistent with the finding
that [1H]+ is rapidly deprotonated by Et3N whereas [3H]+
requires hours to deprotonate by the same strong base. The
relative catalytic activity can be estimated from the acid-
independent regions of the graph of ic/ip vs [CF3CO2H].28 The
values for ic/ip ) 16-20, indicating an acid-independent rate
constant between 50-75 s-1. For the tricarbonyl [1H]+, the ic/ip
of 10 corresponds to an acid-independent rate constant of 20 s-1.
To determine overpotential, electrocatalysis was conducted
in MeCN solution, where the standard reduction potential of
the acid to release H2, E0HA/H2, has been determined.29 For
CF3CO2H in MeCN solution, E0 is -0.89 V. Cyclic voltammetry
of a freshly prepared MeCN solution of [3H]+ displayed the
[3H]+/0 couple at -1.45 V, and the cathodic current dramatically
increased upon addition of CF3CO2H. The overpotential is
estimated at 430 mV (Figure 9). For the pyridinium complex
[4H2]2+, overpotential is significantly smaller, ∼260 mV (Figure
10).
Relevant to the mechanism of H2 production is the finding
that the reversibility of the [2H]+/0 couple is diminished in the
presence of [HNEt3]BF4. This effect is consistent with the
reduced derivative [2H]0 being sufficiently basic to undergo
protonation by [HNEt3]+. Since [HNEt3]BF4 is unable to
protonate 2, the system 2/[HNEt3]+ is catalytically inactive (in
contrast to 2/[CF3CO2H]+). To further clarify the mechanism
of hydrogen production, we found that in the presence of excess
acid, the catalytic current depends linearly on the concentration
of [3H]+. Thus, the catalysis is first-order in [3H]+.
Using cyclic voltammetry, we determined the oxidation
potentials of 1 and 3. Similar to the pKa determination, PhCN
was used to approximate E1/2 values on the MeCN scale. A 1
mM solution of 1 in PhCN displayed two oxidation events, one
reversible at -0.543 V and a second irreversible event at -0.124
V (Table 2). The cyclic voltammogram for 3 was similar,
displaying a reversible oxidation at -0.722 V and an irreversible
oxidation event at -0.191 V. We assign these couples as one-
electron processes. The oxidation of 1 with one equiv of
ferrocenium has been shown to afford the monocation.14
Discussion
The new hydride complexes are confirmed to be robust and
functional models for the [NiFe]-hydrogenases, at least with
respect to certain structural features and their ability to catalyze
hydrogen evolution. Specifically, the hydrides represent struc-
tural mimics of the Ni-R form of these enzymes, an S ) 0
state that is thought to feature an Fe(µ-SR)2(µ-H)Ni core.1 In
the model complexes, the coordination sphere at Ni is square
pyramidal, having rearranged from the tetrahedral geometry of
the Fe(I)Ni(I) precursor. Such a rearrangement is unlikely in
the protein, wherein the Ni center adopts a geometry intermedi-
ate between square-planar and tetrahedral.4
The present paper describes examples of hydride complexes
that sustain at least quasi-reversible redox events. Reversible
redox has been observed in a few bimetallic hydrides32 but is
rare for monometallic hydrides where redox changes the pKa
by many orders of magnitude, which often precludes reversibil-
ity.33,34 Nature’s selection of bimetallic active sites in the [FeFe]-
and [NiFe]-hydrogenases thus provides a way to soften the effect
of redox on the acid-base properties of the hydride. The pKa
and electrochemical data allow us to estimate the homolytic
and heterolytic bond energies for the bond between the NiFe
center and the µ-H ligand. Following the relations in eq 5 and
6, the free energy of H• donation, ∆GH•, is calculated to be 57
kcal/mol. These values indicate relatively weak M-H bonds,34,35
Acidity of Nickel-Iron Hydrides. To better understand the
thermodynamic properties of the catalysts, we sought to
determine both the pKa’s of the new nickel-iron hydrides and
the electrochemical properties of their conjugate bases. Com-
pounds 1 and 2 are poorly soluble in MeCN, a solvent with a
well-established pKa scale,30 We instead employed PhCN as
the solvent for the pKa determinations. Although PhCN does
(27) Wilson, A. D.; Newell, R. H.; McNevin, M. J.; Muckerman, J. T.;
Rakowski DuBois, M.; DuBois, D. L. J. Am. Chem. Soc. 2006, 128,
358–366.
(31) Eilers, G.; Schwartz, L.; Stein, M.; Zampella, G.; de Gioia, L.; Ott,
S.; Lomoth, R. Chem.sEur. J. 2007, 13, 7075–7084.
(28) Fraze, K.; Wilson, A. D.; Appel, A. M.; Rakowski DuBois, M.;
DuBois, D. L. Organometallics 2007, 26, 3918–3924.
(29) Wayner, D. D. M.; Parker, V. D. Acc. Chem. Res. 2002, 26, 287–
294.
(32) Barton, B. E.; Rauchfuss, T. B. Inorg. Chem. 2008, 47, 2261–2263.
(33) Tilset, M. J. Am. Chem. Soc. 1992, 114, 2740–2741. Cheng, T.-Y.;
Szalda, D. J.; Zhang, J.; Bullock, R. M. Inorg. Chem. 2006, 45, 4712–
4720. Hamon, P.; Toupet, L.; Hamon, J.-R.; Lapinte, C. Organome-
tallics 1992, 11, 1429–1431.
(30) Kaljurand, I.; Ku¨tt, A.; Soova¨li, L.; Rodima, T.; Ma¨emets, V.; Leito,
I.; Koppel, I. A. J. Org. Chem. 2005, 70, 1019–1028.
9
14882 J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010

Active Site of the Nickel-Iron Hydrogenases
A R T I C L E S
Table 3. Fe-H and Ni-H Bond Distance (Å) in Diiron and
Nickel-Iron Dithiolato Hydrides
Scheme 3. Catalytic Cycle Proposed for Hydrogen Evolution by
[HNiFe(pdt)(dppe)L2(CO)]BF4
hydride
M-H distances
∆d(M-H)
37
(dppe)dibasal(CO)Fe-H: 0.013
[HFe2(pdt)(CO)4(dppe)]BF4
1.627(3)
(CO)3Fe-H:
1.640(4)
38
(NHC)2(CO)Fe-H:
0.15
0.15
[HFe2(pdt)(CO)4(NHC-chelate)]BF4
1.710
(CO)3Fe-H:
1.562 Å
39
(PR3)apical(CO)2Fe-H:
1.59(1)
unsym-[HFe2(SC2H4PMe2)2(CO)4]BF4
(PR3)basal(CO)2Fe-H:
1.74(1)
40
(PMe3)basal(CO)2Fe-H: 0.07
1.63(1)
cis-[HFe2(pdt)(CN)(CO)4(PMe3)]BF4
(NC)basal(CO)2Fe-H:
1.70(1)
Fe-H: 1.46(6)
Ni-H: 1.64(6)
Fe-H: 1.49(3)
Ni-H: 1.89(3)
14
0.18
0.4
[HNiFe(pdt)(dppe)(CO)3]BF4
[HNiFe(pdt)(dppe)(PPh3)(CO)2]BF4
Table 4. Estimated Catalytic Properties of the Hydrides.
Conditions: see Experimental Section
consistent with the stability of the mixed valence cations.
Substitution of CO by PPh3 affects the [NiFe]I,I/I,II and
[NiFe]I,II/II,II couples as well as the pKa. Comparing [1H]+ to
[3H]+, the ∆pKa of 4.2 corresponds to 5.7 kcal/mol. The ∆E1/2
of the [HNiFe]+/0 couples corresponds to 4.6 kcal/mol.
Consequently, although [3H]+ is more difficult to reduce than
[1H]+, its formation requires weaker acids and thus its
catalytic function operates at a lower overpotential.
a
catalyst
Ecat (V, vs Fc0/+
)
rateb (s-1
)
overpotentialc (V)
[1H]+
[2H]+
[3H]+
[4H]+
-1.20
20
50
50
0.31
0.43
0.41
∼0.4
0.26
045
-1.32
-1.30
∼-1.3
-1.15
50
[4H2]2+
[NiFe]-hydrogenase
(A. Vinosum)
50
-0.345 (pH 7.4)d
500e
a For [1H]+ through [4H]+, potential at ipc/2 for an acid-independent
voltammogram.26 Potentials can be corrected from Fc+/0 (MeCN) to
NHE using the relation ENHE ) EFc + 0.717 V.46 b Estimated from
acid-independent region of ic/ip vs [acid] plot, see ref 28. c Calculated
using the relation Overpotential is the difference between Ecat and -0.89
V.26 d Onset potential, 40 °C, vs NHE.47 e Lower limit estimate, 30 °C,
pH 6.
Although the 10/+ and 30/+ couples are reversible, the couples
1+/2+ and 3+/2+ are not. Both are required to calculate ∆GH ,
-
which is related to the affinities of [1]2+ and [3]2+ for H-. The
∼0.5 V difference observed between the first and second
potentials is expected,36 so these redox potentials were used to
approximate the hydride donor strength for [1H]+ and [3H]+
(Table 2). These data suggest that the [1]2+ and [3]2+ have very
high affinities for hydrides, the high positive charge and
bimetallic structure being contributing factors.
in the [NiFe] enzymes, the Fe-H bond may be described as
having the character of a terminal (vs bridging) hydride.
Catalysis by these complexes is proposed to operate via the
sequence of reactions in Scheme 3. The voltammetric responses
indicate that the reaction is second order in protons and first order
in the bimetallic complex. The rate of protonation depends on both
the basicity of the metal center and steric effects.41 The associated
acid-dependent rate constants are not effective benchmarks for
catalytic efficiency, in part because they depend on the proton
source.42 More useful are the relative rates for catalysis in the acid-
independent regime obtained at high [H+], as well as the overpo-
tential (Table 4).43 This high [H+] regime is proposed to resemble
enzymatic conditions where protons are efficiently provided to the
NiFe center.44
I,I/I,II
1/2
∆GH• ) 1.37pKa + 23.06E
+ 54.9
(5)
(6)
∆GH ) 1.37pKa + 23.06EI,I/I,II + 23.06EI,II/II,II + 79.6
-
1/2
1/2
One of the more striking results is the asymmetry of the
Fe-H-Ni linkage, which we propose is relevant to the mech-
anism by which these complexes reduce protons. In the parent
hydride [HNiFe(pdt)(dppe)(CO)3]BF4, ([1H]BF4) the difference
of the iron-hydride and nickel-hydride bond distances
(∆d(M-H)) is 0.15 Å, whereas in [HNiFe(pdt)(dppe)(PPh3)-
(CO)2]BF4, the Ni-H bond is 0.4 Å longer than the Fe-H bond
(Table 3). Although terminal hydrides are invoked for the
catalytic mechanism of [FeFe]-hydrogenases, catalytically active
µ-hydrido diiron dithiolates have been reported.6 The asymmetry
of the Ni-H-Fe linkage in the present cases suggests that even
(37) Ezzaher, S.; Capon, J.-F.; Gloaguen, F.; Pe´tillon, F. Y.; Schollhammer,
P.; Talarmin, J.; Pichon, R.; Kervarec, N. Inorg. Chem. 2007, 46,
3426–3428.
(38) Morvan, D.; Capon, J.-F.; Gloaguen, F.; Le Goff, A.; Marchivie, M.;
Michaud, F.; Schollhammer, P.; Talarmin, J.; Yaouanc, J.-J. Orga-
nometallics 2007, 26, 2042–2052.
(39) Zhao, X.; Hsiao, Y.-M.; Lai, C.-H.; Reibenspies, J. H.; Darensbourg,
M. Y. Inorg. Chem. 2002, 699–708.
(34) Tilset, M. In ComprehensiVe Organometallic Chemistry III; Crab-
tree, R. H., Mingos, D. M. P., Eds.; Elsevier: Oxford, 2007; pp 279-
305.
(35) Choi, J.; Pulling, M. E.; Smith, D. M.; Norton, J. R. J. Am. Chem.
Soc. 2008, 130, 4250–4252. Wang, D.; Angelici, R. J. J. Am. Chem.
Soc. 1996, 118, 935-942.
(36) Houser, E. J.; Venturelli, A.; Rauchfuss, T. B.; Wilson, S. R. Inorg.
Chem. 1995, 34, 6402.
(40) Gloaguen, F.; Lawrence, J. D.; Rauchfuss, T. B.; Be´nard, M.; Rohmer,
M.-M. Inorg. Chem. 2002, 41, 6573–6582.
(41) Kramarz, K. W.; Norton, J. R. Prog. Inorg. Chem. 1994, 42, 1–65.
(42) Ezzaher, S.; Capon, J.-F.; Dumontet, N.; Gloaguen, F.; Pe´tillon, F. Y.;
Schollhammer, P.; Talarmin, J. J. Electroanal. Chem. 2009, 626, 161–
170.
(43) Rakowski Dubois, M.; Dubois, D. L. Acc. Chem. Res. 2009, 42, 1974–
1982.
9
J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010 14883

A R T I C L E S
Barton and Rauchfuss
of Et2O. Recrystallization from CH2Cl2/Et2O afforded red micro-
crystals. Yield: 1.35 g (1.71 mmol, 96%). H NMR (500 MHz,
Pyridylphosphines confer fascinating properties to certain
catalysts48 and offered the possibility of facilitating proton
transfer.49 When installed on diiron dithiolates, pyridyl phos-
phines undergo N-protonation, which leads to milder Ecat, an
effect attributable to an electrostatic influence.50 We observe
similar effects in this study: the overpotential decreases by ∼140
mV for the dication [4H2]2+ vs [4H]+ (Table 4). The amine
also accelerates the deprotonation of the µ-hydride, which is
otherwise slow. The pathway by which this deprotonation occurs
is suggested by the observation that the hydride exists in
equilibrium with the pyridinium salt. In the protein, the protons
exchange between the metal centers (hydride ligands) and the
terminal thiolate ligands.4
1
CD2Cl2, 20 °C): δ -3.53 (1H, tt: JPH ) 6, JHH ) 0.6 Hz correlates
with signal at δ2.5, HNiFe), 1.57 (1H, qt, axial (SCH2)2CH2), 2.0
(2H, t, axial (SCH2)2CH2), 2.5 (2H, d, equatorial (SCH2)2CH2), 2.65
(1H, dt, equatorial (SCH2)2CH2), 2.78 (4H, m, PCH2CH2P), 7.5 -
8.0 (20H, m, C6H5). 31P NMR (202 MHz, CD2Cl2): δ 71. 13C{1H}
NMR (19 °C, CD2Cl2, 150 MHz): δ 26, 36 (s, 2:1, pdt CH2); 30
2
(t, 1JPC ≈ JPC ) 10 Hz, PCH2CH2P); 130, 134, 134.5 (PPhn); 204
(s, Fe(CO)3), 205 (s, Fe(CO)3). FT-IR (CH2Cl2): νCO ) 2082, 2024
cm-1. Anal. Calcd for C32H31BF4FeNiO3P2S2 (found): C, 50.10
(50.16); H, 4.55 (4.75). Single crystals of [1H]BF4 ·CH2Cl2 were
grown from CH2Cl2-ether.
Reaction of 1 with B(C6F5)3 and H2. Under an inert atmosphere
4.0 mg B(C6F5)3 (0.0078 mmol) and 6.6 mg (0.0094 mmol) of 1
was dissolved with 0.5 mL of CD2Cl2 in a J. Young NMR tube.
The 1H and 31P{1H} NMR spectra were recorded initially showing
16% conversion to the hydride [1H]+, which we attribute to the
action of (H2O)B(C6F5)3. Spectra recorded after 1 h verified that
no change occurred. The sample was then frozen and put under an
H2 atmosphere. 1H and 31P{1H} NMR spectra showed nearly
complete conversion to [1H]+.
Experimental Section
Unless otherwise indicated, reactions were conducted using
Schlenk and cannula-filtration techniques at reduced temperatures.
Solvents for syntheses were HPLC-grade and further purified using
an alumina filtration system (Glasscontour Co., Irvine, CA), NMR
solvents were either dried with CaH2 and stored under nitrogen
over activated molecular sieves or purchased as ampules from
Cambridge Isotope Laboratories. Diiron nonacarbonyl, tetrafluo-
roboric acid in diethyl ether, triphenylphosphite, triphenyphosphine,
and trifluoroacetic acid were purchased from Aldrich and used as
received. Tetrabutylammonium hexafluorophosphate (Aldrich) was
recrystallized from methylene chloride and hexane. NMR spectra
were recorded at room temperature on a Varian Mercury spec-
trometer. NMR chemical shifts are quoted in ppm; spectra are
referenced to TMS for 1H and 85% H3PO4 for 31P{1H} NMR
spectra.
[HNiFe(pdt)(dppe)(P(OPh)3)(CO)2]BF4, [2H]BF4. To a 250-
mL round bottomed Schlenk flask was dissolved 1.245 g (1.58
mmol) of [2H]BF4 in 40 mL of CH2Cl2. To this solution, 414 µL
(1.58 mmol) of P(OPh)3 was added and the mixture was stirred for
6 h at 35 °C. Solvent was then removed under vacuum, and the
product was extracted into a small amount of warm EtOH. Cooling
of this extract to -78 °C precipitated the red product. This process
was repeated 3× followed by recrystallization of the material from
an EtOH solution by the addition of hexane. Yield: 1.06 g (1.0
1
mmol, 62%). H NMR (400 MHz, CD2Cl2): δ 6.6-8.0 (35H, m,
NiFe(pdt)(dppe)(CO)3, 1. To a 500-mL round bottomed Schlenk
flask with stir bar was added 2.25 g (4.01 mmol) of Ni(pdt)(dppe),
1.52 g (4.19 mmol) of Fe2(CO)9, and 40 mL of CH2Cl2. After
stirring the red slurry for 6 h, solvent was removed under vacuum,
and the red residue was washed with four 30 mL portions of MeCN
to remove Fe2(pdt)(CO)6 and Fe(CO)5. The remaining red-green
solid was extracted into ca. 5 mL of CH2Cl2, and this extract was
filtered through 4 × 12 cm plug of silica gel, rinsing with CH2Cl2.
A mobile green product eluted, leaving unreacted Ni(pdt)(dppe). The
green solution was then concentrated and then diluted with 100 mL
of hexane to precipitate green microcrystals. Yield: 0.745 g (1.06 mmol,
Ph’s), 2.88-1.1 (PCH2CH2P, SCH2CH2CH2S), -3.45 (1H, dt, Ni(µ-
H)Fe). 31P{1H} NMR (161 MHz, CD2Cl2): δ 161 (s, P(OPh)3), 65.8
(br, dppe). FT-IR (CH2Cl2): νCO ) 2031, 1981 cm-1
.
[HNiFe(pdt)(dppe)(PPh3)(CO)2]BF4, [3H]BF4. Method A. To
a 100-mL round bottomed Schlenk flask fitted with magnetic stir
bar was added 0.126 g (0.180 mmol) of NiFe(pdt)(dppe)(CO)3 from
the glovebox and dissolved in 25 mL of CH2Cl2. To this green
solution 0.077 g of (0.220 mmol) [HPPh3]BF4 and 0.105 g (0.400
mmol) of PPh3 were added. After 3.5 h photolysis with a Spectroline
black light lamp (365 nm), the FT-IR spectrum showed complete
consumption of [1H]BF4. The solution was then concentrated under
vacuum and addition of 40 mL of Et2O provided a red precipitate.
The product was washed with 3 × 10 mL of Et2O, and dried under
vacuum. Yield: 0.113 g (0.121 mmol, 67%).
1
27%). H NMR (500 MHz, CD2Cl2, 20 °C): δ 1.3 (1H, qt, axial
(SCH2)2CH2), 1.85 (1H, dt, equatorial (SCH2)2CH2), 1.9 (2H, t, axial
(SCH2)2CH2), 2.5 (2H, dt, equatorial (SCH2)2CH2), 2.2 (4H, m,
PCH2CH2P), 7.4 - 7.7 (20H, m, C6H5). 31P{1H} NMR (202 MHz,
Method B. To a 250-mL round bottomed flask fitted with
magnetic stir bar was prepared a solution of 0.262 g (0.333 mmol)
of [1H]BF4 in 50 mL of THF. To this solution 0.98 g (3.74 mmol)
of PPh3 was added. After stirring the solution for 2 h at 40 °C, the
solvent was removed in vacuum yielding an red colored oil, which
was washed with four 20-mL portions of hexane. The remaining
oil was redissolved in 30 mL of CH2Cl2, and the microcrystalline
product was precipitated by addition of 100 mL of hexane. Yield:
0.235 g (0.230 mmol, 70%). 1H NMR (400 MHz, CD2Cl2): δ
6.8-7.9 (35H, m, C6H5), 2.7 (4H, m, PCH2CH2P), 2.7-1.4 (6H,
m, SCH2CH2CH2S), δ -3.08 (1H, dt, Ni(µ-H)Fe). 31P{1H} NMR
(161 MHz, CD2Cl2): δ 69.5 (s, PPh3), 65.8 (br, dppe). 13C{1H}
NMR (19 °C, CD2Cl2, 150 MHz): δ 25.0, 25.75, 36.2 (s, 1:1:1,
pdt CH2’s); 28.6 (br, PCH2CH2P); 128-134 (m br, PPhn); 211 (br,
Fe(CO)2). 13C{1H} NMR (-60 °C, CD2Cl2, 150 MHz): δ 24.95,
25.56, 36.62 (pdt CH2’s), 27.98 + 29.18 (br, dppe PCH2CH2P),
128-134 (m br, C6H5), 211.2 + 211.6 (br, Fe(CO)2). FT-IR
CD2Cl2): δ 63.6. FT-IR (CH2Cl2): νCO ) 2028, 1952 cm-1
.
[HNiFe(pdt)(dppe)(CO)3]BF4, [1H]BF4. To a 100-mL round
bottomed Schlenk flask with magnetic stir bar was added 1.25 g
(1.78 mmol) of 1 and 10 mL of CH2Cl2. To this green solution
was added 0.30 mL (2.078 mmol) of HBF4•Et2O, immediately
producing a red solution. The solution was then concentrated under
vacuum and the product was precipitated by the addition of 20 mL
(44) Ga´lvan, I. F.; Volbeda, A.; Fontecilla-Camps, J. C.; Field, M. J.
Proteins: Struct., Funct., Bioinf. 2008, 73, 195–203.
(45) Pershad, H. R.; Duff, J. L. C.; Heering, H. A.; Duin, E. C.; Albracht,
S. P. J.; Armstrong, F. A. Biochemistry 1999, 38, 8992–8999.
(46) Connelly, N. G.; Geiger, W. E. Chem. ReV. 1996, 96, 877–922.
(47) Leger, C.; Jones, A. K.; Roseboom, W.; Albracht, S. P. J.; Armstrong,
F. A. Biochemistry 2002, 41, 15736–15746.
(48) Hartwig, J. F. Organotransition Metal Chemistry, from Bonding to
Catalysis; University Science Books: New York, 2010. Drent, E.;
Arnoldy, P.; Budzelaar, P. H. M. J. Organomet. Chem. 1994, 475,
57–63.
(CH2Cl2): νCO
)
2016, 1964 cm-1
.
Anal. Calcd for
(49) Rakowski DuBois, M.; DuBois, D. L. Chem. Soc. ReV. 2009, 38, 62–
72. Barton, B. E.; Olsen, M. T.; Rauchfuss, T. B. J. Am. Chem. Soc.
2008, 130, 16834–16835. Grotjahn, D. B. Chem.sEur. J. 2005, 11,
7146–7153. Grotjahn, D. B. Pure Appl. Chem. 2010, 82, 635–647.
(50) Li, P.; Wang, M.; Chen, L.; Liu, J.; Zhao, Z.; Sun, L. Dalton Trans.
2009, 1919–1926.
C49H46BF4FeNiO2P3S2 (found): C, 57.40 (57.48); H, 4.52 (4.36).
NiFe(pdt)(dppe)(PPh3)(CO)2, 3. In a 100-mL round-bottomed
Schlenk flask was dissolved 0.110 g (0.107 mmol) of [3H]BF4 in
5 mL of CH2Cl2 and 2 mL of MeOH. To this red solution 5.8 mg
(0.107 mmol) of NaOMe was added. After stirring for 3 h, the
9
14884 J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010

Active Site of the Nickel-Iron Hydrogenases
A R T I C L E S
reaction mixture was evaporated under vacuum. The residue was
washed with H2O (3 × 5 mL) and MeOH (3 × 5 mL), and the
green powder was dried under vacuum. Yield: 74 mg (0.079 mmol,
signal (δ -3.19). A 31P{1H} spectrum verified the presence of the
deuteride complex ([4D]BF4) with no decomposition.
Kinetics of Deprotonation of [3H]BF4 with NEt3. In a J.Young
NMR tube, a solution of 5.4 mg (0.0052 mmol) of [3H]BF4 in 0.5
mL of CD2Cl2 was treated with 12 µL (0.086 mmol) of NEt3 added
1
74%). H NMR (400 MHz, CD2Cl2): δ 7.2-8.0 (35H, m, C6H5),
2.1 (4H, m, PCH2CH2P), 1.8-0.8 (6H, m, SCH2CH2CH2S). 31P{1H}
NMR (161 MHz, CD2Cl2): δ 55 (t, PPh3), 45 + 77 (br, dppe).
1
by syringe. The tube was then sealed, and H and 31P{1H} NMR
FT-IR (CH2Cl2): νCO 1971, 1916 cm-1
.
Anal. Calcd for
spectra were recorded. The first-order decay plot was constructed
from 31P{1H} NMR spectra, as the ratio of 3/[3H]BF4 could be
readily determined by integration of the respective PPh3 signals.
C49H45FeNiO2P3S2 (found): C, 62.10 (62.78); H, 4.78 (4.45).
[HNiFe(pdt)(dppe)(PPh2Py)(CO)2]BF4 ([4H]BF4) and its
Protonation. To a 250-mL round-bottom Schlenk flask was added
0.400 g [1H]BF4 (0.508 mmol) and dissolved in 30 mL THF. To
this solution 0.150 g (0.570 mmol) of Ph2PyP was added. After
stirring the solution for 3 h at 40 °C, solvent was concentrated,
and the product was precipitated by addition of Et2O. The red solid
was recrystallized from 15 mL of acetone by the addition of 60
mL of EtOAc. The red microcrystalline material was dried under
vacuum and stored in the glovebox. Yield: 0.309 g (0.302 mmol,
59%). 1H NMR (500 MHz, CD2Cl2): δ 6.8-7.9, 8.8 (35H, m, C6H5,
C5H4N), δ 862 (4H, m, PCH2CH2P), δ 2.53-1.49 (6H, m,
SCH2CH2CH2S), δ 3.19 (1H, dt, Ni(µ-H)Fe). 31P{1H} NMR (202
MHz, CD2Cl2): δ 73.7 (s, PPh2Py), 65.7 (br, dppe). FT-IR (CH2Cl2):
νCO ) 2022, 1971 cm-1. Samples of [4H2]2+ were generated by
protonation of degassed CH2Cl2 solutions with ∼5 equiv of
CF3CO2H, however the resulting dication was observed to decom-
pose over the course of minutes. The decomposition mixture
consisted of [1H]BF4, Ni(pdt)(dppe), and [HPPh2Py]+ as indicated
by 31P{1H} and 1H NMR spectra. 1H NMR (400 MHz, CD2Cl2): δ
7.2-7.8 (m), 8.0 (t), 8.3 (t), 8.8 (d) (35H, C6H5, C5H4N), δ 86
(4H, m, PCH2CH2P), δ 2.7-1.5 (6H, m, SCH2CH2CH2S), δ 914
(1H, dt, Ni(µ-H)Fe). 31P{1H} NMR (161 MHz, CD2Cl2): δ 79 (s,
1
The H NMR data confirm the pseudofirst order behavior.
Electrochemistry, General Considerations. As the nickel-iron
hydrides presented in this paper degrade over the course of minutes
in MeCN solution, electrochemistry was mainly performed on
CH2Cl2 solutions. Cyclic voltammetry experiments were carried out
in a 20-mL one compartment glass cell with tight-fitting Teflon lid
with three tight-fitting electrodes and nitrogen gas inlet, interfaced
with a BAS-100 Electrochemical Analyzer. The working electrode
was a glassy carbon disk (diameter: 0.3 cm). A silver wire was
used as a pseudoreference electrode, and the counter electrode was
a Pt wire. The electrolyte was 0.1 M Bu4NPF6 in CH2Cl2. Ferrocene
was added as an internal reference and cyclic voltammograms were
each referenced to this Fc0/+ couple (0.00 V). iR compensation was
applied: solutions were pulsed prior to each scan to determine the cell
resistance, this compensation was applied to the subsequent voltam-
mogram. Between scans electrodes were polishing with alumina.
Overpotentials are estimated from the potential at 0.5(ipc) where
ipc is the peak current in the acid-independent regime (see Figure
11). Using the acidity constant for CF3CO2H (TFA) in MeCN
solution, pKTFAMeCN,of 12.65 (and ignoring homoconjugation),25
E0
was calculated using Evans’ relationship (eq 5).26
TFA/H2
PPh2Py), 66 (br, dppe). FT-IR (CH2Cl2): νCO ) 2032, 1982 cm-1
.
E0TFA ) E0H - (2.303RT/F)pKMeCN ) -0.89 V (5a)
+
TFA
pKa Determination of [3H]BF4. In a J. Young NMR tube, 0.8
mL dry degassed PhCN was added to 10.0 mg (0.0098 mmol) of
[3H]BF4. To this solution 197 µL of a freshly prepared solution of
0.197 M of 4-methoxypyridine in PhCN (pKaMeCN ) 14.23)25 was
added. The 31P NMR spectrum was then recorded after 1, 2, 3, and
5 h. At each interval the ratio of 3:[3H]+ remained unchanged at
2:1, despite steady decomposition of the sample. The ratio was
determined by integration of the respective PPh3 signals.
Cyclic Voltammetry for [2H]BF4 and 2. A solution of 2.6 mg
(0.00243 mmol) of [2H]BF4 in 5 mL of CH2Cl2 was prepared in
the CV cell and was treated with successive aliquots (18 µL, 2
equiv) of freshly prepared 0.268 M CF3CO2H-CH2Cl2 solution.
A solution of 5.9 mg (0.0055 mmol) of [2H]BF4 in 3.0 mL of
MeCN (1.7 mM) with 0.1 M [NBu4]PF6 was prepared in the CV
cell. Cyclic voltammograms were then recorded between 100-1000
mV/s. A solution of 5.1 mg (0.0052 mmol) of 2 in 2.7 mL of MeCN
and 0.3 mL of CH2Cl2 (1.683 mM) with 0.1 M [NBu4]PF6 was
prepared in the CV cell. Cyclic voltammograms were then recorded
between 100-1000 mV/s.
Cyclic Voltammetry for [4H]BF4. A solution of 3.8 mg (0.0037
mmol) of [4H]BF4 in 5 mL of CH2Cl2 was prepared in the CV cell
and was treated with successive aliquots (27.7 µL, 2 equiv) of a
freshly prepared solution of 0.268 M CF3CO2H in CH2Cl2 solution.
Cyclic voltammograms were recorded at 100 mV/s. To determine the
order with respect to [4H]BF4, a solution of 145 µL of CF3CO2H in
5 mL of CH2Cl2 was prepared in the CV cell and treated with
successive amounts of solid [4H]BF4. The result of this titration
indicated that the rate of catalysis is first-order with respect to [4H]BF4.
pKa Determination of [1H]BF4. In a 25 mL Schlenk flask, 4
mL dry degassed PhCN was added to 5.8 mg (0.0073 mmol) of
[1H]BF4. To this solution was added 27.6 µL of a freshly prepared
solution of 0.5 M of aniline in PhCN (pKaMeCN ) 10.7).25 The FT-
IR spectrum was then recorded after 3, 8, and 18 h. At each interval
the ratio of 1:[1H]+ remained unchanged at 1:1.
H/D Exchange of [3H]BF4 with D2O. Under an inert atmosphere
4.3 mg (0.0042 mmol) of [3H]BF4 was dissolved with 0.5 mL d6-
acetone (ampule, Cambridge) in a J. Young NMR tube. A 1H NMR
spectrum was recorded for t ) 0, then 10 µL (0.56 mmol) D2O
was added (in air) and subsequent scans were collected by at 2
min intervals. The intensity of the µ-H signal (δ -3.08), determined
vs phenyl region, decayed in a first order manner. After the complete
disappearance of the hydride signal for [3H]BF4, a 31P{1H} spectrum
was recorded, verifying the presence of the deuteride ([3D]BF4)
with no decomposition.
Acknowledgment. This research was sponsored by NIH. We
thank Dr. Danielle Gray for the crystallographic analysis of [3H]BF4.
H/D Exchange of [4H]BF4 with D2O. A solution of 7.2 mg
(0.00705 mmol) of [4H]BF4 in 0.5 mL of d6-acetone was prepared
in a J.Young NMR tube. The 1H NMR spectrum was recorded for
t ) 0. The sample was then frozen, 10 µL (0.56 mmol) of D2O
was added, and the sample tube evacuated and then the sample
thawed. The 1H NMR spectrum was recorded 5 min after thawing
the sample, showing nearly complete consumption of the hydride
Supporting Information Available: Spectra, electrochemical
data, and illustrative calculations. Crystallographic information
file (cif) for [3H]BF4. This material is available free of charge
via the Internet at http://pubs.acs.org.
JA105312P
9
J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010 14885