Isolation of a mixed valence diiron hydride: Evidence for a spectator hydride in hydrogen evolution catalysis

Article abstract of DOI:10.1021/ja312458f

The mixed-valence diiron hydrido complex (μ-H)Fe₂(pdt)(CO) ₂(dppv)₂ ([H1]⁰, where pdt =1,3- propanedithiolate and dppv = cis-1,2-C₂H₂(PPh ₂)₂), was generated by reduction of the differous hydride [H1]⁺ using decamethylcobaltocene. Crystallographic analysis shows that [H1]⁰ retains the stereochemistry of its precursor, where one dppv ligand spans two basal sites and the other spans apical and basal positions. The Fe - -Fe bond elongates to 2.80 from 2.66 A?, but the Fe-P bonds only change subtly. Although the Fe-H distances are indistinguishable in the precursor, they differ by 0.2 A? in [H1]⁰. The X-band electron paramagnetic resonance (EPR) spectrum reveals the presence of two stereoisomers, the one characterized crystallographically and a contribution of about 10% from a second symmetrical (sym) isomer wherein both dppv ligands occupy apical-basal sites. The unsymmetrical (unsym) arrangement of the dppv ligands is reflected in the values of A(³¹P), which range from 31 MHz for the basal phosphines to 284 MHz for the apical phosphine. Density functional theory calculations were employed to rationalize the electronic structure of [H1]⁰ and to facilitate spectral simulation and assignment of EPR parameters including ¹H and ³¹P hyperfine couplings. The EPR spectra of [H1]⁰ and [D1]⁰ demonstrate that the singly occupied molecular orbital is primarily localized on the Fe center with the longer bond to H, that is, Fe^II- H···Fe^I. The coupling to the hydride is A( ¹H) = 55 and 74 MHz for unsym- amd sym-[H1]⁰, respectively. Treatment of [H1]⁰ with H⁺ gives 0.5 equiv of H₂ and [H1]⁺. Reduction of D⁺ affords D ₂, leaving the hydride ligand intact. These experiments demonstrate that the bridging hydride ligand in this complex is a spectator in the hydrogen evolution reaction.

Full text of DOI:10.1021/ja312458f

Article
pubs.acs.org/JACS
Isolation of a Mixed Valence Diiron Hydride: Evidence for a Spectator
Hydride in Hydrogen Evolution Catalysis
Wenguang Wang,† Mark J. Nilges,† Thomas B. Rauchfuss,*,† and Matthias Stein*,‡
†School of Chemical Sciences, University of Illinois, Urbana, Illinois 61801, United States
‡Max Planck Institute for Dynamics of Complex Technical Systems, Sandtorstraβe 1, 39106 Magdeburg, Germany
S
* Supporting Information
ABSTRACT: The mixed-valence diiron hydrido complex (μ-
H)Fe2(pdt)(CO)2(dppv)2 ([H1]0, where pdt =1,3-propane-
dithiolate and dppv = cis-1,2-C2H2(PPh2)2), was generated by
reduction of the differous hydride [H1]+ using decamethylco-
baltocene. Crystallographic analysis shows that [H1]0 retains
the stereochemistry of its precursor, where one dppv ligand
spans two basal sites and the other spans apical and basal
positions. The Fe---Fe bond elongates to 2.80 from 2.66 Å, but
the Fe−P bonds only change subtly. Although the Fe−H
distances are indistinguishable in the precursor, they differ by
0.2 Å in [H1]0. The X-band electron paramagnetic resonance
(EPR) spectrum reveals the presence of two stereoisomers, the
one characterized crystallographically and a contribution of
about 10% from a second symmetrical (sym) isomer wherein both dppv ligands occupy apical−basal sites. The unsymmetrical
(unsym) arrangement of the dppv ligands is reflected in the values of A(31P), which range from 31 MHz for the basal phosphines
to 284 MHz for the apical phosphine. Density functional theory calculations were employed to rationalize the electronic structure
of [H1]0 and to facilitate spectral simulation and assignment of EPR parameters including 1H and 31P hyperfine couplings. The
EPR spectra of [H1]0 and [D1]0 demonstrate that the singly occupied molecular orbital is primarily localized on the Fe center
with the longer bond to H, that is, FeII−H···FeI. The coupling to the hydride is A(1H) = 55 and 74 MHz for unsym- amd sym-
[H1]0, respectively. Treatment of [H1]0 with H+ gives 0.5 equiv of H2 and [H1]+. Reduction of D+ affords D2, leaving the
hydride ligand intact. These experiments demonstrate that the bridging hydride ligand in this complex is a spectator in the
hydrogen evolution reaction.
example, Fe2(SR)2(CO)6, are poor bases and thus resist
protonation. These species do however reduce at relatively
mild potentials (about −1.65 V, all couples vs Fc+/0 in MeCN
solution, about −1 V vs the standard hydrogen electrode), and
the resulting anionic Fe(I)Fe(0) or Fe(0)Fe(0) species
protonate readily thereby initiating the HER.10,11 Reductions
can be achieved at a cathode12 or in photosensitized systems by
electron transfer from a high potential donor such as
INTRODUCTION
■
Reflecting interest in sustainable fuels and energy storage,1,2 the
hydrogen evolution reaction (HER) is receiving intense
attention. Metallic platinum is a prominent catalyst for the
HER, although commercial electrolyzers rely on nickel oxides.3
Interest in the HER has led to a focus on hydrogenase enzymes
and related bioinspired catalysts.4 Because of their relationship
to active sites of the hydrogenases, Fe−S−CO ensembles have
been of particular interest.5 These clusters are redox active, but
unlike traditional high-spin Fe−S clusters, they are low spin
species and form hydride derivatives, as seems to be required
for metal-catalyzed hydrogen evolution and uptake.
[Ru(bipy)3] (E / = −1.69 V).13 The second major class of
+
1
2
bioinspired HER catalysts are donor-ligand-substituted diiron-
(I) dithiolates, which are sufficiently basic that they undergo
direct protonation at the Fe−Fe bond.5,14 This protonation
produces cationic diferrous μ-hydrides that are susceptible to
reduction and subsequent protonation resulting in the HER.7,15
A third family of bioinspired HER catalysts, which appear to be
truly biomimetic, also feature some donor ligands in place of
CO, but they undergo protonation at a single Fe center, usually
mediated by an amine cofactor.9,16
Guided by the crystallographic characterization of two main
families of hydrogenases, catalysts have been developed that
resemble their active sites to varying extents. Although modest
progress has been reported for functional mimics of the
[NiFe]−H2ases,6 hundreds of catalysts have been developed
based on the active site of the [FeFe]−H2ases.7,8 These
catalysts are of the type Fe2(SR)2(CO)6−xLx, although only a
few exhibit high rates or low overpotentials9 and none exhibits
both.10 Generally, these catalysts can be classified according to
their HER mechanism. The CO-rich diiron compounds, for
Received: December 20, 2012
Published: February 5, 2013
© 2013 American Chemical Society
3633
dx.doi.org/10.1021/ja312458f | J. Am. Chem. Soc. 2013, 135, 3633−3639

Journal of the American Chemical Society
Article
All diiron-based HER catalysts operate by the intermediacy
of mixed valence hydrides, an otherwise rare class of
complexes.11,17 These S = 1/2 species often arise by protonation
of mixed valence Fe(I)Fe(0) species (eqs 1 and 2):
(CO)2(dppv)2]BF4 ([H1]+) in THF−Bu4NPF6 exhibit a
range of reduction potentials each with differing reversibility.
Upon replacing CO with phosphine ligands in this series of
complexes, the reduction potentials shift to more negative
values from −1.27 V for [H3]+/0 to −1.62 V for [H2]+/0 and
−1.77 V for [H1]+/0. As indicated in Figure 1, the redox event
becomes more reversible for the more demanding reductions,
as reflected by changes in ipc/ipa from 0.52 to 0.99. These results
suggested that [H1]0 should be sufficiently stable to allow
isolation.
[FeIFeI] + e− → [Fe0FeI]−
(1)
[Fe0FeI]− + H+ → [HFeIIFeI]
(2)
Alternatively, similar mixed valence hydrides are also produced
by 1 e− reduction of diferrous hydrides (eqs 3 and 4):10,11
Reduction of [H1]+ with decamethylcobaltocene (Cp2*Co,
[FeIFeI] + H+ → [HFeIIFeII]+
E / = −1.94 V in CH2Cl220) afforded the targeted complex
(3)
1
2
[H1]0 (Scheme 1). The conversion is signaled by a change in
[HFeIIFeII]+ + e− → [HFeIIFeI]
(4)
These two schemes highlight the centrality of the hydride
derivatives of FeIIFeI centers. Reduction enhances the
susceptibility of hydride complexes toward protonolysis to
give H2.15,18 Related to the HER process, S = 1/2 diiron
hydrides have been generated in situ and characterized by
electron paramagnetic resonance (EPR) spectroscopy. Specif-
ically, the hydrides [HFe2(SH)2(CO)6]0, [HFe2(pdt)-
(CO)4(PMe3)2]0, and [HFe2(edt)(CO)4(PMe3)2]0 display
EPR signals centered near g = 2.007, with the unpaired
electron spin exhibiting coupling to the nuclear spin of the
bridging hydride and other ligands on these diiron centers
(edt2− = C2H4S22−, pdt2− = C3H6S22−).2,19 In the present work,
we have isolated analytically pure samples of the related mixed
valence hydride [HFe2(pdt)(CO)2(dppv)2]0 ([H1]0, dppv =
cis-1,2-C2H2(PPh2)2).9 In addition to being a source of
Scheme 1
1
crystallographic and spectroscopic information, this S = /2
mixed-valence radical provides an opportunity to probe more
deeply within the catalytic cycle for the reduction of protons to
dihydrogen.
the color of the THF solution from brown to black. After
evaporating the reaction mixture, the product was extracted
into toluene and precipitated by pentane. Black microcrystals of
[H1]0 are stable in the absence of air.
RESULTS
■
Reduction of Diferrous Hydrides. Redox properties were
evaluated for three diiron hydrides of differing basicities, redox
potentials, and steric bulk (Figure 1). The cyclic voltammo-
grams of [HFe2(pdt)(CO)4(dppv)]BF4 ([H3]+), [HFe2(pdt)-
(CO)3(dppv)(PMe3)]BF4 ([H2]+), and [HFe2(pdt)-
The IR spectrum of [H1]0 exhibits a strong νCO band at 1893
cm−1, which is shifted by 57 cm−1 to lower energy compared to
νCO for [H1]BF4. The magnitude of the shift in νCO is in line
with density functional theory (DFT) calculations (Supporting
Information), which predict a shift of 46 cm−1. One electron
reduction of [HFe2(pdt)(CO)4(PMe3)2]+ shifts νCOavg by 80
cm−1.2 Treatment of [H1]0 with one equiv of FcBF4 gave back
[H1]+ (Figure 2).
Figure 1. Cyclic votammograms for [HFe2(pdt)(CO)2(dppv)2]BF4
(solid line), [HFe2(pdt)(CO)3(dppv)(PMe3)]BF4 (dash-dot line),
and [HFe2(pdt)(CO)4(dppv)]BF4 (dashed line) couples. Conditions:
10 mM sample in THF, 0.1 M Bu4NPF6; scan rate, 100 mV s−1;
+/0
potentials vs Fc+/0. Results: E / [H1] = −1.77 V, ipc/ipa = 0.99;
1
2
Figure 2. IR spectra in νCO region for [H1]BF4 in THF solution (a),
the same solution after treatment with Cp*2Co (b), and oxidation of
this solution with 1 equiv of solid [Cp2Fe]BF4 (c).
+/0
+/0
1
1
E / [H2] = −1.62 V, ipc/ipa = 0.72; E / [H3] = −1.27 V, ipc/ipa
=
2
2
0.52.
3634
dx.doi.org/10.1021/ja312458f | J. Am. Chem. Soc. 2013, 135, 3633−3639

Journal of the American Chemical Society
Article
Crystallographic Characterization of HFe2(pdt)-
(CO)2(dppv)2. The considerable stability of [H1]0 allowed us
to grow single crystals suitable for X-ray diffraction. Crystallo-
graphic analysis reveals a neutral complex of the type
Fe2(SR)2L6(μ-X). The framework is similar to unsym-[H1]+
in that one dppv is chelated to the two dibasal sites on Fe(1)
and the other dppv spans apical−basal positions on Fe(2).21
Relative to [H1]+, the Fe---Fe distance in [H1]0 is longer by
0.138 Å (Table 1). The Fe−S bonds are also elongated by
a
Table 1. Bond Distances (Å) in unsym-[H1]BF4 and unsym-
[H1]0
bond
[H1]+
[H1]0
Fe(1)−Fe(2)
Fe(1)−H
Fe(2)−H
Fe(1)−C
Fe(2)−C
Fe(1)-Savg
Fe(2)−Savg
Fe(1)−Pavg
Fe(2)−Pbasal
Fe(2)−Papical
2.6646(6)
1.65(3)
1.68(3)
1.739(4)
1.780(3)
2.262
2.8030(5)
1.61(2)
1.82(3)
1.780(2)
1.746(2)
2.311
2.2736
2.3167
2.2336
2.1892
2.236(9)
2.2287(9)
2.2264(7)
2.3039(7)
Figure 3. Structure of unsym-[H1]0. Thermal ellipsoids are set at the
50% probability level. For clarity, phenyl groups are drawn as lines, and
hydrogen atoms (except the hydride ligand) are omitted.
a
Taken from ref 21.
approximately 0.05 Å. The hydride ligand was located in the
difference map, and its position and isotropic displacement
parameters were refined. In [H1]+, the hydride ligand is
approximately equidistant between two iron centers, whereas in
[H1]0, the hydride ligand is semibridging, with the Fe−H
distances differing by 0.21 Å. The location of the hydride
ligands is consistent with the results of DFT calculations (see
below). Another striking difference is the elongation of the
Fe(2)−Papical bond by 0.07 Å. Overall, the crystallographic data
indicate that Fe(2) is more reduced than Fe(1).
EPR Characterization of HFe2(pdt)(CO)2(dppv)2 and
DFT Analysis. The X-band EPR spectrum of a fluid (toluene
at −30 °C) solution of [H1]0 exhibits a rich set of couplings
(Figure 4 and Table 2). Digital simulation of the spectrum
1
indicates the presence of two S = /2 species in the ratio of
85:15, the same isomer ratio seen for [H1]+.9 The subspectrum
from the major isomer is dominated by a hyperfine interaction
of 284 MHz that is proposed to originate from the unique
apical phosphine ligand on Fe(2) of unsym-[H1]0. The unique
phosphorus donor is attached to the center of greater Fe(I)
character (see below). The remaining hyperfine splittings can
1
be simulated by couplings to four I = /2 centers, originating
from the three basal phosphines (A(31P)basal ∼31 MHz) and the
hydride (A(1H) = 55 MHz). For comparison, in [HFe2(pdt)-
(CO)4(PMe3)2]0, Aiso(31P)basal is 41.7 MHz (absolute values
given for Aiso).2 The EPR spectrum of [H1]0 also consists a
second subspectrum, originating from the symmetrical isomer
of [H1]0, which shows coupling to two apical phosphorus
centers, A(31P)apical = 228 and 192 MHz, and two basal
phosphorus centers, ∼30 MHz (Figure 4). Compared to the
unsym isomer, the hydride displays a larger hyperfine
interaction with A(1H) = 74 MHz. In [HFe2(SH)2(CO)6]0
and [HFe2(pdt)(CO)4(PMe3)2]0, Aiso(1H) values are 90.8 and
75.8 MHz, respectively, which are comparable to that in sym-
[H1]0.2,19 The difference in the A values for the apical versus
basal P atoms indicates that the singly occupied molecular
orbital (SOMO) is mainly localized at the two iron nuclei with
Figure 4. X-band EPR spectra for toluene solutions of [H1]0 (i) and
[D1]0 (ii) at −30 °C in toluene with simulation of unsym-[H1]0 (A);
simulation of sym-H1 (B); simulation of a mixture of unsym- and sym-
[H1]0 (A + B); experimental spectrum of the mixture of [H1]0 (exp);
simulation of unsym-[D1]0 (A′); simulation of sym-[D1]0 (B′);
simulation of [D1]0 (A′ + B′); experimental spectrum of [D1]0 (exp′).
3635
dx.doi.org/10.1021/ja312458f | J. Am. Chem. Soc. 2013, 135, 3633−3639

Journal of the American Chemical Society
Article
a
Table 2. EPR Parameters for unsym- and sym-[H1]0 and [D1]0 from Simulations
b
c
c
c
c
compd
g factor
A(31Papical
)
A(31Pbasal
)
A(1H)
A(2H)
unsym-[H1]0
sym-[H1]0
2.003, 2.015, 2.026
2.002, 2.010, 2.021
2.003, 2.015, 2.026
2.002, 2.010, 2.021
284
32, 31, 31
33, 30
55
74
−
−
−
8.5
228, 192
288
unsym-[D1]0
32, 31, 31
34, 30
sym-[D1]0
232, 194
−
11.6
a
b
c
See Figure 4. g tensors were obtained by simulations of EPR spectra of the frozen sample (Supporting Information). Isotropic hyperfine coupling
constant (absolute value) in MHz also obtained by simulation. The differences in the A values between the spectra for [H1]0 vs [D1]0 are within the
experimental uncertainty of 3 MHz.
Figure 5. Unpaired spin density distribution in HFe2(pdt)(CO)2(dppv)2 with isocontour plots at 0.005 e−/a03 (violet). Atomic spin populations for
unsym-[H1]0 (left): Fe(1) 0.32, Fe(2) 0.64, H −0.03, Papical 0.053, Pbasal 0.001, −0.003, and −0.01. For sym-[H1]0 (right): Fe 0.36 and 0.39, H −0.02,
Papical 0.06 and 0.07, Pbasal 0.01 and 0.003.
lobes of the orbitals pointing toward the apical binding sites.
Phosphorus nuclei in the basal orientation are thus orthogonal
to the SOMO and acquire less spin density.
[HFe2(pdt)(CO)4(PMe3)2]0 where 0.3% unpaired spin density
resides on the bridging hydride, in [H1]0, the hydride ligand
carries −0.03 e− (unsym) and −0.02 e− (sym).
The EPR spectrum of the deuteride [D1]0 was also
examined. The phosphorus couplings relative to [H1]0 are
unchanged. The small 1:1:1 deuterium splitting can be clearly
observed (Figure 4ii). The A(2H) values are 8.5 and 11.6 MHz
for unsym- and sym-D1, respectivly. These diminished hyperfine
couplings agree with the diminished nuclear g factor for
deuterium: gH/gD = 6.514.
In Gibbs free energies, sym-[H1]0 is less stable than unsym-
[H1]0 by 4 kcal/mol (BP86-D) and 1 kcal/mol (B3LYP-D).
The 0.15:0.85 distribution estimated from EPR analysis
corresponds to a ΔG value of 0.8 kcal/mol at −30 °C.
Structural changes upon reduction of [H1]+ to [H1]0 are well
reproduced in the calculations (see Table 1 and the Supporting
Information). The reduction potential of [H1]+/0 was
calculated to be −2.7 V compared to Fc+/0 in the gas phase
and −2.0 V using a COSMO solvation model. The latter value
agrees well with the measured value of −1.8 V.9 The calculated
EPR parameters agree well with the experimentally determined
ones (see the Supporting Information) and allow an assign-
ment of nuclei and signs of hyperfine interactions in the
experimental EPR spectra.
DFT Calculations. DFT caculations are in agreement with
experimental findings and yield a more rhombic g tensor for
unsym-[H1]0 (2.002, 2.005, 2.020) than for sym-[H1]0 (2.002,
2.009, 2.011). The calculated atomic spin populations show
that 96% of the unpaired spin in the unsym-[H1]0 resides on
the Fe centers with 0.32 e− on Fe(1) and 0.64 e− on Fe(2) (see
Figure 3 for atom numbering). The apical phosphorus center
on Fe(2) accounts for 5.3% of the calculated unpaired spin,
whereas the other three Pbasal centers contribute approximately
1%. The apical P ligand gives rise to a large hyperfine
interaction with Aiso of +302 MHz while the isotropic 31P
hyperfine interactions of the basal P ligands are −11 to −30
MHz. The calculations on unsym-[H1]0 support the assignment
of unpaired electron spin mostly residing on Fe(2) as reflected
in the large 31P hyperfine interaction for the apical P atom (see
Figure 5).
Reactions of [HFe2(pdt)(CO)2(dppv)2]. Complex [H1]+ is
an electrocatalyst for the HER in the presence of strong acids.
The catalytic current is observed at −2.1 V versus Fc+/0, which
corresponds to the [H1]+/0 couple.9 As the intermediate in this
catalytic process, [H1]0 was tested for its competency to reduce
protons. As expected, [H1]0 was found to react readily with
strong acids with evolution of H2. However, the details of this
reaction proved to be surprising. Upon addition of 1.1 equiv of
H(OEt2)2BArF4 to THF solution of [H1]0, the color
immediately changed to brown red from black. The exclusive
organometallic product (94% isolated yield) was [H1]+, as
verified by IR, and 1H and 31P{1H} NMR analyses. It is known
that [H1]+ exists as two isomers that interconvert slowly, but
reduction followed by oxidation with H+ did not affect this
ratio. Acids that are weaker than H(OEt2)2BArF4, NH4PF6, and
CF3CO2H, with pKaMeCN values of 8 and 12,22 respectively, also
In sym-[H1]0, the unpaired electron is almost equally
distributed over the two iron atoms with atomic spin
populations of 0.36 and 0.39. Thus, they bear a total of 75%
unpaired spin density; the two apical phosphorus ligands can
acquire 7% each, reflected in their large isotropic hyperfine
interactions of +165 and +202 MHz. The two basal phosphines
only carry 1.3% of the unpaired spin. Compared to
3636
dx.doi.org/10.1021/ja312458f | J. Am. Chem. Soc. 2013, 135, 3633−3639

Journal of the American Chemical Society
Article
led to rapid evolution of H2 from [H1]0. We found however
that [H1]0 is unreactive toward benzoic acid, pKaMeCN = 20.7.
Yields of H2 approached 50% relative to the amount of [H1]0
(three experiments, standard deviation of 0.035). Upon treating
[H1]0 with D(Et2O)2BArF4, the organoiron product was
exclusively [H1]+. The yield of [H1]+ was quantified using
the very stable salt [HFe2(edt)(CO)4(PMe3)2]PF6 as an
integration standard for NMR analysis. The ratio of [H1]+ to
[HFe2(edt)(CO)4(PMe3)2]+ was found to be close to 1:1 by
integration of the corresponding μ-H signals, thus confirming
that no [D1]+ was produced, whereas only [D1]+ was produced
in the reaction of [D1] and H(Et2O)2BArF4, as confirmed by
2H, 1H, and 31P NMR analyses (Supporting Information).
These results are summarized in eq 5 and 6:
[H1]0 has been previously implicated as an intermediate in
hydrogen evolution catalyzed by [H1]+. The new insights into
the fact that it operates using a spectator hydride ligand
provides a suitable opportunity to summarize current
mechanistic schemes. As mentioned in the Introduction, two
broad pathways exist for the HER by bioinspired diiron
complexes, depending on the initiating reaction: reduction for
the hexacarbonyls (mainly) vs protonation for substituted
complexes.10,11 Catalytic cycles initiated by protonation are
always followed by reduction of the diferrous hydrides to give
mixed-valence hydrides (HFeIIFeI). The subsequent step in the
HER can be classified according to four subpathways, beginning
with a bimolecular pathway that is independent of [H+] (eq 7):
2[HFeIIFeI] → 2[FeIFeI] + H2
(7)
HFe2(pdt)(CO)2(dppv)2 + H+
The poor reversibility for the couple [HFe2(pdt)-
30
(CO)4(PMe3)2]0/+
,
[HFe2(pdt)(CO)3(dppv)(PMe3)]0/+
,
→ [HFe2(pdt)(CO)2(dppv)2]+ + 0.5H2
(5)
and [HFe2(pdt)(CO)4(dppv)]0/+ is attributed to this process.
Three proton-induced pathways can also be envisioned for the
HER (eq 8−10):
HFe2(pdt)(CO)2(dppv)2 + D+
→ [HFe2(pdt)(CO)2(dppv)2]+ + 0.5D2
(6)
[HFeIIFeI] + H+ → [FeIIFeI]+ + H2
(8)
In an effort to identify the nucleophilic site on the diiron
complex, we examined the reaction of [H1]0 with CF3SO3Me,
expecting to generate an S-methylated radical cation [H1Me]+.
S-Alkylation of diferrous dithiolates has been demonstrated
previously.23 However, the reaction of [H1]0 and CF3SO3Me in
toluene also produced [H1]+ with high yields. Gas chromato-
graphic analysis of the volatile products revealed small amounts
of methane, but the mass balance was poor.
[HFeIIFeI] + [HFeIIFeII]+ → [FeIIFeI]+ + [FeIFeI] + H2
(9)
2[HFeIIFeI] + 2H+ → 2[HFeIIFeII]+ + H2
(10)
In the case of [H1]0, both the stoichiometry and deuterium
labeling show that the μ-hydride ligand is a spectator,
implicating the new process in eq 10. The dominance of this
pathway reflects the bulky nature of [H1]0.
Aside from the fact that the HER requires protonation of this
mixed-valence species, mechanistic details are uncertain.
Protonation of highly reduced compounds can trigger
intermetallic electron transfer leading to the HER.31 Proto-
nation of [H1]0 at Fe would afford a Fe(II)Fe(III) dihydride
[HH1]+, whereas S-protonation5,32 of [H1]0 would preserve
the oxidation states.
DISCUSSION AND CONCLUSIONS
■
The new diiron hydride [HFe2(pdt)(CO)2(dppv)2]0 represents
a sterically stabilized analogue of the mixed-valence hydride
complexes invoked in many biomimetic HER pro-
cesses.7,8,10,11,17 The low symmetry of the main isomer of
[H1]0 allowed the analysis of the distribution of the SOMO
over the nuclei in the core of the complex. Specifically, the Fe
center with the stronger interaction with the hydride ligand is
less reduced than the other Fe center. This observation is
consistent with the destabilizing influence of hydrides on very
low oxidation states, as is indicated by various ligand electronic
parameters.24 The result implies that reduction of diferrous
complexes containing terminal hydride ligands, as required for
the HER,9 would occur at the nonhydride site.
EXPERIMENTAL SECTION
■
Materials and Methods. Compounds [HFe2(pdt)(CO)4(dppv)]-
BF416 ([H3]BF4), [HFe2(pdt)(CO)3(dppv)(PMe3)]BF4 ([H2]BF4),33
and [HFe2(pdt)(CO)2(dppv)2]BF4 ([H1]BF4)21 were prepared as
previously reported. Cp*2Co was purchased from Sigma−-Aldrich and
used as received.
IR spectra were recorded on a Perkin−Elmer Spectrum 100 Fourier
transform IR (FT-IR) spectrometer. Cyclic voltammetry was
performed under nitrogen at room temperature using a CH
Instruments CHI600D electrochemical analyzer with a glassy carbon
working electrode, Pt wire counter electrode, and the pseudoreference
electrode Ag wire. The sample (1 mM) and [NBu4]PF6 (100 mM)
were dissolved in THF. The scan rate was 0.1 V s−1, and potentials
were reported relative to internal Fc/Fc+. Microanalytical data were
acquired using an Exeter Analytical CE-440 elemental analyzer.
Crystallographic data were collected using a Bruker SMART equipped
with a 3-circle platform diffractometer and an APEX II charge-coupled
device area detector.
[DFe2(pdt)(CO)2(dppv)2]PF6 ([D1]PF6). A solution of 160 mg
(0.15 mmol) of 1 in 20 mL of CH2Cl2 was treated with 0.1 mL (1.29
mmol) of CF3CO2D (99.5 atom %D) in a glovebox. After the reaction
solution stirred for 2 h, the solvent was removed under vacuum. The
residue, a brown solid, was rinsed with Et2O (10 mL × 3) and then
redissolved in 5 mL of MeOH. This solution was treated with 5 mL of
CH3OH saturated with NH4PF6. A brown solid precipitated and was
isolated by filtration. The product was washed with H2O (10 mL × 3),
Despite the ubiquity of iron hydrides and their mechanistic
centrality, radical hydrides of iron are rare.2,19,25 Two examples
of ferric hydrides are [Cp*Fe(dppe)H]+ and L3Fe(μ-NH)(μ-
H)FeL3 (L3 = [PhB(CH2PPh2)3]−).26 Crystallographically
characterized iron(I) hydrides include HFe(dppe)2 and the S
3
=
/
species [(NacNac)FeH]− (dppe = 1,2-C2H4(PPh2)2,
2
NacNac = 2,2,6,6-tetramethyl-3,5-bis(2,6-diisopropylphenyl-
imido)heptyl).27
As d7−d6 systems featuring bridging hydrides and two
thiolates, [H1]0 and the active site of the [NiFe]−hydrogenases
(“Ni−C” state) are potentially related. The EPR parameters
however suggest that the present NiFe system and the present
FeFe hydride are quite different. EPR and ENDOR spectra for
the NiFe enzyme from Desulfovibrio vulgaris Miyazaki F and
Ralstonia eutropha reveal a small isotropic hyperfine couplings
constant Aiso ∼ −3.5 MHz for the bridging hydride.28 In these
2
cases, the SOMO is mainly localized on the Ni 3dz orbital,
which is perpendicular to the Ni---H vector.29
3637
dx.doi.org/10.1021/ja312458f | J. Am. Chem. Soc. 2013, 135, 3633−3639

Journal of the American Chemical Society
Article
MeOH (10 mL × 3), and diethyl ether (10 mL × 3) and then dried
under vacumn. Yield: 120 mg (66%). 1H NMR (500 MHz, CD2Cl2): δ
DFT Calculations. DFT calculations were performed with
ADF2012 and the ZORA Hamiltonian for relativistic effects and
spin−orbit coupling implemented as described in the Supporting
Information.36
7.06−7.09 (m, 40H, 40
× ArH), 6.59−6.61 (m, 2H,
Ph2PCH=CHPPh2), 6.33−6.36 (m, 2H, Ph2PCH=CHPPh2), 2.44−
2.58 (m, 2H, SCH2CH2CH2S), 1.85−2.03(m, 2H, SCH2CH2CH2S),
1.54 (m, 2H, SCH2CH2CH2S). 2H NMR (500 MHz, CH2Cl2 with 1%
CD2Cl2 as internal standard): δ −14.76 (m, μ-D). 31P NMR (500
MHz, CD2Cl2): δ 87, 85.8 (s, sym-[D1]+); δ 86, 82.2, 80.8, and 74.8 (s,
ASSOCIATED CONTENT
■
S
* Supporting Information
−
unsym-[D1]+); −145.0 (septet, PF6 ). FT-IR (CH2Cl2, νCO): 1951
Crystallographic information as a CIF file; additional IR, EPR,
2
and NMR (1H, H, 31P) spectra; computational details; and a
(vs), 1971 (sh) cm−1.
DFe2(pdt)(CO)2(dppv)2 ([D1]0). Compound D1 was prepapred by
reduction of [D1]PF6 following the same procedure as [H1]0. Yield:
76%. FT-IR (toluene, νCO): 1892 (vs) cm−1.
table of optimized structures and calculated EPR parameters.
This material is available free of charge via the Internet at
http://pubs.acs.org.
HFe2(pdt)(CO)2(dppv)2 ([H1]0). To a solution of [H1]BF4 (230
mg, 0.2 mmol) in 50 mL of dry THF was added Cp*2Co (73 mg, 0.22
mmol) as solid in an nitrogen box. The color of the mixture
immediately changed from brown red to black, and the THF was
evaporated under vacumn. The residue was extracted into toluene (20
mL × 3). The black solution was filtered through Celite, concentrated,
and diluted with 20 mL of pentane, resulting in precipitation of a black
solid. Yield: 154 mg (72%). FT-IR (THF, νCO): 1893 (vs) cm−1. Anal.
Calcd (found) for C57H51Fe2O2P4S2: C, 64.12 (63.71); H, 4.81 (5.12).
Single crystals were obtained by slow vapor diffusion of pentane into a
toluene solution of [H1]0 at room temperature.
AUTHOR INFORMATION
Corresponding Author
rauchfuz@illinois.edu; matthias.stein@mpi-magdeburg.mpg.de
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported as part of the ANSER Center, an
Energy Frontier Research Center funded by the U.S.
Department of Energy, Office of Science, Office of Basic
Energy Sciences, under award number DE-SC0001059, and the
Max-Planck Society for the Advancement of Science.
HER Experiments. Reactions were conducted in a tapered VWR
microwave reaction vial containing a small stir bar in a glovebox. For a
typical procedure, into a septum-capped vial containing 2 mL of
solution of [H1]BF4 (1 mM) in THF was injected with 0.1 mL of
THF solution of H(OEt2)2BArF4 (22 mM). The septum was sealed
with grease and removed from the glovebox. Into the vial was injected
60 μL of methane followed by gas chromatographic analysis on a
column packed with 5 Å molecular sieves (carrier gas: Ar), using an
Agilent 7820A instrument equipped with a thermal conductivity
detector. The response factor for H2/CH4 was 3.6 under our
conditions, as established by calibration with known amounts of
standard H2 and CH4. Three samples were run for this experiment.
The yields for H2 generation based on the amount of [H1]BF4 were
quantified as 44%, 43%, and 51%, respectivly. Thus, the average yield
of H2 was calculated to be 46% with a standard deviation σ of 0.035.
To quantify the yield of [H1]+ for the reaction of [H1]0 and
REFERENCES
■
(1) Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
15729.
(2) Jablonskye, A.; Wright, J. A.; Fairhurst, S. A.; Peck, J. N. T.;
̇
Ibrahim, S. K.; Oganesyan, V. S.; Pickett, C. J. J. Am. Chem. Soc. 2011,
133, 18606.
(3) Hauch, A.; Ebbesen, S. D.; Jensen, S. H.; Mogensen, M. J. Mater.
Chem. 2008, 18, 2331. Holladay, J. D.; Hu, J.; King, D. L.; Wang, Y.
Catal. Today 2009, 139, 244.
(4) Artero, V.; Chavarot-Kerlidou, M.; Fontecave, M. Angew. Chem.
Int. Ed. 2011, 50, 7238. Dempsey, J. L.; Brunschwig, B. S.; Winkler, J.
R.; Gray, H. B. Acc. Chem. Res. 2009, 42, 1995. Rakowski DuBois, M.;
DuBois, D. L. Acc. Chem. Res. 2009, 42, 1974.
(5) Zaffaroni, R.; Rauchfuss, T. B.; Gray, D. L.; De Gioia, L.;
Zampella, G. J. Am. Chem. Soc. 2012, 134, 19260.
(6) Barton, B. E.; Rauchfuss, T. B. J. Am. Chem. Soc. 2010, 132,
14877. Carroll, M. E.; Barton, B. E.; Gray, D. L.; Mack, A. E.;
Rauchfuss, T. B. Inorg. Chem. 2011, 50, 9554. Weber, K.; Kramer, T.;
Shafaat, H. S.; Weyhermuller, T.; Bill, E.; van Gastel, M.; Neese, F.;
Lubitz, W. J. Am. Chem. Soc. 2012, 134, 20745.
(7) Tschierlei, S.; Ott, S.; Lomoth, R. Energy Environ. Sci. 2011, 4,
2340.
(8) Felton, G. A. N.; Mebi, C. A.; Petro, B. J.; Vannucci, A. K.; Evans,
D. H.; Glass, R. S.; Lichtenberger, D. L. J. Organomet. Chem. 2009,
694, 2681.
D(Et2O)2BArF ,34 one equiv of [HFe2(edt)(CO)4(PMe3)2]PF6 was
4
1
added to the sample before H NMR analysis. The ratio of [H1]+ to
[HFe2(edt)(CO)4(PMe3)2]PF6 was found to be 0.99:1 by integration
of the corresponding μ-H signal, indicating no [D1]+ was produced.
To quantify the yield of [D1]+ for the reaction of [D1]0 and
H(Et2O)2BArF , one equiv of [HFe2(edt)(CO)4(PMe3)2]PF6 was
4
added to the sample before NMR analysis. The ratio of [D1]+ to
[HFe2(edt)(CO)4(PMe3)2]PF6 was found to be 1.95:1 by integration
of the corresponding 31P signal. The yield of [D1]+ was caculated to be
̈
2
97.5%. The H NMR spectrum showed only μ-D signal of [D1]+,
while the 1H NMR spectrum displayed one μ-H signal for
[HFe2(edt)(PMe3)2(CO)4]+. These results suggest that only [D1]+
was produced and there was no H/D exchanges between [D1]+ and
[HFe2(edt)(CO)4(PMe3)2]+.
Oxidation of [H1]0 with CF3SO3Me. A reaction vial containing 5
mL of solution of [H1]BF4 (47 mg, 0.044 mmol) in toluene was
capped with a septum in a glovebox. The reaction mixture was treated
with 1 equiv of CF3SO3Me (5 μL) by injection, and the spectrum was
sealed with grease. The black solution turned to brown in seconds with
concomitant appearance of a brown-colored solid precipitate. The gas
in the headspace of the vial was sampled by syringe and analyzed by
gas chromatography, which showed only a trace amount of methane.
The solvent was evaporated, and the substance was dissovled in
CH2Cl2. FT-IR (CH2Cl2, νCO): 1950 (vs), 1971 (sh) cm−1. The 31P
NMR spectrum matched that for [H1]+.
(9) Carroll, M. E.; Barton, B. E.; Rauchfuss, T. B.; Carroll, P. J. J. Am.
Chem. Soc. 2012, 134, 18843.
(10) Gloaguen, F.; Rauchfuss, T. B. Chem. Soc. Rev. 2009, 38, 100.
(11) Tard, C.; Pickett, C. J. Chem. Rev. 2009, 109, 2245.
(12) Felton, G. A. N.; Vannucci, A. K.; Okumura, N.; Lockett, L. T.;
Evans, D. H.; Glass, R. S.; Lichtenberger, D. L. Organometallics 2008,
27, 4671.
(13) Eckenhoff, W. T.; Eisenberg, R. Dalton Trans. 2012, 41, 13004.
Wang, M.; Chen, L.; Li, X.; Sun, L. Dalton Trans. 2011, 40. Reisner, E.
Eur. J. Inorg. Chem. 2011, 1005.
(14) Wright, J. A.; Pickett, C. J. Chem. Commun. 2009, 5719.
EPR Experiments. EPR samples were prepared in a MBraun
glovebox. The sample concentration is approximately 2 mM in
toluene. EPR spectra were recorded on a Varian E-line 12″ Century
series X-band CW spectrometer. EPR spectra were simulated using the
program SIMPOW6.35
Jablonskyte, A.; Wright, J. A.; Pickett, C. J. Dalton Trans. 2010, 39,
̇
3026. Liu, C.; Peck, J. N. T.; Wright, J. A.; Pickett, C. J.; Hall, M. B.
Eur. J. Inorg. Chem. 2011, 1080.
(15) Heinekey, D. M. J. Organomet. Chem. 2009, 694, 2671.
3638
dx.doi.org/10.1021/ja312458f | J. Am. Chem. Soc. 2013, 135, 3633−3639

Journal of the American Chemical Society
Article
(16) Barton, B. E.; Olsen, M. T.; Rauchfuss, T. B. J. Am. Chem. Soc.
2008, 130, 16834.
(17) Wang, M.; Chen, L.; Sun, L. Energy Environ. Sci. 2012, 5, 6763.
(18) Capon, J.-F.; Gloaguen, F.; Petillon, F. Y.; Schollhammer, P.;
́
Talarmin, J. Coord. Chem. Rev. 2009, 253, 1476.
(19) Keizer, P. N.; Krusic, P. J.; Morton, J. R.; Preston, K. F. J. Am.
Chem. Soc. 1991, 113, 5454.
(20) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877.
(21) Barton, B. E.; Rauchfuss, T. B. Inorg. Chem. 2008, 47, 2261.
(22) Izutsu, K. Acid-Base Dissociation Constants in Dipolar Aprotic
Solvents; Blackwell Scientific Publications: Oxford, U.K., 1990.
(23) Zhao, X.; Chiang, C.-Y.; Miller, M. L.; Rampersad, M. V.;
Darensbourg, M. Y. J. Am. Chem. Soc. 2003, 125, 518.
(24) Masui, H.; Lever, A. B. P. Inorg. Chem. 1993, 32, 2199. Chatt, J.;
Kan, C. T.; Leigh, G. J.; Pickett, C. J.; Stanley, D. R. J. Chem. Soc.,
Dalton Trans. 1980, 2032.
(25) Krusic, P. J.; Jones, D. J.; Roe, D. C. Organometallics 1986, 5,
456.
(26) (a) Hamon, P.; Toupet, L.; Hamon, J. R.; Lapinte, C.
Organometallics 1992, 11, 1429. (b) Kinney, R. A.; Saouma, C. T.;
Peters, J. C.; Hoffman, B. M. J. Am. Chem. Soc. 2012, 134, 12637.
(27) (a) Olsen, M. T.; Wang, W.; Rauchfuss, T. B. Unpublished
́ ̀ ̂
work. (b) Ekomie, A.; Lefevre, G.; Fensterbank, L.; Lacote, E.;
Malacria, M.; Ollivier, C.; Jutand, A. Angew. Chem., Int. Ed. 2012, 51,
6942. (c) Chiang, K. P.; Scarborough, C. C.; Horitani, M.; Lees, N. S.;
Ding, K.; Dugan, T. R.; Brennessel, W. W.; Bill, E.; Hoffman, B. M.;
Holland, P. L. Angew. Chem., Int. Ed. 2012, 51, 3658.
(28) (a) Foerster, S.; Gastel, M. v.; Brecht, M.; Lubitz, W. J. Biol.
Inorg. Chem. 2005, 10, 51. (b) Brecht, M.; van Gastel, M.; Buhrke, T.;
Friedrich, B.; Lubitz, W. J. Am. Chem. Soc. 2003, 125, 13075.
(29) Pandelia, M.-E.; Infossi, P.; Stein, M.; Giudici-Orticoni, M.-T.;
Lubitz, W. Chem. Commun. 2012, 48, 823.
(30) Ezzaher, S.; Capon, J.-F.; Dumontet, N.; Gloaguen, F.; Petillon,
́
F. Y.; Schollhammer, P.; Talarmin, J. J. Electroanal. Chem. 2009, 626,
161.
(31) Marinescu, S. C.; Winkler, J. R.; Gray, H. B. Proc. Natl. Acad. Sci.
U.S.A. 2012, 109, 15127.
(32) Apfel, U.-P.; Troegel, D.; Halpin, Y.; Tschierlei, S.; Uhlemann,
U.; Gorls, H.; Schmitt, M.; Popp, J.; Dunne, P.; Venkatesan, M.; Coey,
̈
M.; Rudolph, M.; Vos, J. G.; Tacke, R.; Weigand, W. Inorg. Chem.
2010, 49, 10117.
(33) Barton, B. E.; Zampella, G.; Justice, A. K.; De Gioia, L.;
Rauchfuss, T. B.; Wilson, S. R. Dalton Trans. 2010, 39, 3011.
(34) Brookhart, M.; Grant, B.; Volpe, A. F. Organometallics 1992, 11,
3920.
(35) Nilges, M. J.; Matteson, K.; Belford, R. L. SIMPOW6: a software
package for the simulation of ESR powerder-type spectra. In ESR
Spectroscopy in Membrane Biophysics, Biological Magnetic Resonance;
Hemminga, M. A., Berliner, L., Eds.; Springer: New York, 2007.
(36) (a) Scientific Computing & Modelling. ADF; Vrije Universiteit,
Theoretical Chemistry: Amsterdam, The Netherlands, 2012. (b) te
Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Guerra, C. F.; Van
Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. J. Comput. Chem. 2001,
22, 931.
3639
dx.doi.org/10.1021/ja312458f | J. Am. Chem. Soc. 2013, 135, 3633−3639