Catalytic hydrogen evolution by Fe(II) carbonyls featuring a dithiolate and a chelating phosphine

Article abstract of DOI:10.1021/ic5012988

Two pentacoordinate mononuclear iron carbonyls of the form (bdt)Fe(CO)P₂ [bdt = benzene-1,2-dithiolate; P₂ = 1,1′-diphenylphosphinoferrocene (1) or methyl-2- {bis(diphenylphosphinomethyl)amino}acetate (2)] were prepared as functional, biomimetic models for the distal iron (Fe_d) of the active site of [FeFe]-hydrogenase. X-ray crystal structures of the complexes reveal that, despite similar ν(CO) stretching band frequencies, the two complexes have different coordination geometries. In X-ray crystal structures, the iron center of 1 is in a distorted trigonal bipyramidal arrangement, and that of 2 is in a distorted square pyramidal geometry. Electrochemical investigation shows that both complexes catalyze electrochemical proton reduction from acetic acid at mild overpotential, 0.17 and 0.38 V for 1 and 2, respectively. Although coordinatively unsaturated, the complexes display only weak, reversible binding affinity toward CO (1 bar). However, ligand centered protonation by the strong acid, HBF₄·OEt₂, triggers quantitative CO uptake by 1 to form a dicarbonyl analogue [1(H)-CO]⁺ that can be reversibly converted back to 1 by deprotonation using NEt₃. Both crystallographically determined distances within the bdt ligand and density functional theory calculations suggest that the iron centers in both 1 and 2 are partially reduced at the expense of partial oxidation of the bdt ligand. Ligand protonation interrupts this extensive electronic delocalization between the Fe and bdt making 1(H)⁺ susceptible to external CO binding.

Full text of DOI:10.1021/ic5012988

Article
pubs.acs.org/IC
Catalytic Hydrogen Evolution by Fe(II) Carbonyls Featuring a
Dithiolate and a Chelating Phosphine
Souvik Roy,†,‡,§ Shobeir K. S. Mazinani,†,‡ Thomas L. Groy,† Lu Gan,†,‡ Pilarisetty Tarakeshwar,†
Vladimiro Mujica,†,‡ and Anne K. Jones*,†,‡
‡
†Department of Chemistry and Biochemistry and Center for Bio-Inspired Solar Fuel Production, Arizona State University, Tempe,
Arizona 85287, United States
S
* Supporting Information
ABSTRACT: Two pentacoordinate mononuclear iron carbon-
yls of the form (bdt)Fe(CO)P2 [bdt = benzene-1,2-dithiolate;
P2 = 1,1′-diphenylphosphinoferrocene (1) or methyl-2-{bis-
(diphenylphosphinomethyl)amino}acetate (2)] were prepared
as functional, biomimetic models for the distal iron (Fed) of the
active site of [FeFe]-hydrogenase. X-ray crystal structures of
the complexes reveal that, despite similar ν(CO) stretching
band frequencies, the two complexes have different coordina-
tion geometries. In X-ray crystal structures, the iron center of 1
is in a distorted trigonal bipyramidal arrangement, and that of 2
is in a distorted square pyramidal geometry. Electrochemical investigation shows that both complexes catalyze electrochemical
proton reduction from acetic acid at mild overpotential, 0.17 and 0.38 V for 1 and 2, respectively. Although coordinatively
unsaturated, the complexes display only weak, reversible binding affinity toward CO (1 bar). However, ligand centered
protonation by the strong acid, HBF4·OEt2, triggers quantitative CO uptake by 1 to form a dicarbonyl analogue [1(H)-CO]+ that
can be reversibly converted back to 1 by deprotonation using NEt3. Both crystallographically determined distances within the bdt
ligand and density functional theory calculations suggest that the iron centers in both 1 and 2 are partially reduced at the expense
of partial oxidation of the bdt ligand. Ligand protonation interrupts this extensive electronic delocalization between the Fe and
bdt making 1(H)+ susceptible to external CO binding.
understand and reproduce these enzymes.8 However, although
INTRODUCTION
■
natural hydrogenases have turnover frequencies exceeding
Hydrogen produced from solar energy and water offers the
1000 s−1 at potentials close to the thermodynamic reduction
tantalizing opportunity to produce a storable, renewable fuel on
a scale comparable to global energy challenges.1 However,
potential of the proton, synthetic models seldom come close to
this exquisite reactivity.9,10
developing efficient and renewable catalysts for this trans-
formation has proven challenging. Thus, hydrogenases, the
The active site of [FeFe]-hydrogenases, referred to as the H-
cluster, is a unique six-iron cluster consisting of a [4Fe4S]
biological catalysts for reversible proton reduction to hydrogen,
cluster bridged via a cysteinyl thiolate to a diiron subsite.5,6 This
diiron subcluster, although biologically unprecedented, is highly
reminiscent of known organometallic complexes. As shown in
Figure 1, it consists of a dithiolate ligand bridging the two iron
ions as well as the strong π-acceptors CO and CN− at each iron
center. The proximal iron, Fep, so designated due to its relative
proximity to the [4Fe4S] center, is a coordinatively saturated
octahedral site. On the other hand, hydrogen binding or
production occurs at the distal iron, Fed, an electron-deficient,
five-coordinate, pseudosquare pyramidal center featuring a
terminal open coordination site.
have caught the attention of a broad range of researchers.2,3
Because the elucidation of the structures of both [NiFe]- and
[FeFe]-hydrogenases revealed that these enzymes feature
organometallic active sites, including the diatomic ligands CO
and CN− (Figure 1),4−7 inorganic chemists have sought to
produce both structural and functional models in an effort to
Organometallic complexes of the type [(μ-SR2)Fe2(CO)6]
and their derivatives in which one or more of the carbonyls
have been replaced with more strongly σ-donating ligands, such
as phosphines, have been used extensively as both structural
Figure 1. Active site of (A) [FeFe]-hydrogenase (H-cluster) and (B)
[NiFe]-hydrogenase. Fed and Fep denote the distal and proximal iron,
respectively, in the H-cluster.
Received: March 22, 2014
Published: August 11, 2014
© 2014 American Chemical Society
8919
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
and functional mimics of [FeFe]-hydrogenases.3,11,12 Although
the natural enzyme features iron centers in square pyramidal
environments that are inverted relative to each other such that a
terminal open coordination site is available, the irons in most of
these model complexes are in the so-called “eclipsed” geometry
in which the two pyramids have the same orientation.
Complexes with an “inverted” iron center, although known,
remain the exception.13−15 In an eclipsed geometry, the
bridging, as opposed to terminal, position is most reactive,
facilitating the formation of stable but unreactive bridging
hydrides.16−18 Thus, the models tend to be poor catalysts for
proton reduction and require substantial overpotentials for the
catalysis. Development of mononuclear iron complexes with an
open coordination site can, in principle, overcome this difficulty
and mimic the reactivity of the distal iron site of the enzyme if
an appropriate ligand set can be found to simulate the
electronic environment of the second missing metal.
RESULTS
■
Synthesis and Spectroscopic Characterization. Two
pentacoordinate iron(II)−carbonyl complexes, each with a
chelating bis-phosphine and benzene-1,2-dithiol (bdt), were
synthesized starting from FeCl2: (κ2-dppf)Fe(CO)(κ2-bdt) (1)
and (κ2-NP2)Fe(CO)(κ2-bdt) (2) where dppf is 1,1′-bis-
(diphenylphosphino)ferrocene and NP2 is methyl-2-{bis-
(diphenylphosphinomethyl)amino}-acetate (Scheme 1). The
Scheme 1. Synthesis of Complexes 1 and 2 from FeCl2, the
Appropriate Bis-phosphine Ligand, Benzene-1,2-dithiol, and
CO
Although synthetic efforts immediately following the
elucidation of the crystal structure of [FeFe]-hydrogenases
produced a series of coordinatively saturated mononuclear iron
complexes as spectroscopic models for the H-cluster,19−21
relatively few five-coordinate models have been reported. Liaw
and co-workers demonstrated the synthesis of a pentacoordi-
nate, 16-electron Fe(II) complex [Fe(CO)2(CN)(S,NH-
C6H4)]− and showed that it readily reacted to form
hexacoordinate complexes or dimers.22 They did not, however,
investigate the catalytic activity of this compound. Darensbourg
and co-workers have also produced pentacoordinate iron
dicarbonyls using the strongly π-donating, redox noninnocent
ligand 2-amido-thiophenylate as models for the mononuclear
Fe-containing hydrogenases.23,24 Particularly relevant to this
work, Sellmann and co-workers employed the benzene-1,2-
dithiolate (bdt) ligand to construct [Fe(bdt)(PMe3)2(CO)2]
and noted that it had an unexpected tendency to lose CO to
form a 16-electron complex.25 Rauchfuss and co-workers used
that work as an inspiration to create (Et4N)2[Fe(bdt)-
(CN)2(CO)], as a spectroscopic model of the enzyme active
site.26 Only recently did Ott and co-workers report complexes
of bdt together with a chelating phosphine to create
coordinatively unsaturated monocarbonyl models of the distal
iron site of [FeFe]-hydrogenases and to show that these
compounds are active in electrocatalytic proton reduction.27−29
In this Article, we present a new coordinatively unsaturated,
five-coordinate Fe(II)-carbonyl in a P2S2 coordination environ-
ment. The phosphines are provided by 1,1′-{bis(diphenyl)-
phosphino}ferrocene (dppf), and the steric constraints of the
ferrocene moiety in this ligand cause it to have one of the
largest bite angles observed for a chelating phosphine. On the
other hand, benzene-1,2-dithiolate (bdt), well-known for its
redox activity associated with the conjugation of the sulfur
donors to the aromatic ring, provides the sulfur ligands. The
result is that [(κ2-dppf)Fe(CO)(κ2-bdt)] (1) electrocatalyti-
cally reduces protons from the weak acid acetic acid with
unprecedentedly low overpotentials. Furthermore, the proto-
nated complex binds exogenous CO, a reaction seldom seen in
model compounds but well-known for the enzyme. The
electrocatalytic properties of this complex are directly
compared to those of an analogous compound [(κ2-NP2)Fe-
bdt ligand was employed for both its strong π-donor propensity
and its redox activity. Similarly, dppf was chosen because it is
chelating and, among common chelating bis-phosphine ligands,
has one of the widest bite angles. This angle has a major
influence on the structure of the resulting complex and the
corresponding catalytic properties.30−33 For comparison, to
evaluate the impact of dppf on the electronic and catalytic
properties of the FeIIS2P2 center, 2, which features an N-
containing bis-phosphine ligand (NP2) instead of dppf, was
synthesized. The ligand NP2 is easily obtained by reaction of
two equivalents of Ph2PCH2OH with glycine methyl ester in
refluxing ethanol.34 As shown in Scheme 1, treatment of a
methanolic solution of anhydrous FeCl2 with a solution of the
appropriate bis-phosphine in THF and benzene-1,2-dithiol in
the presence of a base (triethylamine) under a CO atmosphere
afforded the desired diamagnetic complexes, 1 and 2, in 65%
and 60% isolated yields, respectively. A single resonance is
observed in the 31P{1H} NMR spectrum at 66.32 ppm for 1
and 50.21 ppm for 2 (1H NMR spectra are available in the
Supporting Information).
In a tetrahydrofuran (THF) solution, 1 is reddish brown with
visible absorbances at 467 (ε = 4433 M−1 cm−1) and 745 nm (ε
= 974 M−1 cm−1; Figure 2). Another very intense band in the
UV range of 250−300 nm arising from the π−π* transition of
the phenyl groups was also present. The 467 nm band features
a prominent shoulder on the lower energy side, and, by
comparison to the spectrum of 2 (vide infra), likely consists of
two distinct charge transfer transitions. The 745 nm band is
thought to arise from a d−d transition. By comparison, despite
a similar coordination environment and metal oxidation state, 2
is dark green in a THF solution with absorption maxima at 437
(ε = 4537 M−1 cm−1) and 579 nm (ε = 2883 M−1 cm−1). These
bands are tentatively assigned as charge transfers involving the
Fe center and the benzene-1,2-dithiolate ligand.
(CO)(κ2 -bdt)] (2) for NP2
= methyl-2-{bis-
IR spectra of both 1 and 2 in dichloromethane consist of a
single peak in the CO-stretching region at 1918 or 1915 cm−1,
respectively (Figure 3A,B, gray traces). The energy of this
absorbance is comparable to that of related FeII complexes such
as (κ2-dppp)Fe(CO)(Cl2bdt) (dppp = diphenylphosphinopro-
(diphenylphosphinomethyl)amino}acetate, and electronic ex-
planations for the differing reactivities of the two pentacoordi-
nate complexes based on density functional theory (DFT)
calculations are considered.
8920
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
slow diffusion of hexane into a dichloromethane (1) or
chloroform (2) solution of complex. The two complexes
feature remarkably different geometries about the central iron
atom. Complex 2 is a distorted square pyramid (SP) with an
axial CO ligand, while complex 1 is a distorted trigonal
bipyramid (TBP) with CO, phosphorus, and sulfur in the
equatorial positions and phosphorus and sulfur in the apical
positions. The geometries of the complexes were further
characterized by Addison’s τ value, defined as τ = (β − α)/60,
for which β is the larger of the angles between the trans ligands
on the basal plane of a SP or the angle between the two axial
ligands of a TBP. The parameter α is defined as the smaller of
the angles between the trans ligands on the basal plane of a SP
or the larger of the basal angles of a TBP.35 For
pentacoordinate complexes, τ is a measure of the degree of
distortion from ideal SP (τ = 0) or ideal TBP (τ = 1) geometry.
The calculated value of τ is 0.099 for 2 (based on β(S1−Fe1−
P2) = 165.1427° and α(S2−Fe1−P1) = 159.2057°) and 0.721
for 1 (based on β(S2−Fe1−P1) = 171.7433° and α(S1−Fe1−
P2) = 128.4776°), corroborating the geometric assignments of
the complexes. As noted above, dppf has a much larger bite
angle (P−Fe−P = 101.17°) than that of NP2 (87.49°), and this
difference is likely responsible for the geometric differences
about the irons of 1 and 2. For complex 1, the three equatorial
ligands show a significant distortion from C3 symmetry with
bond angles of 134.58° (C−Fe−S), 128.48° (S−Fe−P), and
96.69° (P−Fe−C). The two axial ligands, thiolate and
phosphine, are also slightly distorted from a linear arrangement
with a S−Fe−P angle of 171.74°. We note that 1 is a
diamagnetic, formally FeII complex. Perfect TBP geometry does
not permit a diamagnetic ground state for a d6 metal, but the
observed distorted geometry is consistent with the S = 0
ground state.36,37 It is also worth noting that the coordination
geometry may be different in solution. Despite the metal
coordination geometry differences, the Fe−C and C−O bond
lengths of 1 and 2 are very similar to each other. This is
consistent with the similar ν(CO) stretching frequencies
observed for the two complexes. Additionally, as shown in
Table 2, the C−C bond lengths of the benzene-1,2-dithiolato
ligand show an alternating pattern of two shorter C−C bonds
(average 1.37 Å for 1 and 1.38 Å for 2) and four longer ones
(average 1.40 Å for 1 and 2) for both 1 and 2. Moreover, the
two C−S bonds are also not identical. The average C−S bond
lengths, 1.74 Å for 1 and 1.75 Å for 2, are slightly shorter than
typical bond lengths for the C−S single bonds (1.76−1.77 Å)
in benzene-1,2-dithiolate, suggesting that, in the metal-
locomplexes, the C−S bond orders are greater than 1.38−40
The observed distortions of the bdt ligand represent clear
evidence that it is partially oxidized and possesses substantial
1,2-dithiobenzosemiquinonate, π-radical character.40 Concom-
itantly, the charge of the Fe center in both complexes is
expected to be less than +2 (see Computational Studies below).
The distortion of the ligand is more obvious for 1, indicating
that the bdt ligand is more oxidized and the pentacoordinate Fe
center is more reduced in this complex than those in 2.
Figure 2. UV−vis spectra of (solid line) 1 and (dashed line) 2 in THF
at room temperature. Spectra were collected from THF solutions of
approximately 0.1 mM complex.
Figure 3. IR spectra in the presence and absence of CO for (A) 1 and
(B) 2 in the absence of acid. (C) Analogous spectra for 1 in the
presence of 3 equiv of HBF4. Black traces show the IR spectra after
bubbling CO through the solutions of the complexes. Gray traces
show the IR spectra (A and B) after purging the solution with nitrogen
to remove CO and (C) after the addition of NEt3. Spectra were
collected in CH2Cl2.
pane, Cl2bdt = 3,6-dichloro-1,2-benzenedithiolate).27 The
remarkable similarity of the CO-stretching frequencies for the
two complexes indicate that changing the bis-phosphine ligand
has little detectable impact on the Fe-CO bonding interactions.
Crystal Structures. The structures of 1 and 2 were
determined by single-crystal X-ray diffraction and are shown in
Figure 4. Selected bond distances and angles are given in Tables
1 and 2, and additional crystallographic information is available
in Table S1 (Supporting Information). Crystals were grown by
Reactivity toward CO. To investigate whether the open
coordination site on complexes 1 and 2 is accessible for external
ligand binding, we studied reactions of 1 and 2 with CO. Figure
3A shows the Fourier transform infrared spectroscopy (FTIR)
spectrum of a solution of 1 after it was saturated by bubbling
with CO for 10 min. In addition to the 1918 cm−1 signal of the
parent complex, two new CO stretching bands are observed at
1996 and 2020 cm−1, indicating the formation of an
8921
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
Figure 4. Molecular structures of (left) 1 and (right) 2 with thermal ellipsoids drawn at 50% probability level. Hydrogen atoms have been omitted
for clarity.
Table 1. Selected Bond Lengths (Å) and Bond Angles (deg)
for 1 and 2
for 2 was also observed via identification of new CO stretches
at 1995 and 2021 cm−1 (Figure 3B). Complex 2-CO also
reverted reversibly to 2 upon removal of CO. We note,
however, that the formation of 2-CO was less complete than
the formation of 1-CO. The complex 2-CO was not produced
in quantities sufficient to be detected by 31P NMR.
Furthermore, the ratios of the intensities of the CO stretching
bands in the IR spectrum also suggest that the majority of 2
remains unreacted. The difference in reactivities of the two
complexes may be attributable to the geometries about the iron
centers. Formation of a cis dicarbonyl from 2 will require a
substantial distortion with one of the extant ligands moving to a
position trans to a carbonyl. However, such a large rearrange-
ment is not required for the reaction of 1.
The addition of the strong acid HBF4·OEt2 facilitated a
quantitative reaction between 1 and CO. The newly formed
complex, [1(H)-CO]+, is red in solution and has characteristic
CO stretching vibrations at 2089 and 2043 cm−1 (Figure 3C).
Complex 1 could be reversibly regenerated from [1(H)-CO]+
by purging the solution with nitrogen and adding triethylamine.
An analogous reaction of 2 was not observed under the same
conditions. The relative intensities of and energy gap between
the two signals associated with CO stretching in the IR
spectrum of [1(H)-CO]+ are reminiscent of the well-
characterized cis,cis,cis-Fe(CO)2(dppe)(SPh)2, which has anal-
ogous peaks at 2017 and 1970 cm−1.41 Most notable is that the
stretching frequencies of [1(H)-CO]+ are 73 cm−1 higher in
energy compared to those of this reference compound. This
shift can be explained by the requirement for acid for this
reaction. Because bdt complexes of transition metals can have
mixed metal−ligand character frontier orbitals (vide infra), the
complex is likely protonated at one of the sulfurs, resulting in a
substantially weaker ligand and increasing the electrophilicity of
the metal center.42−44 The result is less back-bonding into the
π* LUMO of the CO and a stronger CO bond reflected in a
higher energy stretch. Such enhanced electrophilicity of a d6
metal ion on ligand protonation is not unprecedented.23,45
Hence, we postulate that the protonated dicarbonyl species is
cis,cis,cis-[Fe(CO)2(κ2-dppf)(bdt-H)]+.
bond lengths
1
2
Fe1−S1
Fe1−S2
Fe1−P1
Fe1−P2
Fe1−C41
C41−O41
bond angles
2.1719(7)
2.2243(7)
2.2405(7)
2.2241(7)
1.732(3)
1.162(3)
1
2.2007(12)
2.1767(12)
2.2222(12)
2.2249(12)
1.715(4)
1.154(5)
2
P2−Fe−P1
101.18(2)
89.21(2)
134.57(8)
88.52(8)
90.19(8)
96.69(8)
128.48(3)
171.74(3)
173.4(2)
87.49(4)
89.31(4)
S1−Fe1−S2
C41−Fe1−S1
C41−Fe1−S2
C41−Fe1−P1
C41−Fe1−P2
S1−Fe1−P2
S2−Fe1−P1
O1−C41−Fe1
101.30(14)
106.58(14)
94.11(14)
93.28(14)
165.14(5)
159.2(5)
176.7(4)
Table 2. Bond Distances (Å) within the Benzene-1,2-
dithiolate Ligand in Complexes 1 and 2
bond lengths
1
2
C1−C6
C1−C2
C2−C3
C3−C4
C4−C5
C5−C6
C1−S1
C2−S2
1.398(3)
1.404(3)
1.410(3)
1.365(3)
1.401(4)
1.380(3)
1.745(2)
1.735(2)
1.412(6)
1.386(6)
1.407(6)
1.373(6)
1.394(7)
1.385(6)
1.746(3)
1.757(4)
Fe(CO)2S2P2 complex, 1-CO. The presence of two new bands
indicates that the CO ligands are in a cis orientation. The 31P
NMR spectrum obtained under the same conditions also
includes both the 66.32 ppm resonance of the starting material
and a new signal at 62.59 ppm, providing additional evidence
for the formation of 1-CO. Following removal of CO from the
solution by purging with nitrogen, the signals associated with 1-
CO were no longer present. This demonstrates that binding of
external CO to 1 is a reversible process. The analogous reaction
Electrochemistry and Catalysis. The redox transitions of
the two complexes were probed by cyclic voltammetry.
Electrochemical analysis of 1 and 2 in 0.1 M [NBu4][PF6]/
8922
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
THF was performed under an argon atmosphere. As shown in
Figure 5, cyclic voltammograms of 1 show a reversible
Figure 5. Cyclic voltammograms of 1 (solid line; 0.48 mM) and 2
(dashed line; 1.5 mM) in 0.1 M [NBu4][PF6]/THF at a scan rate of
0.2 V s−1.
a
c
reduction at E1/2 = −1.65 V (ip /ip = 0.99, ΔEpc= 0.144 V)
a
and two reversible oxidations at E1/2 = +0.13 V (ip /ip = 0.90,
c
a
ΔEp= 0.140 V) and +0.44 V (ip /ip = 0.98, ΔEp = 0.137 V)
versus Fc+/0 (Fc = ferrocene). Controlled potential coulometry
shows that the reduction at −1.65 V is a one-electron process
(Figure S1, Supporting Information), and it likely corresponds
to the FeII/I couple of the pentacoordinate Fe center. We note
that a second, smaller reduction peak can be observed on close
inspection, which may indicate that a second, more complicated
reduction process also occurs to a limited extent. The
oxidations at +0.13 and +0.44 V can be assigned to the FeIII/II
couples for the Fe center in dppf and the pentacoordinate Fe
center. The oxidation wave at +0.13 V is likely associated with
the oxidation of the FeII center in dppf because quasi-reversible
oxidation of free dppf occurs at E1/2 = 0.183 V in 1,2-
dichloroethane.46,47 On the other hand, cyclic voltammograms
of 2 reveal that it undergoes a reversible, one-electron reduction
at E1/2 = −1.86 V (ip /ip = 0.96, ΔEp = 0.164 V) and an
irreversible one-electron oxidation at E1/2 = +0.32 V. By
analogy to 1, the reduction and oxidation waves are likely to be
FeII/I and FeIII/II couples, respectively. Notably, reduction of the
pentacoordinate FeII center in 2 occurs at more negative
potential than 1. More interestingly, in contrast to the two
reversible FeIII/II couples in 1, the FeIII/II couple in 2 is
irreversible, indicating that coordination of dppf leads to
improved stability of the oxidized complex in FeIIIFeII and
FeIIIFeIII states. In the case of 2, oxidation of the
pentacoordinate FeII center might be associated with a
geometry change, subsequent reaction with solvent molecules,
or ligand dissociation.
Figure 6. Cyclic voltammograms of 1 (top; 0.56 mM) and 2 (bottom;
1.25 mM) with various concentrations of acetic acid in 0.1 M
[NBu4][PF6]/THF at a scan rate of 0.2 V s−1. The acid concentrations
used are 0.2, 0.4, 0.6, 0.8, 1, 1.2, and 1.4 M for complex 1 and 0.05, 0.1,
0.2, 0.4, 0.6, 0.8, and 1 M for complex 2.
verified by gas chromatography (GC) analysis, and Faradaic
efficiency of hydrogen production was determined to be 93%
for 1 and 96% for 2. The overpotentials for proton reduction by
the two complexes, determined using the method reported by
Artero and co-workers, were relatively small, only 0.17−0.2 and
0.38−0.43 V for 1 and 2, respectively (Table S2, Supporting
Information).51 The half-wave potentials for the catalytic
current, used for the overpotential calculation, were determined
as the potential corresponding to the maximum value of (di/
dE), that is, the first derivative of the current data from the
cyclic voltammograms. For comparison, a mononuclear iron
complex {κ2-(Ph2PCH2N(X)CH2PPh2)}Fe(CO)(κ2-bdt) (X =
1,1-diethoxy-ethyl) similar to 2 was reported to reduce protons
from acetic acid in acetonitrile with overpotential in the range
of 0.23−0.27 V.27 The kinetics of proton reduction were
evaluated by considering the effect of catalyst concentration
and acid concentration on observed activity (Figures S3 and S4,
Supporting Information). Figure S3 (Supporting Information)
shows that catalytic peak current, icat, depends linearly on
catalyst concentration, [cat], for both 1 and 2. This
demonstrates a first-order dependence of the catalytic current
on the concentration of the catalyst at fixed acid concentrations
as described by eq 1, in which n is the number of electrons
involved in the catalytic reaction, A is the area of the electrode,
D is the diffusion coefficient of the catalyst, k is the rate
a
c
The electrocatalytic proton reduction activities of 1 and 2
were investigated in THF in the presence of acetic acid
(pKa(THF) = 24.42)48 and p-toluenesulfonic acid (p-TsOH). As
shown in Figure 6, sequential addition of acetic acid from 0.2 to
1.4 M renders the reduction wave for the FeII/I couple
irreversible and leads to an increase in current. This is
characteristic of electrocatalytic proton reduction, because
direct proton reduction from acetic acid at the glassy carbon
electrode is negligible in this potential range (Figure S2,
Supporting Information).49,50 Production of hydrogen gas was
8923
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
constant, and x is the order of the reaction with respect to
acid.52 Figure S4 (Supporting Information) shows that the ratio
of catalytic current to reductive peak current measured in the
absence of acid, icat/ip, is also linear with respect to acid
concentration for catalyst concentrations in the range of 0.74−
1.13 mM for 1 and 0.47−1.35 mM for 2. This indicates that the
reaction is second-order with respect to acid concentration as
described by eq 2, for a scan rate (υ) of 0.1 V s−1. At the highest
acid concentration investigated (1.6 M), a value of icat/ip of 35
was obtained for 2 (0.6 mM), corresponding to a turnover
frequency of 241 s−1. A much slower rate was observed for 1
(0.74 mM) with a turnover frequency of 10 s−1 in 1.8 M acetic
acid.
(0.28−0.83 mM) and increases linearly with increasing
concentration of p-TsOH (panel A of Figure S5, Supporting
Information). A similar result has been reported for the
mononuclear iron complex analogous to 2 incorporating a
different NP2 ligand, N,N-bis{(diphenylphosphino)methyl}-
2,2-diethoxyethanamine.27 Interestingly, for both 1 and 2, at
low concentration of p-TsOH (1−2 equiv of the catalyst), a
new reduction wave at 150 mV less negative potential is
observed together with the original FeII/I couple and grows with
increasing acid concentration (Figure S6, Supporting Informa-
tion). However, at higher acid concentration (more than 2
equiv), the two peaks merge and produce a single catalytic
wave. The observation that this new peak emerges for both
complexes 1 and 2 excludes the possibility that it is associated
with protonation at the amine group of the NP2 ligand in 2.
Association between either the sulfurs or the aromatic ring of
the bdt ligand and the proton/acid is, however, a distant
possibility.
icat = nFA[cat] D(k[acid]x)
(1)
⎛
⎜
⎞
⎟
n
icat/ip =
RT(k[acid]x )/Fν
⎝
⎠
(2)
0.4463
Computational Studies. To complement the experimental
results, DFT calculations were carried out on 1, 2, 1(H)+,
2(H)+, [1(H)-CO]+, and [2(H)-CO]+. Calculations were
performed using the B3LYP hybrid functional and the 6-
31G* basis set, and the DFT-optimized structures, which agree
well with the corresponding crystal structures for complexes 1
and 2, were confirmed as energy minima. A detailed
comparison of calculated and experimental metric parameters
can be found in Table S3 (Supporting Information). The
calculated frontier molecular orbitals for the complexes 1, 2,
1(H)+, and 2(H)+ are shown in Figure 8. Similarly, Figure 9
shows the frontier molecular orbitals for [1(H)-CO]+ and
[2(H)-CO]+. Complexes 1 and 2 both have highest occupied
molecular orbitals (HOMOs) that are delocalized over the Fe
and much of the bdt ligand. In particular, the HOMOs are a
bonding combination of iron d orbitals and sulfur p(π) orbitals
as well as an antibonding combination of sulfur and the
adjacent carbon atoms, implying strong electron delocalization
over the iron and the bdt ligand (Table S5, Supporting
Information).53−55 It is worth noting, however, that from the
perspective of the frontier molecular orbitals, the Fe−S
interactions are not equivalent for the two complexes. The
HOMO of 1 includes interactions only between the Fe and the
axial sulfur, likely a result of the unusual geometry. On the
other hand, the HOMO of 2 includes substantial contributions
from both sulfurs of the bdt ligand. Furthermore, the HOMOs
of 1 and 2 bear considerable resemblance to the HOMO of the
free bdt2− ligand or the singly occupied molecular orbital
(SOMO) of the free ligand in the π-radical anion form
(bdt−•).56,57 From the molecular orbital approach, the metal
dithiolate interaction in both complexes can best be described
as resulting from transfer of electron density from the HOMO
of the bdt2− ligand to empty Fe d orbitals. On the other hand,
the orbital density profiles shown in Figure 8 and the
percentage orbital contribution given in Table S5 (Supporting
Information) indicate that the lowest unoccupied molecular
orbitals (LUMOs) are dominated primarily by contributions
from the Fe d orbitals with the sulfur and phosphorus atoms
playing a minor role. There is almost no contribution from the
rest of the ring structure to the LUMOs. This suggests that
reduction of the complexes results in substantial accumulation
of charge localized at the metal center yielding a highly basic
iron site for interaction with protons. Furthermore, the
significant iron character of the LUMOs (51 and 43% for 1
The electrocatalytic activities of the two complexes were also
studied in the presence of the stronger acid, p-toluenesulfonic
acid monohydrate (p-TsOH). Irreversible catalytic waves
corresponding to the reduction of protons are observed at
the potentials of the 1/1− and 2/2− couples (Figure 7). Figure
S9 (Supporting Information) shows voltammograms demon-
strating that direct reduction by the electrode surface requires
substantially higher overpotentials. The catalytic peak current is
largely independent of catalyst concentration (panel B of Figure
S5, Supporting Information) for 1 over the investigated range
Figure 7. Cyclic voltammograms of 1 (top; 0.83 mM) and 2 (bottom;
0.88 mM) in the presence of p-TsOH in 0.1 M [NBu4][PF6]/THF at
a scan rate of 0.2 V s−1. Acid concentrations are 2, 4, 6, 8, and 10 mM
for complex 1 and 1, 2, 3, 5, and 10 mM for complex 2.
8924
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
Figure 9. Electron density profiles of the HOMOs and LUMOs of the
cis and trans conformers of [1(H)-CO]+ and [2(H)-CO]+.
Figure 8. Electron density profiles of the HOMOs and LUMOs of 1,
1(H)+, 2, and 2(H)+.
as a protonation site. Calculations for 1(H)+ with the proton
localized on the Fe indicate that it is 13 kcal/mol higher in
energy than a 1(H)+ complex with a protonated thiolate.
Likewise, in the case of 2, an N-protonated species is 6.4 kcal/
mol higher in energy than the S-protonated species. Therefore,
the rest of the computational studies were carried out assuming
that protonation occurs exclusively at the thiolate sulfur. The
geometry optimized structure of 1(H)+ shows that protonation
results in only minor changes about the Fe center; in particular,
the τ value of 1(H)+, 0.65, is not significantly different from
that of 1, 0.72 (Table S3, Supporting Information). In contrast,
upon protonation, 2 undergoes considerable distortion from its
nearly square pyramidal geometry (τ = 0.09) to a hybrid of
square pyramidal and trigonal bipyramidal geometries (τ =
0.42). Furthermore, electron density profiles of the HOMO
and LUMO of 1(H)+ and 2(H)+ reveal stark differences that
are important for understanding their different reactivities
toward CO. The HOMO of 1(H)+ is localized entirely on the
ferrocene moiety, while the HOMO of 2(H)+ is delocalized
over the entirety of the bdt ligand with minimal contribution
from the Fe atom. The bonding pattern in the bdt ligand and
the orbital contributions in the HOMO suggest that in 2(H)+,
the bdt ligand is partially oxidized, resulting in a charge on the
Fe center of less than +2. The Mulliken charge decomposition
analysis also suggests that one of the thiolate sulfurs is
noticeably less negative in 2(H)+ than those in 2, and the
aromatic ring also carries less negative charge. On the other
and 2, respectively) is consistent with the abilities of these
complexes to reversibly bind CO.
Additionally, Mulliken charge decomposition analysis was
used to quantify charge transfer between various fragments
within each complex. The Mulliken charge decomposition
analysis values (Table S7, Supporting Information) are
consistent with the frontier molecular orbital bonding
description above. The ferrocenyl Fe of the phosphine ligand
of 1 has a charge in the narrow range of +0.46 to +0.48
throughout all complexes in which it is present, that is, it serves
as an internal standard for the value expected for Fe(II) using
this computational method. This value is very similar to the
+0.4 determined in a recent Lowdin population analysis.58 The
̈
catalytically active Fe always carries a less positive charge,
indicating a net transfer of charge from the ligands to the metal
to achieve lower than 2+ oxidation state.
To correlate the observed trends in the reactivity of the two
complexes with CO in the presence of acid, we also performed
DFT calculations for the protonated complexes, 1(H)+ and
2(H)+, assuming compositional integrity following protonation.
For complex 1, protonation is most likely to occur at either the
Fe center or the thiolate sulfur with the highest contribution to
the HOMO.59 The possibilities for 2 are more numerous
because both the sulfurs contribute to the HOMO, and it
features an amine group in the NP2 ligand that could also serve
8925
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
hand, the orbital analysis suggests that the addition of a proton
to 1 disrupts the electron delocalization between the bdt ligand
and the Fe center, reinstating the aromaticity of the C6H4 ring
of the protonated bdt ligand. Therefore, 1(H)+ is expected to
behave more like a typical coordinatively unsaturated FeII, d6
complex (i.e., to bind exogenous ligands). This difference in the
electronic structures of 1(H)+ and 2(H)+ is likely responsible
for the fact that protonation induced CO uptake is observed
only for 1.
both in understanding the reactivity of natural enzymes and
functional small molecule mimics and in constructing more
effective functional catalysts.
Two recent examples emphasize the importance of geometric
constraints on the production of efficient proton reduction
catalysts. First, assembly of an active hydrogenase by
incorporation of an inactive, inorganic model into the active
site of an apo-[FeFe]-hydrogenase highlights the importance of
geometric constraints from a protein in forcing unexpected
reactivity at metallocenters.60 Second, the [Ni(PPh2NPh)2]2+
Simple frontier orbital concepts also suggest that the
interaction of CO will be stronger with 1(H)+ than with
2(H)+. The LUMO of 1(H)+ is about 0.1 eV lower in energy
than that of 2(H)+. Because the HOMO of CO has an energy
of ∼−10.1 eV, it is expected to interact more strongly with
1(H)+. The addition of a CO molecule to either 1(H)+ or
2(H)+ requires a change to octahedral geometry to
accommodate the new ligand resulting in sp3d2 hybridization
catalyst of the DuBois group is approximately 2 orders of
magnitude faster than the closely related [Ni(PP2hNC H X )2]
cousins; this is likely due in part to the planarity at the nickel
center imposed by the different geometric requirements of the
ligands influencing the hydricity of the Ni−H bond.52
Complexes 1 and 2 offer another example in which the metal
geometry imposed by the steric constraints of the phosphine
ligand has a dramatic impact on resulting chemical properties.
Both computational studies and single crystal X-ray structures
suggest that the bdt ligand is partially oxidized in both
complexes and, as a result of the extensive π-overlap between
the metal and the ligand, the overall charge on the iron is less
than +2. However, the π-interaction between the iron and the
bdt in 1 is more flexible, allowing the complex to behave as
might be expected for an unsaturated Fe(II) complex and
undergo proton-induced CO uptake to produce an 18-electron
complex. The same unusual electronic environment also allows
1 to reduce protons with less overpotential than 2, albeit at
considerably reduced rates. Unfortunately, this classic trade-off
between rate and overpotential continues to be a problem for
synthetic proton reduction catalysts. Nonetheless, we anticipate
that this combination of unusual ligands geometry, together
with redox active ligands, may prove fruitful in developing
reversible catalysts for hydrogen production and oxidation that
function with minimal overpotentials.
2+
6
4 2
2
2
2
involving the dz and dx −y orbitals. Thus, the enhanced
2
2
2
contributions of the dz and dx −y orbitals to the LUMO of
1(H)+ clearly indicate that the addition of CO to 1(H)+ would
be more facile than to 2(H)+ (Table S5, Supporting
Information). Finally, it is worth noting that for related
pentacoordinate complexes, Ott and co-workers have predicted
that complexes in square pyramidal geometry should be more
reactive than those in trigonal bipyramidal geometry, especially
in the presence of bulky ligands.29 Our experimental results
suggest the opposite for these complexes, but it is important to
remember that geometries observed in the solid state may be
different from those found in solution.
Calculations of the CO bound protonated complexes, [1(H)-
CO]+ and [2(H)-CO]+, show that trans attachment of the
second CO molecule is more favorable than cis attachment by
1.4 and 1.0 kcal/mol, respectively. However, in all reactions, the
cis complex was detected via FTIR spectroscopy. This may
indicate that the reaction is under kinetic control under these
experimental conditions. It is worth noting that in all cases (i.e.,
both cis and trans attachment), charge decomposition analysis
suggests that carbonylation results in a significant increase in
electron density at Fe1. In large part, this electron density
comes at the expense of that in the CO ligands (Table S7,
Supporting Information). The structural data also suggests a
lengthening of the Fe−C bonds and a shortening of the CO
bonds accompanies attachment of the second CO (Table S3,
Supporting Information). These computational results are
consistent with the FTIR experiments, demonstrating a shift
of the CO stretches to higher energy for complexes [1(H)-
CO]+ and [2(H)-CO]+.
EXPERIMENTAL SECTION
■
All reactions were carried out under an atmosphere of nitrogen using
standard Schlenk and vacuum-line techniques unless otherwise noted.
Anhydrous dichloromethane and methanol were purchased from
Sigma-Aldrich, and deuterated solvents were purchased from Cam-
bridge Isotope Laboratories. Tetrahydrofuran was dried by distilling
overnight over sodium and benzophenone. All starting materials were
obtained commercially and used without further purification. H, 13C,
1
and 31P NMR spectra were recorded at room temperature on a Varian
1
Liquid-State NMR spectrometer (400 or 500 MHz for H). NMR
chemical shifts are quoted in ppm; spectra were referenced to
1
tetramethylsilane for H and 13C NMR. The 31P NMR spectra were
referenced to external phosphoric acid at 0 ppm. FTIR spectra were
recorded on a Bruker VERTEX 70 spectrophotometer using a stainless
steel sealed liquid spectrophotometer cell with CaF2 windows. UV−vis
measurements were performed on a Hewlett-Packard 8453 spec-
trophotometer using quartz cuvettes with a 1 cm path length.
CONCLUSION
■
In summary, we have synthesized two pentacoordinate
FeII(CO)P2S2 complexes using benzene-1,2-dithiol and two
different chelating bis-phosphine ligands: NP2 and dppf.
Although the electronic properties of both phosphines are
comparable and the resulting complexes might be expected to
be similar, the differing geometrical constraints of the two
phosphines result in complexes with dramatically different
reactivity. In the solid state, in contrast to the SP complex
formed with NP2, the wider bite angle of dppf results in
formation of a TBP complex. These geometric differences lead
to significant changes in both the electronic and the chemical
properties of the complexes, including reactivity toward CO,
reduction potentials, electrocatalytic activity, and energies of
charge-transfer bands. These observations may prove important
Synthesis of Methyl 2-(Bis(diphenylphosphinomethyl)-
amino)acetate, NP2. To an anaerobic solution of formaldehyde
(37 wt % in water; 1 mL, 12.3 mmol) in absolute ethanol (10 mL) was
added diphenylphosphine (1.9 mL, 10.9 mmol) dropwise under argon.
The reaction mixture was stirred at room temperature for 30 min
followed by the addition of glycine methyl ester hydrochloride (0.7 g,
5.6 mmol) in 40% aqueous ethanol (5 mL). The cloudy reaction
mixture became clear upon stirring for 2 h. Volatile materials were
removed under reduced pressure to afford a colorless oily residue. The
crude product was purified via column chromatography on silica with
hexane/ethyl acetate/triethylamine (66:33:1) as eluent to afford NP2
as a colorless oil. Yield: 2.4 g (85%). Rf = 0.85 (1:1 hexane/ethyl
1
acetate, 1% NEt3). H NMR (400 MHz, CDCl3): δ = 7.38 (m, 8H),
8926
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
7.26 (m, 12H), 3.81 (s, 2H), 3.7 (d, 4H), 3.61 (s, 3H). 13C{1H} NMR
(100 MHz, CDCl3): δ = 171.21, 137.58, 137.46, 133.09, 132.90,
128.57, 128.37, 128.34, 128.30, 58.02, 55.65, 51.28. 13P{1H} NMR
(161.8 MHz, CDCl3): δ = −27.19.
Electrochemistry. Electrochemical experiments were carried out
using either a CHI 1200A electrochemical analyzer or a PG-STAT
128N Autolab electrochemical analyzer. A conventional three-
electrode cell was used for recording cyclic voltammograms. The
working electrode was a 3 mm diameter glassy carbon disk polished
with 1 mm and 0.3 mm deagglomerated alpha alumina, successively,
and sonicated for 15 min in ultrapure water prior to use. The
supporting electrolyte was [NBu4][PF6] (0.1 M in THF). The Ag/Ag+
reference electrode was prepared by immersing a silver wire anodized
with AgCl in a THF solution of 0.1 M [NBu4][PF6]. A platinum wire
was used as counter electrode. Deaeration of the solutions was
performed by bubbling argon through the solution for 15 min, after
which an atmosphere of Ar was maintained during the course of
electrochemical measurements. All potentials are reported relative to
the ferrocene couple (Fc+/Fc) as reference. Concentrations of the
complexes were determined spectrophotometrically on the basis of the
following extinction coefficients: ε(467 nm) = 4433 M−1 cm−1 and
ε(437 nm) = 4537 M−1 cm−1 for 1 and 2, respectively.
Bulk electrolysis experiments were undertaken in a sealed BASi bulk
electrolysis cell under an Ar atmosphere. The working electrode was a
reticulated vitreous carbon electrode (cylinder of 40 mm diameter,
50 mm height, and 5 mm depth). A nonaqueous (THF) Ag/AgCl
reference electrode was placed in a separate compartment and
connected via a fine porosity glass frit. A platinum wire was used as
counter electrode. The electrochemical cell was sealed following
15 min of deaeration with argon. For determination of the quantity of
hyrogen produced, samples were removed from the headspace of the
cell via a Hamilton locking gastight syringe, and an equal volume of
argon was concomitantly added. Following calibration with H2 samples
of known concentration, a Varian CP-3800 gas chromatograph
(thermal conductivity detector, Alltech Porapak Q 80/100 column,
Ar as carrier gas) was used to separate the headspace samples and
quantify the amount of hydrogen present.
Computational Details. DFT calculations were carried out using
the Becke gradient−corrected exchange functional and Lee−Yang−
Parr correlation functional with three parameters (B3LYP) and the 6-
31G* basis set.59,61−69 This level of theory has been found to yield
reliable geometries and vibrational frequencies for a number of first-
row transition metal systems.70−74 Nonetheless, in light of recent
studies indicating the improved performance of the BP86 and TPSS
functionals in describing transition metal containing systems, the
geometries and energies of 2 were also calculated using these
functionals and the larger TZVPP basis sets.75−77 These results are
shown in Table S4 (Supporting Information) and indicate that the
B3LYP/6-31G* level of theory is reliable for the systems investigated
in this study. The “overlap population” parameter listed in Tables S5
and S6 (Supporting Information) is a measure of the nature of the
interaction between the orbitals involved. Thus, a positive overlap
population represents a bonding interaction, a negative overlap
population corresponds to an antibonding interaction, and a zero
overlap population indicates no bonding between the fragments.
Synthesis of (dppf)Fe(CO)(bdt), 1. To an anaerobic solution of
anhydrous FeCl2 (60.3 mg, 0.48 mmol) in anhydrous methanol
(8 mL) was added 1,1′-bis(diphenylphosphino)ferrocene (280 mg,
0.5 mmol) in THF (4 mL) dropwise under a CO atmosphere. After
the reaction mixture stirred for 30 min at room temperature, a solution
of benzene-1,2-dithiol (0.07 mL, 0.6 mmol) and triethylamine
(0.17 mL, 1.2 mmol) in methanol (3 mL) was added dropwise. The
reaction mixture turned to violet and then to dark brown. After the
mixture stirred for 1 h at room temperature, the solvent was removed
under reduced pressure, and the residue was purified via column
chromatography on silica with hexane/dichloromethane (1:2) as
eluent. The product was obtained as dark brown powder. Yield:
1
244 mg (65%). Rf = 0.4 (1:1 hexane/CH2Cl2). H NMR (400 MHz,
CD2Cl2): δ = 8.04 (dd, J = 6.0, 3.2 Hz, 2H), 7.59 (m, 4H), 7.41−7.36
(m, 8H), 7.18−7.10 (m, 10H), 4.89 (s, 2H), 4.48 (s, 2H) 4.24 (s, 2H),
4.12 (s, 2H). 13C{1H} NMR(100 MHz, CD2Cl2): δ = 134.98, 132.71,
130.39, 129.50, 129.32, 127.76, 127.41, 121.38, 76.96, 74.92, 74.79,
72.60, 71.40. 31P{1H} NMR (161.8 MHz, CD2Cl2): δ = 66.32. IR
(CH2Cl2, cm−1): ν(CO) 1918. APCI mass spectrum (positive mode):
751.0198 [(M − CO + H)+].
Synthesis of (NP2)Fe(CO)(bdt), 2. To an anaerobic solution of
anhydrous FeCl2 (75 mg, 0.6 mmol) and NP2 ligand (264 mg,
0.56 mmol) in anhydrous methanol (11 mL) under a CO atmosphere
were added benzene-1,2-dithiol (0.07 mL, 0.6 mmol) and triethyl-
amine (0.17 mL, 1.2 mmol). The color of the solution turned black.
After the mixture was stirred at room temperature for 2 h, the solvent
was removed under reduced pressure, and the residue was purified via
column chromatography on silica with hexane/ethyl acetate (4:1) as
eluent. The desired compound was obtained as a green solid. Yield:
1
240 mg (60%). Rf = 0.3 (3:1 hexane/ethyl acetate). H NMR (400
MHz, CD2Cl2): δ = 7.99 (dd, J = 6, 3.2 Hz 2H), 7.64 (m, 4H), 7.50
(m, 6H), 7.15 (m, 6H), 7.09 (dd, J = 6, 3.2 Hz, 2H), 6.96 (t, J = 7.6
Hz, 4H), 4.05−3.97 (m, 2H), 3.86−3.80 (m, 2H), 3.70 (s, 3H), 3.59
(s, 2H). 13C NMR (100 MHz, CD2Cl2): δ = 133.68 (t), 132.74 (t),
130.46 (s), 129.83 (s), 129.05 (s), 128.42 (t), 127.82 (t), 121.34 (s),
62.00 (t), 56.73 (s) 56.57 (s), 51.64 (s). 31P NMR (161.8 MHz): δ =
50.21. IR (CH2Cl2, cm−1): ν(CO) 1915. APCI mass spectrum
(positive mode): m/z = 710.0799 [(M + H)+], 682.0787 [(M − CO +
H)+].
Reaction of 1 with CO in the Presence of HBF4.OEt2. A
solution of 1 (3.2 mg, 4.1 μmol) in CH2Cl2 (2 mL) was saturated with
CO and HBF4·OEt2 (0.2 mL 0.074 M solution in CH2Cl2, 14.3 μmol,
3.6 equiv) was added dropwise to the reaction mixture. After the
solution stirred for 10 min at room temperature, the color of the
solution changed from dark brown to red. Formation of the CO
adduct [1(H)-CO]+ was indicated by IR and 31P NMR. The addition
of triethylamine (0.04 mL, 28.6 μmol) to the reaction mixture
followed by purging with N2 led to the release of CO and quantitative
regeneration of 1. IR (CH2Cl2): ν(CO) 2089, 2043. 31P {1H} NMR
(161.8 MHz, CD2Cl2): δ = 4.92.
X-ray Crystallography. A representative crystal of each
compound was individually mounted on the end of a thin glass fiber
using Apiezon type N grease and optically centered. Cell parameter
measurements and single-crystal X-ray diffraction data collection were
performed at low temperature (123 K) with a Bruker Smart APEX
diffractometer. Graphite monochromated Mo Kα radiation (λ =
0.71073 Å) in the ω−φ scanning mode was used for the
measurements. The structure was solved by direct methods and
refined by full-matrix least-squares on F2. The following programs were
used: data collection, Bruker Instrument Service v2010.9.0.0; cell
refinement and data reduction, SAINT V7.68A; structure solution and
refinement, SHELXS-97; molecular graphics, XShell v6.3.1; prepara-
tion of material for publication, Bruker APEX2 v2010.9-1.30. Details of
crystal data and parameters for data collection and refinement are
listed in Table S1 (Supporting Information).
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional electrochemical data and 1H NMR spectra. Selected
X-ray crystal structural data. Description of the method for
determining electrocatalytic overpotential. Summary of struc-
tural data, orbital contributions and overlap populations
calculated via DFT and, where appropriate, compared to X-
ray analysis. Benchmark calculations. Mulliken charge decom-
position analysis values. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: jonesak@asu.edu. Tel.: 480-965-0356.
8927
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
(26) Rauchfuss, T. B.; Contakes, S. M.; Hsu, S. C. N.; Reynolds, M.
A.; Wilson, S. R. J. Am. Chem. Soc. 2001, 123, 6933−6934.
(27) Beyler, M.; Ezzaher, S.; Karnahl, M.; Santoni, M.-P.; Lomoth, R.;
Ott, S. Chem. Commun. 2011, 47, 11662−11664.
(28) Gardner, J. M.; Beyler, M.; Karnahl, M.; Tschierlei, S.; Ott, S.;
Hammarstrom, L. J. Am. Chem. Soc. 2012, 134, 19322−19325.
(29) Orthaber, A.; Karnahl, M.; Tschierlei, S.; Streich, D.; Stein, M.;
Ott, S. Dalton Trans. 2014, 43, 4537−4549.
(30) Bandoli, G.; Dolmella, A. Coord. Chem. Rev. 2000, 209, 161−
196.
Present Address
§Dr. Souvik Roy, Laboratoire de Chimie et Biologie des
́
Metaux, CEA-Grenoble, 17 rue des Martyrs, 38054 Grenoble
Cedex 9, France.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported through the Center for Bio-
Inspired Solar Fuel Production, 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-SC0001016.
(31) Bandoli, G.; Trovo, G.; Dolmella, A.; Longato, B. Inorg. Chem.
1992, 31, 45−51.
(32) Butler, I. R.; Cullen, W. R.; Kim, T. J.; Rettig, S. J.; Trotter, J.
Organometallics 1985, 4, 972−980.
(33) Longato, B.; Pilloni, G.; Valle, G.; Corain, B. Inorg. Chem. 1988,
27, 956−958.
REFERENCES
(34) Durran, S. E.; Elsegood, M. R. J.; Hawkins, N.; Smith, M. B.;
Talib, S. Tetrahedron Lett. 2003, 44, 5255−5257.
■
(1) Armstrong, F. A. Science 2013, 339, 658−659.
(2) Vignais, P. M.; Billoud, B. Chem. Rev. 2007, 107, 4206−4272.
(3) Tard, C.; Pickett, C. J. Chem. Rev. 2009, 109, 2245−2274.
(4) Volbeda, A.; Charon, M.-H.; Piras, C.; Hatchikian, E. C.; Frey,
M.; Fontecilla-Camps, J. C. Nature 1995, 373, 580−587.
(5) Peters, J. W.; Lanzilotta, W. N.; Lemon, B. J.; Seefeldt, L. C.
Science 1998, 282, 1853−1858.
(35) Addison, A. W.; Rao, T. N.; Reedijk, J.; van Rijn, J.; Verschoor,
G. C. J. Chem. Soc., Dalton Trans. 1984, 1349−1356.
(36) Nguyen, D. H.; Hsu, H.-F.; Millar, M.; Koch, S. A.; Achim, C.;
Bominaar, E. L.; Munck, E. J. Am. Chem. Soc. 1996, 118, 8963−8964.
(37) Hsu, H.-F.; Koch, S. A.; Popescu, C. V.; Munck, E. J. Am. Chem.
̈
Soc. 1997, 119, 8371−8372.
(38) Ray, K.; Begum, A.; Weyhermuller, T.; Piligkos, S.; van Slageren,
J.; Neese, F.; Wieghardt, K. J. Am. Chem. Soc. 2005, 127, 4403−4415.
(39) Ray, K.; Bill, E.; Weyhermuller, T.; Wieghardt, K. J. Am. Chem.
Soc. 2005, 127, 5641−5654.
(6) Nicolet, Y.; Piras, C.; Legrand, P.; Hatchikian, C. E.; Fontecilla-
Camps, J. C. Struct. Folding Des. 1999, 7, 13−23.
(7) Happe, R. P.; Roseboom, W.; Pierik, A. J.; Albracht, S. P. J.;
Bagley, K. A. Nature 1997, 385, 126−126.
(8) Evans, D. J.; Pickett, C. J. Chem. Soc. Rev. 2003, 32, 268−275.
(9) Jones, A. K.; Sillery, E.; Albracht, S. P. J.; Armstrong, F. A. Chem.
Commun. 2002, 866−867.
(40) Ray, K.; Weyhermuller, T.; Goossens, A.; Craje, M. W. J.;
Wieghardt, K. Inorg. Chem. 2003, 42, 4082−4087.
(41) Takacs, J.; Marko, L.; Parkanyi, L. J. Organomet. Chem. 1989,
361, 109−116.
(42) Waters, T.; Wang, X.-B.; Woo, H.-K.; Wang, L.-S. Inorg. Chem.
2006, 45, 5841−5851.
(43) Benedito, F. L.; Petrenko, T.; Bill, E.; Weyhermuller, T.;
Wieghardt, K. Inorg. Chem. 2009, 48, 10913−10925.
(44) Solis, B. H.; Hammes-Schiffer, S. J. Am. Chem. Soc. 2012, 134,
15253−15256.
(45) Heiden, Z. M.; Gorecki, B. J.; Rauchfuss, T. B. Organometallics
2008, 27, 1542−1549.
(46) Corain, B.; Longato, B.; Favero, G.; Ajo, D.; Pilloni, G.; Russo,
U.; Kreissl, F. R. Inorg. Chim. Acta 1989, 157, 259−266.
(47) Pilloni, G.; Longato, B.; Corain, B. J. Organomet. Chem. 1991,
420, 57−65.
(48) Barron, D.; Buti, S.; Ruiz, M.; Barbosa, J. Phys. Chem. Chem.
Phys. 1999, 1, 295−298.
(49) Felton, G. A. N.; Glass, R. S.; Lichtenberger, D. L.; Evans, D. H.
Inorg. Chem. 2006, 45, 9181−9184.
(50) Felton, G. A. N.; Vannucci, A. K.; Chen, J.; Lockett, L. T.;
Okumura, N.; Petro, B. J.; Zakai, U. I.; Evans, D. H.; Glass, R. S.;
Lichtenberger, D. L. J. Am. Chem. Soc. 2007, 129, 12521−12530.
(51) Fourmond, V.; Jacques, P.-A.; Fontecave, M.; Artero, V. Inorg.
Chem. 2010, 49, 10338−10347.
(52) Helm, M. L.; Stewart, M. P.; Bullock, R. M.; DuBois, M. R.;
DuBois, D. L. Science 2011, 333, 863−866.
(53) Glassey, W. V.; Hoffmann, R. J. Chem. Phys. 2000, 113, 1698−
1704.
(10) Madden, C.; Vaughn, M. D.; Diez-Perez, I.; Brown, K. A.; King,
P. W.; Gust, D.; Moore, A. L.; Moore, T. A. J. Am. Chem. Soc. 2012,
134, 1577−1582.
(11) 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, 2681−2699.
(12) Gloaguen, F.; Rauchfuss, T. B. Chem. Soc. Rev. 2009, 38, 100−
108.
(13) Stoian, S. A.; Hsieh, C.-H.; Singleton, M. L.; Casuras, A. F.;
Darensbourg, M. Y.; McNeely, K.; Sweely, K.; Popescu, C. V. J. Biol.
Inorg. Chem. 2013, 18, 609−622.
(14) Wang, W.; Rauchfuss, T. B.; Moore, C. E.; Rheingold, A. L.; De
Gioia, L.; Zampella, G. Chem.Eur. J. 2013, 19, 15476−15479.
(15) Hsieh, C. H.; Erdem, O. F.; Harman, S. D.; Singleton, M. L.;
Reijerse, E.; Lubitz, W.; Popescu, C. V.; Reibenspies, J. H.; Brothers, S.
M.; Hall, M. B.; Darensbourg, M. Y. J. Am. Chem. Soc. 2012, 134,
13089−13102.
(16) Darensbourg, M. Y.; Lyon, E. J.; Zhao, X.; Georgakaki, I. P. Proc.
Natl. Acad. Sci. U.S.A. 2003, 100, 3683−3688.
(17) Siegbahn, P. E. M.; Tye, J. W.; Hall, M. B. Chem. Rev. 2007, 107,
4414−4435.
(18) Armstrong, F. A. Curr. Opin. Chem. Biol. 2004, 8, 133−140.
(19) Darensbourg, D. J.; Reibenspies, J. H.; Lai, C.-H.; Lee, W.-Z.;
Darensbourg, M. Y. J. Am. Chem. Soc. 1997, 119, 7903−7904.
(20) Hsu, H.-F.; Koch, S. A.; Popescu, C. V.; Munck, E. J. Am. Chem.
̈
Soc. 1997, 119, 8371−8372.
(21) Lai, C.-H.; Lee, W.-Z.; Miller, M. L.; Reibenspies, J. H.;
Darensbourg, D. J.; Darensbourg, M. Y. J. Am. Chem. Soc. 1998, 120,
10103−10114.
(54) Hoffmann, R. Solids and Surfaces: A Chemist’s View of Bonding in
Extended Structures; VCH Publishers: New York, 1988.
(55) Gorelsky, S. I.; Lever, A. B. P. J. Organomet. Chem. 2001, 635,
187−196.
(22) Liaw, W.-F.; Lee, N.-H.; Chen, C.-H.; Lee, C.-M.; Lee, G.-H.;
Peng, S.-M. J. Am. Chem. Soc. 2000, 122, 488−494.
(23) Liu, T.; Li, B.; Popescu, C. V.; Bilko, A.; Perez, L. M.; Hall, M.
́
(56) Ray, K.; Weyhermuller, T.; Neese, F.; Wieghardt, K. Inorg.
̈
Chem. 2005, 44, 5345−5360.
B.; Darensbourg, M. Y. Chem.Eur. J. 2010, 16, 3083−3089.
(24) Shima, S.; Pilak, O.; Vogt, S.; Schick, M.; Stagni, M. S.; Meyer-
Klaucke, W.; Warkentin, E.; Thauer, R. K.; Ermler, U. Science 2008,
321, 572−575.
(57) Ray, K.; Petrenko, T.; Wieghardt, K.; Neese, F. Dalton Trans.
2007, 1552−1566.
(58) Mutter, S. T.; Platts, J. A. Chem.Eur. J. 2010, 16, 5391−5399.
(59) Liu, Y.-C.; Chu, K.-T.; Jhang, R.-L.; Lee, G.-H.; Chiang, M.-H.
Chem. Commun. 2013, 49, 4743−4745.
(25) Sellmann, D.; Kleine-Kleffmann, U.; Zapf, L.; Huttner, G.;
Zsolnai, L. J. Organomet. Chem. 1984, 263, 321−331.
8928
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929

Inorganic Chemistry
Article
(60) Berggren, G.; Adamska, A.; Lambertz, C.; Simmons, T.;
Esselborn, J.; Atta, M.; Gambarelli, S.; Mouesca, J. M.; Reijerse, E.;
Lubitz, W.; Happe, T.; Artero, V.; Fontecave, M. Nature 2013, 499,
66−69.
(61) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon,
M. S.; Defrees, D. J.; Pople, J. A. J. Chem. Phys. 1982, 77, 3654−3665.
(62) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785−
789.
(63) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652.
(64) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J.
Phys. Chem. 1994, 98, 11623−11627.
(65) Rassolov, V. A.; Pople, J. A.; Ratner, M. A.; Windus, T. L. J.
Chem. Phys. 1998, 109, 1223−1229.
(66) Neese, F. WIREs Comput. Mol. Sci. 2012, 2, 73−78.
(67) Harihara, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213−
222.
(68) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery Jr., J. A.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.;
Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Persson,
G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.;
Hasegawa, J.; Ishida, M.; Niakajima, T.; Honda, Y.; Kitao, O.; Nakai,
H.; Klene, M.; Li, X.; Konox, J. E.; Hratchian, H. P.; Cross, J. B.;
Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.;
Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J.
J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.;
Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman,
J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.;
Stefanov, B. B.; Liu, G.; Liashenki, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision
C.02; Gaussian, Inc.: Wallingford, CT, 2004.
(69) Gorelsky, S. I. AOMix program, http://www.sg-chem.net/.
(70) Tarakeshwar, P.; Finkelstein-Shapiro, D.; Hurst, S. J.; Rajh, T.;
Mujica, V. J. Phys. Chem. C 2011, 115, 8994−9004.
(71) Tarakeshwar, P.; Kumar, T. J. D.; Balakrishnan, N. J. Phys. Chem.
A 2008, 112, 2846−2854.
(72) Imbert, C.; Hratchian, H. P.; Lanznaster, M.; Heeg, M. J.;
Hryhorczuk, L. M.; McGarvey, B. R.; Schlegel, H. B.; Verani, C. N.
Inorg. Chem. 2005, 44, 7414−7422.
(73) Greene, S. N.; Richards, N. G. J. Inorg. Chem. 2006, 45, 17−36.
(74) Bruschi, M.; Zampella, G.; Fantucci, P.; De Gioia, L. Coord.
Chem. Rev. 2005, 249, 1620−1640.
(75) Ye, S. F.; Neese, F. Inorg. Chem. 2010, 49, 772−774.
(76) Brothers, S. M.; Darensbourg, M. Y.; Hall, M. B. Inorg. Chem.
2011, 50, 8532−8540.
(77) Wu, H.; Hall, M. B. C. R. Chim. 2008, 11, 790−804.
8929
dx.doi.org/10.1021/ic5012988 | Inorg. Chem. 2014, 53, 8919−8929