Direct probing of photoinduced electron transfer in a self-assembled biomimetic [2Fe2S]-hydrogenase complex using ultrafast vibrational spectroscopy

Article abstract of DOI:10.1021/ic500777d

A pyridyl-functionalized diiron dithiolate complex, [μ-(4-pyCH ₂-NMI-S₂)Fe₂(CO)₆] (3, py = pyridine (ligand), NMI = naphthalene monoimide) was synthesized and fully characterized. In the presence of zinc tetraphenylporphyrin (ZnTPP), a self-assembled 3·ZnTPP complex was readily formed in CH₂Cl₂ by the coordination of the pyridyl nitrogen to the porphyrin zinc center. Ultrafast photoinduced electron transfer from excited ZnTPP to complex 3 in the supramolecular assembly was observed in real time by monitoring the (C≡O) and (C=O)_NMI spectral changes with femtosecond time-resolved infrared (TRIR) spectroscopy. We have confirmed that photoinduced charge separation produced the monoreduced species by comparing the time-resolved IR spectra with the conventional IR spectra of 3^?- generated by reversible electrochemical reduction. The lifetimes for the charge separation and charge recombination processes were found to be τ_CS = 40 ± 3 ps and τ_CR = 205 ± 14 ps, respectively. The charge recombination is much slower than that in an analogous covalent complex, demonstrating the potential of a supramolecular approach to extend the lifetime of the charge-separated state in photocatalytic complexes. The observed vibrational frequency shifts provide a very sensitive probe of the delocalization of the electron-spin density over the different parts of the Fe₂S₂ complex. The TR and spectro-electrochemical IR spectra, electron paramagnetic resonance spectra, and density functional theory calculations all show that the spin density in 3^?- is delocalized over the diiron core and the NMI bridge. This delocalization explains why the complex exhibits low catalytic dihydrogen production even though it features a very efficient photoinduced electron transfer. The ultrafast porphyrin-to-NMI-S₂-Fe₂(CO)₆ photoinduced electron transfer is the first reported example of a supramolecular Fe₂S₂-hydrogenase model studied by femtosecond TRIR spectroscopy. Our results show that TRIR spectroscopy is a powerful tool to investigate photoinduced electron transfer in potential dihydrogen-producing catalytic complexes, and that way to optimize their performance by rational approaches.

Full text of DOI:10.1021/ic500777d

Article
pubs.acs.org/IC
Terms of Use
Direct Probing of Photoinduced Electron Transfer in a Self-
Assembled Biomimetic [2Fe2S]-Hydrogenase Complex Using
Ultrafast Vibrational Spectroscopy
Ping Li,†,⊥ Saeed Amirjalayer,‡,⊥ Frantisek Hartl,*,§ Martin Lutz,|| Bas de Bruin,† Rene Becker,†
̌
́
Sander Woutersen,*,‡ and Joost N. H. Reek*,†
†Homogeneous & Supramolecular Catalysis, van ’t Hoff Institute for Molecular Sciences, University of Amsterdam, Science Park 904,
1098 XH Amsterdam, The Netherlands
‡Molecular Photonics, van ’t Hoff Institute for Molecular Sciences, University of Amsterdam, Science Park 904, 1098 XH Amsterdam,
The Netherlands
§Department of Chemistry, University of Reading, Whiteknights, Reading RG6 6AD, United Kingdom
||Crystal and Structural Chemistry, Bijvoet Center for Biomolecular Research, Utrecht University, Padualaan 8, 3584 CH Utrecht, The
Netherlands
S
* Supporting Information
ABSTRACT: A pyridyl-functionalized diiron dithiolate com-
plex, [μ-(4-pyCH2−NMI-S2)Fe2(CO)6] (3, py = pyridine
(ligand), NMI = naphthalene monoimide) was synthesized
and fully characterized. In the presence of zinc tetraphenylpor-
phyrin (ZnTPP), a self-assembled 3·ZnTPP complex was
readily formed in CH2Cl2 by the coordination of the pyridyl
nitrogen to the porphyrin zinc center. Ultrafast photoinduced
electron transfer from excited ZnTPP to complex 3 in the
supramolecular assembly was observed in real time by
monitoring the ν(CO) and ν(CO)NMI spectral changes
with femtosecond time-resolved infrared (TRIR) spectroscopy.
We have confirmed that photoinduced charge separation
produced the monoreduced species by comparing the time-resolved IR spectra with the conventional IR spectra of 3•−
generated by reversible electrochemical reduction. The lifetimes for the charge separation and charge recombination processes
were found to be τCS = 40 3 ps and τCR = 205 14 ps, respectively. The charge recombination is much slower than that in an
analogous covalent complex, demonstrating the potential of a supramolecular approach to extend the lifetime of the charge-
separated state in photocatalytic complexes. The observed vibrational frequency shifts provide a very sensitive probe of the
delocalization of the electron-spin density over the different parts of the Fe2S2 complex. The TR and spectro-electrochemical IR
spectra, electron paramagnetic resonance spectra, and density functional theory calculations all show that the spin density in 3•−
is delocalized over the diiron core and the NMI bridge. This delocalization explains why the complex exhibits low catalytic
dihydrogen production even though it features a very efficient photoinduced electron transfer. The ultrafast porphyrin-to-NMI-
S2−Fe2(CO)6 photoinduced electron transfer is the first reported example of a supramolecular Fe2S2-hydrogenase model studied
by femtosecond TRIR spectroscopy. Our results show that TRIR spectroscopy is a powerful tool to investigate photoinduced
electron transfer in potential dihydrogen-producing catalytic complexes, and that way to optimize their performance by rational
approaches.
achieve effective electronic communication between the
chromophore and the catalytic site, namely, (a) covalent
linkage, (b) multicomponent systems, and (c) supramolecular
assemblies. Quenching of the excited state in the first family of
covalently bound Ru(terpy)2-Fe2S2 systems was attributed to
fast energy transfer.5,6 Moreover, a reductive quenching by the
electron-rich diiron center was encountered for the Ru2+
INTRODUCTION
■
Artificial mimics of the active site of [FeFe]-hydrogenases have
attracted increasing attention in the past 15 years,1 since the
enzyme structure was deduced from X-ray crystallography.2 To
date, hundreds of diiron dithiolate model complexes have been
explored as electroactive catalysts for proton reduction along
homo- or heterogeneous catalytic paths.3 More recently,
photocatalytic hydrogen production based on Fe2S2 complexes
in a combination with a light-harvesting chromophore has been
demonstrated.4 Three strategies have been developed to
Received: April 3, 2014
Published: April 26, 2014
© 2014 American Chemical Society
5373
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
Chart 1. Schematic Molecular Structures of Selected Diiron Hydrogenase Model Complexes
photosensitizer side-attached via a phosphine ligand.7 In
contrast, an intermolecular electron transfer from photo
reduced [Ru(bpy)3]+ to [(μ-SCH2)2XFe2(CO)6] (X = CH2
or NCH2Ph) was proven with transient absorption (TA)
spectroscopy.8 More recently, the formation of a Fe0FeI species
via one-electron transfer from the photogenerated organic
transfer from Fc to ZnTPP•+.13b These reports have introduced
an ideal model system to study the ultrafast electron transfer
process closely related to photocatalytic dihydrogen produc-
tion. However, the formation of the reduced diiron dithiolate
species was only monitored by transient electronic absorption
with a maximum at 616 nm, most likely corresponding to the
reduced NMI-S2−Fe0FeI(CO)6 species.
̇
radical, Et2NCHCH3, to [(μ-pdt)Fe2(CO)5(PMe3)] (pdt =
propane-1,3-dithiolate), was confirmed by quenching experi-
ments monitored in situ by electron paramagnetic resonance
(EPR) spectroscopy.9
On the other hand, IR spectroscopy has been established as a
powerful tool for the direct characterization of active
intermediates formed during electrochemical redox reactions
and protonation of diiron dithiolate complexes in the ground
state,15 that is, without photoexcitation of a light harvesting
chromophore component in the assembly, benefiting from the
high sensitivity of CO stretching modes to changes in the π-
back-donation from the substituted diiron core. For example, a
mixed-valence paramagnetic FeIFeII hydride-bridged species has
recently been characterized by IR spectroelectrochemistry.16 To
obtain a detailed understanding of photoinduced electron
transfer in potential dihydrogen-producing catalytic complexes,
one ideally would like to combine the high sensitivity of the IR
response with the subpicosecond time resolution that has
already been achieved in the visible wavelength region. To the
best of our knowledge, the electron transfer process from a
photoexcited chromophore to a diiron hydrogenase model
complex has yet not been investigated by transient IR
spectroscopy (TRIR). Here we report such an ultrafast TRIR
spectroscopic study of photoinduced electron transfer in a self-
assembled Fe2S2·ZnTPP system, Fe2S2 standing for [μ-(4-
pyCH2−NMI-S2)Fe2(CO)6] (3, Chart 1), wherein the ZnTPP
chromophore in the selectively populated singlet excited state
transfers an electron to the Fe2S2 complex unit. The photo
reduced species, 3•−, was also characterized separately by IR
and ultraviolet−visible (UV−vis) spectroelectrochemistry, EPR
spectroscopy and density functional theory (DFT) calculations.
Furthermore, we investigated the influence of carbonyl
substitution at the diiron center on its redox properties with
[μ-(4-pyCH2−NMI-S2)Fe2(CO)4(dppv)] (3-dppv, dppv = cis-
1,2-bis(diphenylphosphino)ethylene; Chart 1). This study
shows for the first time by TRIR the rate of electron transfer
from the photoexcited state of the chromophore to the [FeFe]-
Hydrogenase model complex, and that the spin density is
located mainly on the organic NMI ligand rather than on the
metal, thereby explaining the adverse effect on the light-driven
proton reduction.
The first example of a self-assembled porphyrin-diiron
complex, namely, [(μ-SCH2)2NC(O)C5H4N]Fe2(CO)6·
ZnTPP, was reported by Song and co-workers.10
A
luminescence quenching efficiency up to 78% within this self-
assembly supported the proposed electron transfer. Two years
later, the first evidence from time-resolved (TR) spectroscopy
was obtained for the electron transfer from excited ZnTPP to
[{(μ-SCH2)2N(CH2CH2OOCPy)}Fe2(CO)6] in a noncova-
lent assembly. The diiron core photoreduction was verified by
using nanosecond flash photolysis in the Sun’s laboratory.11
Shortly afterward, our group reported a family of self-assembled
Fe2S2−PPh2(pyridyl)-zinc porphyrin complexes that exhibited
photocatalytic activity, notably in the presence of two different
porphyrin chromophores. Infrared (IR) spectroscopic monitor-
ing of the photoreaction has revealed that a disubstituted
catalyst, namely, [(μ-pdt)Fe2(CO)4(PPh2Py)2], was formed via
disproportionation of the photo reduced parent pentacarbonyl
complex.12
More recently, a photodriven ultrafast intramolecular
electron transfer from photoexcited zinc porphyrin to a
hydrogenase-model diiron dithiolate complex was thoroughly
investigated by Wasielewski and co-workers in covalently linked
zinc porphyrin-NMI-S2−Fe2(CO)6 (NMI = naphthalene
monoimide) with femtosecond TA spectroscopy.13 The study
yielded lifetimes for the charge separation and recombination
processes, namely, τCS = 24
1 ps and τCR = 57
1 ps
(CH2Cl2), respectively. Compared to the NMI-lacking
reference complex, that is, [(μ-naphdt)Fe2(CO)6] (1, naphdt
= naphthalene-1,8-dithiolate)14 (Chart 1), the incorporation of
the electron-withdrawing NMI group in [μ-(tol-NMI-S2)-
Fe2(CO)6] (tol = toluene) (2; Chart 1) positively shifted the
first reduction potential by ca. 0.4 V and made the electron
transfer from zinc porphyrin (ZnP) to the diiron part in 2-ZnP
thermodynamically favorable (ΔG = −0.63 V).13a Notably, the
charge recombination process slowed down to τCR = 67 2 ns
by incorporating a ferrocenyl (Fc) substituent on a phenyl
group of ZnTPP, leading to a subsequent second electron
5374
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
isosbestic point at 555 nm and gave the association constant
Kass = 1.2 × 104 mol−1 dm3 (Supporting Information, Figure
S4).
As shown in Figure 2, the luminescence of ZnTPP at 598 and
644 nm was strongly quenched (up to 87%) upon addition of
RESULTS AND DISCUSSION
■
Self-Assembly Study of 3·ZnTPP. Pyridin-4-yl-function-
alized complex 3 was readily prepared in a moderate yield
(50%) by treating the disulfide ligand, 4-pyCH2−NMI-S2 (L,
see Supporting Information, Figure S1 for its X-ray crystal
structure), with 2 equiv of [Fe2(CO)9] in tetrahydrofuran
(THF) at room temperature (RT) for 2 h. Complex 3 was fully
characterized by 1H NMR, Fourier transform (FT) IR
spectroscopies, high-resolution mass spectrometry (HR-MS),
and X-ray crystal structure determination (Figure 1). In
Figure 2. Luminescence quenching observed along the titration of
ZnTPP with [μ-(4-pyCH2−NMI-S2)Fe2(CO)6] (3) in CH2Cl2.
Excitation wavelength: 555 nm. Concentrations: [ZnTPP] = 8.0 ×
10−5 mol dm−3; [3] = 0 (red) to ca. 1.0 × 10−3 mol dm−3 (purple).
ca. 13 equiv of complex 3 to the ZnTPP solution in CH2Cl2.
The Stern−Volmer constant KSV = 6.2 × 103 mol−1 dm3 was
calculated from the linear dependence of F(0)/F (emission
intensity before/after the addition of the quencher) on the
concentration of 3 in the corresponding Stern−Volmer plot
(Supporting Information, Figure S5). In a control experiment
using a pyridyl-free compound, [μ-(PhCH2−NMI-S2)-
Fe2(CO)6] (3a), instead of 3, a much smaller Stern−Volmer
constant KSV = 9.2 × 102 mol−1 dm3 was observed (Supporting
Information, Figure S6), the small residual quenching most
likely being due to occasional complexation (probably mediated
by π-stacking) between 3a and ZnTPP.
Figure 1. (left) Molecular structure of [(4-pyCH2−NMI-S2)-
Fe2(CO)6] (3) in the crystal, with displacement ellipsoids drawn at
the 50% probability level. Hydrogen atoms are omitted for clarity.
Selected bond lengths (Å), angles (deg) and torsion angles (deg):
Fe(1)−Fe(2), 2.50955(16); Fe(1)−S(1), 2.2361(2); Fe(1)−S(2),
2.2442(2); Fe(2)−S(1), 2.2402(2); Fe(2)−S(2), 2.2451(2); Fe(1)−
COap, 1.7894(8); Fe(1)−COba, 1.8057(9); Fe(2)−COap, 1.7993(10);
Fe(2)−COba, 1.8024(8); S(1)−Fe(1)−S(2), 85.225(8); S(1)−
Fe(2)−S(2), 85.107(8); S(1)−Fe(1)−Fe(2)−S(2), −109.494(9).
(right) DFT-optimized (BP86, def2-TZVP) structure of self-
assembled 3·ZnTPP.
Time-Resolved Infrared Spectroscopy, Infrared Spec-
troelectrochemistry, and DFT Calculations. Complex 3 in
CH2Cl2 displays two reversible one-electron reduction steps at
E
1/2 = −1.28 and −1.65 V (vs Fc/Fc+, Supporting Information,
CH2Cl2, complex 3 exhibits three characteristic ν(CO)
bands at 2080, 2046, and 2007 cm−1 (Supporting Information,
Figure S2), which are blue-shifted by ca. 2 cm−1 compared to
[μ-(tol-NMI-S2)Fe2(CO)6] (2; Chart 1).13 Two additional
weak absorption bands are found at 1706 and 1664 cm−1
belonging to ν(CO)NMI vibrations of the monoimide
carbonyls. Similar wavenumbers, namely, 1691 and 1653
cm−1, were also observed for free ligand L in CH2Cl2
(Supporting Information, Figure S2). The single-crystal X-ray
structure determination revealed the characteristic butterfly
Fe2S2 core of 3 featuring an Fe−Fe bond length of 2.50955(16)
Å.
To study the self-assembling process between complex 3 and
ZnTPP in solution, steady-state UV−vis and emission spectra
were recorded during the titration of 8.0 × 10−5 mol dm−3
ZnTPP in CH2Cl2 by complex 3. An apparent red shift of the
Q-bands of ZnTPP was observed upon the addition of complex
3 (Supporting Information, Figure S3), reflecting the axial
coordination of the pyridyl-N to the porphyrin zinc center. The
determined association constant Kass = 4.5 × 103 mol−1 dm3 is
comparable with those for the related examples.11,12 The
titration of ZnTPP with the free disulfide ligand (L) was
accompanied by similar red shift of Q-bands with a well-defined
Figure S7). The cathodic potentials are negatively shifted by ca.
0.14 V compared to the values reported for [μ-(tol-NMI-
S2)Fe2(CO)6] (2).13a In the DFT-optimized self-assembled 3·
ZnTPP (see Supporting Information, Figure S8 for frontier
orbitals), the driving force for the photodriven electron transfer
from photoexcited 1ZnTPP to complex 3 has been estimated as
−ΔGCS = 0.64 eV in CH2Cl2 (see the Supporting Information),
indicating that the charge separation is thermodynamically
feasible.
A femtosecond-to-picosecond TRIR study was conducted
with 2.0 mM 3 and 2.0 mM ZnTPP in CH2Cl2, wherein the
concentration of the dynamic self-assembly, 3·ZnTPP, is ca. 1.4
mol dm−3 according to the association constant calculated
above. ZnTPP was selectively photoexcited by a 553 nm laser
pulse. The difference absorbance transient IR spectra recorded
in the ν(CO) and ν(CO)NMI regions are shown in Figure
3a,b, respectively. The time resolution of the experiment was
∼200 fs.
1
Photodriven ultrafast electron transfer from excited ZnTPP
to complex 3 resulted in the gradual formation of one-electron
reduced 3•− (characterized separately by IR spectroelectro-
chemistry, see below). As shown in Figure 3a, the three positive
ν(CO) bands around 2052, 2019, and 1976 cm−1 of 3•−
5375
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
Figure 3. (top) TRIR spectra recorded in different wavenumber regions (a) ν(CO) and (b) ν(CO)NMI for 2.0 mM [μ-(4-pyCH2−NMI-
S2)Fe2(CO)6] (3) and 2.0 mM ZnTPP in CH2Cl2 at 293 K, following the 553 nm fs laser excitation. The insets show the corresponding kinetic
traces at 1976 and 2008 cm−1 (a), and at 1628 and 1668 cm−1 (b), respectively. The curves through the data points are the result of a global least-
squares fit. (bottom) Absorbance difference IR SEC spectra recorded in different wavenumber regions (c) ν(CO) and (d) ν(CO)NMI during
the one-electron reduction of 1.0 mM 3 in CH2Cl2 (0.1 M nBu4NPF6) within an OTTLE cell at 293 K (see Supporting Information, Figure S9 for
the thin-layer cyclic voltammogram response).
arose simultaneously with the bleaching of the ground state
absorption of the parent compound around 2075, 2044, and
2008 cm−1. The IR pattern remained as three well-defined
peaks after the charge separation, indicating that no significant
geometry change occurred at the diiron dithiolate hexacarbonyl
moiety during its photo reduction. The maximum ground state
bleach was achieved in ca. 100 ps, followed by the decay of the
charge separated state. The zero absorbance line was restored
within 1.5 ns, indicating that the ground state was fully
recovered by the charge recombination. Global kinetics curve
fitting yielded the lifetimes for the charge separation and charge
acceptor is different, because of the additional methylene
linker; (2) the relative orientations of the photosensitizer and
the electron acceptor are significantly different in the covalently
linked and supramolecular complex. In 2-ZnP, the porphyrin
ring is approximately parallel to the NMI symmetry axis,
whereas in 3·ZnTPP it is at an angle of ∼22°. Both the distance
and the relative orientation of the electron donor and acceptor
have a strong influence on the electron-transfer rate. Addition-
ally, the coordination sphere of the zinc atom is different in the
two structures as the coordination number of Zn changed from
4 to 5 due to axial coordination of the pyridine in 3·ZnTPP,
which could also change the driving force for the electron
transfer. To test the influence of the axial-pyridine ligand on the
redox potentials of ZnTPP, cyclic voltammetry of ZnTPP (1.0
mM) was conducted in the absence and presence of 4-
ethylpyridine (10 mM; see Supporting Information, Figure
S10). The energy level analysis (Supporting Information,
recombination as τCS = 40
3 ps and τCR = 205
14 ps,
respectively. Compared to the covalently bonded model 2-
ZnP,13a both processes, but in particular the charge
recombination (from 57 to 205 ps), were significantly slowed
down in 3·ZnTPP, which was most likely caused by two effects:
(1) the bond distance between the photosensitizer and
5376
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
Figure S11) revealed that the coordination of the pyridine to
ZnTPP increases the driving force for charge separation while it
decreases the driving force for charge recombination. This is in
line with the lifetimes observed experimentally, explaining the
stronger inhibiting effect being for the charge recombination
rather than for the charge separation.
In contrast to the supramolecular system, the control TRIR
experiment with 2.0 mM [μ-(PhCH2−NMI-S2)Fe2(CO)6] (3a)
(the pyridyl-lacking analogue of 3, Chart 1) and 2.0 mM
ZnTPP under identical conditions revealed no electron transfer
(Supporting Information, Figure S12). All these results indicate
that the photoinduced electron transfer requires the formation
of the self-assembly between the pyridyl-N and the porphyrin
zinc center. To confirm this, we performed control TRIR
experiment with 2.0 mM 3 in the absence of ZnTPP, in which
we observed an ultrafast spectral change within 1 ps, which
decayed within ca. 200 ps while the maximum intensity of the
transient signal reached only ca. 5% compared to that obtained
for the 3·ZnTPP assembly (see Supporting Information, Figure
S13 for a direct comparison). These small transient signals are
not associated with the one-electron reduction of 3, but with
CO dissociation and reorganization of the FeFe bond.17,18
Notably, the ν(CO) frequencies of the carbonyls at the
diiron center shifted in average by ca. 27 cm−1 to lower energy
upon the one-electron transfer from the photoexcited
porphyrin to complex 3. At the same time, the ν(CO)NMI
frequencies of the monoimide group shifted from 1705 and
1668 to 1628 cm−1, as shown in Figure 3b, in line with the
dominant photoreduction of the monoimide. The IR wave-
number changes obtained with TRIR provide an important
information for analyzing the electronic structure of the
reduced species.
In the ground state, the reduction of several diiron dithiolate
complexes was studied previously by IR spectroelectrochemis-
try (IR SEC) to characterize the unstable one-electron reduced
core, Fe0FeI.3b,15b,19 For a direct comparison with the TRIR
spectra obtained above, we carried out the electrochemical
reduction of complex 3 in CH2Cl2 within an OTTLE cell.
The IR spectrum of stable electro-generated 3•− (Figure
3c,d) matched well that of the photoproduct encountered in
the TRIR experiments, directly confirming that the one-
electron transfer from photoexcited ZnTPP to complex 3 took
place during the TRIR measurement. Upon electrochemical
reduction, the carbonyl stretching frequencies of the diiron core
and imide group both shifted to lower energy, by 28 cm−1 for
ν(CO) and 80 cm−1 for ν(CO)NMI (the second band
accompanying that at 1629 cm−1 was observed at 1583 cm−1 for
3•−, see Supporting Information, Figure S14), respectively.
Importantly, the red shift (Δνav(CO) = 28 cm−1) of the
carbonyls on iron was much smaller compared to values
reported for the one-electron reduction FeIFeI → Fe0FeI in
hexacarbonyl complexes, for example, [(μ-S2C3H6)Fe2(CO)6]
(Δνav(CO) ≈ 70 cm−1),15b [(μ-SEt)2Fe2(CO)6] (Δνav(C
O) ≈ 76 cm−1)3b and [(μ-bpdt)Fe2(CO)6] (bpdt = biphenyl-
2,2′-dithiolate, Δνav(CO) ≈ 80 cm−1).19 The reported 70−
80 cm−1 red shift corresponds to the one-electron reduction of
FeFe core, strongly suggesting the dominant reduction of the
NMI group over the diiron center in complex 3 during both
photoinduced and electrochemical one-electron reduction. The
electron density at the π-conjugated NMI bridge in 3•− is in
line with the DFT-calculated spin density plot shown in Figure
4. A comparable localization of spin density on the non-
innocent quinone bridge in [(μ-S2 C1 0 H4 O2 )-
Figure 4. Spin density plot (BP86, def2-TZVP) for [μ-(4-pyCH2−
NMI-S2)Fe2(CO)6]•− (3•−).
Fe2(CO)4(PPh3)2]•− was recently reported by Glass and co-
workers20 during the preparation of this manuscript. In a
previous report, very small spin density delocalization over the
diiron core (Δνav(CO) ≈ 5 cm−1) was observed for the one-
electron-reduced [(μ-bdt)Fe2(CO)5(C60(H)PPh2)]•− bearing a
fullerene-functionalized ligand.21 The spin density localized in
the diiron orbital in 3•− was calculated as 0.23 e−1. In the DFT-
optimized 3•−, the Fe−Fe distance is 2.557 Å, that is ca. 0.05 Å
longer than that of 3 in the solid state, in line with the electron
density distributed over the diiron center.
IR Spectroelectrochemical and DFT Study of Com-
pound 1. As a reference complex bearing a naphthalene ring
without monoimide group, [(μ-naphdt)Fe2(CO)6] (naphdt =
naphthalene-1,8-dithiolate, 1),14 was used for the IR SEC
monitoring of the reduction under identical conditions (Figure
5, top). The ν(CO) red shift (Δνav(CO) ≈ 79 cm−1)
observed for the reduction of 1 to 1•− closely resembles the
reported examples outlined above, being indeed much larger
than that observed for complex 3, where the acceptor NMI
group strongly determines the cathodic path. In the DFT-
optimized 1•−, the Fe−Fe bond is considerably weakened and
the distance reduced to 2.81 Å, which is 0.3 Å longer compared
to 1 in the solid state.14 The spin density localized on the diiron
center is 0.75 e−1 (Figure 5, bottom), being equally distributed
over two iron atoms. The dominant reduction of the Fe−Fe
bond in 1, which is different from the actual reduction of an
Fe−S bond in [(μ-bpdt)Fe2(CO)6] (bpdt = biphenyl-2,2′-
dithiolate),19 was further supported by EPR spectroscopy and
theoretical calculations discussed below in the EPR spectros-
copy section.
The spin density at the Fe2 centers in 1•− and 3•− exhibits a
nice correlation with the IR ν(CO) shifts, in agreement with
the reduction of different moieties (the different lowest
unoccupied molecular orbitals characters) revealed by the
theoretical calculations. Figure 6 shows the accurate linear
correlation between the experimentally measured average low-
energy IR shifts of the terminal carbonyls, Δνav(CO),
reflecting the one-electron reduction, and the theoretically
calculated spin density on the Fe2 core; the data for 3-dppv
5377
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
and 1 equiv of Cp2Co (Cp = cyclopentadienyl) (E1/2 = −1.33 V
vs Fc/Fc+)22 gave no redox reaction as revealed by IR
spectroscopy (temperature- and solvent-dependent reduction
of 3 by Cp2Co was observed, see Figures S19−22 and
corresponding discussion in the Supporting Information), and
the yellow color of parent 3 was maintained. In contrast, the
solution color turned deep blue immediately after the addition
of 1 equiv of Cp*2Co (E1/2 = −1.94 V vs Fc/Fc+)22 to 3; IR
spectroscopy at the same time confirmed the formation of 3•−.
In addition, strong absorption at 602 and 790 nm was observed
for 3•− by UV−vis SEC experiment (Supporting Information,
Figure S20), in agreement with the color turning blue. The
freshly prepared blue solution of 3•− was used for EPR
spectroscopic measurement at RT. As shown in Figure 7 (top)
Figure 5. (top) Absorbance difference IR SEC spectra recorded during
the one-electron reduction of 1.0 mM [(μ-naphdt)Fe2(CO)6] (1) in
CH2Cl2 (0.1 M nBu4NPF6) within an OTTLE cell at 293 K. The
dashed purple line was extracted to show the IR absorption of 1•−
,
whereof especially the ν(CO) band at 2008 cm−1 that was largely
hidden in the resulting difference spectra. (bottom) Spin density plot
(BP86, def2-TZVP) for [(μ-naphdt)Fe2(CO)6]•− (1•−).
Figure 7. (top) Experimental and simulated X-band EPR spectrum of
[μ-(4-pyCH2−NMI-S2)Fe2(CO)6]•− (3•−). Experimental conditions:
T = 293 K, microwave power of 0.2 mW, field modulation amplitude
of 0.1 G, modulation frequency of 9.383 319 GHz. The sample was
prepared by chemical reduction of 3 with 1 equiv of Cp*2Co in
toluene at 293 K in a glovebox. (bottom) Experimental and simulated
X-band EPR spectrum of [(μ-naphdt)Fe2(CO)6]•− (1•−). Exper-
imental conditions: T = 10 K, microwave power of 2.0 mW, field
modulation amplitude of 4 G, modulation frequency of 9.364 522
GHz. The sample was prepared by chemical reduction of 1 with 1
equiv of Cp*2Co in dichloromethane at 195 K under Ar.
Figure 6. Linear correlation between Δν(CO) and the spin density
population at the Fe2 core. Values for IR shifts were determined from
the IR SEC monitoring of the reduction of [(μ-naphdt)Fe2(CO)6] (1,
green), [μ-(4-pyCH2−NMI-S2)Fe2(CO)6] (3, blue), and [μ-(4-
pyCH2−NMI-S2)Fe2(CO)4(dppv)] (3-dppv, red).
(Supporting Information, Figure S15) and zero point were
included in the linear fitting. The ν(CO) values serve as a
convenient quantitative probe to evaluate the degree of the
reduction/spin density localization on the Fe2 center in these
diiron hexacarbonyl complexes. The full set of DFT-calculated
data for the frontier orbitals and spin density plots for
complexes 1, 3, and 3-dppv can be found in the Supporting
Information (Figures S16−18).
a well-defined, relatively sharp isotropic EPR signal lacking a
hyperfine structure was detected at 293 K. The signal reveals an
isotropic g-value of 2.013, in a good agreement with the DFT-
calculated giso value of 2.012. This is in line with the DFT-
calculated electronic structure showing that the unpaired
electron of 3•− mainly located in an NMI-centered orbital.
EPR Spectroscopy. In toluene at 293 K, mixing of [μ-(4-
pyCH2−NMI-S2)Fe2(CO)6] (3) (E1/2 = −1.28 V vs Fc/Fc+)
5378
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
Table 1. IR Spectroscopic Data for 1, 3,3a, 3-dppv and Their Radical Anions
a
b
b
compound
[(μ-naphdt)Fe2(CO)6] (1)
[(μ-naphdt)Fe2(CO)6]•− (1•−
ν(CO)
Δνav(CO)
ν(CO)NMI
Δνav(CO)NMI
2074, 2038, 1999
2008, 1950, 1915
2080, 2046, 2007
2054, 2018, 1976
2079, 2045, 2006
2051, 2016, 1977
2031, 1964, 1920
2005, 1934, 1898
)
79
28
28
[μ-(4-pyCH2−NMI-S2)Fe2(CO)6] (3)
[μ-(4-pyCH2−NMI-S2)Fe2(CO)6]•− (3•−
[μ-(PhCH2−NMI-S2)Fe2(CO)6] (3a)
1706, 1666
1629, 1583
1703, 1663
1627, 1582
1699, 1659
1616, 1573
)
80
79
84
[μ-(PhCH2−NMI-S2)Fe2(CO)6]•− (3a•−
)
[μ-(4-pyCH2−NMI-S2)Fe2(CO)4(dppv)] (3-dppv)
[μ-(4-pyCH2−NMI-S2)Fe2(CO)4(dppv)]•− [(3-dppv)•−
]
26
b
a
Experiments were conducted in CH2Cl2, data for radical anions were extracted from IR SEC. The average IR bands shift during the one-electron
reduction of the corresponding neutral compound.
Attempts to record an isotropic EPR spectrum of [(μ-
naphdt)Fe2(CO)6]•− (1•−) in solution were not successful. At
10 K, a rather broad near isotropic signal was detected at g =
2.0045, together with a weak signal (g ≈ 2.027) stemming from
an unknown impurity (Figure 7, bottom). The signal of 1•− is
much broader than that of 3•−, and spectral simulation points
to a small rhombicity of the g-tensor (g11 = 2.0066, g22 = 2.0045,
summarized the IR data for the compounds related to this
study.
Single-crystal X-ray diffraction (Figure 8, left) reveals that the
Fe−Fe bond in 3-dppv is significantly lengthened to 2.5510(5)
g33 = 2.0000; gav = 2.0045). The DFT-calculated g-tensor (g11
=
2.0066, g22 = 2.0045, g33 = 2.0017; gav = 2.0043) corresponds
quite well with the measurement, giving support to the
electronic structure presented in Figure 5, showing that the
unpaired electron is mainly located at the diiron center of the
mixed-valent Fe0FeI system.
Photoinduced H2 Production. TRIR spectroscopy
documented the photoinduced electron transfer from ZnTPP
to complex 3, which can be the first step in the catalytic cycle of
H2 formation. Photogeneration of molecular hydrogen was
attempted by irradiating complex 3 (4 μmol) with a 500 W Xe
lamp (λ ≤ 530 nm was removed by a cutoff filter to prevent
direct photoexcitation and consequent degradation of the Fe2S2
complex) in the presence of the light harvesting chromophore
(ZnTPP, 4 μmol), a proton source (HOAc, 22 μmol) and a
bulky sacrificial electron donor (iPr2NEt, 22 μmol) in toluene
(5 mL). Only a very small amount (ca. 5 μL) of dihydrogen was
detected by an online gas chromatograph system (Supporting
Information, Figure S23). The formation of molecular
hydrogen was likely due to a noncatalytic photolysis process.
The spin density in 3•− largely delocalized over the NMI group
after the electron transfer to the Fe2S2 complex, prevented the
following reaction with protons. This behavior is consistent
with the noncatalytic nature of the cathodic cyclic voltammo-
gram of 3 recorded in the presence of acetic acid (Supporting
Information, Figure S24).
Figure 8. (left) Molecular structure of [μ-(4-pyCH2−NMI-S2)-
Fe2(CO)4(dppv)] (3-dppv) in the crystal with displacement ellipsoids
drawn at the 50% probability level. Disordered CH2Cl2 solvent
molecules, hydrogen atoms and phenyl ellipsoids on dppv have been
omitted for clarity. Selected bond lengths (Å) and angles (deg):
Fe(1)−Fe(2), 2.5510(5); Fe(1)−S(11), 2.2331(6); Fe(1)−S(21),
2.2270(5); Fe(2)−S(11), 2.2631(6); Fe(2)−S(21), 2.2594(6);
Fe(2)−COap, 1.795(3); Fe(2)−COba, 1.782(3); Fe(1)−P(12),
2.1767(6); Fe(1)−P(22), 2.2159(6); Fe(1)−COba, 1.754(2);
S(21)−Fe(2)−S(11), 83.68(2); S(21)−Fe(1)−S(11), 85.12(2);
S(11)−Fe(2)−Fe(1)−S(21), 109.40(3). (right) Spin density plot
CO Substitution with dppv in 3. In an attempt to make
the Fe2S2 complexes more reactive toward protonation, a CO
substitution reaction was carried out to make the iron centers
more electron rich. Complex [μ-(4-pyCH2−NMI-S2)-
(BP86, def2-TZVP) for (3-dppv)•−
.
Å from 2.50955(16) Å in 3 due to the electron donating ligand,
dppv, indicating the electron rich character of the Fe−Fe bond
in 3-dppv (see Supporting Information, Figure S18 for the
frontier orbitals). As a result, the average Fe−COba bond was
shortened from 1.8057(9) and 1.8024(8) in 3 to 1.754(2) Å in
3-dppv by the enhanced π back-donating effect. The rotation of
the chelating dppv ligand in solution was proven by variable
temperature (VT)-NMR measurements (Supporting Informa-
tion, Figure S25). As shown in Figure 8 (right), the dppv
substitution imposed slight changes in the spin density
distribution in (3-dppv)•− (0.21 e−1 on Fe2), which is further
supported by the IR SEC result (Δν(CO) = 26 cm−1, shown
Fe2 (CO)4 (dppv)] (3-dppv, dppv
= cis-1,2-bis-
(diphenylphosphino)ethylene) was readily obtained (86%
yield) by the reaction of complex 3 with 1 equiv of dppv in
CH2Cl2 at RT. Complex 3-dppv displays three ν(CO) bands
at 2030, 1962, 1920 cm−1, and two ν(CO)NMI bands at 1697
and 1658 cm−1 (Supporting Information, Figure S15 and Table
1). The substitution of two carbonyl ligands with chelating
dppv leads to an average shift of ν(CO) bands by 73 cm−1 to
lower energy, which is comparable with the literature data
reported for [(μ-pdt)Fe2(CO)4(dppv)] and [(μ-edt)-
Fe2(CO)4(dppv)] by Rauchfuss and co-workers.23 Table 1
5379
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
in Supporting Information, Figure S15). The correlation
between spin density distribution and IR shift for (3-dppv)•−
was incorporated in the linear fitting shown in Figure 6.
A steady state fluorescence titration of ZnTPP with complex
3-dppv revealed the formation of self-assembled 3-dppv·ZnTPP
with an association constant as Kass = 5.5 × 103 mol−1 dm3
(Supporting Information, Figure S26), which is comparable to
that for 3·ZnTPP. The Stern−Volmer quenching constant was
calculated to be KSV = 1.6 × 104 mol−1 dm3, strongly suggesting
that the ZnTPP was quenched by the electron transfer to 3-
dppv (Supporting Information, Figure S27).
The first reduction potential of 3-dppv was found at −1.48 V,
that is, negatively shifted by ca. 200 mV compared to that for
parent complex 3 (Table 2). The potential shift caused by the
Table 2. Cathodic Potentials of 3, 3a, and 3-dppv
Figure 9. Experimental and simulated X-band EPR spectrum of [μ-(4-
pyCH2−NMI-S2)Fe2(CO)4(dppv)]•− [(3-dppv)•−]. Experimental
conditions: T = 293 K, microwave power of 0.2 mW, field modulation
amplitude of 0.1 G, modulation frequency of 9.380 295 GHz. The
sample was prepared by chemical reduction of 3 with 1 equiv of
Cp*2Co in toluene at RT in a glovebox.
1
2
E1/2 (V),
E1/2 (V),
(ΔEp (mV))
a
a
compound
(ΔEp (mV))
b
[(μ-naphdt)Fe2(CO)6] (1)
−1.67, (−)
−2.00, (−)
−1.65, (84)
[μ-(4-pyCH2−NMI-S2)Fe2(CO)6]
(3)
−1.28, (92)
−1.29, (93)
−1.48, (88)
[μ-(PhCH2−NMI-S2)Fe2(CO)6]
(3a)
−1.65, (102)
Table 3. Experimental and DFT-Calculated EPR Parameters
of (3-dppv)•−
c
[μ-(4-pyCH2−NMI-S2)
−2.02
Fe2(CO)4(dppv)] (3-dppv)
a
b
exp.
DFT
a
All potentials are reported vs Fc/Fc+ used as an internal standard.
Reference 14. Epc for irreversible reduction.
g-value
2.0140
2.0121
b
c
c
c
c
AH1
13.5
7.3
iso
c
iso
c
c
AH2
6.5
1.6
AP1
AP2
6.0
9.5
4.7
9.6
dppv ligation is smaller than that expected for a chelating
diphosphine ligand, while comparable to that resulted from the
CO-to-PPh3 monosubstitution (the first reduction potential
negatively shifted by 210 mV) in [(μ-pdt)Fe2(CO)6].24 The
spin density in (3-dppv)•−, again mainly distributed over the
NMI group, explains the decreased effect of the dppv/CO
substitution on the redox potentials. Cyclic voltammetry in the
presence of HOAc showed no catalytic proton reduction
activity for 3-dppv around its first reduction wave (−1.48 V)
(Supporting Information, Figure S28). A slightly larger amount
of H2 (20 μL, turnover number (TON) ≈ 0.2) was detected in
the presence of 3-dppv (Supporting Information, Figure S23),
which proved to be more stable than 3 on irradiation under the
same conditions. The activity (up to 0.2 TON) achieved by this
assembly is significantly lower compared to the previously
reported efficient photodriven H2 evolution systems based on
earth-abundant metal catalysts.4,25
iso
iso
a
b
On the basis of spectral simulation; see Figure 9. Orca, B3LYP, def2-
c
TZVP. Two equivalent protons.
model in a self-assembled system. The lifetimes of the charge
separation (CS) and charge recombination (CR) processes
were calculated from curve fitting as τCS = 40 3 ps and τCR
=
205 14 ps, respectively. Compared to the covalently bound
model 2-ZnP, both processes but especially the charge
recombination were dramatically slowed down in the supra-
molecular assembly 3·ZnTPP, as a result of the difference in
relative orientation of the photosensitizer with respect to the
electron acceptor and the coordination of a pyridyl group to the
Zn(TPP). The electronic structure of reduced 3•− was studied
by IR spectroelectrochemistry (IR SEC), EPR spectroscopy,
and DFT calculations. The combined results reveal that the
NMI group is noninnocent during the first reduction of
complex 3, which explains the low activity in photodriven
hydrogen production. The current results suggest a clear
outlook for future directions in this research. New complexes
should be prepared with redox innocent bridges that tune the
reduction potentials of Fe2S2 mimics properly while avoiding
delocalization of the spin density. Such systems are anticipated
to lead to active bioinspired Fe2S2 catalysts for photodriven
hydrogen evolution. This study shows that TRIR is a powerful
tool for the characterization of the ultrafast electron transfer
process between the photoexcited chromophore and the Fe2S2
hydrogenase model.
The EPR spectrum of (3-dppv)•−, generated from 3-dppv via
one-electron reduction with 1 equiv of Cp*Co in toluene
solution as 293 K, reveals an isotropic signal with a g-value of
2.0140. Hyperfine couplings are partially resolved. Simulation
of the spectrum was possible after inspection of the DFT-
calculated EPR parameters. The spectrum is dominated by the
hyperfine coupling with two sets of equivalent protons from the
NMI moiety (H1 and H2) and two nonequivalent phosphorus
nuclei (P1 and P2; labeling shown in Figure 9). A satisfactory
simulation was obtained using similar hyperfine couplings as in
the DFT calculations, although it was necessary to make the
proton hyperfine coupling significantly larger to obtain a good
fit (see Figure 9 and Table 3).
CONCLUSIONS
EXPERIMENTAL SECTION
■
■
TRIR spectroscopic study of complex 3 was successfully
conducted to characterize for the first time an ultrafast electron
transfer process between excited ZnTPP and a Fe2S2 H2-ase
General Procedures. All manipulations were performed
under a nitrogen or argon atmosphere, using standard Schlenk
techniques, or in a glovebox. Commercially available chemicals,
5380
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
4-(aminomethyl)pyridine, Fe2(CO)9, and cis-1,2-bis-
(diphenylphosphino)ethene were used as received from
Aldrich. The supporting electrolyte nBu4NPF6 was recrystal-
lized from methanol and dried under vacuum at 80 °C
overnight. 1,8-Naphthalic anhydride-4,5-disulfide26 and [(μ-
naphdt)Fe2(CO)6] (1, naphdt = naphthalene-1,8-dithiolate)14
was stirred for 1 h at RT, the mixture was filtered through
Celite. The solution was collected and dried by vacuum. The
resulting dark solid was purified by column chromatography on
Silica gel (60−200 μm) with CH2Cl2 and hexane (3:2, v/v) as
eluent. The analytically pure solid of 3a was obtained as an
1
orange powder. Yield: 115 mg (64%). H NMR (400 MHz,
were synthesized by literature procedures. The 1H, 13C, and 31
P
acetone-d6): δ 8.62 (d, 2H, 3JHH = 7.7 Hz, Naph), 8.45 (d, 2H,
3
3JHH = 7.7 Hz, Naph), 7.46 (d, 2H, JHH = 7.5 Hz, Ph), 7.31−
NMR spectra were collected with a Bruker AVANCE 400
spectrometer. IR spectra were recorded with a Nicolet Nexus
FT-IR spectrometer. Elemental analyses were performed with a
PerkinElmer 2400 elemental analyzer. HR-MS were obtained
on a time-of-flight JEOL AccuTOF LC-plus mass spectrometer
(JMS-T100LP).
7.21 (m, 3H, Ph) ppm. 13C NMR (101 MHz, acetone-d6): δ
208.09, 163.79, 138.34, 134.90, 134.20, 130.37, 130.15, 129.15,
129.12, 128.06, 126.45, 126.30, 44.18 ppm. IR (CH2Cl2, cm−1):
ν(CO) 2079 (s), 2045 (s), 2006 (s); ν(CO)NMI 1703 (w),
1663 (w). Anal. Calcd for C25H11Fe2NO8S2: C 47.72, H 1.76, N
2.23; found: C 47.67, H 1.47, N 2.19%.
[4-pyCH2−NMI-S2] (L). 4-(Aminomethyl)pyridine (2.7 g, 25
mmol) was added to a suspension of 1,8-naphthalic anhydride-
4,5-disulfide (1.1 g, 4.2 mmol) in 2-methoxyethanol (60 mL) at
RT. The mixture was refluxed for 48 h under N2. After the
mixture cooled to RT, golden plate crystals that formed rapidly
were collected by filtration, washed with diethyl ether (3× 20
mL), and dried under vacuum. Yield: 0.8 g (54%). Crystals
suitable for X-ray diffraction were obtained by recrystallization
in hot chlorobenzene. 1H NMR (400 MHz, CDCl3): δ 8.52 (d,
3JHH = 6.1 Hz, 2H, Py), 8.45 (d, 3JHH = 8.1 Hz, 2H, Naph), 7.51
(d, 3JHH = 8.1 Hz, 2H, Naph), 7.34 (d, 3JHH = 6.0 Hz, 2H, Py),
5.36 (s, 2H, CH2) ppm. 13C NMR (101 MHz, CDCl3): δ
163.26, 154.14, 148.14, 134.17, 133.56, 130.67, 124.00, 116.95,
116.72, 42.83 ppm. IR (CH2Cl2, cm−1): ν(CO)NMI 1692
(w), 1655 (w). HR-MS time-of-flight electrospray ionization
(TOF-ESI+) Calcd for C18H11N2O2S2 [M + H]+: 351.0262;
found 351.262. Anal. Calcd for C18H10N2O2S2: C 61.70, H 2.88,
N 7.99; found: C 61.34, H 3.07, N 8.00%.
[μ-(4-pyCH2−NMI-S2)Fe2(CO)4(dppv)] (3-dppv). To a sol-
ution of 3 (90 mg, 0.14 mmol) in CH2Cl2 (10 mL) was added
cis-1,2-bis(diphenylphosphino)ethene (dppv) (60 mg, 0.15
mmol) in CH2Cl2 (5 mL). The mixture was stirred for 12 h
at RT. The resulting greenish brown solution was concentrated
to ca. 1 mL, and then hexane (10 mL) was added to precipitate
the crude product as light brown powders. Yield: 120 mg
(86%). Crystals suitable for X-ray diffraction were obtained by
1
diffusion of pentane into the CH2Cl2 solution of 3-dppv. H
3
NMR (400 MHz, CDCl3): δ 8.70 (d, 2H, JHH = 6.5 Hz,
Naph), 8.52 (s, 2H, Py), 8.2−7.5 (m, 20H, Ph), 7.46 (d, 2H,
3JHH = 6.3 Hz, Naph), 5.32 (s, 2H, CH2), 5.27 (s, 2H, PCH)
ppm. 13C NMR (101 MHz, acetone-d6): δ 164.01, 150.73,
147.36, 134.39, 134.29, 132.31, 131.89, 131.77, 131.64, 129.84,
129.71, 129.62, 128.51, 128.43, 123.72, 123.30, 43.02 ppm. 31P
NMR (162 MHz, acetone-d6): δ 96.76 ppm. IR (CH2Cl2,
cm−1): ν(CO) 2031(s), 1965 (s), 1923 (w); ν(CO)NMI
=
[PhCH2−NMI-S2] (La). A similar procedure for the synthesis
1699 (w), 1659 (w). HR-MS (TOF-atomospheric pressure
chemical ionization (APCI)+) Calcd for C48H33Fe2N2O6P2S2
[M + H]+: 970.9956; found 970.9924. Anal. Calcd for
C48H32Fe2N2O6P2S2·0.5CH2Cl2: C 57.50, H 3.28, N 2.77;
found: C 57.31, H 3.03, N 2.77%.
of L was applied by using benzylamine instead of 4-
1
(aminomethyl)pyridine. Yield: 0.8 g (66%). H NMR (400
3
MHz, DMSO-d6): δ 8.33 (d, JHH = 8.1 Hz, 2H, Naph), 7.85
(d, 3JHH = 8.1 Hz, 2H, Naph), 7.38−7.16 (m, 5H, Ph), 5.24 (s,
2H, CH2) ppm. 13C NMR (101 MHz, CDCl3): δ 163.40,
153.32, 137.64, 134.01, 133.25, 129.88, 129.03, 128.55, 127.53,
117.34, 116.77, 43.71 ppm. IR (CH2Cl2, cm−1): ν(CO)NMI
1690 (w), 1653 (w). Anal. Calcd for C19H11NO2S2: C 65.31, H
3.17, N 4.01; found: C 65.02, H 3.02, N 4.00%.
Spectroelectrochemistry. Cyclic voltammograms were
recorded using an Autolab PGSTAT302N electrochemical
workstation, and an airtight three-electrode cell under dry N2.
The working electrode was a carefully polished Pt microodisc
(diameter 0.5 mm). A coiled Pt wire was used as a counter
electrode and a coiled Ag wire as the pseudoreference
electrode. All electrode potentials reported in this work are
referenced versus the internal standard ferrocene/ferrocenium
(Fc/Fc+) couple.21 The measurements were performed on 1.0
mM complexes in CH2Cl2 containing 0.1 M nBu4NPF6 as a
supporting electrolyte. Cyclic voltammograms in the presence
of HOAc were recorded using the same setup while a polished
glassy carbon disc (diameter 3 mm) was used as a working
electrode. IR SEC was performed in an optically transparent
thin-layer (200 μm) electrochemical (OTTLE) cell27 equipped
with CaF2 optical windows and a platinum minigrid working
electrode. The difference absorbance IR spectra were recorded
on a Nicolet Nexus FT-IR spectrometer in the course the thin-
layer cyclic voltammetry scanning process (v = 2 mV s−1
controlled by a PGSTAT 10 (Eco-Chemie) potentiostat (see
Supporting Information, Figure S9 for a thin-layer cyclic
voltammetry response).
[μ-(4-pyCH2−NMI-S2)Fe2(CO)6] (3). THF (50 mL) was
added to a mixture of L (100 mg, 0.28 mmol) and Fe2(CO)9
(208 mg, 0.57 mmol) under N2 at RT. After it was stirred for 1
h at RT, the mixture was filtered through Celite. The solution
was collected and dried by vacuum. The resulting dark solid
was purified by column chromatography on Silica gel (60−200
μm) with CH2Cl2 and MeOH (100:0.3, v/v) as eluent. The
analytically pure solid of 3 was obtained as an orange powder.
Yield: 90 mg (50%). Crystals suitable for X-ray diffraction were
collected by diffusion of pentane into the CH2Cl2 solution of 4
3
at −20 °C. 1H NMR (400 MHz, acetone-d6) δ 8.67 (d, JHH
=
3
6.7 Hz, 2H, Py), 8.63 (d, JHH = 7.7 Hz, 2H, Naph), 8.44 (d,
3JHH = 7.7 Hz, 2H, Naph), 7.54 (d, 3JHH = 6.7 Hz, 2H, Py), 5.37
(s, 2H, CH2) ppm. 13C NMR (101 MHz, acetone-d6) δ 208.06,
163.91, 158.39, 150.67, 149.76, 135.27, 134.22, 130.44, 126.49,
126.15, 125.83, 43.14 ppm. IR (CH2Cl2, cm−1): ν(CO) 2080
(s), 2046 (s), 2007 (s); ν(CO)NMI 1706 (w), 1666 (w). HR-
MS (TOF-ESI+) Calcd for C24H11Fe2N2O8S2 [M + H]+:
630.8657; found 630.8638%.
Steady-State UV−vis Absorption and Emission. UV−
vis absorption spectra in the titration experiments were
measured on a Shimadzu UV-2700 spectrophotometer, using
2 mm quartz cuvettes. Emission spectra were measured on a
Spex Fluorolog 3 spectrofluorimeter, equipped with double
[μ-(PhCH2−NMI-S2)Fe2(CO)6] (3a). THF (50 mL) was
added to a mixture of La (100 mg, 0.28 mmol) and
Fe2(CO)9 (208 mg, 0.57 mmol) under N2 at RT. After it
5381
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
grating monochromators in the excitation and emission
channels. The excitation light source was a 450W Xe lamp
and the detector a Peltier cooling R636−10 (Hamamatsu)
photomultiplier tube. The CH2Cl2 solution of the ligand or
complexes 3, 3-dppv (1.6 × 10−3 M) were prepared by
dissolving the corresponding compound in the ZnTPP solution
(8.0 × 10−5 M) to keep the concentration of ZnTPP constant
during the titration experiment.
ACKNOWLEDGMENTS
■
We are grateful to The Netherlands’ Organization for Scientific
Research (NWO−CW; ECHO Grant No. 700.57.042; F.H.
and J.N.H.R.) and the German National Academy of Sciences
(Leopoldina research fellowship, Grant No. LPDS 2011-18;
S.A.) for financial support. Helpful discussions with Dr.
Volodymyr Lyaskovskyy, Mr. Hungcheng Chen, and Dr.
Rene
́
Williams (University of Amsterdam) are acknowledged.
Time-Resolved Infrared Spectroscopy. Tunable visible
pump and mid-IR probe were generated using a Ti:sappahire
laser (Spectra-Physics Hurricane, 600 μJ) with a repetition rate
of 1 kHz pumping two commercial BBO-based OPAs (Spectra-
Physics OPA-800C). Visible pump pulses (553 nm) were
generated by sum-frequency mixing the Ti:sapphire pump and
idler (centered at 1791 nm) of one of the OPAs in BBO (UV
pulse energy = 3 μJ); IR probe pulses were generated by
difference-frequency mixing signal and idler from the other
OPA in AgGaS2, for details see ref 28. The sample cell with
CaF2 windows spaced by 500 μm was replaced in the IR focus.
Using a Newport ESP300 translation stage, the delay positions
were scanned by mechanically adjusting the beam-path of the
UV pump. The temporal resolution of 200 fs has been obtained
from full width at half-maximum of the pump probe cross-
correlate function. The transient spectra were obtained by
subtracting nonpumped absorption spectra from the pumped
absorption spectra that were recorded by a custom built 30
pixel MCT detector coupled to an Oriel MS260i spectrograph.
EPR Spectroscopy and Simulation. Experimental X-band
EPR were recorded on a Bruker EMX plus spectrometer
equipped with a He temperature control cryostat system
(Oxford Instruments). The spectra were simulated by iteration
of the anisotropic g-values, (super)hyperfine coupling constants
and line widths using the W95EPR program (obtained from
Prof. Frank Neese, the University of Bonn).
REFERENCES
■
(1) (a) Tard, C.; Liu, X.; Ibrahim, S. K.; Bruschi, M.; Gioia, L. D.;
Davies, S. C.; Yang, X.; Wang, L.-S.; Sawers, G.; Pickett, C. J. Nature
2005, 433, 610−613. (b) Tard, C.; Pickett, C. J. Chem. Rev. 2009, 19,
2245−2274. (c) Gloaguen, F.; Rauchfuss, T. B. Chem. Soc. Rev. 2009,
38, 100−108. (d) Capon, J.-F.; Gloaguen, F.; Petillon, F. Y.;
́
Schollhammer, P.; Talarmin, J. Coord. Chem. Rev. 2009, 253, 1476−
1494.
(2) Peters, J. W.; Lanzilotta, W. N.; Lemon, B. J.; Seefeldt, L. C.
Science 1998, 282, 1853−1858.
(3) (a) Gloaguen, F.; Lawrence, J. D.; Rauchfuss, T. B. J. Am. Chem.
Soc. 2001, 123, 9476−9477. (b) Chong, D.; Georgakaki, I. P.; Mejia-
Rodriguez, R.; Sanabria-Chinchilla, J.; Soriaga, M. P.; Darensbourg, M.
Y. Dalton Trans. 2003, 4158−4163. (c) Capton, J.-F.; Gloaguen, F.;
Schollhammer, P.; Talarmin, J. J. Electroanal. Chem. 2004, 566, 241−
247. (d) Duan, L.; Wang, M.; Li, P.; Na, Y.; Wang, N.; Sun, L. Dalton
Trans. 2007, 1277−1283. (e) Nann, T.; Ibrahim, S. K.; Woi, P.-M.; Xu,
S.; Ziegler, J.; Pickett, C. J. Angew. Chem., Int. Ed. 2010, 49, 1574−
1577.
(4) (a) Lomoth, R.; Ott, S. Dalton Trans. 2009, 9952−9959.
(b) Wang, M.; Chen, L.; Li, X.; Sun, L. Dalton Trans. 2011, 40,
12793−12800. (c) Wang, F.; Wang, W.-G.; Wang, H.-Y.; Si, G.; Tung,
C.-H.; Wu, L.-Z. ACS Catal. 2012, 2, 407−416. (d) Streich, D.; Astuti,
Y.; Orlandi, M.; Schwartz, L.; Lomoth, R.; Hammarstrom, L.; Ott, S.
̈
Chem.Eur. J. 2010, 16, 60−63. (e) Yu, T.; Zeng, Y.; Chen, J.; Li, Y.-
Y.; Yang, G.; Li, Y. Angew. Chem., Int. Ed. 2013, 52, 5631−5635.
(5) (a) Ott, S.; Kritikos, M.; Åkermark, B.; Sun, L. Angew. Chem., Int.
Computational Details. The geometries of 1, 1•−, 3, 3•−,
3-dppv, and 3-dppv•−, were optimized with Turbomole 6.4
program29 using the BP8630−32 functional and the def2-
TZVP33 basis set. These calculations also yielded the frontier
orbitals and spin density plots EPR parameters34 were
subsequently calculated with ORCA,35 using the coordinates
from the structures optimized in Turbomole as input. In the
Orca calculations, we used Ahlrich’s def2-TZVP33 basis set and
the B3LYP functional.36
Ed. 2003, 42, 3285−3288. (b) Wolpher, H.; Borgstrom, M.;
̈
Hammarstrom, L.; Bergquist, J.; Sundstrom, V.; Styring, S.; Sun, L.;
̈
̈
Åkermark, B. Inorg. Chem. Commun. 2003, 989−991. (c) Ott, S.;
Borgstrom, M.; Kritikos, M.; Lomoth, R.; Bergquist, J.; Åkermark, B.;
̈
Hammarstrom, L.; Sun, L. Inorg. Chem. 2004, 43, 4683−4692.
(6) Sun, L.; Åkermark, B.; Ott, S. Coord. Chem. Rev. 2005, 249,
1653−1663.
(7) Ekstrom, J.; Abrahamsson, M.; Olson, C.; Bergquist, J.; Kaynak,
̈
F. B.; Eriksson, L.; Sun, L.; Becker, H.-C.; Åkermark, B.;
Hammarstrom, L.; Ott, S. Dalton Trans. 2006, 4599−4606.
̈
(8) Na, Y.; Pan, J.; Wang, M.; Sun, L. Inorg. Chem. 2007, 46, 3813−
3815.
ASSOCIATED CONTENT
■
(9) Li, X.; Wang, M.; Chen, L.; Wang, X.; Dong, J.; Sun, L.
S
* Supporting Information
ChemSusChem 2012, 5, 913−919.
Details on X-ray crystal analyses of L, 3, and 3-dppv, as well as
results for chemical reduction of complex 3, DFT calculations,
spectroscopic and electrochemical analyses. This material is
available free of charge via the Internet at http://pubs.acs.org.
(10) Song, L.-C.; Tang, M.-Y.; Mei, S.-Z.; Huang, J.-H.; Hu, Q.-M.
Organometallics 2007, 26, 1575−1577.
(11) Li, X.; Wang, M.; Zhang, S.; Pan, J.; Na, Y.; Liu, J.; Åkermark, B.;
Sun, L. J. Phys. Chem. B 2008, 112, 8198−8202.
(12) Kluwer, A. M.; Kapre, R.; Hartl, F.; Lutz, M.; Spek, A. L.;
Brouwer, A. M.; van Leeuwen, P. W. N. M.; Reek, J. N. H. Proc. Natl.
Acad. Sci. U.S.A. 2009, 106, 10460−10465.
AUTHOR INFORMATION
■
(13) (a) Samuel, A. P. S.; Co, D. T.; Stern, C. L.; Wasielewski, M. R.
J. Am. Chem. Soc. 2010, 132, 8813−8815. (b) Poddutoori, P.; Co, D.
T.; Samuel, A. P. S.; Kim, C. H.; Vagnini, M. T.; Wasielewski, M. R.
Energy Environ. Sci. 2011, 4, 2441−2450.
Corresponding Authors
*E-mail: f.hartl@reading.ac.uk. (F.H.)
*E-mail: s.woutersen@uva.nl. (S.W.)
*E-mail: j.n.h.reek@uva.nl. (J.N.H.R.)
(14) Wright, R. J.; Lim, C.; Tilley, T. Don Chem.Eur. J. 2009, 15,
8518−8525.
Author Contributions
⊥P.L. and S.A. contributed equally to this work.
(15) (a) Best, S. P. Coord. Chem. Rev. 2005, 249, 1536−1554.
(b) Borg, S. J.; Behrsing, T.; Best, S. P.; Razavet, M.; Liu, X.; Pickett,
C. J. J. Am. Chem. Soc. 2004, 126, 16988−16999. (c) Tschierlei, S.;
Ott, S.; Lomoth, R. Energy Environ. Sci. 2011, 4, 2340−2352.
Notes
The authors declare no competing financial interests.
5382
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383

Inorganic Chemistry
Article
(16) Jablonskyte, 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, 18606−18909.
(17) (a) Bingaman, J. L.; Kohnhorst, C. L.; Van Meter, G. A.;
McElroy, B. A.; Rakowski, E. A.; Caplins, B. W.; Gutowski, T. A.;
Stromberg, C. J.; Webster, C. E.; Heiweil, E. J. J. Phys. Chem. A 2012,
116, 7261−7271. (b) Johnson, M.; Thuman, J.; Letterman, R. G.;
Stromberg, C. J.; Webster, C. E.; Heilweil, E. J. J. Phys. Chem. B 2013,
117, 15792−15803.
(18) (a) Stewart, A. I.; Clark, I. P.; Ibrahim, S. K.; Parker, A. W.;
Pickett, C. J.; Hunt, N. T. J. Phys. Chem. B 2008, 112, 10023−10032.
(b) Stewart, A. I.; Wright, J. A.; Greetham, G. M.; Kaziannis, S.;
Santabarbara, S.; Towrie, M.; Parker, A. W.; Pickett, C. J.; Hunt, N. T.
Inorg. Chem. 2010, 49, 9563−9573. (c) Kaziannis, S.; Santabarbara, S.;
Wright, J. A.; Greetham, G. M.; Towrie, M.; Parker, A. W.; Pickett, C.
J.; Hunt, N. T. J. Phys. Chem. B 2010, 114, 15370−15379.
(19) Singh, P. S.; Rudbeck, H. C.; Huang, P.; Ezzaher, S.; Eriksson,
L.; Stein, M.; Ott, S.; Lomoth, R. Inorg. Chem. 2009, 48, 10883−
10885.
(20) Hall, G. B.; Chen, J.; Mebi, C. A.; Okumura, N.; Swenson, M.
T.; Ossowski, S. E.; Zakai, U. I.; Nichol, G. S.; Lichtenberger, D. L.;
Evans, D. H.; Glass, R. S. Organometallics 2013, 32, 6605−6612.
(21) Liu, Y.-C.; Yen, T.-H.; Tseng, Y.-J.; Hu, C.-H.; Lee, G.-H.;
Chiang, M.-H. Inorg. Chem. 2012, 51, 5997−5999.
(22) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877−910.
(23) Justice, A. K.; Zampella, G.; Gioia, L. D.; Rauchfuss, T. B.; van
der Vlugt, J. I.; Wilson, S. R. Inorg. Chem. 2007, 46, 1655−1664.
(24) Li, P.; Wang, M.; He, C.; Li, G.; Liu, X.; Chen, C.; Åkermark, B.;
Sun, L. Eur. J. Inorg. Chem. 2005, 2506−2513.
(25) (a) Nocera, D. G.; Teets, T. S. Chem. Commun. 2011, 47, 9268−
9274. (b) Artero, V.; Chavarot-Kerlidou, M.; Fontecave, M. Angew.
Chem., Int. Ed. 2011, 50, 7238−7266. (c) Du, P.; Eisenberg, R. Energy
Environ. Sci. 2012, 5, 6012−6021.
(26) Tesmer, M.; Vahrenkamp, H. Eur. J. Inorg. Chem. 2001, 1183−
1188.
(27) Krejcí
179−187.
̌ ̆
k, M.; Danek, M.; Hartl, F. J. Electroanal. Chem. 1991, 317,
(28) Meuzelaar, H.; Marino, K. A.; Huerta-Viga, A.; Panman, M. R.;
Smeenk, L. E. J.; Kettelarij, A. J.; van Maarseveen, J. H.; Timmerman,
P.; Bolhuis, P. G.; Woutersen, S. J. Phys. Chem. B 2013, 117, 11490−
11501.
(29) Ahlrichs, R.; Bar, M.; Barton, H.-P.; Bauernschmitt, R.; Bocher,
̈
̈
S.; Rhrig, M.; Eichkorn, K.; Elliott, S.; Furche, F.; Haase, F.; M. Haser,
̈
C. Hattig, Horn, H.; Huber, C.; Huniar, U.; Kattannek, M.; A. Kohn,
C. Kolmel, Kollwitz, M.; May, K.; Ochsenfeld, C.; H. Ohm, A. Schafer,
̈
̈
̈
̈
̈
Schneider, U.; Treutler, O.; Tsereteli, K.; Unterreiner, B.; von Arnim,
M.; Weigend, F.; Weis, P.; Weiss, H. Turbomole, Version 6.4;
University of Karlsruhe: Karlruhe, Germany, January 2002.
(30) Becke, A. D. Phys. Rev. A 1988, 38, 3098−3110.
(31) Perdew, J. P. Phys. Rev. B 1986, 33, 8822−8824.
(32) Perdew, J. P. Phys. Rev. B 1986, 34, 7406−7406.
(33) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7,
3297−3305.
(34) Some references and reviews on DFT approaches to EPR
parameters: (a) van Lenthe, E.; van der Avoird, A.; Wormer, P. E. S. J.
Chem. Phys. 1997, 107, 2488−2498. (b) van Lenthe, E.; van der
Avoird, A.; Wormer, P. E. S. J. Chem. Phys. 1998, 108, 4783−4796.
(c) Neese, F. Curr. Opin. Chem. Biol. 2003, 7, 125−135. (d) Neese, F.;
Solomon, E. In MagnetoscienceFrom Molecules to Materials; Miller, J.
S.; Drillon, M., Eds.; Wiley: New York, 2003; pp 345−466. (e) Peng,
G.; Nichols, J.; McCullough, E. A.; Spence, J. Inorg. Chem. 1994, 33,
2857−2864.
(35) Neese, F. ORCAAn ab initio, Density Functional and
Semiempirical program package, Version 2.9; Max Planck Institute for
Chemical Energy Conversion: Mulheim an der Ruhr, Germany, 2010.
(36) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1998, 37, 785−
789. (b) Becke, A. D. J. Chem. Phys. 1993, 98, 1372−1377. (c) Becke,
A. D. J. Chem. Phys. 1993, 98, 5648−5652.
5383
dx.doi.org/10.1021/ic500777d | Inorg. Chem. 2014, 53, 5373−5383