Light-driven electron transfer between a photosensitizer and a proton-reducing catalyst Co-adsorbed to NiO

Article abstract of DOI:10.1021/ja3082268

While intermolecular hole-hopping along the surface of semiconductors is known, there are no previous examples of electron-hopping between molecules on a surface. Herein, we present the first evidence of electron transfer from the photoreduced sensitizer Coumarin-343 (C343) to complex 1, both bound on the surface of NiO. In solution, 1 has been shown to be a mononuclear Fe-based proton-reducing catalyst. The reduction of 1 is reversible and occurs within 50 ns after excitation of C343. Interfacial recombination between the reduced 1^(-) and NiO hole occurs on a 100 μs time scale by non-exponential kinetics. The observed process is the first essential step in the photosensitized generation of H₂ from a molecular catalyst in the absence of a sacrificial donor reagent.

Full text of DOI:10.1021/ja3082268

Communication
pubs.acs.org/JACS
Light-Driven Electron Transfer between a Photosensitizer and a
Proton-Reducing Catalyst Co-adsorbed to NiO
James M. Gardner,† Maryline Beyler, Michael Karnahl, Stefanie Tschierlei, Sascha Ott,*
and Leif Hammarstrom*
̈
Ångstrom Laboratory, Department of Chemistry, Uppsala University, Box 523, 75120 Uppsala, Sweden
̈
S
* Supporting Information
ping is the likely mechanism to drive the cumulative hole-
transfer chemistry.4,5,10 However, to the best of our knowledge
there exist no examples in the literature of light-driven
“electron-hopping” between molecules on the surface of any
semiconductor. We report on the first evidence for “electron-
hopping” from a photosensitizer to a complex 1 co-adsorbed to
a thin film of NiO. In solution, complex 1 is a proton-reducing
catalyst.11 All components are based on abundant elements.
The described process is a first step in the photosensitized
generation of H2 from a molecular catalyst in the absence of a
sacrificial reagent.
ABSTRACT: While intermolecular hole-hopping along
the surface of semiconductors is known, there are no
previous examples of electron-hopping between molecules
on a surface. Herein, we present the first evidence of
electron transfer from the photoreduced sensitizer
Coumarin-343 (C343) to complex 1, both bound on the
surface of NiO. In solution, 1 has been shown to be a
mononuclear Fe-based proton-reducing catalyst. The
reduction of 1 is reversible and occurs within 50 ns after
excitation of C343. Interfacial recombination between the
reduced 1(−) and NiO hole occurs on a 100 μs time scale
by non-exponential kinetics. The observed process is the
first essential step in the photosensitized generation of H2
from a molecular catalyst in the absence of a sacrificial
donor reagent.
Designing chemical systems to take advantage of intermo-
lecular charge-hopping has many benefits: (i) many sensitizers
may transfer charge to a single acceptor; (ii) modular design
allows the screening of different catalysts without elaborate
synthesis; (iii) decreased reliance on robust bonds between a
photosensitizer and catalyst; and (iv) decoupling light
absorption events by a photosensitizer and the resulting
catalysis.
ye-sensitized solar cells (DSSCs) have been investigated
Dand developed extensively for use in photovoltaic
devices.1,2 However, their potential for solar fuels production
is a very recent and still little explored subject.3−6 A critical
challenge is matching the rate of light-driven charge separation
and transport with the catalytic steps of fuel formation and
water oxidation. Catalysts for water oxidation have been
attached directly to a TiO2-bound sensitizer3 or deposited in
a separate Nafion layer on top of the TiO2/dye film.4,5 The
former approach is more synthetically demanding, while the
latter gives rather poor electron transport between the separate
layers. For the corresponding H2-forming reaction on p-type
materials, there is only one study that used a dye/NiO film and
a cobaloxime catalyst,7 but the cobaloxime was free in solution
and may therefore also act as a redox shuttle to the counter
electrode to create current instead of H2. In this paper, we
investigate a different approach where we co-sensitize a
mesoporous NiO film with both dye and catalyst. This relies
on the possibility that, after electron transfer from NiO to the
excited dye, efficient surface electron-hopping occurs between
dye molecules and the catalyst.
For the proton-reducing catalyst 2 in Figure 1, molecular
turnover frequencies >10 000 s−1 are reported.11 However, the
Figure 1. Complex 1 was functionalized with a carboxylate group for
anchoring to NiO; the structurally analogous complex 211 has a higher
solubility in acetonitrile and is, in solution, a proton-reducing catalyst
that operates at low overpotential.
visible photon flux from sunlight is <1000 s−1 nm−2.12 To
match the catalytic efficiency using sunlight to photo-initiate
the reaction would require absorbing all visible photons on an
area >10 nm2. This is impossible if the photosensitizer and
catalyst are covalently linked as a single molecule. Allowing for
electron-hopping from a photosensitizer to the catalyst
dramatically increases the number of photosensitizers that can
contribute to the chemistry of a single catalyst.
The laser dye Coumarin-343 (C343) was used as the
photosensitizer, since it has a high reduction potential in its
one-electron-reduced state (C3430/−; −2.1 V vs Fc+/0) and is
known to rapidly inject holes into NiO.13−15 Complex 1 is
From the field of DSSCs, hole transfer between molecules
along the interface of semiconductors is known.8,9 This occurs
as a “hole-hopping” mechanism from donor molecules to
acceptor molecules, allowing for charges to migrate over
significant distances at the surface of semiconductors. This
hole-transfer mechanism is mediated solely by non-covalent
interactions between adjacent molecules. In dye-sensitized
water oxidation, photosensitizer-to-photosensitizer hole-hop-
Received: August 19, 2012
Published: November 9, 2012
© 2012 American Chemical Society
19322
dx.doi.org/10.1021/ja3082268 | J. Am. Chem. Soc. 2012, 134, 19322−19325

Journal of the American Chemical Society
Communication
identical to the previously reported complex 2 in the first and
second coordination spheres around Fe,11 but 1 contains a
pendant carboxylate to bind to the surface of NiO (see Figure
1). Sequentially, C343 and 1 were bound to the surface of
mesoporous NiO thin films. In transient absorption measure-
ments for the co-adsorbed films, it was seen that, by exciting
C343 with pulsed laser light, the radical anion of 1 was visible
within 50 ns and recombined with NiO(+) on a time scale of
100 μs.
NiO films were sensitized by immersion of films in Ar-
saturated ethanolic solutions of C343 in the dark over a span of
2 days. After sensitization with C343, the NiO films were co-
adsorbed with 1 by immersion in 3:1 methanol/acetonitrile
solutions for 2 days in the dark and under an Ar atmosphere. As
a control for sensitizing order, other films were sensitized first
with 1 and second with C343. Through the co-adsorption
process, a notable decrease in the UV−visible absorption of
C343 on the surface of NiO films was observed as well as a
small shift in the C343 absorbance to the red (Figure 2). A very
Figure 3. Raman spectra of NiO films adsorbed with C343 (black line)
and C343/1 (red line) acquired at 514 nm excitation. Binding 1 to
NiO resulted in new vibrations at 1925 and 1650 cm−1; exposure to air
for 2 min (blue line) resulted in the loss of those vibrations.
visible absorption spectra (430 and 580 nm), as seen in Figures
4 and S4. Single-electron reduction of 2 (E0 = −1.66 V vs
Fc+/0) led to an increase in the absorbance at ∼500 nm and a
decrease for the peaks of the neutral species.
Figure 2. Absorption spectra of C343/NiO (solid) and C343/1/NiO
(dashed). Inset: Magnification of the region of 1 absorption.
Figure 4. Reducing complex 2 by one electron (2(−), red dashed line)
resulted in decreases in the absorbance of the neutral state (2, black
solid line). The difference between the 2(−) and 2 spectra (blue
dotted line) represents the contributions of 1 to the transient spectra
in Figure 5.
basic calculation for the surface coverage of C343 on NiO,
which is detailed in the Supporting Information, suggested 20%
monolayer coverage. At 20% of a monolayer, the average
C343−C343 distance is 6 Å. However, true C343−C343
distances may be much smaller due to clustering of C343 on
the surface of NiO. We estimate the ratio of 1 to C343 on the
surface of NiO to be between 1:2 and 1:4. Due to the small
extinction coefficient of 1 in the visible, its absorption on the
surface of NiO by UV−visible was weak (Figure 1 inset).
To verify that 1 binds to NiO, Raman measurements were
made of C343-sensitized NiO films before and after slides were
immersed in a solution of 1. The IR spectrum of 1 in
dichloromethane shows a CO stretch at 1908 cm−1 (Figure
S3). As seen in Figure 3, after immersion of NiO films in
solutions of 1, a new feature is observed at 1925 cm−1.
We attribute the small differences in stretching frequencies
between bound and unbound 1 to the differing environments.
The CO stretch is a characteristic vibration for 1, and thus
confirms its presence on the NiO films. The samples did not
appear to be photolabile and showed no damage from
prolonged laser excitation at 514 nm (see Figure S4). However,
exposure of C343/1 co-adsorbed films to air resulted in the loss
of the Raman feature at 1925 cm−1 in <2 min, indicating
oxidative degradation of 1 with concomitant CO loss.
To confirm light-driven electron transfer from photoreduced
C343(−) to 1 co-adsorbed to the surface of NiO, nanosecond
transient absorption measurements were performed. For NiO
thin films that were co-adsorbed with both C343 and 1, the
transient absorption spectra after excitation of C343 (460 nm, 9
mJ cm−2, 8 ns fwhm) resulted in a bleach at 430 nm, a peak at
500 nm, and a broad absorbance throughout the near-IR within
the time resolution of our instruments, 50 ns (Figure 5). When
comparing these spectral features with the spectroelectrochem-
istry on 2 and literature values for NiO,16,17 these transient
absorbance signals are consistent with the formation of 1(−)
and NiO(+). The signals returned to baseline over the span of
100 μs following nonexponential kinetics (Figure S5). The
order in which C343 and 1 were bound to the surface of NiO
did not affect observation of these signals (Figures 5 and S5).
The transient spectra showed a high degree of reproducibility
over the length of the experiment: about 1800 laser flashes
(Figure S6). The absorbance of 1 is insignificant at the
excitation wavelength (460 nm) and does not contribute to the
observed signal.
Due to poor solubility of 1, electrochemistry and
spectroelectrochemistry were performed on the previously
studied analogue 2.11 Both 1 and 2 have two peaks in their
As a control experiment, we measured transient absorption
spectra from the excitation of C343 bound to NiO under an Ar
19323
dx.doi.org/10.1021/ja3082268 | J. Am. Chem. Soc. 2012, 134, 19322−19325

Journal of the American Chemical Society
Communication
adjacent to 1/NiO sites. We propose that this may be one
reason for the much slower recombination with NiO(+) from
1(−) than from most reduced dyes investigated before.14,15,18,20
It is of note that the transient spectra in Figure 5 are similar
to previously published spectra for the triplet excited state of
C343 (3C343).15 However, the 3C343 state has not been
observed on the surface of NiO for longer than 2 ns,15 the yield
14
3
of C343 is very small both on NiO and in solution, the
transient spectra that we measure are not observed in the
absence of 1, and we are unaware of any mechanism whereby 1
3
could increase the lifetime of the C343 state. It is therefore
highly unlikely that the C343 triplet state contributes to the
observed spectral features.
In conclusion, we have observed charge recombination
between singly reduced H2-generating catalysts and valence
band holes in NiO on the 100 μs time scale. From the transient
absorption, we infer that the reduced form of 1 results from
electron transfer from neighboring reduced C343 molecules,
which were generated from the excitation of C343 photo-
sensitizers bound to NiO and potentially in close contact with
1. We believe this represents the first instance of “electron-
hopping” between molecules along the surface of a semi-
conductor and the first evidence of reduction of a hydrogen-
generating catalyst on the surface of NiO.
Figure 5. Transient absorption spectra after excitation of NiO thin
films co-adsorbed with C343 and 1 with pulsed laser light (9 mJ cm−2,
8 ns, 460 nm) results in the formation of 1(−) and NiO(+) within 50
ns of excitation; recombination occurs on a 100 μs time scale.
atmosphere but without 1 co-adsorbed. After the excitation of
C343 on NiO by itself, no signals were observed at our signal
detection limit on a time scale of 50 ns or longer.
A plausible explanation for the observed formation of 1(−)
and NiO(+) from the excitation of C343 is presented in
Scheme 1. As previously stated in the literature, excitation of
Scheme 1. Reduction of Complex 1 Co-adsorbed to NiO
from the Selective Excitation of C343/NiO
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthesis and characterization of 1 and 2; experimental details
for sample preparation and measurements. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
leif.hammarstrom@kemi.uu.se; sascha.ott@kemi.uu.se
C343 on NiO results in the formation of C343(−)/NiO(+)
within femtoseconds of light absorption, (1) and (2) in Scheme
1.14,15 In the absence of an accessible electron acceptor,
recombination between C343(−) and NiO(+) occurs on the
picosecond time scale.14,15 Due to this rapid charge
recombination, it has been previously proposed that charge
transfer from C343(−) bound to NiO to electron acceptors
requires pre-association between the electron acceptor and
C343.14,18 Based on the driving force (∼500 mV), a bound
C343(−) may directly reduce a neighboring bound complex 1,
(3) in Scheme 1.11,13 The initial state is reset by interfacial
charge recombination between 1(−) and NiO(+), (4) in
Scheme 1.
Present Address
†KTH Royal Institute of Technology, Department of
Chemistry, Division of Applied Physical Chemistry
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Funding for this project comes from The K&A Wallenberg
Foundation, The Swedish Energy Agency, and the Swedish
Research Council. M.K. and S.T. thank the Wenner-Gren
Foundation and the German Academic Exchange Service
(DAAD), respectively, for providing post-doctoral fellowships.
Due to the bulky phenyl groups at the phosphorus ligands,
the iron center in 1 is relatively shielded from the surface of
NiO. The slow recombination kinetics between 1(−) and
NiO(+) may be due in part to the large electron tunneling
distance between the iron center and the NiO surface through
five saturated bonds. It has been previously proposed that NiO
holes have low charge mobility and are more localized than
TiO2 conduction band electrons.16,17 To aid in the migration of
electrons across the surface of NiO, a self-exchange mechanism
between neighboring C343 molecules may be active on a sub-
nanosecond time scale. This self-exchange mechanism has been
previously observed in molecular stacks of perylene diimide.19
If a self-exchange mechanism exists, then the true distance
between 1(−) and NiO(+) may be significantly larger due to
the generation of charge carriers on C343/NiO sites non-
REFERENCES
■
(1) Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H.
Chem. Rev. 2010, 110, 6595.
(2) Yella, A.; Lee, H.-W.; Tsao, H. N.; Yi, C.; Chandiran, A. K.;
Nazeeruddin, M. K.; Diau, E. W.-G.; Yeh, C.-Y.; Zakeeruddin, S. M.;
Gratzel, M. Science 2011, 334, 629.
(3) Youngblood, W. J.; Lee, S.-H. A.; Kobayashi, Y.; Hernandez-
Pagan, E. A.; Hoertz, P. G.; Moore, T. A.; Moore, A. L.; Gust, D.;
Mallouk, T. E. J. Am. Chem. Soc. 2009, 131, 926.
(4) Brimblecombe, R.; Koo, A.; Dismukes, G. C.; Swiegers, G. F.;
Spiccia, L. J. Am. Chem. Soc. 2010, 132, 2892.
(5) Li, L.; Duan, L.; Xu, Y.; Gorlov, M.; Hagfeldt, A.; Sun, L. Chem.
Commun. 2010, 46, 7307.
̈
(6) Reisner, E.; Powell, D. J.; Cavazza, C.; Fontecilla-Camps, J. C.;
Armstrong, F. A. J. Am. Chem. Soc. 2009, 131, 18457.
19324
dx.doi.org/10.1021/ja3082268 | J. Am. Chem. Soc. 2012, 134, 19322−19325

Journal of the American Chemical Society
Communication
(7) Li, L.; Duan, L.; Wen, F.; Li, C.; Wang, M.; Hagfeldt, A.; Sun, L.
Chem. Commun. 2012, 48, 988.
(8) Ardo, S.; Meyer, G. J. J. Am. Chem. Soc. 2010, 132, 9283.
(9) Ardo, S.; Meyer, G. J. J. Am. Chem. Soc. 2011, 133, 15384.
(10) Moore, G. F.; Blakemore, J. D.; Milot, R. L.; Hull, J. F.; Song,
H.-e.; Cai, L.; Schmuttenmaer, C. A.; Crabtree, R. H.; Brudvig, G. W.
Energy Environ. Sci. 2011, 4, 2389.
(11) Beyler, M.; Ezzaher, S.; Karnahl, M.; Santoni, M.-P.; Lomoth, R.;
Ott, S. Chem. Commun. 2011, 47, 11662.
(12) NREL. Reference Solar Spectral Irradiance: Air Mass 1.5,
http://rredc.nrel.gov/solar/spectra/am1.5/.
(13) Mori, S.; Fukuda, S.; Sumikura, S.; Takeda, Y.; Tamaki, Y.;
Suzuki, E.; Abe, T. J. Phys. Chem. C 2008, 112, 16134.
(14) Morandeira, A.; Boschloo, G.; Hagfeldt, A.; Hammarstrom, L. J.
̈
Phys. Chem. C 2008, 112, 9530.
(15) Morandeira, A.; Boschloo, G.; Hagfeldt, A.; Hammarstrom, L. J.
̈
Phys. Chem. B 2005, 109, 19403.
(16) Boschloo, G.; Hagfeldt, A. J. Phys. Chem. B 2001, 105, 3039.
(17) Dare-Edwards, M. P.; Goodenough, J. B.; Hamnett, A.;
Nicholson, N. D. J. Chem. Soc., Faraday Trans. 2 1981, 77, 643.
(18) Qin, P.; Wiberg, J.; Gibson, E. A.; Linder, M.; Li, L.; Brinck, T.;
Hagfeldt, A.; Albinsson, B.; Sun, L. J. Phys. Chem. C 2010, 114, 4738.
(19) Tauber, M. J.; Kelley, R. F.; Giaimo, J. M.; Rybtchinski, B.;
Wasielewski, M. R. J. Am. Chem. Soc. 2006, 128, 1782.
(20) Borgstrom, M.; Blart, E.; Boschloo, G.; Mukhtar, E.; Hagfeldt,
̈
A.; Hammarstrom, L.; Odobel, F. J. Phys. Chem. C 2005, 109, 22928.
̈
19325
dx.doi.org/10.1021/ja3082268 | J. Am. Chem. Soc. 2012, 134, 19322−19325