Remarkably efficient photocurrent generation based on a [60] fullerence-triosmium cluster/Zn-porphyrin/boron-dipyrrin triad SAM

Article abstract of DOI:10.1002/chem.201000215

A new artificial photosynthetic triad array, a [60]fullerene-triosmium cluster/zinc-porphyrin/boron-dipyrrin complex (1, Os₃C ₆₀/ZnP/ Bodipy), has been prepared by decarbonylation of Os ₃(CO)₈(CN(CH₂)₃Si(OEt) ₃)(μ₃-η²:η²: η²-C₆₀) (6) with Me₃NO/MeCN and subsequent reaction with the isocyanide ligand CNZnP/Bodipy (5) containing zinc porphyrin (ZnP) and boron dipyrrin (Bodipy) moieties. Triad 1 has been characterized by various spectroscopic methods (MS, NMR, IR, UV/Vis, photoluminescence, and transient absorption spectroscopy). The electrochemical properties of 1 in chlorobenzene (CB) have been examined by cyclic voltammetry; the general feature of the cyclic voltammogram of 1 is nine reversible one-electron redox couples, that is, the sum of those of 5 and 6. DFT has been applied to study the molecular and electronic structures of 1. On the basis of fluorescence-lifetime measurements and transient absorption spectroscopic data, 1 undergoes an efficient energy transfer from Bodipy to ZnP and a fast electron transfer from ZnP to C₆₀; the detailed kinetics involved in both events have been elucidated. The SAM of triad 1 (1/ITO; ITO = indium-tin oxide) has been prepared by immersion of an ITO electrode in a CB solution of 1 and diaza bicyclo-octane (2:1 equiv), and characterized by UV/Vis absorption spectroscopy, water contact angle, X-ray photoelectron spectroscopy, and cyclic voltammetry. The photoelectrochemical properties of 1/ITO have been investigated by a standard three-electrode system in the presence of an ascorbic acid sacrificial electron donor. The quantum yield of the photoelectrochemical cell has been estimated to be 29% based on the number of photons absorbed by the chromophores. Our triad 1 is unique when compared to previously reported photoinduced electron-transfer arrays, in that C₆₀ is linked by π bonding with little perturbation of the C₆₀ electron derealization.

Full text of DOI:10.1002/chem.201000215

DOI: 10.1002/chem.201000215
Remarkably Efficient Photocurrent Generation Based on a [60]Fullerene–
Triosmium Cluster/Zn–Porphyrin/Boron–Dipyrrin Triad SAM
Chang Yeon Lee,[a] Jae Kwon Jang,[a] Chul Hoon Kim,[b] Jaehoon Jung,[c]
Bo Keun Park,[a] Jihee Park,[d] Wonyong Choi,[d] Young-Kyu Han,*[c] Taiha Joo,*[b] and
Joon T. Park*[a]
5586
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 5586 – 5599

FULL PAPER
Abstract: A new artificial photosyn-
thetic triad array, a [60]fullerene–trios-
mium cluster/zinc–porphyrin/boron–di-
reversible one-electron redox couples,
that is, the sum of those of 5 and 6.
DFT has been applied to study the mo-
lecular and electronic structures of 1.
On the basis of fluorescence-lifetime
measurements and transient absorption
spectroscopic data, 1 undergoes an effi-
cient energy transfer from Bodipy to
ZnP and a fast electron transfer from
ZnP to C60; the detailed kinetics in-
volved in both events have been eluci-
dated. The SAM of triad 1 (1/ITO;
ITO=indium–tin oxide) has been pre-
pared by immersion of an ITO elec-
trode in a CB solution of 1 and diaza-
bicyclo-octane (2:1 equiv), and charac-
terized by UV/Vis absorption spectros-
copy, water contact angle, X-ray photo-
electron spectroscopy, and cyclic
voltammetry. The photoelectrochemi-
cal properties of 1/ITO have been in-
vestigated by a standard three-elec-
trode system in the presence of an as-
corbic acid sacrificial electron donor.
The quantum yield of the photoelectro-
chemical cell has been estimated to be
29% based on the number of photons
absorbed by the chromophores. Our
triad 1 is unique when compared to
previously reported photoinduced elec-
tron-transfer arrays, in that C60 is
linked by p bonding with little pertur-
bation of the C60 electron delocaliza-
tion.
pyrrin
Bodipy), has been prepared by decar-
bonylation of Os3(CO)8(CN(CH2)3Si-
(OEt)3)(m3-h2:h2:h2-C60)
(6) with
complex
(1,
Os3C60/ZnP/
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
Me3NO/MeCN and subsequent reac-
tion with the isocyanide ligand
CNZnP/Bodipy (5) containing zinc por-
phyrin (ZnP) and boron dipyrrin
(Bodipy) moieties. Triad 1 has been
characterized by various spectroscopic
methods (MS, NMR, IR, UV/Vis, pho-
toluminescence, and transient absorp-
tion spectroscopy). The electrochemi-
cal properties of 1 in chlorobenzene
(CB) have been examined by cyclic
voltammetry; the general feature of
the cyclic voltammogram of 1 is nine
Keywords: electron transfer · mon-
olayers
· photoelectrochemistry ·
self-assembly · triads
Introduction
component, photosynthetic model systems to transmit and
process solar energy.[2] Functionalized C60 molecular systems
for PET processes often consist of electron donors such as
porphyrin,[4] ferrocene,[5] thiophene,[6] and carbon nano-
tubes[7] with additional light-harvesting antenna molecules.
C60 functions as an efficient acceptor because of its intrigu-
ing characteristics that allow acceleration of photoinduced
charge separation and slowing down of charge recombina-
tion in the electron-transfer process. This is mainly due to
its relatively low-lying LUMO and small reorganization en-
ergy.[4c] Various molecular arrays containing D–A sites and
antenna molecules, from dyads to hexads, have been pre-
pared to address the detailed kinetics of energy and electron
transfer and, in a few cases, these have been successfully ap-
plied to solar energy conversion.[8] Fullerene–porphyrin ar-
chitectures for photosynthetic antenna and reaction center
models have been reviewed by Guldi.[9] DꢁSouza and co-
workers have recently reported a fullerene/zinc–porphyrin/
boron–dipyrrin triad based on self-assembled supramolec-
ular methodology, which has revealed photoinduced sequen-
tial energy and electron transfers among the triad compo-
nents.[10] Imahori and co-workers have reported extremely
long-lived charge-separated states of 380 or 530 ms in tet-
rad[8c] and pentad[8d] systems containing ferrocene, porphy-
rin, and fullerene. These systems involve a cascade of photo-
induced energy transfers and multistep electron transfers
within the molecular arrays and mimic natural photosynthet-
ic reaction systems.
Photoinduced electron transfer (PET) in donor–acceptor
(D–A) molecular systems continues to receive considerable
attention for potential applications such as optoelectronics,
photonics, sensors, and other advanced nanostructured devi-
ces.[1] Recent progress in [60]fullerene functionalization has
led to the development of spectacular surface-immobilized
molecular architectures on a wide variety of surfaces and, in
particular, to highly successful application of the prominent
artificial photovoltaic devices by design of integrated, multi-
[a] Dr. C. Y. Lee,+ J. K. Jang,+ Dr. B. K. Park, Prof. J. T. Park
Department of Chemistry and School of Molecular Science
Korea Advanced Institute of Science and Technology
Daejeon, 305-701 (Korea)
Fax : (+82)42-350-5826
E-mail: joontpark@kaist.ac.kr
[b] Dr. C. H. Kim, Prof. T. Joo
Department of Chemistry
Pohang University of Science and Technology
Pohang 790-784 (Korea)
Fax : (+82)54-279-8127
E-mail: thjoo@postech.ac.kr
[c] J. Jung, Dr. Y.-K. Han
Computational Chemistry Laboratory, Corporate R&D
LG Chem Ltd Research Park, Daejeon, 305-380 (Korea)
Fax : (+82)42-865-3160
E-mail: ykhan@kbsi.re.kr
[d] J. Park, Prof. W. Choi
School of Environmental Science and Engineering
Pohang University of Science and Technology
Pohang 790-784 (Korea)
SAMs are highly ordered molecular assemblies that form
spontaneously by chemisorption of linkage-functionalized
molecules onto surfaces and by lateral intermolecular van
der Waals interactions.[11] SAMs and multilayer thin films
with potential gradients have shown promise in imitating
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000215.
Chem. Eur. J. 2010, 16, 5586 – 5599
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5587

J. T. Park, Y.-K. Han, T. Joo et al.
natural photosynthetic systems and in fabricating efficient
and well-defined photovoltaic cells. A variety of SAMs, in-
cluding porphyrin/fullerene,[2a,b] a ferrocene/porphyrin/fuller-
ene triad,[2c,d] a ferrocene/porphyrin/fullerene/boron–dipyrrin
tetrad,[2e] and organometallic fullerene[2j] on gold and
indium–tin oxide (ITO) surfaces have previously been stud-
ied with respect to solar energy conversion. A comprehen-
sive review of the chemistry of supramolecular [60]fullerene
on surfaces has appeared recently.[3] However, there have
been few studies devoted to a detailed examination of pho-
toinduced energy and electron transfers in these artificial
photovoltaic devices.
In our previous work, we described a highly ordered
[60]fullerene triosmium cluster–zinc–porphyrin dyad SAM
on an ITO surface[12] that achieved nearly full coverage with
the aid of diazabicyclooctane (DABCO). This system exhib-
ited the highest photocurrent generation efficiency, at
19.5%, that had been reported for dyad photoelectrochemi-
cal cells based on SAMs. Our C60–metal cluster complexes
are unique in that the C60–metal interaction in m3-h2:h2:h2-
C60 bonding mode perturbs the C60 hybridization very little,
as shown by our SAM, photovoltaic cell application, and X-
ray structural characterization evidence. These earlier stud-
ies also revealed remarkable thermal and electrochemical
stabilities and electronic communication between C60 and
metal cluster centers, which proved uniquely suitable for
various electronic device applications.
Scheme 1. Structures of 1, 2, 5, and 7.
Results and Discussion
Imahori and co-workers have described an extremely high
quantum yield (50%) photoelectrochemical cell that uses a
mixed SAM composed of a boron dipyrrin moiety and a fer-
rocene/porphyrin/fullerene triad.[2e] The boron dipyrrin
(Bodipy; 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene) mole-
cule has often been employed as an antenna light-harvesting
molecule because it has many favorable properties such as
high fluorescence quantum yield, low rates of intersystem
crossing, large molar absorption coefficients, long excited-
state lifetimes, and excellent photostability.[13] Its structures
can also be tailored by a variety of substitutions to tune its
fluorescence characteristic, allowing the development of ap-
plications in the fields of biological labeling,[14] molecular
wires,[15] and artificial photosynthetic systems.[2e]
Synthesis and characterization of 1–7: The synthetic routes
to 1 are summarized in Scheme 2. Compound 2 (Bodipy-
Ald) was prepared in 87% yield by a published procedure,
as described in the Experimental Section. The key synthetic
intermediate 3 was prepared in 8% yield from a mixed con-
densation of the meso-phenyldipyrromethane, 4-iodobenzal-
dehyde, and Bodipy-Ald (2) in the presence of BF3–diethyl
etherate, followed by an in situ Pd-mediated coupling of
crude porphyrin mixtures with N-(4-ethynylphenyl)forma-
mide. Dehydration of 3 by diphosgene afforded isocyanide-
functionalized compound 4 in 75% yield. Metallation of 4
with Zn
ACHTUNGTRENNUNG
bonylation of Os3(CO)8(CN
N
ACHTUNGTRENNUNG
(6) with Me3NO/MeCN, followed by treatment with 5 in
chlorobenzene (CB) at room temperature, gave triad 1 in
35% yield.
In this study, we report full details of the preparation and
characterization of a new triad artificial photosynthetic mol-
ecule,
[60]fullerene–triosmium
cluster/zinc–porphyrin/
Compounds 2–5, were characterized by NMR (1H, 11B,
and 19F), IR, and MALDI-TOF mass spectroscopy, and by
elemental analyses. Formulation of triad 1 is based on the
molecular ion isotope multiplets (m/z (highest peak): 2868)
in the MALDI-TOF mass spectrum and on spectroscopic
and microanalytical data. The IR spectrum of 1 exhibits two
NC stretches: one at 2187 cmꢀ1, due to the silyl-substituted
isocyanide ligand; and one at 2139 cmꢀ1, due to the porphy-
rin-substituted isocyanide ligand 5. An essentially identical
CO stretching pattern to that seen for the structurally char-
boron–dipyrrin (1, Os3C60/ZnP/Bodipy), and describe its ap-
plication to a photoelectrochemical cell based on a SAM of
1 with DABCO on an ITO electrode (1/ITO). Our triad 1 is
unique compared with the previously reported PET arrays,
in that the C60 cage is linked through p bonding. Complexa-
tion between 1 and DABCO in solution has been monitored
1
by H NMR spectroscopy. DFT has been used to study the
molecular and electronic structures of 1. The detailed kinet-
ics of energy and electron transfer have been investigated
by fluorescence-lifetime measurements and transient ab-
sorption (TA) spectroscopy. Structures of the title molecule
(1) and other related key compounds (2, 5, and 7) are de-
picted in Scheme 1.
acterized
1,2-cis,trans-disubstituted
Os3(CO)7-
AHCTUNGTRENNUNG
which implies that the structure of 1 is isomorphous with
that of 8 (see Figure S1 in the Supporting Information). The
5588
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 5586 – 5599

Photocurrent Generation from a Triad SAM
FULL PAPER
Scheme 2. The synthetic routes to 1.
shift of the terminal CO stretching bands of 1 to the low-
energy region by approximately 20 cmꢀ1 compared with
those of 6 reflects an additional electron–donor effect of the
isocyanide ligand 5. The 11B NMR spectra of 1–5 reveal a
triplet resonance for each at 0.40, 0.39, ꢀ0.46, 0.44, and
ꢀ0.44 ppm, respectively. These are due to a single boron
atom coupled to the two equivalent fluorine atoms in 1–5.
The 19F NMR spectra of 1–5 exhibit a quartet resonance for
each at ꢀ146.7, ꢀ146.7, ꢀ146.7, ꢀ146.7, and ꢀ149.3 ppm, re-
spectively. These are due to the two equivalent fluorine
atoms coupled to a boron atom in these compounds. All
1H NMR spectral data for 1–5 and 7 are consistent with the
corresponding structures shown in Schemes 1 and 2
505 nm), and C60 (lmax =350 nm). The spectra of reference
compounds 2 and 7 show absorptions due to Bodipy and
ZnP, respectively. The spectrum of reference compound 5 is
Steady-state absorption and fluorescence spectroscopy of 1,
2, 5, and 7: The UV/Vis spectrum of 1 in CB (Figure 1) ex-
hibits characteristic absorption bands for ZnP (Soret band:
lmax =427 nm; Q bands: lmax =550, 590 nm), Bodipy (lmax
=
Figure 1. UV/Vis absorption spectra of 1, 2, 5, and 7 in CB.
Chem. Eur. J. 2010, 16, 5586 – 5599
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5589

J. T. Park, Y.-K. Han, T. Joo et al.
essentially a sum of the two absorption spectra of 2 and 7,
as expected. The absorption band of Bodipy does not over-
lap that of the ZnP or the C60 moiety, indicating that Bodipy
can be excited selectively in 1 and 5.
The steady-state fluorescence emission spectra of 1 and 5
are provided in Figure 2a. Bodipy excitation of 5 at 505 nm
results in a typical Bodipy fluorescence with an emission
Electrochemistry of 1, 2, 5, 6, and 7: The electrochemical
properties of free C60, 1, 2, 5, 6, and 7 have been examined
by cyclic voltammetry in CB solutions, with tetrabutylam-
monium perchlorate as the supporting electrolyte (Figure 3)
and the corresponding half-wave potentials (E1/2) are sum-
marized in Table 1.
The data for free C60 are included for comparison. Com-
pounds 2 and 7 show a reversible one-electron redox wave
at E1/2 =ꢀ1.70 V and four reversible one-electron redox
waves at E1/2 =0.65, 0.33, ꢀ1.87, and ꢀ2.17 V, respectively,
which are attributed to redox reactions at Bodipy and ZnP
centers. The cyclic voltammogram (CV) of 5 is the sum of
those of 2 and 7, as expected, with redox waves at E1/2
=
0.64, 0.32, ꢀ1.68, ꢀ1.87, and ꢀ2.08 V. The CV of 6 reveals
four reversible one-electron redox waves, with the third and
fourth waves overlapping. These have been observed simi-
larly in various known complexes, including Os3(CO)8-
ACHUTNERGUNNG 9ꢀnHCAUTNGTREN(NGUN CNCH2Ph)n(m3-
(PMe3)(m3-h2:h2:h2-C60)17 and Os3(CO)
h2:h2:h2-C60) (n=2, 3, and 4),[16] due to C60 reductions com-
bined with electronic communication between the C60 and
the metal-cluster centers. The general feature of the CV of 1
is again the sum of those of 5 and 6, as expected. It shows
nine reversible one-electron redox couples with the sixth,
seventh, and eighth waves overlapping. On the basis of
these observations, the nature of each redox wave of triad 1
is assigned unambiguously in Figure 3 and in Table 1. The
first and second redox waves at E1/2 =0.51 and 0.34 V corre-
spond to the redox reactions of ZnP1+/2+ and ZnP0/1+, re-
spectively. The third and fourth redox waves at E1/2 =ꢀ1.09
0/1ꢀ
and ꢀ1.43 V are due to the redox reaction of C60
and
C601ꢀ/2ꢀ, respectively. The fifth couple at E1/2 =ꢀ1.69 V is at-
tributed to the redox reaction of Bodipy0/1ꢀ. The sixth to
eighth waves at E1/2 =ꢀ1.82, ꢀ1.82, and ꢀ1.90 V overlap,
and correspond to C602ꢀ/3ꢀ, C603ꢀ/4ꢀ, and ZnP0/1ꢀ, respectively.
The ninth redox wave for ZnP1ꢀ/2ꢀ appears at E1/2 =ꢀ2.12 V.
The molecular and electronic structures of 1 have been
studied by applying DFT using the BPW91 functional.
Figure 4 displays the optimized structure, the HOMO, and
the LUMO. The two isocyanide ligands are coordinated in a
cis,trans fashion on the triosmium cluster. The triosmium tri-
angle, ZnP, and Bodipy moieties are almost coplanar and
perpendicular to the four coplanar phenyl rings in the
spacer. The majority of the HOMO is located on the ZnP
moiety, whereas the LUMO is found to be entirely on the
Figure 2. a) Steady-state fluorescence spectra of 1 and 5 in CB. b) Steady-
state fluorescence spectra of 2, 7, and a 1:1 mixture of 2 and 7 in CB. The
concentration (0.8 mm) was kept constant for each of the species. Samples
were excited at 505 nm.
maximum at 518 nm, along with ZnP emissions at 601 and
649 nm, implying that a definite intramolecular energy
transfer takes place from Bodipy to ZnP in 5. To address
the nature of this energy transfer, fluorescence measure-
ments were performed for 2, 7, and a 1:1 mixture of 2 and 7
(Figure 2b). Excitation of 2 at 505 nm reveals a Bodipy
emission, whereas that of 7 shows no emission, which indi-
cates that ZnP is not excited by radiation at 505 nm. Fur-
thermore, excitation of a 1:1 mixture of 2 and 7 at 505 nm
gives a simple sum of the emission spectra of 2 and 7, indi-
cating no intermolecular energy transfer from 2 to 7; this
observation agrees with studies reported earlier.[10] Howev-
er, when triad 1 is excited at 505 nm, ZnP emission is
quenched by over 96% compared with that of 5 (Figure 2a).
There is no appreciable C60 emission in the 700–750 nm
range, so these results show clearly that an efficient electron
transfer occurs from ZnP to C60 in 1. Overall, triad 1 under-
goes an energy transfer from Bodipy to ZnP and an electron
transfer from ZnP to C60.
C
60 moiety.
Time-resolved fluorescence spectroscopy of 1, 2, 5, and 7:
To investigate the kinetics of the energy and electron trans-
fers in triad 1, time-resolved fluorescence (TRF) studies for
1, 2, 5, and 7 were performed by time-correlated single-
photon counting and fluorescence up-conversion methods.
Multiexponential fit results for all the TRF signals are sum-
marized in Table 2.
TRF of Bodipy (2) or of ZnP (7) alone displays single-ex-
ponential decay, to give the S1 state lifetime of 1.9 or 1.8 ns,
respectively (Figure S2 in the Supporting Information).
When the Bodipy moieties in 5 and 1 are excited by 500 nm
5590
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 5586 – 5599

Photocurrent Generation from a Triad SAM
FULL PAPER
550 nm pulses leads to a slow
1.6 ns single-exponential decay,
without a rise component for
the ZnP emission, which sup-
ports the assumption that the
23 ps rise component is due to
energy transfer. The 1.6 ns life-
time of the ZnP moiety in 5
matches well with the lifetime
of 1.8 ns found in 7.
In contrast to 5, TRF of the
ZnP in the triad 1 shows a fast
decay in 35–50 ps, indicating a
fast electron transfer from ZnP
to C60 as confirmed by TA spec-
troscopy (see below). Interest-
ingly, the electron-transfer rate
in triad 1 is somewhat depen-
dent on the excitation wave-
length, as shown in Table 2. The
small slow-decay components
(t3) comprising about 1% of
the total decay may be due to
an impurity and/or to the triad
in a conformation that is not fa-
vorable for charge transfer.
Free-energy changes of 1 and 5
for charge recombination (CR),
charge separation (CS), and
energy transfer (ENT) in CB
were calculated by using Weller
equations.[18] The center-to-
center distance (RCC) between
ZnP and C60 moieties in the
Weller equation is calculated to
be 21 ꢂ from the optimized
structure (see above). These re-
sults are summarized in Table 3,
as are the calculated rate con-
stants for CR, CS, and ENT
processes. The magnitudes of
rate constants of 1 and 5 for CS
and ENT are comparable to the
values reported for other simi-
lar artificial photosynthetic sys-
tems.[4] Energy levels, ENT,
electron-transfer (ET), and re-
laxation processes of 1 are sum-
Figure 3. CVs of 1, 2, 5, 6, and 7 in dry deoxygenated CB (supporting electrolyte 0.1m [(nBu)4N]ACTHNUTRGNE[UGN ClO4]; scan
rate 10 mVsꢀ1).
Table 1. Half-wave potentials (E1/2 vs. EAFc/Fc +) of C60, 1, 2, 5, 6, and 7 in CB.
0/1ꢀ
1ꢀ/2ꢀ
2ꢀ/3ꢀ
3ꢀ/4ꢀ
Compound
ZnP1+/2+
ZnP0/1+
C60
C60
B0/1ꢀ
C60
C60
ZnP0/1ꢀ
ZnP1ꢀ/2ꢀ
[a]
C60
–
–
0.65
0.64
–
–
–
0.33
0.32
–
ꢀ1.06
ꢀ1.43
ꢀ1.91
ꢀ2.38
–
–
–
–
2[b]
7[b]
5[b]
6[b,c]
1[b]
–
–
ꢀ1.70
–
–
–
–
–
–
–
–
–
–
–
ꢀ1.87
ꢀ1.87
–
ꢀ1.90
ꢀ2.17
ꢀ2.08
–
ꢀ2.12
ꢀ1.68
ꢀ1.05
ꢀ1.41
–
ꢀ1.84
0.51
0.34
ꢀ1.09
ꢀ1.43
ꢀ1.69
ꢀ1.82
[a] See reference [17]. [b] This work. [c] See reference [24].
marized
Figure 6.
schematically
in
pulses, much faster decays of the Bodipy emission at
525 nm, and the corresponding rises of the Q-band emission
of ZnP at 650 nm, are observed (Figure 5). This observation
clearly demonstrates the occurrence of photoinduced energy
transfer from the Bodipy to ZnP moieties in 5 and 1. The
energy-transfer time constants are 31 and 41 ps in 5 and 1,
respectively. Direct excitation of the ZnP moiety in 5 by
Transient absorption spectroscopy: TA spectra of 1 were
measured to verify the formation of the charge-separated
states in triad 1 (Figure 7). They show a negative peak at
525 nm at an early stage due to the stimulated emission of
the Bodipy S1 state. More importantly, photoinduced ab-
sorption develops immediately after the excitation over the
Chem. Eur. J. 2010, 16, 5586 – 5599
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5591

J. T. Park, Y.-K. Han, T. Joo et al.
Figure 4. Compound 1: a) Optimized structure; b) HOMO; c) LUMO.
Figure 5. Time-resolved fluorescence signals measured at 525 nm and
650 nm in CB for a) 5 and b) 1. Insets: TRF signals at early times (excita-
tion wavelength 500 nm).
Table 2. Multiexponential fit for the TRF of 1, 2, 5, and 7 in CB detected
at the Bodipy (525 nm) and ZnP (650 nm) emission wavelengths.[a]
Bodipy (525 nm)
t1 [ps]
ZnP (650 nm)
t1 [ps] t2 [ps] t3 [ns]
Compound lex [nm]
Table 3. Energy and electron-transfer rate constants and free-energy
changes of 1 and 5 in CB.
2
7
500
550
500
550
500
550
1900
–
31
–
41
–
–
–
–
–
–
–
–
[a]
[b]
[a]
[b]
[a]
[b]
1.8
1.6
1.6
Compound DGENT
kENT
[sꢀ1
DGCS
[eV]
kCS
DGCR
[eV]
kCR
ꢀ23
[eV]
]
[sꢀ1
]
[sꢀ1
]
5
1
–
5
1
ꢀ0.33
ꢀ0.33
3.1ꢃ1010
–
–
–
–
ꢀ31
–
50 (99.4) 1.6 (0.6)
35 (98.7) 0.94 (1.3)
2.4ꢃ1010 ꢀ0.75
1.9ꢃ1010 ꢀ1.31
2.9ꢃ109
[a] DGENT is the energy difference between the excited singlet state of
Bodipy and that of ZnP in 1 and 5. The DGCS and DGCR values of 1 are
[a] A negative time constant indicates a rise. Values in parentheses are
relative amplitudes [%].
calculated from the Weller equations (DGCS =E
(C60)ꢀE0ꢀEc and DGCR =EredA(C60)ꢀEoxA
(ZnP)ꢀEc). Eox and Ered are ZnP0/
(0.34 V) and C60 (1.09 V), respectively. E0 is the energy of the 0–0
oxA
-
A
E
CHTUNGTRENNUNG
1+
0/1ꢀ
whole detection wavelength region, and it increases gradual-
ly to 500 ps. It is well known that charge separation and re-
combination dynamics can be monitored directly by probing
the ZnP radical cation (ZnP+·) and the C60 radical anion
(C60ꢀ·), which show transient absorption bands centered at
approximately 650 and 1000 nm, respectively.[4c,e,h] Unfortu-
nately, however, the S1 state of ZnP is also known to absorb
strongly near the same wavelength region.[19] The TAs of
Bodipy (2) and ZnP (7) (Figures S3 and S4, respectively, in
the Supporting Information) show clearly that both absorb
strongly in the same wavelength region. The lifetimes of the
Bodipy and ZnP moieties in 1 are 41 and 50 ps, respectively,
and, therefore, the photoinduced absorptions can be safely
assigned within approximately 50 ps to the excited-state ab-
sorptions of the S1 states of Bodipy and ZnP. The photoin-
duced absorptions in the visible and near-infrared (NIR) re-
gions, which show a slow increase of up to 500 ps, should
transition between the lowest excited state and ground state of ZnP. Ec
(coulombic energy) is calculated as e2/(4pe0esRcc), in which Rcc is the
center-to-center distance between ZnP and C60 in 1 and es is the dielectric
constant of the solvent (5.7 for CB). [b] Rate constants for ENT, CS, and
ꢀ1
CR are calculated from the equations: kENT(5)=t(CNZnP-Bodipy*) ꢀt(Bodipy-
ꢀ1
ꢀ1
ꢀ1
,
kENT(1)=t
ꢀt(Bodipy-Ald*)
,
kCS =t
(Os3C60/ZnP*/Bodi-
Ald*)
(Os3C60/ZnP/Bodipy*)
ꢀt(CNZnP*)ꢀ1, and kCR =t
.
ꢀ1
ꢀ1
py)
(Os3C60/ZnP+/Bodipy)
Figure 8 shows the TA signals of 1, in which Bodipy and
ZnP moieties are excited by 500 and 550 nm pulses, respec-
tively, and the subsequent transient species are probed at
two different wavelengths, 650 and 1000 nm. Note that all
the monomer moieties included in triad 1 show strong excit-
ed-state absorptions at both detection wavelengths. To de-
scribe the TA signals properly, a kinetic scheme that in-
cludes all the chemically and spectroscopically distinct spe-
cies needs to be considered. To check the possibility that the
excited state of the fullerene moiety (C60*) is generated, we
measured the TRF signal at 720 nm (Figure S5 in the Sup-
porting Information) at which the C60* emits.[4e,20] We did
ꢀ·
then be assigned to the ZnP+· and C60 species, respectively.
The charge separation and recombination kinetics in triad
1 were investigated by TA at single detection wavelengths.
5592
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 5586 – 5599

Photocurrent Generation from a Triad SAM
FULL PAPER
Figure 6. Energy levels, ENT, ET, and relaxation processes of 1.
not observe any decay component that may be responsible
for the C60* emission when the Bodipy (500 nm) or the ZnP
Q band (550 nm) is excited. In fact, the TRF signal was
almost the same as the Q-band emission at 650 nm, which
indicates that the possibility of the generation of the charge-
separated products from the C60*, and a kinetic scheme in-
volving the C60* can be ignored. Most of the charge-separat-
ed products should, therefore, be generated from the singlet
excited state of ZnP. Any con-
Figure 7. Transient absorption spectra of the triad 1 in CB (excitation
wavelength 500 nm). Each curve is displaced by 0.01 OD (OD=optical
density) from the preceding curve.
k1
k2
k3
k4
k5
A
B
C
D
E
groundstate
ð1Þ
ꢀ! ꢀ! ꢀ! ꢀ! ꢀ!
tribution of the unreactive ZnP
can be ignored because of its
negligible amplitude in the
TRF measurement. All the TA
signals in Figure 8 have been
fitted by the global fitting
method. They, and especially
those probed at 1000 nm, show
a prominent rise that is much
slower than the lifetime of the
S1 state of ZnP, which clearly
requires an intermediate before
the charge-separated state un-
dergoes the charge recombina-
tion process.
We have included two such
intermediate charge-separated
states to fit all the TA data sat-
isfactorily. Thus, energy trans-
fer, charge separation, and re-
combination reaction schemes
can be proposed as in [Eq. (1)],
in which
Bodipy*,
Bodipy, and
A
B
is Os3C60/ZnP/
is Os3C60/ZnP*/
(ꢀ)
C
is Os3C60
/
ZnP(+)/Bodipy. D and E can be
regarded as the charge-separat-
ed states before the charge-re-
combination process.
Figure 8. Transient absorption signal of 1 in CB. Excitation and probe wavelengths (nm): a) 500, 650; b) 500,
1000; c) 550, 650; d) 550, 1000, respectively.
Chem. Eur. J. 2010, 16, 5586 – 5599
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5593

J. T. Park, Y.-K. Han, T. Joo et al.
D and E are introduced to represent the conformational
and other relaxation processes of the nascent charge-sepa-
rated state C, before it undergoes the charge recombination
process, and they may not represent distinct chemical spe-
cies. Thus, k5 should be the charge recombination rate. The
rate equations for each component can be expressed by the
matrix given in [Eq. (2)].
excited state could be one of the possible candidates respon-
sible for the relaxation, because the 400 ps timescale may
correspond to the slow structural changes. Although the
conformational dynamics in the ground state may not be
feasible in our triad system owing to the rigid structure, the
conformational dynamics in the excited state may become
possible because the conjugation in the excited state is much
weaker than that in the ground state. Further studies are
needed to identify the origin of the species D and E, which
are beyond the scope of this work.
All the TA signals follow the proposed reaction scheme
qualitatively. The relative contribution of each component
to the TA signal shows a dependence on the excitation
wavelength. When Bodipy is excited, the contribution of the
ZnP absorption is higher than that of the charge-separated
products (C, D, and E), whereas when the Q band of ZnP is
excited, this is reversed, indicating that charge-separation
dynamics and subsequent relaxation processes depend sub-
stantially on the energetics between the Q band of ZnP and
C60. Although chemical properties of the charge-separated
intermediates C, D, and E, as well as the kinetics between
these species, have not been identified, the decay rate (k5)
of E may correspond directly to the charge-recombination
process.
A complete set of rate equations from [A(t)] to [E(t)] can
be solved from [Eq. (2)] by using the determinant
method.[21] All the transient absorption signals are fitted
with a model function [Eq. (3)], in which Xm denotes a rela-
tive amplitude for each X(t).
X
S ¼
Xm½XðtÞꢁ
ð3Þ
X¼AꢀꢀE
Compared to the results of recent TA experiments for
similar dyad and triad systems,[10,4h] charge recombination ki-
netics with a time constant of about 350–400 ps seem to be
quite fast, although similar experimental results for the
dyads were reported in picosecond TA experiments by Ima-
hori and co-workers.[4c] DꢁSouza et al. reported a charge-re-
combination time of 5 ns in a supramolecular triad in o-di-
chlorobenzene,[10] although this value has a large uncertainty
because a flash-photolysis technique with nanosecond time
resolution was employed. In contrast, Schuster et al. report-
ed time constants in the range 145–435 ns for an azoben-
zene-linked dyad in tetrahydrofuran.[4h] However, they also
reported dramatically decreased lifetimes in the range
600 ps–1.2 ns for the same sample in benzonitrile. Therefore,
if we consider that the kinetics of the recombination process
may be highly solvent-dependent, our results are quite simi-
lar to the results for these systems. Our system may possibly
show slower recombination rates in other polar solvents,
such as tetrahydrofuran.
In the fitting procedure, the initial concentration is con-
strained to [A]0 =1 or [B]0 =1, depending on the excitation
wavelength, and k1 and k2 are fixed to the values obtained
from the TRF measurements. The fit results were overlaid
with the experimental curve in Figure 8, and the amplitude
of each component (Xm) and the rate constants are listed in
Table 4. When the intermediates D and E are not included
Table 4. Fitted amplitudes and rate constants for the transient absorption
signals of triad 1 in chlorobenzene.[a]
ꢀ1
lex [nm] lprobe [nm] k1ꢀ1 [ps] k2ꢀ1 [ps] k3ꢀ1 [ps] k4ꢀ1 [ps]
k5 [ps]
650
1000
650
41 (6)
41 (5)
–
–
50 (31) 350 (12) 340 (20)
350 (31)
500
550
50 (10) 290 (15) 310 (ꢂ0) 340 (71)
35 (35) 350 (20) 400 (13)
35 (15) 430 (20) 430 (17)
370 (32)
440 (48)
1000
[a] Values in parentheses are relative amplitudes [%].
in the model, the fitting is not satisfactory because the slow
rise component of about 400 ps shown in all the TA signals
cannot be described by a three- or four-step model. Hence,
inclusion of the relaxation processes of the nascent charge-
separated state is required, and they may change the absorp-
tion spectrum of the charge-separated state. Note that the
nascent charge-separated state formed from the singlet ex-
cited state of ZnP may be in a vibrationally excited state,
which may undergo vibrational relaxation. This process can
be ignored, however, because the vibrational relaxation
occurs in a few tens of picoseconds, much faster than the ob-
served 400 ps.[22] Conformational dynamics such as the twist-
ing and bending motions around the chromophores in the
Interaction of 1 with DABCO: To elucidate the interaction
of 1 with DABCO, the complexation between 1 and
DABCO in solution was monitored by H NMR spectrosco-
1
py (Figure 9). When DABCO (0.5 equiv) was added to a so-
lution of 1 in chloroform (5.5 mm) two resonances in a 1:1
ratio at ꢀ1.88 and ꢀ4.61 ppm, shifted upfield relative to free
DABCO at 2.80 ppm, were observed. These were due to the
six inner methylene hydrogen atoms (denoted by a red dot)
and six outer methylene hydrogen atoms (denoted as a blue
dot), respectively, indicating that a square-pyramidal 1:1
complex between 1 and DABCO is formed with 0.5 equiv of
free 1 (Figure 9a). The upfield shift and the chemical shift
difference between the inner and outer hydrogen atoms
5594
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 5586 – 5599

Photocurrent Generation from a Triad SAM
FULL PAPER
Figure 9. DABCO hydrogen resonances of 1–DABCO complexes in a) 5.5 mm, b) 8.8 mm, and c) 17 mm 1 in CDCl3.
were attributed to the well-known ring-current effect of the
porphyrin ring. At a higher concentration (17 mm) of 1,
however, a single resonance at ꢀ1.88 ppm was observed for
the 12 methylene hydrogen atoms of DABCO, which implies
formation of a 2:1 complex between 1 and DABCO (Fig-
ure 9c). At the intermediate concentration of 8.8 mm, the in-
crease in the intensity ratio of the two resonances at ꢀ4.61
and ꢀ1.88 ppm is indicative of a higher concentration of the
2:1 complex than the 1:1 complex, as expected (Figure 9b).
Upon formation of the 1:1 or 2:1 complex, the resonance of
the b-hydrogen atoms of the pyrrole ring in ZnP shifts up-
field by 0.15–0.35 ppm, reflecting the electron-donor effect
of the DABCO ligand. Similar concentration-dependent
complexations and the upfield shift of ZnP hydrogen atoms
have been reported previously in the interaction of zinc–por-
phyrin derivatives and DABCO forming the 1:1 or 2:1 com-
plex.[23]
Formation of 1/ITO SAM: The SAM of 1 (1/ITO) has been
prepared by immersion of an ITO electrode in a CB solu-
tion of 1 and DABCO (2:1 equiv) at 1008C for 7 h, followed
by washing with CB and dichloromethane (DCM;
Scheme 3). The absorption spectrum of 1/ITO (Figure 10)
shows a characteristic Soret band at lmax =434 nm, which is
broader and red-shifted by approximately 7 nm relative to
that of 1 in solution. The broadening and bathochromic shift
indicate moderate perturbation within SAMs, mainly due to
DABCO binding and porphyrin aggregation.
Scheme 3. Formation of 1/ITO SAM.
Si(2p), Zn(2p), and Os(4f) peaks involved in the SAM, as
shown in Figure S7 in the Supporting Information.
The water contact angle (qH O) of 708 for 1/ITO is in-
The CVs of 1/ITO in the ZnP oxidation region, with and
without DABCO, reveal two well-resolved one-electron
redox waves at E1/2 =0.41, and 0.69 V for the ZnP oxidations
(Figure 11); in the CV of 1/ITO there is also a reversible
one-electron redox wave at E1/2 =ꢀ1.20 V for the first re-
duction of the C60 cage (Figure 11, inset). Other redox cou-
2
creased substantially from qH O =458 for bare ITO (Fig-
2
ure S6 in the Supporting Information), which is consistent
with adsorption of hydrophobic 1 on the ITO surface. The
X-ray photoelectron spectra (XPS) further confirm the for-
mation of 1/ITO by detection of C(1s), N(1s), O(1s), F(1s),
Chem. Eur. J. 2010, 16, 5586 – 5599
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5595

J. T. Park, Y.-K. Han, T. Joo et al.
appeared immediately upon irradiation of the ITO electrode
with l=440 nm light (0.546 mWcmꢀ2) at an applied poten-
tial of 0.1 V versus Ag/AgCl (3m NaCl) and the response
was repeated reversibly (Figure 12). The anodic photocur-
Figure 10. UV/Vis absorption spectra of 1 and 1/ITO in CB.
Figure 12. Photocurrent response of 1/ITO irradiated with 440 nm light
(0.546 mWcmꢀ2) and at an applied potential of 0.1 V vs. Ag/AgCl (3m
NaCl).
rent increased in proportion to an increase in the positive
bias of the ITO electrode (Figure S8 in the Supporting Infor-
mation). The action spectrum of the 1/ITO/AsA/Pt cell,
which is compared to the UV/Vis absorption spectra of 1/
ITO and 1 in CB in Figure 13, reveals a maximum current
at 440 nm, indicating that porphyrin is a dominant photoac-
tive sensitizer for photocurrent generation in this triad pho-
toelectrochemical cell.
Figure 11. Cyclic voltammogram of the ZnP oxidation region of 1/ITO in
DCM with 0.1m Bu4NPF6 as electrolyte (scan rate 0.1 Vsꢀ1). Inset: the
reduction region of 1/ITO.
ples, however, are broad and are not as well resolved as
those of 1 in solution. Presumably this is because of the het-
erogeneous environment of electrochemically active species
in the SAM. The monolayer surface coverage (G) of 1/ITO
with and without DABCO is estimated to be 1.2ꢃ10ꢀ10 and
0.5ꢃ10ꢀ10 molcmꢀ2, respectively, from the integrated charge
of the first oxidation peak at 0.41 V. This result indicates
more than 85% surface coverage of 1/ITO, based on the
theoretical value of 1.4ꢃ10ꢀ10 molcmꢀ2 calculated for a
close-packed monolayer of 1 in perpendicular orientation to
the ITO surface. Well-ordered structural confinement,
through strong interaction of ZnP with a bifunctional base
DABCO, may be responsible for this high surface coverage,
as shown in Scheme 3. The red shift of the Soret band in the
UV/Vis spectrum of 1 (see above) implies that the axial
DABCO binding to the dz2 orbital of zinc lowers the energy
of the porphyrin excited singlet state and also enhances the
surface coverage of 1 on the ITO electrode.
Figure 13. Absorption spectra of 1 and 1/ITO and the action spectrum of
a 1/ITO/AsA/Pt cell irradiated with 0.546 mWcmꢀ2 light at 0.1 V bias vs.
Ag/AgCl.
The estimated quantum yield for the 1/ITO/AsA/Pt cell is
29%, based on the number of photons absorbed by the
chromophores, and its incident photon-to-current efficiency
(IPCE) at 440 nm is 2.0%. The present triad 1/ITO/AsA/Pt
photoelectrochemical cell exhibits a remarkably high quan-
tum yield, which is much higher than that of our previously
reported dyad cell (19.5%) based on the Os3C60/ZnP array.
This superior performance of the 1/ITO/AsA/Pt photoelec-
trochemical cell might be attributed to the higher conjuga-
Photoelectrochemical properties: Photocurrent measure-
ments for 1/ITO were carried out in 0.1m Na2SO4 solution
containing ascorbic acid (AsA; 50mm) as a sacrificial elec-
tron donor, by using the three-electrode system of the modi-
fied ITO substrate as a working electrode, a Pt counter elec-
trode, and a Ag/AgCl (3m NaCl) reference electrode (de-
noted as a 1/ITO/AsA/Pt cell). A stable anodic photocurrent
5596
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 5586 – 5599

Photocurrent Generation from a Triad SAM
FULL PAPER
spectrophotometer. 1H (400 MHz), 11B (96 MHz), and 19F (282 MHz)
NMR spectra were recorded on Bruker Avance-400, AM 300, and
DRX 300 spectrometers, respectively. Chemical shifts of 1H, 11B, and
19F NMR spectra were referenced to tetramethylsilane, boron trifluoride
etherate, and chlorofluoromethane, respectively. MALDI-TOF mass
spectra were obtained on a Voyager DE-STR spectrometer. Elemental
analyses were performed on a Fisons EA1110 elemental analyzer. The
syntheses of 1, 2, 3, 4, and 5 are described in the Supporting Information.
tion of the present triad 1 than of the previous dyad. In par-
ticular, the 1/ITO/AsA/Pt cell has a much better light-har-
vesting property than the previously reported Os3C60–ZnP
array, with absorption between 500–550 nm (Figure S9 in
the Supporting Information) indicating efficient energy
transfer from Bodipy to ZnP and electron transfer from
ZnP to C60 in the 1/ITO/AsA/Pt cell. This triad photochemi-
cal cell has great potential for solving the poor light-harvest-
ing properties of porphyrin[24] molecules in the blue-green
(450–600 nm) solar spectral regions and for mimicking natu-
ral photosynthetic systems through sequential energy and
electron transfer.
Preparation of 1/ITO: An ITO glass was cleaned with acetone and dried
by blowing a N2 stream over the surface. The ITO electrode was im-
mersed in a CB solution containing 1 and DABCO (total concentration
1mm; molar ratio 1/DABCO=2:1) at 1008C for 7 h under an argon at-
mosphere. The electrode was rinsed and sonicated in CB to remove phys-
isorbed 1, then washed three times with CB and DCM.
Electrochemical measurements: Cyclic voltammetry was carried out on
an Autolab PGSTAT 10 electrochemical analyzer (Eco Chemie, The
Netherlands) using the conventional three-electrode system of a platinum
working electrode (1.6 mm diameter disk, Bioanalytical Systems), a plati-
num-wire counter electrode (wire length 5 cm, diameter 0.5 mm ), and a
Ag/Ag+ reference electrode (0.1m AgNO3/Ag in acetonitrile with a Vy-
corTM salt bridge). All measurements were performed at ambient temper-
ature under a nitrogen atmosphere in a deoxygenated [(nBu)4N]ClO4
(0.1m) solution in dry CB. The analyte concentration was approximately
3ꢃ10ꢀ4 m. All potentials were referenced to the ferrocene/ferrocenium
(Fc/Fc+) standard. The relative number of electrons involved in each re-
duction process was obtained from a graph of current versus timeꢀ1/2 ac-
cording to the Cottrell equation.[30]
Conclusion
We have successfully constructed a highly ordered, nearly
fully covered [60]fullerene–triosmium cluster/zinc–porphy-
rin/boron–dipyrrin triad SAM on an ITO surface with the
aid of DABCO. This cell exhibits the highest photocurrent
generation efficiency, at 29%, ever reported for triad photo-
electrochemical cells based on SAMs. The detailed kinetics
involved in energy and electron transfer of the triad cluster
have been fully elucidated by fluorescence-lifetime measure-
ments and TA spectroscopic data. The remarkable quantum
yield of our triad photoelectrochemical cell is attributed to
the unique molecular characteristics of the cluster arrays:
1) they are thermally and electrochemically very stable,
2) they undergo facile electronic communication between
C60 and the metal cluster moieties, 3) the C60–metal interac-
tion in the m3-h2:h2:h2-C60 bonding mode hardly perturbs the
C60 hybridization, 4) well-ordered structural confinement on
the surface through strong interaction between the Zn2+ ion
and the bifunctional DABCO base results in an extremely
high surface coverage of the SAM, and 5) efficient energy
transfer from Bodipy to ZnP and electron transfer from
ZnP to C60 occur due to the excellent light-harvesting prop-
erty of Bodipy in the blue-green region. The present suc-
cessful application of C60–metal cluster complexes in photo-
electrochemical cells promises other useful technological ap-
plications of C60–metal-cluster-based SAMs for molecular
electronic device fabrication.
Electrochemical measurement of 1/ITO: Cyclic voltammetry was carried
out with an AUTOLAB PGSTAT 10 electrochemical analyzer, using the
conventional three-electrode system of a modified ITO working elec-
trode (electrode area 0.39 cm2), a platinum-wire counter electrode (wire
length 5 cm, diameter 0.5 mm ), and a Ag QRE reference electrode. All
measurements were performed at ambient temperature under a nitrogen
atmosphere in a [(nBu)4N]PF6 solution (0.1m) in dry DCM. All potentials
were referenced to the Fc/Fc+ standard. Surface coverage (G molcmꢀ2
)
was calculated from G=Q/nF. The charge density Q (mCcmꢀ2) was calcu-
lated by integrating the faradaic current peak after subtracting a baseline
of the charging current at a unit area of electrode. F is the Faraday con-
stant and n is the number of electrons transferred. The real surface area
of the ITO electrode (0.39 cm2) was determined by the electrochemical
method based on mass-transfer and adsorption processes.
Photoelectrochemical measurements: Photoelectrochemical measure-
ments were performed on 1/ITO in a specially designed Teflon cell illu-
minated with monochromatic excitation light through a monochromator
(Thermo Oriel, model 77250) by a 300 W xenon lamp (Thermo Oriel,
model 6259). The photocurrent was measured in a three-electrode ar-
rangement (Gamry, Reference 600), a modified ITO working electrode, a
platinum-wire counter electrode, and a Ag/AgCl (3m NaCl) reference
electrode. The light intensity was monitored by an optical power meter
(Newport 1830-C) and corrected. The quantum yield and IPCE of the
photocurrent generation were obtained from f=(i/e)/[I
ACHTUNGTRENNUNG
IPCE (%)=100ꢃ1240ꢃIsc/(Wl), in which I=(Wl)/(hc), i is the photo-
current density, e is the elementary charge, I is the number of photons
per unit area and unit time, l is the wavelength of light irradiation, A is
the absorbance of adsorbed dyes at l nm, W is the light power irradiated
at l nm, c is the velocity of light, h is the Planck constant, and Isc is the
short-circuit photocurrent.
Experimental Section
General: All reactions were carried out under a nitrogen atmosphere by
using standard Schlenk techniques. Solvents were dried over appropriate
drying agents and distilled immediately before use. C60 (99.5%, Southern
Chemical Group LLC) was used without further purification. Anhydrous
trimethylamine N-oxide (m.p. 255–2308C) was obtained from
Me3NO·2H2O (98%, Aldrich) by sublimation (three times) at 90–1008C
under vacuum. Silica gel (Fuji Silysia BW-200T) was used for flash
column chromatography. 4-Ethynylbenzaldehyde,[25] 4,4-difluoro-8-(4’-
iodophenyl)-1,3,5,7-tetramethyl-4-bora-3a,4a-diaza-s-indacene,[26] meso-
Time-resolved fluorescence measurements: Each of the samples was dis-
solved in CB (Aldrich, Spectroscopic Grade). Sample concentrations
were approximately 0.5 mm. In time-resolved experiments, a cuvette
(200 mm thick) containing the sample solution was mounted on a home-
made shaking stage, with a speaker to minimize photodamage. All the
experiments were carried out at ambient temperature. Picosecond TRF
was measured by the time-correlated single-photon counting (TCSPC)
method. Light sources were a homebuilt, cavity-dumped optical paramet-
ric oscillator (OPO)[31] and a homebuilt optical parametric amplifier
(OPA), which provided visible pulses at 550 and 500 nm, respectively. In
TCSPC, a singlet lens was used to focus the excitation beam onto the
phenyldipyrromethane,[27]
Os3(CO)8(CN(CH2)3Si
(OEt)3)(m3-h2:h2:h2-C60) (6),[29] and Os3(CO)7-
(CNCH2Ph)2(m3-h2:h2:h2-C60) (8)[16] were prepared according to published
methods. Infrared spectra were obtained with a Bruker Equinox-55 FTIR
N-(4-ethynylphenyl)formamide,[28]
G
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
Chem. Eur. J. 2010, 16, 5586 – 5599
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5597

J. T. Park, Y.-K. Han, T. Joo et al.
sample and the fluorescence was collected in a back-scattering geometry
by using a parabolic mirror. The emission was sent to a monochromator
and detected with a thermoelectrically cooled MCP-PMT (Hamamatsu
R3809U-51). Magic-angle detection was used to avoid the effect of polar-
ization. The full width at half maximum (FWHM) of the instrumental re-
sponse was 40 ps, to provide time resolution better than 10 ps after de-
convolution. Femtosecond time resolution in TRF was achieved by fluo-
rescence up-conversion. The apparatus with time resolution of 50 fs has
been described in detail previously.[32]
2005, 21, 6385–6391; h) S. Kang, T. Umeyama, M. Ueda, Y. Matano,
H. Hotta, K. Yoshida, S. Isoda, M. Shiro, H. Imahori, Adv. Mater.
2006, 18, 2549–2552; i) T. Hasobe, H. Imahori, P. V. Kamat, T. K.
Ahn, S. K. Kim, D. Kim, A. Fujimoto, T. Hirakawa, S. Fukuzumi, J.
Am. Chem. Soc. 2005, 127, 1216–1228; j) Y. Matsuo, K. Kanaizuka,
K. Matsuo, Y. Zhong, T. Nakae, E. Nakamura, J. Am. Chem. Soc.
2008, 130, 5016–5017.
[3] a) H. Imahori, S. Fukuzumi, Adv. Funct. Mater. 2004, 14, 525–536;
b) D. Bonifazi, O. Enger, F. Diederich, Chem. Soc. Rev. 2007, 36,
390–414.
Transient absorption measurements: All the samples were prepared and
handled similarly to those described for the TRF measurements. Light
sources were based on a homebuilt multipass Ti:sapphire amplifier and
an OPA system operating at 10 kHz. The OPA outputs of wavelengths
500 and 550 nm were used as pump pulses. For the single-color probe TA
experiment, a white-light continuum was generated by self-phase modu-
lation in a sapphire window 1 mm thick. Probe beams at 650 and
1000 nm were obtained from the white-light continuum by using interfer-
ence band-pass filters, and these were focused directly into the sample
without pulse compression. Time resolution from the cross-correlation
measurement was about 200 fs. In the case of TA spectra measurement,
white light continuum pulses in the visible and NIR regions were gener-
ated by self-phase modulation in a sapphire window 1 mm thick and a
photonic crystal fiber, respectively. The continuum pulses were divided
into two portions, the probe and reference beams, and sent to a mono-
chromator (SP300i, Acton) through a bifurcated optical fiber, to be de-
tected simultaneously by a CCD (charge-coupled device) detector (DV-
420-0E, Andor).
[4] a) D. M. Guldi, C. Luo, M. Prato, A. Troisi, F. Zerbetto, M. Sche-
loske, E. Dietel, W. Bauer, A. Hirsh, J. Am. Chem. Soc. 2001, 123,
9166–9167; b) T. D. M. Bell, T. A. Smith, K. P. Ghiggino, M. G. Ra-
nasinghe, M. J. Shephard, M. N. Paddon-Row, Chem. Phys. Lett.
1997, 268, 223–228; c) H. Imahori, K. Hagiwara, T. Akiyama, M.
Aoki, S. Taniguchi, T. Okada, M. Shirakawa, Y. Sakata, Chem. Phys.
Lett. 1996, 263, 545–550; d) H. Imahori, H. Yamada, D. M. Guldi,
Y. Endo, A. Shimomura, S. Kundu, K. Yamada, T. Okada, Y.
Sakata, S. Fukuzumi, Angew. Chem. 2002, 114, 2450–2453; Angew.
Chem. Int. Ed. 2002, 41, 2344–2347; e) A. Lembo, P. Tagliatesta,
D. M. Guldi, J. Phys. Chem. A 2006, 110, 11424–11434; f) L. San-
chez, M. Sierra, N. Martꢄn, A. J. Myles, T. J. Dale, J. Rebek, Jr., W.
Seitz, D. M. Guldi, Angew. Chem. 2006, 118, 4753–4757; Angew.
Chem. Int. Ed. 2006, 45, 4637–4641; g) F. DꢁSouza, G. R. Devipra-
sad, M. E. Zandler, V. T. Hoang, K. Arkady, M. VanStipdonk, A.
Perera, M. E. El-Khouly, M. Fujitsuka, O. Ito, J. Phys. Chem. A
2002, 106, 3243–3252; h) D. I. Schuster, K. Li, D. M. Guldi, A.
Palkar, C. S. Echegoyen, R. J. Cross, M. Niemi, N. V. Tkachenko, H.
Lemmetyinen, J. Am. Chem. Soc. 2007, 129, 15973–15982.
[5] a) D. M. Guldi, M. Maggini, G. Scorrano, M. Prato, J. Am. Chem.
Soc. 1997, 119, 974–980; b) M. Fujitsuka, N. Tsuboya, R. Hamasaki,
M. Ito, S. Onodera, O. Ito, Y. Yamamoto, J. Phys. Chem. A 2003,
107, 1452–1458; c) G. A. Rajkumar, A. S. D. Sandanayaka, K. Ike-
shita, Y. Araki, Y. Furusho, T. Takata, O. Ito, J. Phys. Chem. B 2006,
110, 6516–6525; d) F. Caporossi, B. Floris, P. Galloni, E. Gatto, M.
Venanzi, Eur. J. Org. Chem. 2006, 4362–4366.
Computational details: The calculations were based on DFT at the gener-
alized gradient approximation (GGA) level (Beckeꢁs 1988 functional for
exchange and Perdew-Wangꢁs 1991 functional for correlation,
BPW91[33]). The energy-consistent relativistic effective core potential
(RECP) was used for Zn and Os atoms.[34] Double numerical plus polari-
zation (DNP) basis sets were used for the C, H, B, N, O, and F atoms,
and the valence electrons for Zn and Os were also expanded by using the
DNP basis set. All the structures were optimized without any symmetry
restriction by using the analytical gradients of the energies. The DMol3
program of the software package Material Studio from Accelrys[35] was
used for the geometry optimization and the MO analysis.
[6] K. Matsumoto, M. Fujitsuka, T. Sato, S. Onodera, O. Ito, J. Phys.
Chem. B 2000, 104, 11632–11638.
[7] F. DꢁSouza, R. Chitta, A. S. D. Sandanayaka, N. K. Subbaiyan, L.
DꢁSouza, Y. Araki, O. Ito, J. Am. Chem. Soc. 2007, 129, 15865–
15871.
[8] a) C. Luo, D. M. Guldi, H. Imahori, K. Tamaki, Y. Sakata, J. Am.
Chem. Soc. 2000, 122, 6535–6536; b) H. Imahori, K. Tamaki, D. M.
Guldi, C. Luo, M. Fujitsuka, O. Ito, Y. Sakata, S. Fukuzumi, J. Am.
Chem. Soc. 2001, 123, 2607–2617; c) H. Imahori, D. M. Guldi, K.
Tamaki, Y. Yoshida, C. Luo, Y. Sakata, S. Fukuzumi, J. Am. Chem.
Soc. 2001, 123, 6617–6628; d) H. Imahori, Y. Sekiguchi, Y. Kashiwa-
gi, T. Sato, Y. Araki, O. Ito, H. Yamada, S. Fukuzumi, Chem. Eur. J.
2004, 10, 3184–3196; e) D. Kuciauskas, P. A. Liddell, S. Lin, T. E.
Johnson, S. J. Weghorn, J. S. Lindsey, A. L. Moore, T. A. Moore, D.
Gust, J. Am. Chem. Soc. 1999, 121, 8604–8614; f) F. DꢁSouza, O. Ito,
Coord. Chem. Rev. 2005, 249, 1410–1422.
Acknowledgements
This work was supported by the Korea Research Foundation (Grant no.
KRF-2005-201-C00021), the Korean Government (MOEHRD), the
Korea Science and Engineering Foundation (KOSEF) Nano R&D pro-
gram (Grant no. 2005-02618), the Korean Ministry of Science & Technol-
ogy (MOST), and by the National Research Foundation of Korea Grant
(20100001484), the Korean Government. T.J. acknowledges the support
from KOSEF grant funded by the Korea Government (MEST; Grant no.
R11-2008-012-01001-0).
[9] a) D. M. Guldi, Chem. Soc. Rev. 2002, 31, 22–36; b) D. M. Guldi,
Phys. Chem. Chem. Phys. 2007, 9, 1400–1420.
[10] F. DꢁSouza, P. M. Smith, M. E. Zandler, A. L. McCarty, M. Itou, Y.
Araki, O. Ito, J. Am. Chem. Soc. 2004, 126, 7898–7907.
[1] a) M. R. Wasielewski, Chem. Rev. 1992, 92, 435–461; b) D. Gust,
T. A. Moore, A. L. Moore, Acc. Chem. Res. 2001, 34, 40–48.
[2] a) T. Akiyama, H. Imahori, A. Ajavakom, Y. Sakata, Chem. Lett.
1996, 907–908; b) H. Yamada, H. Imahori, Y. Nishimura, I. Yamaza-
ki, T. K. Ahn, S. K. Kim, D. Kim, S. Fukuzumi, J. Am. Chem. Soc.
2003, 125, 9129–9139; c) H. Imahori, H. Yamada, Y. Nishimura, I.
Yamazaki, Y. Sakata, J. Phys. Chem. B 2000, 104, 2099–2108; d) H.
Imahori, M. Kimurs, K. Hosomizu, T. Sato, T. K. Ahn, S. K. Kim, D.
Kim, Y. Nishimura, Y. Iwao, Y. Araki, O. Ito, S. Fukuzumi, Chem.
Eur. J. 2004, 10, 5111–5122; e) H. Imahori, H. Norieda, H. Yamada,
Y. Nishimura, I. Yamazaki, Y. Sakata, S. Fukuzumi, J. Am. Chem.
Soc. 2001, 123, 100–110; f) A. Ikeda, T. Hatano, S. Shinkai, T.
Akiyama, S. Yamada, J. Am. Chem. Soc. 2001, 123, 4855–4856;
g) V. Chukharev, T. Vuorinen, A. Efimov, N. V. Tkachenko, M.
Kimura, S. Fukuzumi, H. Imahori, H. Lemmetyinen, Langmuir
[11] a) A. Ulman, An Introduction to Ultrathin Organic Films: From
Langmuir–Blodgett to Self-Assembly, Academic Press, San Diego,
1991; b) A. Ulman, Chem. Rev. 1996, 96, 1533–1554; c) C. A.
Mirkin, Inorg. Chem. 2000, 39, 2258–2272; d) A. N. Shipway, E.
Katz, I. Willner, ChemPhysChem 2000, 1, 18–52; e) A. P. Alivisatos,
P. F. Barbara, A. W. Castleman, J. Chang, D. A. Dixon, M. L. Klein,
G. L. McLendon, J. S. Miller, M. A. Ratner, P. J. Rossky, S. I. Stupp,
M. E. Thompson, Adv. Mater. 1998, 10, 1297–1336; f) S. Flink,
F. C. J. M. Veggel, D. N. Reinhoudt, Adv. Mater. 2000, 12, 1315–
1328; g) T. P. Sullivan, W. T. S. Huck, Eur. J. Org. Chem. 2003, 17–
29.
5598
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 5586 – 5599

Photocurrent Generation from a Triad SAM
FULL PAPER
[12] Y.-J. Cho, T. K. Ahn, H. Song, K. S. Kim, C. Y. Lee, W. S. Seo, K.
Lee, S. K. Kim, D. Kim, J. T. Park, J. Am. Chem. Soc. 2005, 127,
2380–2381.
[21] J. I. Steinfeld, J. S. Francisco, W. L. Hase, Chemical Kinetics and Dy-
namics, 2nd ed., Prentice Hall, Portage, 1999.
[22] T. Elsaesser, W. Kaiser, Annu. Rev. Phys. Chem. 1991, 42, 83–107.
[23] a) C. A. Hunter, M. N. Meah, J. K. M. Sanders, J. Am. Chem. Soc.
1990, 112, 5773–5780; b) D. M. Guldi, C. Luo, T. D. Ros, M. Prato,
E. Dietel, A. Hirsh, Chem. Commun. 2000, 375–376.
[24] H. Imahori, T. Umeyama, S. Ito, Acc. Chem. Res. 2009, 42, 1809–
1818.
[25] T. Sasaki, J-F. Morin, M. Lu, J. M. Tour, Tetrahedron Lett. 2007, 48,
5817–5820.
[26] T. A. Golovkova, D. V. Kozlov, D. C. Neckers, J. Org. Chem. 2005,
70, 5545.
[27] C-H. Lee, J. S. Lindsey, Tetrahedron 1994, 50, 11427–11440.
[28] J. M. Tour, A. M. Rawlett, M. Kozaki, Y. X. Yao, R. C. Jagessar,
S. M. Dirk, D. W. Price, M. A. Reed, C. W. Zhou, J. Chen, W. Y.
Wang, I. Campbell, Chem. Eur. J. 2001, 7, 5118–5134.
[29] Y.-J. Cho, H. Song, K. Lee, K. Kim, J. Kwak, S. Kim, J. T. Park,
Chem. Commun. 2002, 2966–2967.
[13] a) R. P. Haugland in The Handbook A Guide to Fluorescent Probes
and Labeling Technologies, 10th ed. (Ed.: M. T. Z. Spence), Molecu-
lar Probes, Eugene, 2005; b) H. C. Kang, R. P. Haugland, U.S.
Patent 5451663, 1995; c) A. Burghart, H. Kim, M. B. Welch, L. H.
Thoresen, J. Reibenspies, K. Burgess, J. Org. Chem. 1999, 64, 7813–
7819; d) J. Chen, A. Burghart, A. Derecskei-Kovacs, K. Burgess, J.
Org. Chem. 2000, 65, 2900–2906; e) B. Turfan, E. U. Akkaya, Org.
Lett. 2002, 4, 2857–2859; f) A. Coskun, E. U. Akkaya, J. Am. Chem.
Soc. 2005, 127, 10464–10465; g) W. Qin, T. Rohand, M. Baruah, A.
Stefan, M. Van der Auweraer, W. Dehaen, N. Boens, Chem. Phys.
Lett. 2006, 420, 562–568.
[14] H. Martin, H. Rudolf, F. Vlastimil, Fluorescence Spectroscopy in
Biology: Advanced Methods and their Applications to Membranes,
Proteins, DNA, and Cells, Springer, Heidelberg, 2005.
[15] a) R. W. Wagner, J. S. Lindsey, J. Am. Chem. Soc. 1994, 116, 9759–
9760; b) R. W. Wagner, J. S. Lindsey, J. Seth, V. Palaniappan, D. F.
Bocian, J. Am. Chem. Soc. 1996, 118, 3996–3997.
[30] A. J. Bard, L. R. Faulkner in Electrochemical Methods, Wiley, New
York, 1980, pp. 142–146.
[16] C. Y. Lee, B. K. Park, J. H. Yoon, C. S. Hong, J. T. Park, Organome-
tallics 2006, 25, 4634.
[31] C.-K. Min, T. Joo, Opt. Lett. 2005, 30, 1855–1857.
[32] H. Rhee, T. Joo, Opt. Lett. 2005, 30, 96–98.
[17] H. Song, K. Lee, J. T. Park, M. G. Choi, Organometallics 1998, 17,
4477–4483.
[33] a) A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100; b) J. P. Perdew,
Y. Wang, Phys. Rev. B 1992, 45, 13244–13249.
[18] a) A. Z. Weller, Phys. Chem. 1982, 132, 93; b) D. Rehm, A. Weller,
Isr. J. Chem. 1970, 10, 529.
[34] D. Andrae, U. Hꢅubermann, M. Dolg, H. Stoll, H. Preub, Theor.
Chim. Acta 1990, 77, 123–141.
[19] J. Rodriguez, C. Kirmaier, D. Holten, J. Am. Chem. Soc. 1989, 111,
6500–6506.
[35] a) B. Delley, J. Chem. Phys. 1990, 92, 508–517; b) B. Delley, J.
Chem. Phys. 2000, 113, 7756–7764.
[20] D. Kim, M. Lee, Y. D. Suh, S. K. Kim, J. Am. Chem. Soc. 1992, 114,
4429–4430.
Received: January 26, 2010
Published online: April 16, 2010
Chem. Eur. J. 2010, 16, 5586 – 5599
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5599