Ultrafast photoinduced electron transfer in directly linked porphyrin - Ferrocene dyads

Article abstract of DOI:10.1021/jp071546b

The ultrafast electron transfer occurring upon Soret excitation of three new porphyrin-ferrocene (XP-Fc) dyads has been studied by femtosecond up-conversion and pump-probe techniques. In the XP-Fc dyads (XP-Fcs) designed in this study, the ferrocene moiet

Full text of DOI:10.1021/jp071546b

5136
J. Phys. Chem. A 2007, 111, 5136-5143
Ultrafast Photoinduced Electron Transfer in Directly Linked Porphyrin-Ferrocene Dyads
Minoru Kubo,*,†,‡ Yukie Mori,§ Masana Otani,† Masataka Murakami,† Yukihide Ishibashi,†
Masakazu Yasuda,† Kohei Hosomizu,| Hiroshi Miyasaka,† Hiroshi Imahori,*,§,| and
Satoru Nakashima*,†
Department of Materials Engineering Science, Graduate School of Engineering Science, Osaka UniVersity,
Toyonaka 560-8531, Japan, Fukui Institute for Fundamental Chemistry, Kyoto UniVersity, Kyoto 606-8103,
Japan, and Department of Molecular Engineering, Graduate School of Engineering, Kyoto UniVersity,
Kyoto 615-8510, Japan
ReceiVed: February 24, 2007; In Final Form: April 17, 2007
The ultrafast electron transfer occurring upon Soret excitation of three new porphyrin-ferrocene (XP-Fc)
dyads has been studied by femtosecond up-conversion and pump-probe techniques. In the XP-Fc dyads
(XP-Fcs) designed in this study, the ferrocene moiety is covalently bonded to the meso positions of 3,5-
di-tert-butylphenyl zinc porphyrin (BPZnP-Fc), pentafluorophenyl zinc porphyrin (FPZnP-Fc), and 3,5-
di-tert-butylphenyl free-base porphyrin (BPH2P-Fc). Charge separation and recombination in the XP-Fcs
were confirmed by transient absorption spectra, and the lifetimes of the charge-separated states were estimated
from the decay rate of the porphyrin radical anion band to be approximately 20 ps. The charge-separation
rates of the XP-Fcs were found to be >1013 s-1 from the S2 state and 6.3 × 1012 s-1 from the S1 state.
Charge separation from the S2 state was particularly efficient for BPZnP-Fc, whereas the main reaction
pathway was from the S1 state for BPH2P-Fc. Charge separation from the S2 and S1 states occurred at virtually
the same rate in benzene and tetrahydrofuran and was much faster than their solvation times. Analysis of
these results using semiquantum Marcus theory indicates that the magnitude of the electronic-tunneling matrix
element is rather large and far outside the range of nonadiabatic approximation. The pump-probe data show
the presence of vibrational coherence during the reactions, suggesting that wavepacket dynamics on the adiabatic
potential energy surface might regulate the ultrafast reactions.
Introduction
tetrahydrofuran (THF) and benzene. ET from the S2 state is
not observed in this system because of the rapid IC.
The kinetics and dynamics of electron-transfer (ET) reactions
are fundamental issues in fields ranging from physical chemistry
to biology, and the theory of ET reactions in condensed media
has been developed on the basis of Marcus theory.1-9 According
to this theory, the ET rate is determined by the donor-acceptor
electronic coupling, the free-energy gap, and the reorganization
energies of the medium and the intramolecular vibrations. Apart
from these energy parameters, the ET rate is also regulated by
dynamic parameters such as the relaxation times of solvent
polarization and intramolecular vibrations. To date, Marcus
theory has been applied to a wide range of ET reactions.10-12
On the other hand, with recent progress in femtosecond
spectroscopy, research attention has shifted toward ET reactions
that are faster than diffusive solvation dynamics and/or vibra-
tional relaxation.13-19 We previously reported an example of
such an ET reaction in a porphyrin-ferrocene dyad with a
pentafluorophenyl free-base porphyrin directly linked to fer-
rocene (abbreviated as FPH2P-Fc; FP ) pentafluorophenyl, H2P
) free-base porphyrin, Fc ) ferrocene).20 After Soret excitation
of this system, rapid ET occurs from the S1 state with a time
constant of 110 fs after S2-S1 internal conversion (IC) in both
In the present study, the photodynamics in other types of
porphyrin-ferrocene dyads was studied with femtosecond up-
conversion (UC) and pump-probe (PP) techniques. We de-
signed three porphyrin-ferrocene (XP-FC) dyads: BPZnP-
Fc (BP ) 3,5-di-tert-butylphenyl, ZnP ) zinc porphyrin),
FPZnP-Fc, and BPH2P-Fc (Figure 1), which are denoted
hereafter as XP-Fcs. The differences between these XP-Fcs
are the presence of either electron-donating (BP) or electron-
withdrawing (FP) substituents and metal (Zn) or metal-free (H2)
centers in the porphyrin component. In this article, we discuss
the ultrafast ET from both the S2 state and the S1 state that
occurs in these XP-Fc systems. ET from the S2 state was found
to occur in under 100 fs and to be most efficient in BPZnP-
Fc. The femtosecond UC and PP data demonstrate that the rates
of ET from the S2 and S1 states are higher than those of the
solvent relaxation processes and are independent of the solvent
polarity and the free-energy gap. The PP data also indicate the
presence of vibrational coherence during the reactions. On the
basis of these results and the analysis with semiquantum Marcus
theory, possible reaction mechanisms are discussed for the
ultrafast ET occurring in the XP-Fc systems.
* Corresponding authors. E-mail:
M.Kubo@neu.edu (M.K.),
Experimental Section
satoru@chem.es.osaka-u.ac.jp (S.N.), imahori@scl.kyoto-u.ac.jp (H.I.).
† Osaka University.
Synthesis. Pyrrole, trifluoroacetic acid, Et3N, and ferrocen-
ecarboxyaldehyde were purchased from Kanto Chemicals
(Tokyo, Japan). Zn(OAc)2‚2H2O and p-chloranil were purchased
from Nacalai Tesque (Kyoto, Japan). All solvents were obtained
‡ Present address: Department of Physics, Northeastern University,
Boston, Massachusetts 02115.
§ Fukui Institute for Fundamental Chemistry, Kyoto University.
| Graduate School of Engineering, Kyoto University.
10.1021/jp071546b CCC: $37.00 © 2007 American Chemical Society
Published on Web 05/27/2007

Electron Transfer in Porphyrin-Ferrocene Dyads
J. Phys. Chem. A, Vol. 111, No. 24, 2007 5137
Figure 2. Steady-state absorption and fluorescence (excited at 390
nm) spectra in benzene. The spectra of the XP-Fcs and XPs are shown
in solid and dashed lines, respectively. Enlarged traces multiplied by
the indicated factors are also shown.
Figure 1. Structural formulas of the three XP-Fcs and their reference
XPs.
from Wako Pure Chemical Industries Ltd. (Osaka, Japan) and
were used without further purification. H NMR spectra were
recorded on an EX-400 spectrometer (JEOL, Tokyo, Japan).
Matrix-assisted laser desorption/ionization (MALDI) mass
spectra (MS) were measured on a KOMPACT MALDI II
instrument (Shimadzu Corporation, Kyoto, Japan).
stirred for 3 h under Ar at 60 °C. After being cooled, the mixture
was washed with distilled water and brine. The organic phase
was then dried over MgSO4, and the solvent was removed in
vacuo. The residue was purified by flash column chromatog-
raphy over SiO2 eluted with hexane/CH2Cl2 (2/1-1/1) to yield
BPZnP-Fc (32 mg, 28 µmol, 56%) as a purple solid. 1H NMR
(CDCl3, 400 MHz) δ 10.23 (2H, d, J ) 4.8 Hz), 8.94 (2H, d,
J ) 4.8 Hz), 8.90-8.93 (4H, ABq), 8.08 (4H, d, J ) 1.6 Hz),
8.05 (2H, d, J ) 1.6 Hz), 7.79 (2H, t, J ) 1.6 Hz), 7.76 (1H,
t, J ) 1.6 Hz), 5.58 (2H, s), 4.83 (2H, s), 4.29 (5H, s), 1.55
(37H, s), 1.54 (18H, s). MS (MALDI-TOF) m/z ) 1122 (M +
H+).
FPH2P-Fc. Pyrrole (403 mg, 6 mmol), ferrocenecarboxy-
aldehyde (642 mg, 3 mmol), and pentafluorobenzaldehyde (588
mg, 3 mmol) were dissolved in 300 mL of CHCl3. Trifluoro-
acetic acid (0.47 mL) was then injected through a syringe under
Ar, and the resulting mixture was stirred for 2 h at room
temperature. To the reaction mixture was added p-chloranil (1.23
g, 5 mmol), and the mixture was stirred at room temperature
overnight. The reaction was quenched with Et3N (1.0 g, 9.9
mmol), and the solvent was removed in vacuo. The residue was
purified by flash column chromatography over SiO2 eluted with
CH2Cl2/hexane (1/2-1/1) to yield FPH2P-Fc (65 mg, 0.065
mmol, 4.3%) as a purple solid. 1H NMR (CDCl3, 400 MHz) δ
10.12 (2H, d, J ) 4.5 Hz), 8.77 (4H, s), 8.74 (2H, d, J ) 4.5
Hz), 5.59 (2H, s), 4.92 (2H, s), 4.23 (5H, s), -2.28 (2H). MS
(MALDI-TOF) m/z ) 993 (M + H+).
1
XP-Fcs were synthesized by procedures reported for similar
A3B-type meso-substituted porphyrins, i.e., by acid-catalyzed
condensation of 2-formylferrocene, 3,5-di-tert-butylbenzalde-
hyde (or pentafluorobenzaldehyde), and pyrrole, followed by
oxidation with chloranil.21 Flash column chromatography of the
reaction mixture on neutral silica gel (Kanto Chemicals, silica
gel 60N, spherical) with hexane/dichloromethane as the eluent
gave the desired dyads as the second eluted bands. Zinc was
inserted by addition of an excess of zinc acetate in a mixture of
CHCl3 and methanol. Reference XPs were prepared according
to the method in the literature.21
BPH2P-Fc. Pyrrole (268 mg, 4 mmol), ferrocenecarboxy-
aldehyde (429 mg, 2 mmol), and 3,5-di-tert-butylbenzaldehyde
(437 mg, 2 mmol) were dissolved in 200 mL of CHCl3.
Trifluoroacetic acid (0.3 mL) was then injected through a syringe
under Ar, and the resulting mixture was stirred for 2 h at room
temperature. To the reaction mixture was added p-chloranil (740
mg, 3 mmol), and the mixture was stirred at room temperature
overnight. The reaction was then quenched with Et3N (0.75 g,
7.4 mmol), and the solvent was removed in vacuo. The residue
was purified by flash column chromatography over SiO2 eluted
with CH2Cl2 to yield BPH2P-Fc (45 mg, 0.042 mmol, 4.2%)
as a purple-red solid. 1H NMR (CDCl3, 400 MHz) δ 10.00 (2H,
s), 8.78-8.81 (6H, m), 8.07 (4H, s), 7.79 (2H, s), 7.76 (1H, s),
5.59 (2H, s), 4.84 (2H, s), 4.22 (5H, s), 1.54 (36H, s), 1.51
(18H, s), -2.26 (2H, s). MS (MALDI-TOF) m/z ) 1060 (M +
H+).
FPZnP-Fc. FPH2P-Fc (35 mg, 35 µmol) was dissolved in
CHCl3 (5 mL), and saturated Zn(OAc)2‚2H2O in MeOH (0.5
mL) was added to the mixture through a syringe. The resulting
mixture was refluxed for 3 h under Ar. After being cooled, the
mixture was washed with distilled water and brine. The organic
phase was dried over MgSO4, and the solvent was removed in
vacuo. The residue was purified by flash column chromatog-
raphy over SiO2 eluted with hexane/CH2Cl2 (2/1-1/1) to yield
BPZnP-Fc. BPH2P-Fc (53 mg, 50 µmol) was dissolved in
CHCl3 (15 mL), and saturated Zn(OAc)2‚2H2O in MeOH (0.6
mL) was added through a syringe. The resulting mixture was
1
FPZnP-Fc (25 mg, 0.028 mmol, 70%) as a purple solid. H

5138 J. Phys. Chem. A, Vol. 111, No. 24, 2007
Kubo et al.
Figure 3. Transient absorption spectra of FPZnP and FPZnP-Fc in THF after excitation with a 532-nm laser pulse with a fwhm of 15 ps.
NMR (CDCl3, 400 MHz) δ 10.30 (2H, d, J ) 4.8 Hz), 8.87
(4H, s), 8.83 (2H, d, J ) 4.8 Hz), 5.56 (2H, s), 4.88 (2H, s),
4.21 (5H, s). MS (MALDI-TOF) m/z ) 1055 (M + H+).
Cyclic Voltammetry. Cyclic voltammograms were obtained
using a 0.1 M n-Bu4NPF6 solution in N2-saturated THF with a
scan rate of 0.05 V s-1 on a CH Instruments model 660 A
electrochemical workstation at room temperature. Glassy carbon,
Pt wire, and Ag/Ag+ (0.01 M AgNO3, 0.1 M n-Bu4NClO4 in
MeCN) electrodes were used as the working, counter, and
reference electrodes, respectively. Ferrocene was used as the
internal standard.
Spectroscopy. The details of the UC measurements were
described previously.22 The cross-correlation trace between the
fundamental and its second harmonic pulses gave e50 fs fwhm
(full width at half-maximum) at the sample position. The time-
resolved S2 fluorescence excited at 390 nm was monitored at
440 nm, whereas the time-resolved S1 fluorescence excited at
410 nm was monitored at 650 nm for BPZnP, 627 nm for
BPZnP-Fc, and 642 nm for the other compounds.
In the single-color PP measurements, the output of the second
harmonics of a homemade cavity-dumped Ti:sapphire laser at
200-kHz repetition rate, centered at 424 nm with a 15-nm
bandwidth, was divided into pump and probe beams.20 The
fwhm of the autocorrelation trace at the sample position was
∼35 fs. In the two-color PP measurements, on the other hand,
a dual OPA femtosecond laser system23 was employed, in which
the wavelengths of the pump and probe beams were set to 400
and 705 nm, respectively. The fwhm of the cross-correlation
trace between the pump and probe pulses was 170 fs at the
sample position. Transient absorption spectra were obtained
using a 532-nm excitation pulse with 15-ps fwhm and a white-
light continuum probe pulse with the PP system described
previously.24
peaks of the XP-Fcs relative to those of the XPs were roughly
60%, 70%, and 90% for BPZnP-Fc, FPZnP-Fc, and BPH2P-
Fc, respectively. The S1 fluorescence of the XP-Fcs was
quenched more dramatically. The S1 fluorescence intensities of
the XP-Fcs were only <5% of those of the XPs. In particular,
the S1 fluorescence of BPZnP-Fc was found to be completely
quenched (<0.5%). These results for S2 and S1 fluorescence
quenching suggest the involvement of ET reaction pathways
from the S2 and S1 states.
Subpicosecond-Picosecond Transient Absorption Changes.
To directly probe the ET reactions, transient absorption spec-
troscopy was applied to FPZnP and FPZnP-Fc in THF. Figure
3a shows the transient absorption spectra of FPZnP excited with
a 532-nm laser pulse with a fwhm of 15 ps. The spectra are
characterized by a strong absorption band at 457 nm and dip
signals corresponding to the bleaching of the Q bands at 550
and 586 nm, which are ascribed to the S1-Sn transition. The
bleaching at 648 nm might be due to the induced emission from
the S1 state. In contrast to the spectra of FPZnP, the spectra of
FPZnP-Fc contain characteristic bands of the porphyrin radical
anion, including a strong, sharp band around 450 nm (ꢀ ≈ 105
M-1 cm-1) and weak, broad bands in the range 700-900 nm
(ꢀ ≈ 104 M-1 cm-1) (Figure 3b).25 The ferrocenium cation band
near 800 nm could not be distinguished because of its small
molar extinction coefficient (ꢀ ≈ 103 M-1 cm-1). The porphyrin
radical anion bands, which emerge within the apparatus response
function, decay as a result of charge recombination (CR) on a
time scale of several tens of picoseconds.
To elucidate the dynamic behavior of the ET process for
shorter times, the time profile of the transient absorbance at
705 nm was measured after Soret excitation at 400 nm, using
laser pulses with a fwhm of 170 fs (Figure 4). An ultrafast rise
of the transient absorbance within the apparatus response
function is followed by a rapid decay with a time constant of
approximately 0.2 ps (0.18-0.25 ps). The ultrafast rise might
involve S2-S1 IC and charge separation (CS) from the S2 state.
The subsequent rapid decay is due to CS from the S1 state,
because this time constant is comparable to the S1 fluorescence
lifetime (vide infra). Note that the decay absorbance change
due to CS from the S1 state implies that the extinction coefficient
of the S1 state at 705 nm is larger than that of the charge-
separated state. Following this subpicosecond decay, a slight
rise occurs with a time constant of approximately 6 ps (4.0-
6.7 ps), and then there is a decay with a time constant of
approximately 20 ps (22-25 ps). The slow decay stems from
Results
Steady-State Absorption and Fluorescence Spectra. The
absorption spectra of the XPs and XP-Fcs are shown in Figure
2. The Soret bands of the XP-Fcs were found to be similar to
those of the reference XPs, although the peak for each XP-Fc
is slightly red-shifted with respect to and broader than that for
each XP. In contrast, the Q bands of the XP-Fcs are
significantly different in shape and position from those of the
XPs. The fluorescence spectra of the XPs and XP-Fcs are also
shown in Figure 2. Very weak S2 fluorescence was observed
for the XPs and XP-Fcs. The intensities of the S2 fluorescence

Electron Transfer in Porphyrin-Ferrocene Dyads
J. Phys. Chem. A, Vol. 111, No. 24, 2007 5139
Figure 4. Absorbance changes at 705 nm in THF after excitation at
400 nm. The changes in the early-time region are shown in the inset.
Observed signals (circles) were fitted with exponential functions (solid
lines). The apparatus response function was taken into account.
CR as shown in Figure 3, whereas the picosecond rise might
be due to vibrational cooling and/or structural relaxation in the
charge-separated state.
Time Profiles of the S2 and S1 Fluorescence. The S2 and
S1 fluorescence decays of the XPs and XP-Fcs were measured
in benzene using the UC method with an apparatus response
function of e50 fs. The S2 fluorescence decays of the XP-Fcs
were found to be very rapid, with time constants of e80 fs,
and are appreciably faster than those of the corresponding
reference XPs, particularly for BPZnP-Fc. This result strongly
suggests that rapid CS from the S2 state occurs in competition
with the S2-S1 IC. The S1 fluorescence decays of the XP-Fcs
were also found to be rapid, with a time constant of ap-
proximately 160 fs, which is consistent with the extremely weak
intensity of the S1 fluorescence in the steady-state measurements
(Figure 2). The rapid decay of the S1 fluorescence results from
CS from the S1 state, as is evident in Figures 3 and 4. Time-
resolved fluorescence emission could not be detected in THF
because of the strong Raman scattering from THF.26
PP Measurements in the Femtosecond Time Region. To
shed light on the early dynamics after Soret excitation, transient
absorption changes were measured at 424 nm using the PP
technique with a higher time resolution of 35 fs. The transient
absorption spectra (Figure 3) feature absorptions at 424 nm in
the S1 state of the XPs and in the charge-separated states of the
XP-Fcs. Thus, the S1 state absorption and the charge-separated
state absorption as well as the ground-state bleaching contribute
to the present PP signals. Figure 6 below shows the PP signals
of the XPs and XP-Fcs in THF and benzene. The time constants
of the dynamics are very similar in THF and benzene, and are
in agreement with the data for fluorescence in benzene displayed
in Figure 5.
Figure 5. UC signals in benzene. Observed signals (circles) were fitted
with exponential functions (solid lines). The apparatus response function
was taken into account.
The signal of BPZnP consists of two components with
subpico- and picosecond time constants (see the top panel of
Figure 6). The slow component (1.2-1.5 ps) is assigned to S2-
S1 IC, as this time constant is in agreement with the S2
fluorescence lifetime. The fast component (0.38-0.39 ps) cannot
be assigned unambiguously as yet, but might involve a relaxation
process in the S2 state. The signal of FPZnP also results from
subpicosecond and picosecond dynamics. The fast component
(0.39-0.48 ps) matches the S2 fluorescence lifetime and is
assigned to S2-S1 IC, whereas the slow component (2.7-3.7
ps) is due to a relaxation process in the S1 state. Similarly, the
signal of BPH2P consists of ultrafast S2-S1 IC (80-87 fs) and
a picosecond relaxation in the S1 state (2.6-2.8 ps).
On the other hand, the dynamics of all of the XP-Fcs feature
an ultrafast rise with a time constant of ∼60-70 fs, followed
by a decay with a time constant of ∼160 fs and a slow decay
on the picosecond time scale. By comparing these time constants
with those in Figures 3-5, we attribute the first two components
to the population dynamics of the S2 and S1 states, respectively.

5140 J. Phys. Chem. A, Vol. 111, No. 24, 2007
Kubo et al.
Figure 6. PP signals at 424 nm in THF and benzene. Observed signals (circles) were fitted with exponential functions (solid lines). The apparatus
response function was taken into account. The coherent spikes were fitted with a Gaussian and an exponential rise. The insets show the oscillatory
components obtained by subtracting the fitted population components from the PP signals.

Electron Transfer in Porphyrin-Ferrocene Dyads
J. Phys. Chem. A, Vol. 111, No. 24, 2007 5141
Figure 7. Fourier-transform frequency spectra of BPH2P and BPH2P-
Fc in benzene. The effect of the pulse width was corrected.
TABLE 1: ET Rates and Free-Energy Gaps of the XP-Fcs
in THF
BZnP-Fc FPZnP-Fc BPH2P-Fc FPH2P-Fca
τ
τ
τ
CS(S2) (fs)
CS(S1) (fs)
CR (ps)
63
160
24
89
160
22
-
-
Figure 8. Semilogarithmic energy gaps for kCS(S1) (O), kCS(S2) (0),
and kCR (4) in THF. The curves were calculated with semiquantum
Marcus theory for V ) 2.3, 40, 100, 200, and 300 meV. T ) 298 K,
λv ) 0.5 eV, p ω ) 0.15 eV, λs ) 0.79 eV (RA ) 10 Å, RD ) 2 Å,
and RDA ) 6.5 Å).
160
25
0.79
110
no data
0.79
1.52
λS (eV)
0.79
0.79
Eox - Ered (eV)
∆GS (eV)
-∆GCR (eV)
∆E(S2) (eV)
∆E(S1) (eV)
-∆GCS(S2) (eV)
-∆GCS(S1) (eV)
1.97
-0.29
1.68
1.52
-0.29
1.23
1.75
-0.29
1.46
-0.29
1.23
2.86
2.87
2.88
2.89
1.95
1.93
1.80
1.79
1.18
1.64
1.42
1.66
0.27
0.70
0.34
0.56
a Data for FPH2P-Fc were obtained from ref 20.
The picosecond component is probably due to structural
relaxation in the charge-separated state.
In addition to the ultrafast dynamic behavior, vibrational
coherence results in oscillatory modulation of the PP signals of
the XPs and the XP-Fcs (see the insets in Figure 6). Fourier-
transform frequency spectra of the oscillation data for BPH2P
and BPH2P-Fc in benzene are shown in Figure 7. The most
intense peak in both spectra is at 320 cm-1, which is assigned
to the porphyrin-ring breathing mode (ν8).27 The second
strongest peak in the spectrum of BPH2P is at 170 cm-1, which
arises from the phenyl translation mode (φ10).27 For BPH2P-
Fc, on the other hand, the signal-to-noise level did not allow us
to determine lower-frequency peaks. Pronounced vibrational
coherence was observed for FPH2P-Fc at frequencies of 130,
200, 250, and 330 cm-1 with a better signal-to-noise ratio; the
detailed analysis of this coherence has been reported elsewhere.20
Rates and Free-Energy Gaps of the CS reactions. The rate
of CS from the S2 state [kCS(S2) ) τCS(S2)-1] can be roughly
estimated with the equation τCS(S2)-1 ) τXP-Fc(S2)-1
-
Figure 9. Reaction scheme after Soret excitation of the XP-Fcs. The
primary and secondary reaction pathways are shown as solid and dashed
arrows, respectively. The data for FPH2P-Fc were obtained from ref
20.
τ
XP(S2)-1, where τXP-Fc(Sn) and τXP(Sn) are the Sn-state lifetimes
of the XP-Fc and the XP, respectively. However, we note that
τCS(S2) of BPH2P-Fc cannot be estimated because the τXP-Fc(S2)
and τXP(S2) values in this system are very similar. The rate of
CS from the S1 state [kCS(S1) ) τCS(S1)-1] can be well
approximated as τXP-Fc(S1)-1 because τXP(S1)-1 is orders of
magnitude smaller. The τCS(S2) and τCS(S1) values of the XP-
Fcs in THF are summarized in Table 1. The CS rates in benzene
could not be distinguished from those in THF on the basis of
the UC and PP data.
oxidation and first reduction potentials, respectively. ∆GS is the
correction term that includes the effects of the solvent and the
Coulombic interaction between the charged donor and acceptor.
The third equation for ∆GS is known as the Born equation.29
RD and RA are the radii of the charged donor and acceptor,
respectively, and RDA is the center-to-center distance between
them. ꢀr is the dielectric constant of the solvent in which Eox
and Ered are determined (THF), and ꢀs is that of the solvent
used in the spectroscopic measurements (THF or benzene). The
Born equation is appropriate for polar solvents such as THF
and has yielded satisfactory results in many cases.1b,12,30
However, for less polar or almost nonpolar solvents (e.g.,
benzene), ∆GS is often overestimated with this equation under
the continuum approximation. The estimation of the solvent
reorganization energy λs also has a similar problem
The free-energy gap for CS (-∆GCS) can be estimated using
the following equations28
-∆GCS(Sn) ) ∆E(Sn) - (-∆GCR
-∆GCR ) Eox - Ered + ∆GS
)
(1)
(2)
∆GS ) e2[1/(2RD) + 1/(2RA)](1/ꢀs - 1/ꢀr) - e2/(ꢀsRDA) (3)
Here, ∆E(Sn) is the excitation energy of the Sn state; -∆GCR is
the free-energy gap for CR; and Eox and Ered are the first
λs ) e2[1/(2RD) + 1/(2RA) - 1/RDA](1/ꢀ∞ - 1/ꢀs) (4)

5142 J. Phys. Chem. A, Vol. 111, No. 24, 2007
Kubo et al.
The λs value for benzene calculated with eq 4 is negligibly
small because the optical dielectric constant ꢀ∞ is very similar
to the static dielectric constant ꢀs. Thus, we estimated -∆GCS
of the XP-Fcs in THF (Table 1) and plotted the CS rates against
-∆GCS in THF (Figure 8).
potential of porphyrin in the S2 state is also lower than that of
ferrocene in the S1 state because the excitation energies are
nearly the same. Thus, ferrocene cannot accept the excited
electron from porphyrin via Dexter-type electron exchange,
although electron transfer does occur from the HOMO of
ferrocene to the HOMO of excited porphyrin, as observed in
the present study.
The relationship between kCS and ∆GCS was analyzed with
the semiquantum Marcus expression
It is worth mentioning that, in some donor-bridge-acceptor
systems, the acceptor containing iron d electrons was found to
enhance intersystem crossing of the donor, depending on the
coordination and spin states of the iron.39 In the XP-Fcs,
however, no apparent role of spin-orbit coupling was found
in the ultrafast processes. Actually, there is no triplet-state signal
in the transient absorption spectra.
kCS ) (2π/p)V2(4πλskBT)-1/2 [e-S(Sn/n!)] exp[-(∆G
+
CS
∑
n
λs + np ω )2/4λskBT] (5)
where V is the donor-acceptor electronic-tunneling matrix
element, S ) λv/p ω is the electron-vibration coupling
constant, and λv is the reorganization energy associated with
the average angular frequency ω of the intramolecular high-
frequency vibrations. Marcus theory is well applicable to directly
linked donor-acceptor systems despite the short donor-
acceptor separation compared to their molecular sizes, as
confirmed for porphyrin-imide dyads.30b-d We found that fitting
the ultrafast CS rates to the Marcus parabola required V > 40
meV, for an adiabaticity parameter4 of κA > 1. This indicates
that the CS reactions in the XP-Fcs are adiabatic. On the other
hand, the plots for the CR rates were placed on the V ) 2.3
meV curve in the nonadiabatic region, which is comparable to
the results obtained for other porphyrin-linked systems.30a,31
The reaction scheme for the series of XP-Fcs including
FPH2P-Fc20 is summarized in Figure 9. ET from the S2 state
is the most efficient for BPZnP-Fc, whereas ET mainly occurs
in BPH2P-Fc from the S1 state after the rapid S2-S1 IC.
FPZnP-Fc is an intermediate case between BPZnP-Fc and
BPH2P-Fc, i.e., both pathways from the S2 and S1 states are
effective for ET.
The energy-gap dependences of the CR rates of the XP-Fcs
are not obvious because the CR rates occur near the top region
of the Marcus parabola (Figure 8), which indicates that the CR
processes are almost barrierless. It should be noted that the
inclusion of the effect of intramolecular high-frequency vibra-
tions (p ω ) 0.15 eV, λv ) 0.5 eV) is essential to reproduce
the observed CR rates. This means that the reaction is facilitated
by quantum-mechanical vibrational channels.
Discussion
The present time-resolved spectroscopy reveals the ultrafast
ET reactions following Soret excitation of the XP-Fcs. The
relatively long lifetime of the S2 state of ZnP enabled us to
investigate the ET reaction pathways from the S2 state as well
as from the S1 state. The time constant for ET from the S2 state
was as high as ∼60 fs, which is comparable to the highest time
constant ever reported for intermolecular ET in condensed
media.15,16,19b
On the other hand, the ultrahigh CS rates cannot be explained
solely in terms of the effect of the intramolecular high-frequency
vibrations. Ultrafast ET on the subpicosecond time scale in
porphyrin-imide dyads has previously been successfully in-
terpreted using semiquantum Marcus theory with p ω ) 0.15
eV, λv ) 0.3 eV, and the electronic-tunneling matrix element
V ≈ 20 meV.30b-d Because this V value is the upper limit of
the nonadiabatic approximation,28 the ET in these systems still
obeys semiquantum Marcus theory. On the other hand, the V
value for ET from both the S2 and S1 states in the XP-Fcs has
to be 40 meV or higher to achieve the observed ET rates. This
magnitude of V is far outside the range of semiquantum Marcus
theory under the nonadiabatic approximation. Moreover, the ET
from the S2 and S1 states of the XP-Fcs is insensitive to the
solvent polarity and much faster than the initial solvation time
(0.43 ps for THF and 0.53 ps for benzene).40 These results
demonstrate that the solvent dynamics are not the driving force
for ET in the XP-Fcs.
The single-color PP data show that there is vibrational
coherence during the ET reactions (Figure 6), which indicates
that the reactions proceed under nonequilibrium conditions
before vibrational relaxation is completed. In a previous study,20
we demonstrated that ultrafast ET in FPH2P-Fc is controlled
by translational and rotational motions of the ferrocene moiety
relative to the porphyrin plane in the coherent regime, which
modulate the π-orbital overlap between porphyrin and ferrocene
to achieve the adiabatic electronic coupling. This coherent
reaction mechanism might be common in the XP-Fcs. How-
ever, the vibrational coherence presented here is much more
complicated than that in FPH2P-Fc, because of the involvement
of the wavepacket motions in the S2 state as well as in the S1
state, the charge-separated state, and the ground state. Quantum-
dynamical simulations of the wavepacket motion in the excited
states41,42 are needed to determine the role of vibrational
coherence in the ultrafast ET in the XP-Fcs.
Because the S0-S2 excitation energy of porphyrin (∼2.9 eV)
is very close to the S0-S1 excitation energy of ferrocene (∼2.8
eV),32,33 energy transfer from porphyrin to ferrocene is an
alternative explanation for the ultrashort S2 lifetimes of the XP-
Fcs. However, based on the Fo¨rster dipole-dipole model,34 the
energy-transfer rate is estimated to be <3.7 × 1011 s-1 for
BPZnP-Fc. Here, because there are no structural data for the
orientation of ferrocene relative to the porphyrin plane, the
maximum possible value of the orientation factor (κ) was used
2
to estimate the upper limit of the rate constant [ κ e (4/π)2
when Jablonski’s two-dimensional transition dipole moment is
assumed for ZnP].35 The low rate constant of energy transfer is
due to the small extinction coefficient of ferrocene (<96 M-1
cm-1 33
and the small S2 fluorescence quantum yield of
)
porphyrin (3.7 × 10-4 for ZnTPP).36 It should be noted that,
when the distance between donor and acceptor is short compared
to their molecular sizes, the dipole-dipole model is not accurate
because of the contribution of dipole-quadrupole interaction;37
however, it is still useful for an order-of-magnitude estimation
of energy-transfer rate, which was confirmed for porphyrin-
fullerene dyads.38 The S2 fluorescence quenching of BPZnP-
Fc is two orders of magnitude faster than the estimated upper
limit of energy-transfer rate. Hence, energy transfer by the
Fo¨rster Coulombic mechanism is ruled out.
Energy transfer by the Dexter electron-exchange mechanism37
is also ruled out. The oxidation potential Eox of BPZnP was
found to be ∼0.4 eV (vs Fc/Fc+ in THF), indicating that the
energy level of the highest occupied molecular orbital (HOMO)
of porphyrin is lower than that of ferrocene. The oxidation

Electron Transfer in Porphyrin-Ferrocene Dyads
J. Phys. Chem. A, Vol. 111, No. 24, 2007 5143
(19) (a) Kambhampati, P.; Son, D. H.; Kee, T. W.; Barbara, P. F. J.
Phys. Chem. A 2000, 104, 10637-10644. (b) Son, D. H.; Kambhampati,
P.; Kee, T. W.; Barbara, P. F. J. Phys. Chem. A 2002, 106, 4591-4597.
(20) Kubo, M.; Mori, Y.; Otani, M.; Murakami, M.; Ishibashi, Y.;
Yasuda, M.; Hosomizu, K.; Miyasaka, H.; Imahori, H.; Nakashima, S. Chem.
Phys. Lett. 2006, 429, 91-96.
(21) (a) Tkachenko, N. V.; Lemmetyinen, H.; Sonoda, J.; Ohkubo, K.;
Sato, T.; Imahori, H.; Fukuzumi, S. J. Phys. Chem. A 2003, 107, 8834-
8844. (b) Kashiwagi, Y.; Imahori, H.; Araki, Y.; Ito, O.; Yamada, K.; Sakata,
Y.; Fukuzumi, S. J. Phys. Chem. A 2003, 107, 5515-5522.
(22) Bhasikuttan, A. C.; Suzuki, M.; Nakashima, S.; Okada, T. J. Am.
Chem. Soc. 2002, 124, 8398-8405.
(23) Miyasaka, H.; Murakami, M.; Okada, T.; Nagata, Y.; Itaya, A.;
Kobatake, S.; Irie, M. Chem. Phys. Lett. 2003, 371, 40-48.
(24) Miyasaka, H.; Moriyama, T.; Itaya, A. J. Phys. Chem. 1996, 100,
12609-12615.
(25) Felton, R. H. The Porphyrins; Dolphin, D., Ed.; Academic Press:
New York, 1978; Vol. V, Chapter 3.
In summary, we have designed three porphyrin-ferrocene
dyads with different substituents and different centers in the
porphyrin component. The variation in these systems of the S2-
S1 IC rates has a large impact in determining whether the main
reaction pathway of ET is from the S2 or S1 state. The ET
processes in these systems were found to be ultrafast and
independent of the solvation dynamics and free-energy gap. A
vibrationally coherent reaction mechanism on the adiabatic
potential surface would be plausible for the ultrafast photoin-
duced ET in the directly linked donor-acceptor systems.
Acknowledgment. This work was partially supported by a
Grant-in-Aid for the JSPS Fellowship for Young Scientists to
M.K. This work was also supported by Grants-in-Aid [Molecular
Nano Dynamics (432) to H.M. and 21st Century COE on Kyoto
University Alliance for Chemistry to H.I.] from the Ministry
of Education, Culture, Sports, Science and Technology (MEXT),
Japan.
(26) Cadioli, B.; Gallinella, E.; Coulombeau, C.; Jobic, H.; Berthier, G.
J. Phys. Chem. 1993, 97, 7844-7856.
(27) (a) Stein, P.; Ulman, A.; Spiro, T. G. J. Phys. Chem. 1984, 88,
369-374. (b) Rush, T. S., III; Kozlowski, P. M.; Piffat, C. A.; Kumble,
R.; Zgierski, M. Z.; Spiro, T. G. J. Phys. Chem. B 2000, 104 5020-5034.
(28) Kavarnos, G. J. Fundamentals of Photoinduced Electron Transfer;
VCH: New York, 1993.
References and Notes
(1) (a) Marcus, R. A. J. Chem. Phys. 1956, 24, 966-978. (b) Marcus,
R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265-322.
(2) Zusman, L. D. Chem. Phys. 1980, 49, 295-304.
(29) Weller, A. Z. Phys. Chem. 1982, 133, 93-98.
(30) (a) Asahi, T.; Ohkohchi, M.; Matsusaka, R.; Mataga, N.; Zhang,
R. P.; Osuka, A.; Maruyama, K. J. Am. Chem. Soc. 1993, 115, 5665-
5614. (b) Mataga, N.; Chosrowjan, H.; Shibata, Y.; Yoshida, N.; Osuka,
A.; Kikuzawa, T.; Okada, T. J. Am. Chem. Soc. 2001, 123, 12422-12423.
(c) Mataga, N.; Chosrowjan, H.; Taniguchi, S.; Shibata, Y.; Yoshida, N.;
Osuka, A.; Kikuzawa, T.; Okada, T. J. Phys. Chem. A 2002, 106, 12191-
12201. (d) Yoshida, N.; Ishizuka, T.; Yofu, K.; Murakami, M.; Miyasaka,
H.; Okada, T.; Nagata, Y.; Itaya, A.; Cho, H. S.; Kim, D.; Osuka, A. Chem.
Eur. J. 2003, 9, 2854-2866.
(31) DeGraziano, J. M.; Liddell, P. A.; Leggett, L.; Moore, A. L.; Moore,
T. A.; Gust, D. J. Phys. Chem. 1994, 98, 1758-1761.
(32) Sohn, Y. S.; Hendrickson, D. N.; Gray, H. B. J. Am. Chem. Soc.
1971, 93, 3603-3612.
(33) Bozak, R. E. AdVances in Photochemistry; Pitts, J. N., Jr.,
Hammond, G. S., Noyes, W. A., Jr., Eds.; Wiley-Interscience: New York,
1971; Vol. 8, pp 227-244.
(34) Fo¨rster, T. Discuss. Faraday Soc. 1959, 27, 7-17.
(35) Mårtensson, J. Chem. Phys. Lett. 1994, 229, 449-456.
(36) Kurabayashi, Y.; Kikuchi, K.; Kokubun, H.; Kaizu, Y.; Kobayashi,
H. J. Phys. Chem. 1984, 88, 1308-1310.
(3) Calef, D. F.; Wolynes, P. G. J. Phys. Chem. 1983, 87, 3387-3400.
(4) Rips, I.; Jortner, J. J. Chem. Phys. 1987, 87, 2090-2104.
(5) Hynes, J. T. J. Phys. Chem. 1986, 90, 3701-3706.
(6) Jortner, J.; Bixon, M. J. Chem. Phys. 1988, 88, 167-170.
(7) Bagchi, B.; Gayathri, N. AdV. Chem. Phys. 1999, 107, 1-80.
(8) (a) Sumi, H.; Marcus, R. A. J. Chem. Phys. 1986, 84, 4894-4914.
(b) Nadler, W.; Marcus, R. A. J. Chem. Phys. 1987, 86, 3906-3924.
(9) (a) Walker, G. C.; Akesson, E.; Johnson, A. E.; Levinger, N. E.;
Barbara, P. F. J. Phys. Chem. 1992, 96, 3728-3736. (b) Barbara, P. F.;
Walker, G. C.; Smith, T. P. Science 1992, 256, 975-981. (c) Kliner, D. A.
V.; Tominaga, K.; Walker, G. C.; Barbara, P. F. J. Am. Chem. Soc. 1992,
114, 8323-8325. (d) Tominaga, K.; Kliner, D. A. V.; Johonson, A. E.;
Levinger, N. E.; Barbara, P. F. J. Chem. Phys. 1993, 98, 1228-1243.
(10) Wasielewski, M. R. Chem. ReV. 1992, 92, 435-461.
(11) Bixon, M.; Jortner, J. AdV. Chem. Phys. 1999, 106, 35-202.
(12) Mataga, N.; Miyasaka, H. AdV. Chem. Phys. 1999, 107, 431-496.
(13) (a) Vos, M. H.; Lambry, J. C.; Robles, S. J.; Youvan, D. C.; Breton,
J.; Martin, J.-L. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 8885-8889. (b)
Vos, M. H.; Rappaport, F.; Lambry, J.-C.; Breton, J.; Martin, J.-L. Nature
1993, 363, 320-325.
(14) Wynne, K.; Reid, G. D.; Hochstrasser, R. M. J. Chem. Phys. 1996,
105, 2287-2297.
(15) (a) Seel, M.; Engleitner, S.; Zinth, W. Chem. Phys. Lett. 1997, 275,
363-369. (b) Engleitner, S.; Seel, M.; Zinth, W. J. Phys. Chem. A 1999,
103, 3013-3019. (c) Baigar, E.; Gilch, P.; Zinth, W.; Sto¨ckl, M.; Ha¨rter,
P.; von Feilitzsch, T.; Michel-Beyerle, M. E. Chem. Phys. Lett. 2002, 352,
176-184.
(37) Dexter, D. L. J. Chem. Phys. 1953, 21, 836-850.
(38) Kesti, T.; Tkachenko, N.; Yamada, H.; Imahori, H.; Fukuzumi, S.;
Lemmetyinen, H. Photochem. Photobiol. Sci. 2003, 2, 251-258.
(39) (a) Kilså, K.; Kajanus, J.; Larsson, S.; Macpherson, A. N.;
Mårtensson, J.; Albinsson, B. Chem. Eur. J. 2001, 7, 2122-2133. (b)
Pettersson, K.; Kilså, K.; Mårtensson, J.; Albinsson, B. J. Am. Chem. Soc.
2004, 126, 6710-6719.
(40) Horng, M. L.; Gardecki, J. A.; Papazyan, A.; Maroncelli, M. J.
Phys. Chem. 1995, 99, 17311-17337.
(41) (a) Jean, J. M.; Friesner, R. A.; Fleming, G. R. J. Chem. Phys.
1992, 96, 5827-5842. (b) Jean, J. M. J. Phys. Chem. A 1998, 102, 7549-
7557.
(42) (a) Wolfseder, B.; Seidner, L.; Domcke, W.; Stock, G.; Seel, M.;
Engleitner, S.; Zinth, W. Chem. Phys. 1998, 233, 323-334. (b) Egorova,
D.; Ku¨hl, A.; Domcke, W. Chem. Phys. 2001, 268, 105-120. (c) Egorova,
D.; Domcke, W. Chem. Phys. Lett. 2004, 384, 157-164.
(16) (a) Yoshihara, K.; Tominaga, K.; Nagasawa, Y. Bull. Chem. Soc.
Jpn. 1995, 68, 696-712. (b) Rubtsov, I. V.; Shirota, H.; Yoshihara, K. J.
Phys. Chem. A 1999, 103, 1801-1808. (c) Rubtsov, I. V.; Yoshihara, K. J.
Phys. Chem. A 1999, 103, 10202-10212.
(17) (a) Fiebig, T.; Stock, K.; Lochbrunner, S.; Riedle, E. Chem. Phys.
Lett. 2001, 345, 81-88. (b) Pandurski, E.; Fiebig, T. Chem. Phys. Lett.
2002, 357, 272-278.
(18) Nakashima, S.; Nagasawa, Y.; Seike, K.; Okada, T.; Sato, M.;
Kohzuma, T. Chem. Phys. Lett. 2000, 331, 396-402.