Design, synthesis, and uniquely electron-spin-polarized quartet photo-excited state of a π-conjugated spin system generated via the ion-pair state

Article abstract of DOI:10.1039/b714868b

A functionalized verdazyl radical, 1, was synthesized, which consists of a bodipy acceptor (A), a phenyl-anthracene donor (D), and a stable verdazyl radical (R). Optical measurements, electron spin resonance (ESR), time-resolved ESR (TRESR), and laser-exc

Full text of DOI:10.1039/b714868b

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Design, synthesis, and uniquely electron-spin-polarized quartet photo-
excited state of a p-conjugated spin system generated via the ion-pair state{
Yoshio Teki,*a Hirotaka Tamekuni,a Kotarou Haruta,a Jun Takeuchib and Yozo Miurab
Received 26th September 2007, Accepted 13th November 2007
First published as an Advance Article on the web 30th November 2007
DOI: 10.1039/b714868b
A functionalized verdazyl radical, 1, was synthesized, which consists of a bodipy acceptor (A), a
phenyl-anthracene donor (D), and a stable verdazyl radical (R). Optical measurements, electron
spin resonance (ESR), time-resolved ESR (TRESR), and laser-excitation pulsed ESR experiments
were carried out. The efficient intra-molecular energy transfer (EnT) from the anthracene moiety
to the bodipy functional component was observed by time-resolved fluorescence spectroscopy and
TRESR measurements. A unique dynamic electron-spin polarization (DEP) was detected for the
quartet photo-excited state of 1 by TRESR and pulsed ESR. Such a unique DEP was not detected
for either the parent verdazyl radical 2 without the bodipy acceptor or new compound 3 with an
anthraquinone moiety instead of the bodipy component. The spectral simulation of the quartet
spectrum of 1 revealed that the unique DEP was generated by competition between a mechanism
involving the intra-molecular ion pair *A2–D+–R and spin–orbit intersystem crossing. The dynamic
electron-spin polarization via the ion-pair state and the mechanism were discussed based on the
experimental data and theoretical calculations. The electronic structure and the spin delocalization
in the photo-excited quartet state of 1 were discussed by comparison with the results of 3.
pendant stable p-radicals. Such p-conjugated spin systems are
1. Introduction
ideal model compounds to study spin alignment in photo-
Spin alignment and magnetic properties of p-radicals are the
most important issue in the field of organic magnetism.1 The
systems are of interest as one of the topics of ‘‘fast photo-
excited states. The photo-excited states of p-conjugated spin
switching of the spin multiplicity’’2 or from the viewpoint of
novel photo-magnetic devices.16
intra-molecular spin alignment and exchange interactions via
p-conjugation in organic stable radicals are quite important as
well as the inter-molecular exchanges in the crystals. However,
works,8,9 photo-induced spin alignment between two pendant
most studies are limited to the ground state. Switching of
In the lowest photo-excited quintet state in our previous
physical properties by external stimuli has recently attracted
radicals was achieved through the triplet excited state of a
much attention in investigations of functional materials.1,2 In
diphenylanthracene moiety. The effective exchange coupling
inorganic compounds, the most representative example of the
between the two radicals changes from antiferromagnetic to
photo-control of magnetism through external stimuli is the
ferromagnetic on photo-excitation. This is the first example of
Prussian blue analogues.2–4 Manipulation of the magnetic
spin manipulation of p radicals in the photo-excited state. In
properties of organic molecules in the electronic ground state
their p-topological isomers, the photo-excited high-spin states
were not observed.8b We also reported a unique lowest triplet
has also been intensively studied by using electron–hole
doping5 and photo-chromic molecules.6 Very recently, photo-
photo-excited state with a low-lying quintet state, which was
switching between diamagnetic and paramagnetic phases
was realized in a 1,3,5-trithia-2,4,6-triazapentalenyl (TTTA)
organic crystal.7
generated by the antiferromagnetic spin coupling between two
iminonitroxide radicals through the photo-excited triplet spin
coupler.12 The role of the p-topology in the spin alignment of
the photo-excited states was clarified in these experiments.
Their electronic structures and the spin alignments were also
theoretically investigated by ourselves9b,10 and other groups.16,17
These p-conjugated spin systems have the following
advantages from the viewpoint of organic magnetism and
materials science. (1) p-Conjugated spin systems lead to robust
spin alignment compared to other triplet/radical-pair systems
(s-bonded systems18 and coordination complexes19). (2) In the
p-conjugated systems, the molecules with the desired spin state
can be well designed by taking the topology of the p-electron
network (p-topology) into account.8–15 One can synthesize
them using well-established synthetic chemistry. (3) An
enhanced intersystem crossing (ISC) mechanism arising from
the attachment of the radical species is available.8b
Studying spin alignment in photo-excited states will provide
key knowledge for photo-control of magnetic properties. We
have reported the lowest photo-excited quartet (S = 3/2) and
quintet (S = 2) high-spin states of p-conjugated organic
molecules8–15 constructed from aromatic hydrocarbons and
aDepartment of Material Science, Graduate School of Science, Osaka
City University, 3-3-138 Sugimoto, Sumiyoshi-ku, Osaka, 558-8585,
Japan. E-mail: teki@sci.osaka-cu.ac.jp; Fax: (+81) 6-6605-2559;
Tel: (+81) 6-6605-2559
bDepartment of Applied Chemistry, Graduate School of Engineering,
Osaka City University, 3-3-138 Sugimoto, Sumiyoshi-ku, Osaka, 558-
8585, Japan
{ Electronic supplementary information (ESI) available: ESR spectra
of 2; electronic structures in the ground state calculated by ab initio
MO method; TRESR spectra of 2. See DOI: 10.1039/b714868b
This journal is ß The Royal Society of Chemistry 2008
J. Mater. Chem., 2008, 18, 381–391 | 381

View Article Online
Fig. 1 Molecular structures of p-radicals 1, 2, 3 and functional
component A (molecule 5) studied in this work.
Scheme 1 Synthesis of 1.
Coupling spin alignment to photo-induced electron transfer
(PET) or energy transfer (EnT) is the next important
fundamental research target for the photo-control of organic
magnetism and for photo-magnetic devices. As a model
compound, we designed and synthesized 1 (A–D–R) shown
in Fig. 1, in which a bodipy20 acceptor moiety A21 is covalently
linked via an anthracene moiety as donor D to the verdazyl
radical R. In our previous work, we reported that the parent p
radical 2 (D–R) (without the bodipy component) has a quartet
(S = 3/2) photo-excited state.9
Previously, we reported briefly the first evidence of the
photo-excited quartet states of 1 with unique electron spin
polarization generated via an ion-pair state.15 Although the
bodipy moiety is well known as an efficient through-bond
energy acceptor for anthracene,21 we have observed that it will
work also as an electron donor for the anthracene moiety in
the photo-excited state. In this work, we describe in detail the
spin alignment and the electronic structures both in the ground
state and in the photo-excited state. We will discuss the
generation mechanism of the unique spin polarization.
Synthesis, UV/Vis absorption, time-resolved fluorescence
measurements, electron spin resonance (ESR), time-resolved
ESR (TRESR), and laser-excitation pulsed ESR experiments
were carried out for 1. For comparison, p-radical 3 (with an
anthraquinone moiety instead of the bodipy component) was
also synthesized and the photo-excited state was investigated.
The ab-initio molecular orbital calculations were carried out in
order to clarify the spin alignment in the photo-excited states as
well as the ground states. The details of the dynamic electron-
spin polarization via the ion-pair state and the generation
pathway were discussed based on the experimental data and
the theoretical calculations. The electronic structure and the
spin delocalization in the photo-excited quartet state of 1 were
discussed by comparison to the results of 3. In this report,
we will focus on the photo-excited states and the related
phenomena as well as the proper physical characterization of 1.
Scheme 2 Synthesis of 3.
component has no EnT character as well as no charge-transfer
(CT) character for the anthracene as discussed later. Therefore,
3 is a suitable compound to estimate the effect of the
delocalization of p-electrons under the conditions without
CT and EnT nature. Stable radical 2 was also synthesized
according to the procedures shown in our previous work.9b
Compound 5 was prepared according to a similar method to
that reported in the literature.21 The synthesis of 9-(p-formyl-
phenyl)-10-bromoanthracene was reported previously.14a
2-Ethynyl-anthraquinone 8 was also prepared according to a
similar method to that reported in the literature.22 Carbonic
acid bis(1-methylhydrazide) was prepared according to the
literature method.23 Other reagents were used as purchased.
Column chromatography was performed on silica gel (Merk
Silica 60) or alumina (Merk Alum. Ox. 60). Melting points
were recorded on a Yanagimoto micro melting-point appara-
tus and are uncorrected. 1H NMR spectra were measured with
a JEOL a-400 (400 MHz) spectrometer; chemical shifts are
expressed in ppm values (d) using Me4Si as an internal
standard. MS spectra were recorded on a JEOL JMS-700T
spectrometer. Hereafter, we describe the details of the
syntheses of 1 and 3; the synthesis of stable 2 was already
reported in our previous work.9b
2. Synthesis and materials
5-(p-Ethynylphenyl)-1,3,7,9-tetramethyl-dipyrromethene (4)
The stable radical 1 was synthesized by means of palladium-
mediated cross-coupling reactions as shown in Scheme 1. To
compare the effect on the attachment of the functional com-
ponent, we also synthesized p-radical 3 as shown in Scheme 2,
in which the anthraquinone component was linked to 2
through triple bonds similar to 1. The anthraquinone
To a solution of 0.843 g (6.48 mmol) of 4-ethylbenzaldehyde
and 1.29 g (13.6 mmol) of 2,4-dimethylpyrrole in 60 mL
anhydrous CH2Cl2 was added one drop of trifluoroacetic
acid and stirred for 24 h under nitrogen. A solution of 1.59 g
(6.48 mmol) of p-chloranil in 20 mL CH2Cl2 was added
dropwise to the mixture and stirred for 30 min. The organic
382 | J. Mater. Chem., 2008, 18, 381–391
This journal is ß The Royal Society of Chemistry 2008

View Article Online
layer was extracted by CH2Cl2, washed with brine, and dried
with MgSO4. Evaporation, column chromatography (alumina,
hexane–CH2Cl2 (2 : 1)) and recrystallization from hexane–
NH, 2H), 5.24 (t, J = 9.9 Hz, CH, 1H), 6.03 (s, pyrrolyl, 2H),
7.40–7.44 (m, ArH, 4H), 7.49 (d, J = 8.3 Hz, ArH, 2H), 7.61–
7.66 (m, ArH, 4H), 7.78 (d, J = 7.8 Hz, ArH, 2H), 7.94 (d,
J = 8.5 Hz, ArH, 2H), 8.76 (d, J = 8.5 Hz, ArH, 2H). Mp
234 uC (decomp.).
1
CH2Cl2 (2 : 1) gave 4 in 12% yield (0.241 g, 0.80 mmol). H
NMR (CDCl3): d 1.31 (s, CH3, 6H), 2.34 (s, CH3, 6H), 3.15 (s,
CCH, 1H), 5.89 (s, pyrrolyl, 2H), 7.28 (d, J = 7.8 Hz, ArH,
2H), 7.57 (d, J = 7.8 Hz, ArH, 2H). m.p.169–171 uC. HRMS
(EI): m/z: calcd for C21H20N2: 300.1626; found 300.1641. Anal.
calcd for C21H20N2: C, 83.96; H, 6.71; N, 9.33. Found: C,
83.34; H, 6.55; N, 9.19%.
9-[4-(6-Oxo-1,5-dimethylverdazyl)phenyl]-10-[4,4-difluoro-
1,3,5,7-tetramethyl-8-(p-ethynylphenyl)-4-bora-3a,4a-diaza-
s-indacene]anthracene (1)
To a vigorously stirred mixture of 0.051 g (0.07 mmol) of 7 and
K2CO3 (0.071 g) in 10 mL of 1,2-dichloroethane was added
PbO2 and the mixture was stirred for 1 h at room temperature.
After the reaction mixture was filtered and concentrated
under reduced pressure, the residue was chromatographed on
alumina with hexane–1,2-dichloroethane (1 : 2). Recrystalliza-
tion from hexane–1,2-dichloroethane gave 1 in 37% yield
(0.019 g, 0.03 mmol). Mp 248 uC (decomp.). HRMS (FAB+,
3-NBA matrix): m/z: calcd for [M+] C45H36BF2N6O: 725.3012;
found 725.3007. Anal. calcd for C45H36BF2N6O: C, 74.49; H,
5.00; N, 11.58. Found: C, 73.95; H, 4.79; N, 11.14%.
4,4-Difluoro-1,3,5,7-tetramethyl-8-(p-ethynylphenyl)-4-bora-
3a,4a-diaza-s-indacene (5)
To a solution of 0.241 g (0.80 mmol) of 4 in 70 mL anhydrous
CH2Cl2 was added 4.47 mL (25.7 mmol) of N,N-diisopropyl-
ethylamine, stirred for 20 min under nitrogen, and added
4.03 mL (32.1 mmol) of BF3OEt2 and stirred for 1 h. The
reaction mixture was washed by 2 N NaOH aq. The organic
layer was extracted by CH2Cl2, washed with brine, and dried
with MgSO4. Evaporation, column chromatography (silica gel,
CH2Cl2) and recrystallization from hexane–CH2Cl2 (1 : 1) gave
1
5 in 66% yield (0.186 g, 0.53 mmol) as red crystals. H NMR
9-(4-Formylphenyl)-10-(2-anthraquinonylethynyl)anthracene (9)
(CDCl3): d 1.40 (s, CH3, 6H), 2.25 (s, CH3, 6H), 3.18 (s, CCH,
1H), 5.98 (s, pyrrolyl, 2H), 7.28 (d, J = 8.3 Hz, ArH, 2H), 7.63
(d, J = 8.3 Hz, ArH, 2H). m.p. 252–254 uC. HRMS(EI): m/z:
calcd for C21H19BF2N2: 348.1609; found 348.1613. Anal. calcd
for C21H19BF2N2:C, 72.44; H, 5.50; N, 8.05. Found: C, 71.63;
H, 5.50; N, 7.90%.
2-Ethynylanthraquinone 8 was synthesized according to a
similar procedure to that reported in the literature.22
A
mixture of 0.300 g (1.29 mmol) of 8, 0.607 g (1.68 mmol) of
9-(p-formylphenyl)-10-bromoanthracene, 0.149 g (0.13 mmol)
of Pd(PPh3)4, and 3 mL of Et3N in 7.5 mL THF was refluxed
for 24 h under nitrogen. The organic layer was extracted by
CH2Cl2, washed with brine, and dried with MgSO4. Evapora-
tion, column chromatography (silica gel, CH2Cl2), and recrys-
tallization from hexane–CH2Cl2 (1 : 1) gave 9 in 51% yield
(0.340 g, 0.66 mmol) as red crystals. 1H NMR (CDCl3): d 7.44–
7.48 (m, ArH, 2H), 7.60–7.71 (m, ArH, 6H), 7.85–7.87 (m, ArH,
2H), 8.15 (d, J = 8.3 Hz, ArH, 2H), 8.18 (dd, J = 8.0 Hz and
1.6 Hz, ArH, 1H), 8.36–8.40 (m, ArH, 2H), 8.43 (d, J = 8.0 Hz,
ArH, 1H), 8.70 (d, J = 12.0 Hz, ArH, 1H), 8.77 (d, J = 8.3 Hz,
ArH, 2H), 10.22 (s, CHO, 1H). Mp 269–270 uC. HRMS (EI):
m/z: calcd for [M+] C37H20O3: 512.1412; found 512.1410. Anal.
calcd for C37H20O3: C, 86.70; H, 3.93. Found: C, 86.41; H, 4.24%.
9-(p-Formylphenyl)-10-[4,4-difluoro-1,3,5,7-tetramethyl-8-
(p-ethynylphenyl)-4-bora-3a,4a-diaza-s-indacene]anthracene (6)
A mixture of 0.140 g (0.40 mmol) of 5, 0.217 g (0.60 mmol) of
9-(p-formylphenyl)-10-bromoanthracene, 0.046 g (0.04 mmol)
of Pd(PPh3)4, and 2 mL of Et3N in 5 mL THF was refluxed
for 24 h under nitrogen. The organic layer was extracted
by CH2Cl2, washed with brine, and dried with MgSO4.
Evaporation, column chromatography (silica gel, hexane–
CH2Cl2 (1 : 1)) gave 6 in 58% yield (0.145 g, 0.23 mmol) as
1
orange crystals. H NMR (CDCl3): d 1.53 (s, CH3, 6H), 2.59
(s, CH3, 6H), 6.03 (s, pyrrolyl, 2H), 7.41–7.46 (m, ArH, 4H),
7.59–7.66 (m, ArH, 6H), 7.94 (d, J = 8.4 Hz, ArH, 2H), 8.14
(d, J = 7.8 Hz, ArH, 2H), 8.78 (d, J = 8.4 Hz, ArH, 2H), 10.21
(s, CHO, 1H). Mp 247 uC (decomp.). HRMS (FAB+, 3-NBA
matrix): m/z: calcd for C42H31BF2N2O: 628.4948; found
628.2504. Anal. calcd for C42H31BF2N2O: C, 80.26; H, 4.97;
N, 4.46. Found: C, 80.37; H, 5.07; N, 4.38%.
9-[4-(6-Oxo-1,5-dimethyl-1,2,4,5-tetrazinan-3-yl)phenyl]-10-(2-
anthraquinonyl-ethynyl)anthracene (10)
To a hot solution of 0.228 g (1.93 mmol) of carbonic acid
bis(1-methylhydrazide) in MeOH (5 ml)–THF (5 mL), was
added dropwise a solution of 0.760 g (1.48 mmol) of 9 in
MeOH (80 ml)–THF (80 mL), and the mixture was refluxed
for 15 h. The reaction mixture was evaporated under reduced
pressure and the crude residue was washed by MeOH gave 10
(Crude) in 92% yield (0.838 g, 1.37 mmol). 1H NMR (CDCl3):
d 3.26 (s, CH3, 6H), 4.56 (d, J = 9.8 Hz, NH, 2H), 5.25 (t, J =
10.0 Hz, CH, 1H), 7.46 (m, ArH, 4H), 7.64–7.67 (m, ArH, 4H),
7.78 (d, J = 7.6 Hz, ArH, 2H), 7.85–7.87 (m, ArH, 2H), 8.17
(dd, J = 8.0 Hz and 1.6 Hz, ArH, 2H), 8.36–8.40 (m, ArH, 2H),
8.42 (d, J = 8.0 Hz, ArH, 2H), 8.69 (d, J = 1.6 Hz, ArH, 1H), ),
8.75 (d, J = 8.8 Hz, ArH, 2H). Mp 227 uC (decomp.). HRMS
(FAB+, T-Gly matrix): m/z: calcd for [M + H] C40H29N4O3:
613.2234; found 613.2227.
9-[4-(6-Oxo-1,5-dimethyl-1,2,4,5-tetrazinan-3-yl)phenyl]-10-
[4,4-difluoro-1,3,5,7-tetramethyl-8-(p-ethynylphenyl)-4-bora-
3a,4a-diaza-s-indacene]anthracene (7)
A mixture of 0.150 g (0.24 mmol) of 6 and 0.127 g (1.07 mmol)
of carbonic acid bis(1-methylhydrazide) in MeOH (10 ml)–
THF (20 mL) was stirred for 15 h at room temperature. The
reaction mixture was evaporated under reduced pressure and
the crude residue was washed by MeOH gave 7 (Crude) in 84%
1
yield (0.147 g, 0.20 mmol). H NMR (CDCl3): d 1.55 (s, CH3,
6H), 2.58 (s, CH3, 6H), 3.25 (s, CH3, 6H), 4.55 (d, J = 9.9 Hz,
This journal is ß The Royal Society of Chemistry 2008
J. Mater. Chem., 2008, 18, 381–391 | 383

View Article Online
9-[4-(6-Oxo-1,5-dimethylverdazyl)phenyl]-10-(2-
anthraquinonylethynyl)anthracene (3)
configuration: a glassy carbon disk electrode served as the
working electrode, a Pt wire and Ag/Ag+ were used as the
counter electrode and the reference electrode, respectively.
0.1 M of (t-Bu)4NBF6 was used as the electrolyte. Ferrocene
was used as internal standard.
A mixture of 0.100 g (0.16 mmol) of 10 and 0.026 g (0.24 mmol)
of p-benzoquinone in 30 mL benzene was refluxed for 4 h.
After the reaction mixture was filtered and concentrated
under reduced pressure, the residue was chromatographed
on alumina with benzene, and freeze-dried to give 3 in 36%
yield (0.036 g, 0.059 mmol) as orange powders. Mp 264 uC
(decomp.). HRMS (FAB+ 3-NBA matrix): m/z: calcd for
[M + H] C40H26N4O3: 610.2005; found 610.1990. Anal. calcd
for C40H25N4O3: C, 78.80; H, 4.13; N, 9.19. Found: C, 77.71;
H, 4.09; N, 8.87%. Anal. calcd for C40H25N4O3?0.5H2O: C,
77.66; H, 4.24; N, 9.05. Found: C, 77.71; H, 4.09; N, 8.87%.
4. Results and discussion
4.1 UV/Vis absorption and emission properties
The UV/Vis spectrum of a solution of p-radicals 1, 2, 3 and
bodipy component A (the molecule 5) in toluene at room
temperature are shown in Fig. 2. A sharp band at 505 nm of 1
is assigned to the characteristic absorption band of the bodipy
acceptor A from the comparison between their spectra.21 The
characteristic absorption bands of the verdazyl moiety are not
seen clearly in 1–3, because the absorption coefficients are
much smaller than those of the anthracene or bodipy moieties
on the same scale. The weak n A p* transition of the verdazyl
radical R overlaps with the characteristic absorption band of
the bodipy moiety (see the scale-up (650) spectrum of 2). The
absorption bands (360–460 nm bands) characteristic of the
anthracene moiety are red-shifted by about 40 nm compared
with those of 2 (320–420 nm bands), indicative of an
interaction between the D and A components. The magnitude
of the interaction between D and A is estimated to be about
2300 cm21 from the shift. Interestingly, the band of the
anthracene moiety is enhanced about 3 times by the attach-
ment of the bodipy component. At the moment, we can not
show the detail mechanism of this enhancement.
3. Measurements
3.1 Optical
UV/Vis spectra were measured on a JASCO V-570 at
room temperature. Time-resolved emission spectra were taken
at 77 K using an optical cryostat (Oxford Optistat-CF).
Excitation was carried out at 355 nm light by an Nd:YAG
laser (Continuum Surelite II-10) or at 505 nm light from an
optical parametric oscillator (OPO) system (Continuum
Surelite OPO) pumped by the YAG laser. The time-resolved
emission spectra were measured by an imaging mono-
chromator (Roper Scientific SP-300) followed by a CCD
detector with a high-speed gating system (Roper Scientific
PI-MAX 512SS-FG-43). Toluene glass matrix was used for the
emission experiments.
A similar red-shift of the absorption band characteristic of
the anthracene moiety was also observed in the case of 3,
showing the delocalization of p-electrons toward the anthra-
quinone component. Judging from the magnitude of the shift,
the amount of the delocalization is similar to that of 1.
However, in contrast to the case of 1, enhancement of the
absorption intensity was not observed in this case. Therefore,
these findings show that 3 is a suitable compound to judge
only the effect of the delocalization of p-electrons in the
electronic state.
3.2 ESR, TRESR, and pulsed ESR measurements
A conventional X-band ESR spectrometer (JEOL TE300) was
used for the cw-ESR experiment. In the TRESR measure-
ments, the same spectrometer was used without field modula-
tion. Signals were amplified by a wide-band preamplifier,
transferred to a high-speed digital oscilloscope (LeCroy
9350C), and accumulated for each point. Excitation of 1 was
carried out with 505 nm or 447 nm light from the OPO system
pumped by the YAG laser. The temperature was controlled by
an Oxford ESR910 cold He gas flow system. The pulsed ESR
measurements were performed on an X-band ESR spectro-
meter equipped with a pulsed microwave unit (JEOL ES-
PX1150) and a high-speed digital oscilloscope (Tektronix
TDS5034). The pulsed microwaves were amplified with a 1 kW
travelling wave tube amplifier (TWTA). Hahn’s p/2–p pulse
sequence was used for spin-echo detection. The microwave
pulse was synchronized with the laser excitation by using a
delay-pulse generator (Stanford Research DG535). In the
echo-detected nutation experiment, the first microwave pulse
length was varied. All TRESR and pulsed ESR experiments
were carried out with toluene as glass matrix or solvent.
Samples were degassed by repeated freeze–pump–thaw cycles.
Fig. 3 shows fluorescence spectra of 1 observed upon
excitation at different wavelengths. Similar fluorescence
spectra were observed upon both 355 nm (shoulder of the
optical band of the anthracene moiety) and 505 nm (the optical
3.3 Cyclic voltammetry measurements
Cyclic voltammograms were measured in CH2Cl2 or aceto-
nitrile solution using a Hokuto Dennko HSV-100. The
measurements were made with a standard three-electrode
Fig. 2 UV/Vis absorption spectra of 1, 2, 3 and 5 (component A).
This journal is ß The Royal Society of Chemistry 2008
384 | J. Mater. Chem., 2008, 18, 381–391

View Article Online
Fig. 4 Conventional ESR spectra of 1 and 3 in toluene solution at
room temperature. (a) and (b) are the solution ESR spectra of 1 and 2,
respectively (upper one; observed spectrum, lower one; simulation).
Fig. 3 Fluorescence spectra and decay profile of 1 in toluene glass
matrix at 77 K (solid line: 355 nm excitation, broken line: 505 nm
excitation). The inset shows the time decay profile of the fluorescence.
HOMO of the A component (Eox = +0.85 V) becomes a half
filled orbital (a partially vacant) upon photo-excitation.
Therefore, A* will act as an electron acceptor for D, and
PET occurs immediately in 1 through p conjugation. In the
photo-excited state of 1, the bodipy component will act as an
‘‘electron acceptor’’ (A*), and the phenylanthracene moiety
plays the role of an electron donor (D).
band of the component A) excitation. Both spectra are very
similar to the fluorescence spectrum characteristic of bodipy
reported in the literature.21 Thus, the fluorescence charac-
teristic of bodipy was obtained by excitation of D as well as of
A. This finding shows that efficient energy transfer occurs
from the anthracene moiety D to the bodipy component A.
The time-decay profile of the fluorescence is also shown in
Fig. 3. The decay profile was well analyzed as a single-
exponential decay, and the lifetime of the photo-excited
doublet state (A–1D*–R) generated immediately after the
pulsed laser excitation was determined to be about 6 ns.
4.3 Electronic ground states
To clarify the electronic ground state, we measured conven-
tional ESR spectra of 1, 2 and 3 in toluene solution. Fig. 4
shows the solution ESR spectra of 1 and 3 at room tempera-
ture and the spectral simulations. A similar spectrum was
observed for 2 (Fig. S1 in ESI{). The solution ESR spectra
were well analyzed by the spectral simulation. Their electronic
ground states are the doublet (S = 1/2) state. The g values, the
hyperfine coupling constants (hfcc) of the four nitrogen atoms,
and hfcc of the hydrogen atoms in CH3 groups in the verdazyl
radical moiety are listed in Table 2. These values are very close
to those of the phenylverdazyl radical reported by Neugebauer
and Fisher.25 This means that the attachments of the func-
tional groups (bodipy or anthraquinone) have less effect on the
electronic structures of the radical moieties in the ground
states. Thus, the electronic structures of the verdazyl radical
moieties in 1 and 3 are close to that of the phenylverdazyl
radical. This finding is also supported by the ab-initio mole-
cular orbital calculations26 in their electronic ground states
(Fig. S2 in ESI{).
4.2 Electro-chemical properties
The electro-chemical behaviour of 1, 2, 3 and A (the
molecule 5) was studied in CH2Cl2 or acetonitrile solution by
the CV method. Ferrocene was used as internal standard.
Redox potentials of 1, 2, 3 and A are listed in Table 1. These
oxidation potentials shows that the p-radicals 1 and 2 are
oxidized at lower potentials than phenylanthracene (D). This is
probably because the energies of the HOMOs of 1 and 2
become higher (+0.80 eV A +0.27 eV in 1 and +0.80 eV A
+0.24 eV in 2) than in phenylanthracene (D) upon attachment
of the verdazyl radical component.24 The oxidation potential
of phenylanthracene is slightly higher than that of component
A.21 The reduction potential of component A is Ered
=
21.75 V, showing the LUMO of A is much higher than the
HOMO of 1. Thus, there is no CT character in the ground
state. However, if one electron photo-excitation is carried out
from the HOMO of component A, the orbital becomes lower
in energy, because on-site Coulomb repulsion is removed. The
4.4 TRESR of photo-excited states
To clarify more about the photo-excited state, we measured
time-resolved ESR (TRESR) and pulsed ESR spectra syn-
chronized to pulsed laser excitation. Fig. 5a shows the TRESR
spectrum of 1 together with the spectral simulation, which was
observed 0.3 ms after laser excitation of the absorption band
Table 1 Redox potentials of p-radicals 1, 2, 3, phenylanthracene (D)
and bodipy component (A) (ferrocene was used as internal standard).
The energies of the HOMOs of 1, 2 and 3 become higher than those of
phenylanthracene upon attachment of the verdazyl radical component
Table 2 g Values and hyperfine coupling constants (A) of 1 and 3 in
the electronic ground states. Numbers in parentheses show the number
of equivalent nuclei
E
ox2/V
E
ox1/V
E
red/V
1
2
3
+0.85
+0.79
+0.47
—
+0.27
+0.24
+0.26
+0.80
+0.85
—
—
—
—
g
A1N(2)/mT
A2N(2)/mT
AH(6)/mT
1
2
3
2.0033
2.0040
2.0038
0.65
0.64
0.64
0.53
0.53
0.53
0.53
0.53
0.53
D
A (5)
—
21.75
This journal is ß The Royal Society of Chemistry 2008
J. Mater. Chem., 2008, 18, 381–391 | 385

View Article Online
Fig. 6 Details of the spectral simulation of 1. (a) Simulation obtained
by the superimposition of (b), (c), and (d); (b) simulation obtained by
assuming the selective population to the high-field wavefunctions; (c)
simulation obtained by assuming the selective population to the
zero-field wavefunctions; (d) simulation of polarized doublet state
with g = 2.005.
Fig. 5 TRESR of 1 and 2 at 30 K in glass matrix. a: absorptions;
e: emissions of microwaves. a) Upper: spectrum of 1 observed upon
excitation with l = 505 nm, lower: simulation spectrum; b) spectrum of
1 observed upon excitation with l = 447 nm; c) observed spectrum of 2.
Table 3 Spin quantum numbers S, g values, and zero-field splitting
parameters of the lowest photo-excited states of 1, 2 and 3
(l = 505 nm) of the bodipy component A. As shown in Fig. 5b,
the almost same spectrum was obtained by excitation of the
absorption band (l = 447 nm) of the anthracene moiety. This
is the result of the efficient energy transfer from the anthracene
moiety D to the component A as pointed out in the emission
experiments. As shown in Fig. 3, the fluorescence spectrum
characteristic of the bodipy was obtained by excitation of the
absorption band of D as well as of A, showing the efficient
EnT occurs from D to A component. Fig. 5c is the TRESR
spectrum of 2, which was observed using 2-MTHF glass
matrix upon laser excitation of the absorption band (l =
355 nm) of the anthracene moiety. Comparison with the
TRESR spectrum of 2 (Fig. 5c) clearly shows that the phase
pattern of the dynamic electron polarization (DEP) of 1
(aeeaae; a/e: absorption/emission of microwaves) is different to
that of 2 (aaaeee). This unique DEP is generated for the first
time by the attachment of the bodipy functional group.
S
g
D/cm21
E/cm21
1
2
3
3/2
3/2
3/2
2.0035
2.0035
2.0035
0.0215
0.0230
0.0215
0.0010
0.0004
0.0017
wavefunctions (the mechanism through IP state discussed
later) and by the selective population to the zero-field
wavefunctions (spin-orbit intersystem crossing mechanism,
SO-ISC), respectively. Fig. 6d is the simulation spectrum of the
polarized doublet (S = 1/2) state with g = 2.005.
Judging from the g value, this doublet state may be the
electronic ground state of 1, which was polarized by the energy
relaxation from the photo-excited states. The actual amount of
the contribution of the doublet state is very little (0.14),
because the doublet signals assemble at the same resonance
field position. Wasielewski et al. also reported donor/acceptor
systems with pendant radicals in solution30 at almost the same
time as our previous report.15 Although they did not detect the
photoexcited quartet spectra directly,31 they reported the
doublet ground state with a net spin polarization as similar
to that shown in Fig. 6d, which is also pointed out in our
previous paper.32 They gave an explanation for the polarized
ground state by assuming D–Q (doublet–quartet) mixing in the
photo-excited state. In our system, the sign of the exchange
coupling between the radical component and the excited triplet
moiety is expected to be ferromagnetic and the magnitude of
the exchange interaction is much larger in the observed quartet
state. Apart from the sign of the net polarization, the
mechanism generating the net polarization in the doublet
ground state of 1 may resemble that of Wasielewski’s systems,
Fig. 5a (lower) shows the spectral simulation of the TRESR
spectrum of 1. We show the details of the spectral simulation
in Fig. 6. The spectral simulations were carried out using the
hybrid eigenfield27/exact diagnonalization method,28 taking
dynamic electron polarization (DEP) into account.8,29 The
spin Hamiltonian parameters of 1 were accurately determined
by the spectral simulation. The determined parameters are
listed in Table 3 together with those of 29 and 3. The simula-
tion was obtained by the superimposition of Fig. 6b, c and d
(maximum intensity peaks of (a)–(d) are normalized in these
figures) with the weight of 0.64 : 0.36 : 0.82 (actual weights of
each spectra are 0.55, 0.45, and 0.14), respectively. Here, the
total population of the quartet state is normalized to unity.
Fig. 6b and c are the simulation spectra of the quartet (S = 3/2)
state obtained by the selective population to the high-field
386 | J. Mater. Chem., 2008, 18, 381–391
This journal is ß The Royal Society of Chemistry 2008

View Article Online
Fig. 7 TRESR and simulation of 3 at 30 K in toluene glass matrix.
(a) Spectrum of 3 observed upon excitation with l = 480 nm; (b)
spectral simulation.
Fig. 8 Time profile of the TRESR signals of 1.
because a charge separated IP state is concerned in our system,
and in the IP state, D–Q mixing is expected as discussed later.
Fig. 7 is the TRESR spectrum and the spectral simulation of
compound 3. This spectrum was observed using toluene glass
matrix by excitation (l = 480 nm) corresponding to the
absorption band of the anthracene moiety. The characteristic
absorption band of anthraquinone is 325 nm, which is much
higher in energy than that of anthracene. Therefore, the EnT
can not occur in this situation. In this case, the DEP phase
pattern was aaaeee, which was similar to the case of 2.
Thus, the unique DEP phase pattern (aeeaae) observed in the
case of 1 was not obtained for 3, in which there is no CT and
no EnT nature.
Fig. 9 Spin-echo-detected pulsed ESR of 1 at 30 K in toluene
glass matrix.
The TRESR spectra of 2 and 3 were well simulated by
assuming the selective population only to the zero-field
wavefunctions (SO-ISC mechanism) as shown in Fig. S3
(ESI{) and Fig. 7b, respectively. Thus, the clear difference in
the generation pathway of the photo-excited quartet states
was proven between 1 and the others (2 and 3). The observed
spectrum of 3 quite resembles the simulation spectrum of
Fig. 6c (the simulation assuming the SO-ISC mechanism). The
magnitude of D of 1 is about 7% smaller than that of 29 and
indicates delocalization of the unpaired electron toward the
acceptor moiety A. The magnitude of D is very close to that of
1. Thus, the delocalization of the unpaired p-electrons via the
triple bond in 3 is similar to that of 1.
MS = 21/2 « +1/2 transition arising from the ground and
photo-excited states. Other signals come from the photo-
excited state, as was confirmed by the spectrum without
photo-excitation. The spectral pattern of the signals due to the
photo-excited state is almost the same as that of the TRESR
spectrum, in that all signals have similar phase memory times
and spin–lattice relaxation times.12b In the echo-detected
nutation experiment, the pulse length of the first microwave
pulse was changed and the second pulse was set to p pulse. The
rotation angle of the magnetic moment depends on the length
of the first microwave pulse. Since the rate constant of the
rotation is directly proportional to the ESR transition
probability under proper conditions, we can determine the
spin quantum number directly from the period (2p/vTN) of
the rotation of the magnetic moment as described below.
Under the condition v1 % vZFS, the TN frequency vTN of the
MS « MS + 1 transition is given by eqn (1):33
The time profiles of the TRESR spectrum of 1 are shown in
Fig. 8. The wing signals decay within about 1 ms. In contrast,
the signal at ca. 330 mT remains long after laser excitation.
Only this signal is not reproduced by the spectral simulation.
This finding shows that this signal may be arising from
different species, such as long-lived radical pairs. Other signals
come from the same high-spin quartet species.
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
vTN
~
SðSz1Þ{MSðMSz1Þv1
(1)
where v1 = gbB1/h, vZFS = D/h, and B1 is the microwave-field
strength. Therefore, in the quartet state, the expected TN
frequencies for MS = ¡1/2 « ¡3/2 and MS = 21/2 « +1/2
allowed transitions are !3v1 and 2v1, respectively. Fig. 10a
and Fig. 10b–e shows typical nutation behaviour for the
signals arising from the ground state and the photo-excited
states, respectively. For the ground state, vGTN is equal to v1,
because 1 has a doublet ground state (S = 1/2). The observed
4.5 Pulsed ESR of 1
To confirm the spin state, spin-echo-detected transient
nutation (TN) spectroscopy33 was carried out. Fig. 9 depicts
the electron spin-echo-detected ESR spectrum observed 0.3 ms
after pulsed laser excitation (l = 505 nm). Fig. 10 shows typical
TN behaviour of the photo-excited state and the electronic
ground state. The strong center signal is a superposition of the
This journal is ß The Royal Society of Chemistry 2008
J. Mater. Chem., 2008, 18, 381–391 | 387

View Article Online
calculations of the photo-excited states were carried out based
on the density functional theory (DFT) using Gaussian 98.26
In the case of the parent molecule 2, the molecular structure
was optimized by the DFT calculation (Ubecke 3LYP/STO-
3G). For molecule 3, the structure optimization was carried
out using a semi-empirical method (MNDO/AM1), because of
the large molecular size. For the optimized molecular struc-
tures, more accurate calculations ((Ubecke 3LYP/6-31G(p,d))
were carried out, in order to clarify the role of the spin
polarization effect. Fig. 11a and b show the calculated total
spin density distributions in the lowest photo-excited quartet
states of 2 and 3, respectively. The mechanism of the spin
alignment was clarified based on the spin distribution.
Unfortunately, the ab-initio MO calculation using the same
level was not successful in the case of 1, because the structure
optimization was not converged even in the semi-empirical
calculation (MNDO/AM1).34 In the excited quartet states of 2
and 3 the spin density distributions of the anthracene moieties
are nearly symmetrical as shown in Fig. 11. Almost two net
unpaired spins exist on the anthracene moiety and the
remaining one unpaired spin exists on the pendant radical
component. This shows that the excited quartet states are
generated by a one-electron transition from the b-HOMO (p)
to the a-LUMO (p*), the same as those reported pre-
viously.8–10 These results are consistent with our experimental
results, in which the magnitude of the observed fine-structure
splitting has been well interpreted as an exchange coupled
system35 constructed from the pp* excited triplet state of the
anthracene moiety and the doublet state of the pendant radical.
In the anthracene moiety, positive spin densities overcome
the negative spin densities in the quartet photo-excited state.
Thus, the spin delocalization overcomes the spin polarization
effect on the anthracene moiety. In contrast, alternating signs
of the spin densities are realized on the pendant phenyl groups
in the phenylanthracene moiety. This sign alternation of the
spin densities shows that the spin polarization mechanism
overcomes the spin delocalization within the phenyl group
even in the excited state. Thus, the unpaired spin in the
pendant radical couples ferromagnetically via the static spin
polarization effect in the phenyl group to the excited triplet
spins delocalized on the anthracene moiety, and this leads to
the excited quartet high-spin state.
Fig. 10 Electron spin transient nutation (TN) behaviour of 1. The
transitions I–IV and G correspond to the transition indicated by
arrows in Fig. 9. The TN behaviour of the ground state was measured
without photo-excitation at the magnetic field position corresponding
to the G transition.
Table 4 The ratios of TN frequencies (vEX/vG) for each transition
indicated in Fig. 9 by arrows
I
II
III
IV
vEXTN/vG
1.75
1.66
1.76
1.74
TN
Fig. 11b shows that, in the anthraquinone moiety, the
delocalization of the unpaired electrons is restricted to the
phenyl group, which is linked to the ethynyl group and further
delocalization does not occur. This restricted slight delocaliza-
tion in the functional group is consistent with the experimental
result of the magnitude of D of 3 being about 7% smaller than
ratios vEX/vG for each transition indicated in Fig. 9 by
arrows are listed in Table 4. The ratios for the transitions I–IV
are close to !3, which is expected for the ratio of the TN
frequencies between the quartet and the doublet states.
The TN experiments show unambiguously that all signals
indicated by arrows are assigned to MS = ¡1/2 « ¡3/2
transitions of the quartet state. The center signal attributed
to MS = 21/2 « +1/2 transitions showed complicated TN
behaviour as shown in Fig. 10d. Fourier transformation of
the TN behaviour gave a power spectrum with multi-
frequencies, consisting of 2v1, v1, and unknown lower-
frequency signals.33b
4.6 Spin alignment in the photo-excited quartet state of 1
Fig. 11 Spin density distributions in the photo-excited quartet states
of 2 and 3 calculated by the ab-initio MO calculations. (a) Spin density
distribution of 2, (b) spin density distribution of 3.
To obtain the physical picture of the photo-induced intra-
molecular spin alignment, the ab-initio molecular orbital
388 | J. Mater. Chem., 2008, 18, 381–391
This journal is ß The Royal Society of Chemistry 2008

View Article Online
that of 2. The magnitude of the D value of the photo-excited
quartet state of 1 agrees well with that of 3 as shown in Table 3.
This finding shows that the similar situation occurs in the
quartet state of 1. Thus, the almost the same D values for 1 and
3 in their photo-excited quartet states mean that the spin
delocalization of 1 is also similar to that of 3. These findings
confirm that the difference observed in the TRESR spectra
originated from the difference of the generation pathway of the
photo-excited high-spin states and that the photo-excited
quartet state of 1 is generated through the ion-pair state
obtained by the photo-induced electron transfer (PET) from
the anthracene moiety to the bodipy functional group as
discussed later.
Fig. 13 Spin density distributions of the triplet state of 2+ and the
constituent components calculated by the ab-initio MO calculations.
(a) Triplet state of 2+, (b) phenylanthracene cation radical, (c)
phenylverdazyl radical.
2
?
where A is the doublet state. Fig. 13 shows the spin density
distribution of the triplet state of 2+ (D+–R) together with
those of the ground states of the phenylanthracene cation
radical (D+) and the phenylverdazyl radical (Ph–R) obtained
by the ab-initio MO calculations. The signs of spin densities in
the phenyl groups in the phenylanthracene cation radical (D+)
are coincident with those of the phenylverdazyl radical (Ph–R).
Therefore, according to the p-topological rule established well
in the electronic ground state, the cationic state of 2 (D+–R)
constructed from D+ and R is expected to be the triplet ground
state from the spin density distributions. In the charge
separated IP state of 1, the electron spins of A2 and D+ are
far apart and their exchange coupling is very weak. As a
consequence, the wavefunctions of the doublet and quartet
states are mixed, leads to the quantum mixing states (w1–w6)
given in eqn (2a)–(2f):
4.7 Mechanism of the DEP generation of 1
The quartet photo-excited states of triplet/radical pairs
reported so far in the solid phase are formed by spin–orbit
intersystem crossing (SO-ISC) of the parent triplet state,18,19 or
by enhanced SO-ISC via p conjugation with the pendant
radical.8–14 In both cases, the SO-ISC mechanism generates
selective population of the zero-field (ZF) wavefunctions of the
quartet spin states, and this leads to an A/E (aaaeee) or E/A
(eeeaaa) pattern. In contrast, for the two-spin system, it is well-
known that the radical-pair (RP) mechanism36 in solution
leads to selective population of the high-field (HF) wavefunc-
tions of the MS sublevels by singlet–triplet mixing S–T0 (or
S–T2). The DEP depends on the pathway that leads to the
observed state. Thus, the DEP pattern gives evidence for the
dynamic process generating the observed state. Spectral
simulation shown in Fig. 5 was carried out by assuming
selective population for both the ZF and HF wavefunctions
(ZF : HF = 0.45 : 0.55). Thus, a competition between SO-ISC
and other mechanisms occurs in 1. In the case of the triplet
state, similar competition between SO-ISC and RP mechan-
isms has been observed only in the reaction centers of photo-
systems I and II37 and in their model systems via the ion-pair
(IP) state.38
ꢀ
ꢀ
w1&jTz1aT~ Q3=2T
(2a)
(2b)
rffiffi
rffiffiffi
ꢀ
ꢀ
ꢀ
1
2
ꢀ
w2&jTz1bT~
Q1=2Tz
D1=2T
3
3
rffiffi
rffiffi
ꢀ
ꢀ
ꢀ
2
1
ꢀ
w3&jT0aT~
w4&jT0bT~
w5&jT{1aT~
Q1=2T{
D1=2
T
(2c)
(2d)
3
3
rffiffi
rffiffi
rffiffi
ꢀ
ꢀ
ꢀ
2
1
ꢀ
Q{1=2Tz
D{1=2
T
T
3
3
We propose a model to understand the unique electron
spin polarization of the quartet photo-excited state of 1.
Immediately after photo-excitation, component A reaches the
singlet photo-excited state (S1). As already mentioned in the
section on electro-chemical properties, PET is expected to
occur immediately in this excited state from D to A component
through the p conjugation in 1, as shown in Fig. 12.39 This
rffiffi
ꢀ
ꢀ
ꢀ
1
2
ꢀ
Q1=2T{
D{1=2
(2e)
(2f)
3
3
ꢀ
ꢀ
w6&jTz1aT~ Q3=2T
Fig. 14 shows the schematic diagram of the most plausible
mechanism of the unique DEP generation observed in 1. As
mentioned above, the wavefunctions of the doublet states
|D¡1/2T and the quartet states |QMsT are mixed in the charge
separated IP state of 1, because their exchange coupling is very
weak. This IP state will be generated through the PET from the
excited doublet state generated by the photo-excitation of the
anthracene moiety (D) as shown in Fig. 12. The PET process
conserves the spin quantum numbers, S and MS. The initial
state generated by photo-excitation is the doublet excited state,
because 1 has a doublet ground state. Therefore, the sublevels
w2–w5 of the quantum mixing states will be selectively
populated, because these sublevels have doublet character
|D¡1/2T. This selective population transfer to w2–w5 spin
sublevels with doublet character is reasonable, because this is
a kind of the opposite case (spin selective population) to the
2
leads to the charge-separated ion-pair (IP) state (A –D +–R),
?
?
Fig. 12 Generation of the ion-pair state of 1 by photo-induced
electron transfer (PET) followed by charge recombination and
intersystem crossing processes.
This journal is ß The Royal Society of Chemistry 2008
J. Mater. Chem., 2008, 18, 381–391 | 389

View Article Online
the charge-separated IP state moves to the MS = ¡1/2 spin
sublevels |D¡1/2T and |Q¡1/2T, leading to DEP. This mechan-
ism via the intramolecular doublet/triplet quantum mixing IP
state followed by the charge-recombination intersystem cross-
ing process will occur in solid, liquid, and gas phases and
compete with the enhanced SO-ISC mechanism derived from
the pendant radical. Unfortunately, direct detection of the IP
state by transient absorption spectroscopy was unsuccessful,
because strong emission from the bodipy component masks the
absorption band of the anthracene radical cation. However,
the unique spin polarization pattern observed by both TRESR
and pulsed ESR spectroscopy gives clear evidence of the IP
2
state A –D +–R as the photo-excited state. This is the first
observation of a photo-excited quartet state generated via the
IP state of an organic molecule with unpaired spins.
?
?
In contrast to the case of 1, the TRESR spectrum of the
photo-excited quartet state of 3 was well simulated by the
enhanced SO-ISC mechanism showing no contribution from
other mechanisms. As mentioned already, 3 have no EnT and
no CT character. This finding supports that the unique DEP
of 1 is generated as the result of the EnT nature of the bodipy
component followed by the PET as discussed above. The DEP
patterns of 1 and 3 are different, because the DEP pattern
depends on the generation pathway of the observed spin states.
In contrast, the magnitudes of the fine-structure parameters of
1 and 3 are similar, because the parameters represent the
distribution of the unpaired electrons of the observed states
being independent of the generation pathway.
Fig. 14 Mechanism of DEP generation in the quartet excited state of
1 through the IP state. J is the magnitude of the effective exchange
coupling between unpaired electrons. The magnitude of the population
transfer from the weakly coupled wavefunction (ZF wavefunction) to
the strongly coupled ones (HF wavefunctions) giving pure quartet
and doublet states depends on the duration times of the charge-
recombination process and the Dg values.
spin selective depopulation reported in a weakly coupled trip-
quartet and trip-doublet system.40 During the charge-recom-
bination process, the separation of the unpaired electrons
decreases. Therefore, these weakly coupled wavefunctions
w2–w5 change to the strongly exchange-coupled pure doublet
(|D¡1/2T) and quartet wavefunctions (|QMsT) through the
charge recombination process, as shown in Fig. 14. As a results
of MS conservation during the first charge-recombination
process, the population of the charge-separated IP state (left
side of Fig. 14) can selectively move to the MS = ¡1/2 spin
sublevels |D¡1/2T and |Q¡1/2T (right side of Fig. 14), which
leads to a non-Boltzmann population (DEP). ISC and the spin
selective population (DEP) of the high-field wavefunctions of
the quartet state are also expected due to D–Q¡1/2 mixing
driven by the difference in g values Dg and in hyperfine
interactions, similar to S–T0 mixing of the triplet/radical pair.36
In a paper30 published at almost same time as our previous
report,15 Wasielewski et al.discussed this D–Q¡1/2 mixing
simply as a mechanism similar to S–T0 mixing in radical pairs.
We pointed out the possibility of the above-mentioned
mechanism, although we did not rule out the contribution of
the D–Q¡1/2 mixing by Dg and hyperfine mechanism as
described in our previous paper.15 In the p-conjugated spin
system, the charge-recombination process occurs first and the
D–Q mixing time may be short. Even in such a situation,
the mechanism proposed here will work well, because only the
generation of the quantum mixed D–Q¡1/2 state is important,
which are selectively populated as the result of the doublet
nature.40 The rapid charge-recombination process conserves
the MS values. As mentioned above, the selective population of
5. Conclusions
We synthesized two photo-excited high spin verdazyl radicals 1
and 3, in which the bodipy acceptor (A) or the anthraquinone
component, respectively, is linked to a phenylanthracene
verdazyl radical (D–R). The efficient intra-molecular energy
transfer (EnT) from the anthracene moiety to the bodipy
functional component was observed by time-resolved fluores-
cence spectroscopy and TRESR measurements. A unique
dynamic electron-spin-polarization (DEP) was detected for the
quartet photo-excited state of 1 by TRESR and pulsed ESR.
Such unique DEP was not detected for either 2 or 3 without a
bodipy component. The spectral simulation of the quartet
spectrum of 1 revealed that the unique DEP was generated by
a competition between the mechanism involving the intra-
molecular ion pair *A2–D+–R and spin–orbit intersystem
crossing. A most plausible model is proposed for the formation
of the uniquely spin-polarized quartet photo-excited state
observed for 1. This is the first detection of a photo-excited
quartet state generated via the IP state of an organic molecule
with unpaired spins. This means that ‘‘photo-synthetic DEP
pattern’’ reported hitherto only in the reaction center of photo-
systems I and II36 and their model systems37 has been first
observed in the quartet high-spin state. The observation of the
new DEP pattern is of great value in the field of the spin
chemistry. In addition, although there is no evidence of the
contribution of the high spin (S ¢ 3/2) photo-excited states in
natural photosynthesis, the findings of this work will open the
possibility of the utilization of the high spin states toward the
efficient artificial photosynthesis.
390 | J. Mater. Chem., 2008, 18, 381–391
This journal is ß The Royal Society of Chemistry 2008

View Article Online
electron spin distribution and redox properties of the verdazyl
radicals is discussed.
Acknowledgements
25 F. Neugebauer and H. Fisher, Angew. Chem., Int. Ed. Engl., 1980,
19, 725.
This work was financially supported by the Grant-in-Aid for
Scientific Research on General Research (B) (No. 13440211)
from the Japan Society of the Promotion of Science (JSPS) and
for Scientific Research on Priority Areas ‘‘Application of
Molecular Spins’’ (Area No. 769, Proposal No.15087208) and
‘‘Molecular Theory for Real Systems’’ (Area No. 461,
Proposal No.19029039) from the Ministry of Education,
Culture, Sports, Science, and Technology (MEXT), Japan.
26 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr.,
R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, P. Salvador, J. J. Dannenberg,
D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman,
J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu,
A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts,
R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. G.
Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez,
M. Head-Gordon, E. S. Replogle and J. A. Pople, GAUSSIAN 98
(Revision A.11), Gaussian, Inc., Pittsburgh, PA, 2001.
27 G. G. Belford, R. L. Belford and J. F. Bukhalter, J. Magn. Reson.,
1973, 11, 251.
References
1 See: Proceedings of the 9th International Conference on Molecule-
based Magnets: Polyhedron, 2005, 24, 2063–2912.
2 O. Sato, J. Tao and Y.-Z. Zhang, Angew. Chem., Int. Ed., 2007, 46,
2152.
3 (a) O. Sato, T. Iyoda, A. Fujishima and K. Hashimoto, Science,
1996, 272, 704; (b) N. Shimamoto, S. Ohkoshi, O. Sato and
K. Hashimoto, Inorg. Chem., 2002, 41, 678.
28 Y. Teki, I. Fujita, T. Takui, T. Kinoshita and K. Itoh, J. Am.
Chem. Soc., 1994, 116, 11499.
4 (a) M. Verdaguer, Science, 1996, 272, 698; (b) A. Bleuzen,
C. Lomenech, V. Escax, F. Villain, F. Varret, C. Cartier
dit Moulin and M. Verdaguer, J. Am. Chem. Soc., 2000, 122, 6648.
5 A. Izuoka, M. Hiraishi, T. Abe, T. Sugawara, K. Sato and
T. Takui, J. Am. Chem. Soc., 2000, 122, 3234.
6 (a) K. Matsuda and M. Irie, J. Am. Chem. Soc., 2001, 123, 9896; (b)
K. Takayama, K. Matsuda and M. Irie, Chem.–Eur. J., 2003, 9,
5605.
29 The details of the spectral simulation are similar to those described
in ref. 8b, with eqn (6) from ref. 8b being modified as follows:
PMs = v1PzMersofield + v2PhMigshfield, v1 = 0.45, and v2 = 0.55, as
zerofield
Ms
described in the text. P is the expectation value of the density
matrix, which represents the populations of the zero-field spin
highfield
Ms
sublevels and P
is that of the high-field spin sublevels
(|S, MST). The doublet ground state was independently simulated
using the g value determined experimentally and superimposed
onto the quartet spectrum.
7 H. Matsuzaki, W. Fujita, K. Awaga and H. Okamoto, Phys. Rev.
Lett., 2003, 91, 017403.
30 (a) E. T. Chemick, Q. Mi, R. F. Kelley, E. A. Weiss, B. A. Jones,
T. J. Marks, M. A. Ratner and M. R. Wasielewski, J. Am. Chem.
Soc., 2006, 128, 4356; (b) Q. Mi, E. T. Chemick, D. W. McCamant,
E. A. Weiss, T. J. Marks, M. A. Ratner and M. R. Wasielewski,
J. Phys. Chem. A, 2006, 110, 7323.
31 Although, in ref. 30b, Wasielewski et al. reported a e/a type
radical-pair like signal without any structures in solution, the spin
state was not confirmed.
8 (a) Y. Teki, S. Miyamoto, K. Iimura, M. Nakatsuji and Y. Miura,
J. Am. Chem. Soc., 2000, 122, 984; (b) Y. Teki, S. Miyamoto,
K. Iimura, M. Nakatsuji and Y. Miura, J. Am. Chem. Soc., 2001,
123, 294.
9 (a) Y. Teki, M. Nakatsuji and Y. Miura, Int. J. Mod. Phys. B,
2001, 15, 4029; (b) Y. Teki, M. Nakatsuji and Y. Miura, Mol.
Phys., 2002, 100, 1385.
10 Y. Teki, Polyhedron, 2001, 20, 1163.
11 Y. Teki, M. Kimura, S. Narimatsu, K. Ohara and K. Mukai, Bull.
Chem. Soc. Jpn., 2004, 77, 95.
12 (a) Y. Teki and S. Nakajima, Chem. Lett., 2004, 33, 1500; (b)
Y. Teki, T. Toichi and S. Nakajima, Chem.–Eur. J., 2006, 12, 2329.
13 Y. Teki, Polyhedron, 2005, 24, 2299.
32 See p. 4668 in ref. 15 and the ESR in ref. 15.
33 (a) J. Isoya, H. Kanda, J. R. Norris, J. Tang and M. K. Bowman,
Phys. Rev. B, 1990, 41, 3905; (b) A. V. Astashkin and A. Schweiger,
Chem. Phys. Lett., 1990, 174, 595.
34 Although we tried the increase of the SCF cycles and the HOMO–
LUMO mixing method to force the convergence, we could not
obtain enough SCF convergency. This may be attributed to the
intrinsic electronic structure of BF2 in the bodipy component and
the large molecular size of 1.
14 (a) Y. Teki and S. Nakajima, Polyhedron, 2005, 24, 2185; (b)
T. Toichi and Y. Teki, Polyhedron, 2005, 24, 2337.
15 Y. Teki, H. Tamekuni, J. Takeuchi and Y. Miura, Angew. Chem.,
Int. Ed., 2006, 45, 4666.
35 A. Bencini and D. Gatteschi, EPR of Exchange Coupled Systems,
Springer-Verlag, Berlin, 1990.
36 F. J. Adrian, J. Chem. Phys., 1974, 54, 3918.
37 H. Levanon and J. R. Norris, Chem. Rev., 1978, 78, 185.
38 (a) K. Hasharoni, H. Levanon, S. R. Greenfield, D. J. Gosztola,
W. A. Svec and M. R. Wasielewski, J. Am. Chem. Soc., 1995, 117,
8055; (b) K. Hasharoni, H. Levanon, S. R. Greenfield, D. J.
Gosztola, W. A. Svec and M. R. Wasielewski, J. Am. Chem. Soc.,
1996, 118, 10228.
16 I. Ciofini, P. P. Laine, M. Zamboni, C. A. Daul, V. Marvaud and
¯
C. Adamo, Chem.–Eur. J., 2007, 13, 5360.
17 (a) P. Huai, Y. Shiomi and S. Abe, Phys. Rev. B, 2005, 72, 094413/1;
(b) P. Huai, Y. Shiomi and S. Abe, Synth. Met., 2003, 137, 1255.
18 (a) C. Corvaja, M. Maggini, M. Prato, G. Scorrano and M. Venzin,
J. Am. Chem. Soc., 1995, 117, 8857; (b) J. Fujiwara, Y. Iwasaki,
Y. Ohba, S. Yamauchi, N. Koga, S. Karasawa, M. Fuhs,
K. Mo¨bius and S. Weber, Appl. Magn. Reson., 2001, 21, 483 and
references therein.
39 From the absorption spectrum shown in Fig. 2, the magnitude
of the interaction between D and A is estimated to be about
2300 cm21 as shown in the text. Such a large interaction cannot be
a through-space interaction. Therefore, it is clear that PET is a
through-bond effect in this case.
40 M. Asano, A. van Est, U. Kru˝ger, D. Stehlik, Y. Kaizu and
H. Levanon, J. Phys. Chem. A, 1999, 103, 6704. In this paper, they
discussed the selective depopulation arising from the doublet
nature in a weakly coupled trip–quartet and trip–doublet system
toward the ground state. This discusstion can be reasonably
applied to the selective population in the quantum mixed states
by the ‘‘inter-system crossing’’ from the higher energy excited
doublet state.
19 (a) K. Ishii, J. Fujiwara, Y. Ohba and S. Yamauchi, J. Am. Chem.
Soc., 1996, 118, 13079; (b) K. Ishii, Y. Hirose and N. Kobayashi,
J. Phys. Chem. A, 1999, 103, 1986.
20 bodipy: 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene.
21 C. W. Wan, A. Burghart, J. Chen, F. Bergestroem, L. Johansson,
M. F. Wolford, T. G. Kim, M. R. Topp, R. M. Hochstrasser and
K. Burgess, Chem.–Eur. J., 2003, 9, 4430.
22 R. A. Abou-Elkhair and T. L. Netzel, Nucleosides, Nucleotides
Nucleic Acid, 2005, 24, 85.
23 C. L. Barr, P. A. Chase, R. G. Hicks, M. T. Lemaire and
C. L. Stevens, J. Org. Chem., 1999, 64, 8893.
24 J. B. Gilroy, S. D. J. McKinnon, B. D. Koivisto and R. G. Hicks,
Org. Lett., 2007, 9, 4838. In this paper, a correlation between the
This journal is ß The Royal Society of Chemistry 2008
J. Mater. Chem., 2008, 18, 381–391 | 391