Twisted molecular geometry and localized electronic structure of the triplet excited gem-diphenyltrimethylenemethane biradical: Substituent effects on thermoluminescence and related theoretical calculations

Article abstract of DOI:10.1016/j.tet.2011.06.052

Substituent effects on the energies of electronic transitions (ETs) between the triplet excited and ground states of gem-diphenyltrimethylenemethane biradicals (³2a) were explored by using thermoluminescence (TL) spectroscopy and density functi

Full text of DOI:10.1016/j.tet.2011.06.052

Tetrahedron 67 (2011) 7431e7439
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Twisted molecular geometry and localized electronic structure of the triplet
excited gem-diphenyltrimethylenemethane biradical: substituent effects on
thermoluminescence and related theoretical calculations
,
c
a,d
Yasunori Matsui a, Hayato Namai b, Ikuko Akimoto c , Ken-ichi Kan’no , Kazuhiko Mizuno
,
*
,d,
Hiroshi Ikeda a
*
a Department of Applied Chemistry, Graduate School of Engineering, Osaka Prefecture University, Sakai, Osaka 599-8531, Japan
b Department of Chemistry, Graduate School of Science, Tohoku University, Sendai 980-8578, Japan
c Department of Material Science and Chemistry, Faculty of Systems Engineering, Wakayama University, Wakayama 640-8510, Japan
d Research Institute for Molecular Electronic Devices (RIMED), Osaka Prefecture University, Sakai, Osaka 599-8531, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Substituent effects on the energies of electronic transitions (ETs) between the triplet excited and ground
states of gem-diphenyltrimethylenemethane biradicals (32aꢀꢀ) were explored by using thermolumines-
cence (TL) spectroscopy and density functional theory (DFT) including time-dependent (TD) DFT. Linear
Received 30 April 2011
Received in revised form 14 June 2011
Accepted 18 June 2011
3
free energy (Hammett) analyses of TL energies of a variety of para-substituted aryl derivatives of 2ꢀꢀ*
Available online 7 July 2011
gave reasonable correlations with the substituent constant, s
ꢀ. The slope of Hammett plots of the data are
nearly identical to one obtained from a similar analysis of the photoluminescence (PL) energies of the
This paper is dedicated to the memory of
Emeritus Professor Yuho Tsuno (Kyushu
University)
3
structurally-related 1,1-diarylethyl radicals (3ꢀ*). The results suggest that TL of 2ꢀꢀ* and PL of 3ꢀ* derive
from a common diarylmethyl radical fluorophore. This interpretation is also supported by the DFT and
3
TDDFT calculated electronic structures and ET energies of 2ꢀꢀ and 3ꢀ. Thermodynamic and kinetic
þ
analyses of the charge recombination (CR) process between 2ꢀ and 1ꢀꢁ, which generates 32ꢀꢀ*, revealed
Keywords:
that substituents not only alter the TL energies but also the TL intensities of 32ꢀꢀ*. The observations made
in this effort demonstrate that 32ꢀꢀ* as well as 32ꢀꢀ and 2ꢀþ have greatly twisted molecular geometries and
highly localized electronic structures.
Hammett equation
Electronic structure
Molecular geometry
Electron transfer
Charge recombination
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
Substituent effects on short-lived organic radicals3 also have
been studied both theoretically4 and experimentally5,6 to analyze
Linear free energy relationships, exemplified by the Hammett
equation,1 are perhaps the most fundamentally important tools in
physical organic chemistry. Although the Hammett equation was
originally based on thermodynamic relationships associated with
the acidities of para-substituted benzoic acids, and applied to ki-
netic analyses of reactions, it is also used to probe features of
electronic transitions (ETs) of interesting molecules or in-
termediates. Studies of substituent effects on ETs associated with
light absorption and emission can provide information about
electronic structures, molecular geometries, reactivities, and other
properties of the excited states of molecules.2 Hammett analyses
that utilize substances containing electron-donating (EDGs) and/or
electron-withdrawing groups (EWGs) is a sophisticated strategy for
this purpose.
reaction mechanisms, especially in the context of materials
chemistry.7 As compared with closed-shell molecules, organic
radicals have some advantages, such as ideal reversible redox
properties and low energy (i.e., longer wavelength) ETs caused by
the presence of characteristic SOMO energy levels that exist be-
tween HOMOs and LUMOs.
Thus, organic radicals hold great potential in the fields of or-
ganic electronic devices, such as photovoltaics,8 batteries,9 and
light-emitting diodes (OLEDs).10 Unfortunately, few studies aimed
at exploring the substituent effects on ET energies of organic rad-
icals have been carried out.
Previously, we described the observation of an intense green
(
lTL¼501 nm) thermoluminescence (TL) during annealing of
g-
10a or
X-irradiated10c methylcyclohexane (MCH) matrices containing 2,2-
diphenyl-1-methylenecyclopropane (1a, Scheme 1).10 The TL was
assigned to fluorescence associated with decay of the triplet excited
state of the gem-diphenyltrimethylenemethane biradical (32aꢀꢀ*) to
* Corresponding authors. E-mail addresses: akimoto@sys.wakayama-u.ac.jp (I. Aki-
moto), ikeda@chem.osakafu-u.ac.jp (H. Ikeda).
0040-4020/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2011.06.052

7432
Y. Matsui et al. / Tetrahedron 67 (2011) 7431e7439
as follows: colorless needles from ethanol (mp 49e50 ꢂC); 1H NMR
(300 MHz, CDCl3) dppm 1.99 (2H, dd, J¼2.7, 2.0 Hz), 5.69 (1H, td,
J¼2.3, 2.0 Hz), 5.84 (1H, td, J¼2.7, 2.3 Hz), 7.36 (4H, AA0BB0,
J¼8.2 Hz), 7.55 (4H, AA0BB0, J¼8.2 Hz); 13C NMR (75 MHz, CDCl3)
dppm 22.3, 32.8, 105.5, 122.3, 125.3 (q, 2C), 128.2 (4C), 128.4 (4C),
137.6 (2C), 146.6 (2C); IR (KBr) 1616, 1325, 1163, 1117, 1069, 1013,
901, 858 cmꢁ1; Anal. Calcd for C18H12F6: C 63.16, H 3.53, F 33.30,
found: C 63.45, H 3.78, F 33.56.
3
2.3. TL of 2ꢀꢀ* induced by using the
method
g-irradiationeannealing
A solution of MCH (1 mL) containing 1 (5 mM) in a flat vessel
[suprasil, 2ꢃ10ꢃ40 mm] was degassed by using five freeze (77 K)e
pump (0.1 mmHg)ethaw (room temperature) cycles. After sealing
the tube, a glassy MCH matrix was prepared by immersion in
a Dewar vessel filled with liquid nitrogen. g-Irradiation was carried
out with 4.0 TBq 60Co source in liquid nitrogen at 77 K for 40 h at
Tohoku University. The matrix was rapidly annealed at room tem-
perature. TL spectra during the annealing process were recorded by
using the photonic multichannel spectral analyzer (PMA-11,
Hamamatsu Photonics).
Scheme 1. Mechanism for TL emission from 32ꢀꢀ* (photo: 32aꢀꢀ*): (i) charge separation
induced by g
- or X-irradiation, (ii) bond fission, (iii) CR between 2ꢀþ and 1ꢀꢁ on an-
nealing, (iv) deactivation accompanied by TL (fluorescence), (v) regeneration of 1 by
bond formation.
3
2.4. TL of 2ꢀꢀ* induced by using the X-irradiationeannealing
its triplet ground state. In this report, we proposed a new OLED
3
method12
strategy utilizing 2ꢀꢀ* as an emitter, which we named Organic
Radical Light-Emitting Diode (ORLED), and suggested that this
approach has great potential for overcoming significant problems
associated with normal fluorescent emitters used in typical
A solution of MCH (0.1 mL) containing 1 (20 mM) in a quartz ESR
tube (f¼5 mm) was degassed by using five freeze (77 K)epump
3
(0.1 mmHg)ethaw (room temperature) cycles. After sealing the
tube, a glassy MCH matrix was prepared by immersion in a quartz
Dewar vessel filled with liquid nitrogen. X-Irradiation (40 kV, 1 mA)
was carried out by using equipment consisting of an X-ray tube
(Oxford, Series 5000/75, W target, lMAX¼0.05 nm) and power
supply (SPELLMAN, XLG). The matrix was annealed slowly in the
same Dewar vessel. TL spectra during annealing process were
recorded at 2 s intervals with an exposure time of 2 s by using a CCD
detector (Roper Scientific, Spec-10: 256TE-A) and spectrofluorom-
eter (Acton Research Corp., Spectrapro-150i). Temperature moni-
toring of the matrix was carried out with a Au:FeeChromel
thermocouple.
OLEDs.10a TL and electroluminescence (þEL) of 2ꢀꢀ* are induced by
charge recombination (CR) between 2ꢀ and a radical anion of 1
(1ꢀꢁ) and an electronþ(eꢁ), respectively. UVevis absorption spec-
troscopic studies of 2ꢀ and structurally-related 1,1-diarylethyl cat-
þ
ions (3þ, Scheme 1) by Ikeda and co-workers11 revealed that 2ꢀ
have a unique, largely twisted molecular geometry and separated
electronic structure containing isolated radical and cation centers.
To develop ORLEDs using the 1a/32aꢀꢀ* system, studies of sub-
stituent effects on the ETs of 32ꢀꢀ and 32ꢀꢀ* are important so that their
molecular geometries and electronic structures can be clarified. For
this purpose, we have carried out an investigation in which sub-
3
stituent effects on TL energy of 2ꢀꢀ* were examined by employing
derivatives 32befꢀꢀ* that bear CH3, CH3O, F, Cl, and CF3 groups on the
C4 positions of two benzene rings. The photoluminescence (PL)
energies of the structurally-related 3aefꢀ* were also investigated.
2.5. PL of 3ꢀ*
3
The 1,1-diarylethyl radical derivatives 3aefꢀ were generated by
-irradiation (40 h, 4.0 TBq) of the degassed glassy MCH matrices
containing the corresponding 1,1-diarylethanol (4aef, Scheme 2) in
the quartz vessels at 77 K. PL spectra of 3aefꢀ* were then recorded
by using PMA-11 on photoexcitation (lEXw340 nm) at 77 K.
Finally, DFT and TDDFT calculations on 2aefꢀꢀ and 3aefꢀ were
g
performed. Below, we present the results of this effort, which show
3
that 2ꢀꢀ* possess twisted molecular geometries and separated
electronic structures and provide the basis forþthermoꢁdynamic and
kinetic analyses of the CR process between 2ꢀ and 1ꢀ
2. Experimental section
.
2.1. General methods
All melting points were obtained with a Yanako (MP-500) ap-
paratus and are reported uncorrected. Satisfactory elemental
analyses were obtained for a new compound, 1f. 1H NMR spectra
were recorded at 300 MHz 13C NMR spectra were obtained at
75 MHz. MCH (spectroscopic grade) was dried over molecular
Scheme 2. Generation of the 1,1-diarylethyl radicals (3ꢀ).
ꢀ
sieves 4 A. Other solvents used for syntheses were purified and
dried by usual methods.
2.6. Quantum chemical calculations
Geometry optimizations of 32ꢀꢀ and 3ꢀ were performed with the
cc-pVDZ basis set,13 using UB3LYP method.14,15 ETs were computed
using the TD-UB3LYP method16 with cc-pVDZ basis set. Spin con-
taminations are negligible. All the calculations were performed on
Gaussian 98 program.18
2.2. Preparation of substrates
Physical data of 1aee are reported in Ref. 11a, while those of 2,2-
bis(40-trifluoromethylphenyl)-1-methylenecyclopropane (1f) are

Y. Matsui et al. / Tetrahedron 67 (2011) 7431e7439
7433
In this effort, ab initio MO calculations, such as those using the
second-order MøllerePlesset (MP2) method, were not carried out
because overestimation of spin densities and spin contaminations
often occur in case of open-shell species. In contrast, DFT methods
often give more satisfactory results for these species. Stephens and
co-workers reported that compared with the self-consistent field
or MP, DFT/B3LYP accurately predicts vibrational absorption and
circular dichromism spectra.19 Our experience also shows that the
DFT method is sufficiently accurate to give reliable molecular ge-
TLspectra obtained under
g
-irradiation conditionswere relatively
broad, as compared with those under X-irradiation conditions.
Annealing of the -irradiated matrix in a flat vessel results in altered
g
shape of emission band at ca. 540 nm (Fig. 1a) that results from the
existence of a range of conformers of 32aꢀꢀ* owing to the wide tem-
perature distribution in the relativelylarge flat vessel andthe fast rate
of annealing. In contrast, when X-irradiation conditions are used,
annealing does not result in a change of the shape of TL bands since
3
only a few conformers of 2aꢀꢀ* owing to the narrow temperature
þ
ometries and electronic structures of 2ꢀ (UB3LYP/cc-pVDZ)11 and
distribution in the relatively small ESR tube and the slow rate of
annealing. Unfortunately, however, a direct comparison is not pos-
the 1,4-diarylcyclohexane-1,4-diyl radical cations (UB3LYP/cc-
pVDZ),20 as well as other species.
sible due to the characteristics of radiation sources that a
diates from thesource(60Co) in a radial fashion, while a sourceof XRD
equipment radiates X-ray in a small range (
¼5 mm). Thus, the flat
vessel and ESR tube are suitable for - and X-irradiation, respectively.
Although the - and X-rays have respective high (2.8 MeV) and
g-ray ra-
f
3. Results and discussion
g
g
3.1. Time-dependent TL spectra
low (ca. 1 keV) photon energies, the same phenomenon (i.e., ioni-
zation of substrates followed by TL emission) occurs. Note that
photon energies or ionizing power to finally produce 2aꢀꢀ* does
not affect on TL wavelengths. Consequently, the differences in the
shapes of bands in the TL spectra coming from annealing g- and X-
A
g-irradiated MCH matrix containing 1a in a flat vessel
3
(2ꢃ10ꢃ40 mm) was annealed at a rate of ca. 70 K minꢁ1 until the
melting point of MCH (147 K21) was reached (within ca. 50 s).
Emission spectra associated with TL, observed during the course of
the annealing process, are shown in Fig. 1a. TL, which began at the
early stage of annealing (tA¼30 s) and was complete by tA¼80 s,
displayed bands with maxima at lTL¼501 and 533 nm. The bands
are assigned to TeT fluorescence of 32aꢀꢀ*.10 TL emission spectra of
irradiated samples appear to be caused by differences in the shapes
of the vessels and the annealing environments.
3
3.2. Hammett analysis of TL energies of 2ꢀꢀ*
an X-irradiated MCH matrix containing 1a in an ESR tube (
f
¼5 mm)
Annealing of g-irradiated glassy MCH matrices containing 1aef
during annealing are displayed in Fig. 1b. TL in this case, where the
matrix was slowly annealed (ca. 20 K minꢁ1) in the Dewar vessel,
began relatively slowly (tA¼90 s) and finished at tA¼170 s, as
from 77 K to ca. 147 K give rise to emission spectra with TL bands at
lTL
(g
)¼501, 515, 529,22 503, 525, and 515 nm, respectively [Fig. 2aef
(green), Table 1]. Similarly, annealing of X-irradiated glassy MCH
matrices containing these substances from 77 K to ca. 147 K yields
spectra with TL bands at lTL(X)¼501, 514, 503, 521, and 515 nm,
respectively [Fig. 2aef (blue), Table 1]. Under the X-irradiation
conditions, the halogenated derivatives 1d and 1e were found to
exhibit very weak TL [Fig. 2d and e (blue)]. Moreover, no TL was
observed for an X-irradiated matrix containing 1c. Therefore,
compared to when the
g-irradiation method was employed.
Hammett analyses (see below) were performed by using TL ener-
3
gies of 2ꢀꢀ* obtained employing the
g
-irradiation method.
Although a variety of substituent constants (
s
ꢀ) have been de-
fined for radical reactions, sꢀ for diarylmethyl radical moieties or
excited states of radicals have not yet been proposed. As a result, we
have used the three representative sꢀ parameters that are in-
dependently defined by Cheng,4 Arnold,5 and Creary6 for the
monoaryl-substituted radicals or related compounds (Table 1).
Cheng derived sꢀ (Cheng) constants based on DFT-calculated spin
densities (rX) of the benzylic carbon of para-substituted benzyl
radicals (5ꢀ, Scheme 3a).4 By using ESR hyperfine coupling con-
stants (aX) of the two benzylic hydrogen atoms of 5ꢀ, Arnold derived
sꢀ (Arnold) (Scheme 3b).5 Finally, Creary derived sꢀ (Creary) by
using rate constants (kX) of the methylenecyclopropane rear-
rangement of 2-aryl-3,3-dimethyl-1-methylenecyclopropanes (6,
Scheme 3c) that take place via transition states 7z and trimethy-
lenemethane (TMM)-type intermediates 8ꢀꢀ.6
From analyses of TL spectra, recorded upon annealing following
3
g
-irradiation, the relative emission energies of 2befꢀꢀ* were de-
termined based on
D
ETL(32ꢀꢀ*)¼ETL(32befꢀꢀ*)ꢁETL(32aꢀꢀ*). These
values display a reasonable correlation [correlation coefficient
R2¼0.93, see Fig. 3a (solid line) and Eq. 1a] with sꢀ (Cheng). As
described below, the PL energy data of the CH3O-substituted de-
rivative of the structurally-related 1,1-diarylethyl radicals (3cꢀ*,
Scheme 1) deviated from the fitted line (Fig. 3d). However, the
3
energies of the corresponding species 2cꢀꢀ* [
D
ETL(32cꢀꢀ*)] did not.
Even when
tion between
change [see Fig. 3a (dotted line) and Eq. 1b]. On the other hand, the
fits of
ETL(32ꢀꢀ*) using the other substituent constants,
ꢀ (Arnold)
D
ETL(32cꢀꢀ*) is excluded from the analysis, the correla-
ETL(32ꢀꢀ*) and sꢀ (Cheng) does not significantly
D
3
Fig. 1. Time-dependent changes of TL spectra of 2aꢀꢀ* in (a)
irradiated MCH matrices obtained every 10 s following tA¼0.
g
-irradiated and (b) X-
D
s

7434
Y. Matsui et al. / Tetrahedron 67 (2011) 7431e7439
Table 1
3
Substituent constants s
ꢀ, TL, and PL wavelengths of 2ꢀꢀ* and 3ꢀ*, and calculated ET
wavelengths of 32ꢀꢀ and 3ꢀ
X
sꢀ
sꢀ
sꢀ
32ꢀꢀ
*
32ꢀꢀ 3ꢀ*
3ꢀ
(Cheng)a (Arnold)a (Creary)a
b
c
d
c
lTL
(
g
)
lTL (X)b lET
lPL
lET
nm
nm
nm nm nm
H
CH3
0.00
þ0.11
0.000
þ0.002
þ0.018
ꢁ0.011
þ0.011
ꢁ0.009
0.00
þ0.16
þ0.27
ꢁ0.06
þ0.11
þ0.05
501
515
529
503
525
515
501
514
de
412 522 429
423 545 444
435 551 459
413 536 432
431 549 449
429 538 446
CH3O þ0.22
F
þ0.03
þ0.14
þ0.08
503
521
515
Cl
CF3
a
sꢀ (Cheng), sꢀ (Arnold), and sꢀ (Creary) values are quoted from Refs. 4e6,
respectively.
b
c
Recorded during annealing of MCH matrices.
Calculated at the TD-UB3LYP/cc-pVDZ level.
Recorded in MCH matrices at 77 K, each excitation wavelength is 340ꢅ10 nm.
Not obtained.
d
e
(a)
ρ
σ
ρ
(b)
α
σ
α
(c)
σ
Scheme 3. Species and reactions for definitions of the substituent constants by (a) sꢀ
(Cheng) defined with the spin densities (rX) at benzylic carbon of 5ꢀ calculated by DFT,
(b) s
ꢀ (Arnold) defined with ESR hyperfine coupling constants (aX) of the two benzylic
hydrogen atoms of 5ꢀ, (c) sꢀ (Creary) defined with the rate constants (kX) of methyl-
enecyclopropane rearrangements of 6.
and sꢀ (Creary), are poor (R2¼0.60 and 0.75, see Fig. 3b and c and
Eqs. 2 and 3).
3
D
D
D
D
ETLð 2ꢀꢀꢄÞ ¼ ꢁ14:8sꢀðChengÞ ꢁ 0:06; R2 ¼ 0:93
(1a)
(1b)
(2)
3
ETLð 2ꢀꢀꢄÞ ¼ ꢁ14:4sꢀðChengÞ ꢁ 0:17; R2 ¼ 0:93
Fig. 2. (aef) TL spectra of 32aefꢀꢀ* induced by
g-(green, 40 h at 77 K) and X-irradiation
3
ETLð 2ꢀꢀꢄÞ ¼ ꢁ84:4sꢀðArnoldÞ ꢁ 1:33; R2 ¼ 0:60
(blue, 1 h at 77 K) of MCH matrices containing 1aef.22 (gel) PL spectra of 3aefꢀ* in
MCH matrices at 77 K, each excitation wavelengths is lEX¼340ꢅ10 nm.
3
ETLð 2ꢀꢀꢄÞ ¼ ꢁ8:91sꢀðCrearyÞ ꢁ 0:70; R2 ¼ 0:75
(3)

Y. Matsui et al. / Tetrahedron 67 (2011) 7431e7439
7435
2
ꢀ
D
D
ð
3
ꢄÞ
sꢀ
ð
Þ
EPL
¼ ꢁ55:0 Arnold ꢁ 1:29; R ¼ 0:47
(5)
(6)
EPLð3ꢀ Þ ¼ ꢁ6:11sꢀðCrearyÞ ꢁ 0:85; R2 ¼ 0:67
ꢄ
3
The relatively large ratio of
r
values for 2ꢀꢀ* and 3ꢀ* [
r
(32ꢀꢀ*)/
r
(3ꢀ*)¼1.38] suggests that differences exist in the substituent effects
on the TL and PL energies of 32ꢀꢀ* and 3ꢀ*, respectively. However, the
PL energy of 3cꢀ* deviates significantly from the line obtained from
3
other derivatives [see Fig. 3d and Eq. 4b]. If the data for 2cꢀꢀ* and
3cꢀ* are excluded from the analysis, the ratio of
r values is close to
unity [
r
(32ꢀꢀ*)/
r
(3ꢀ*)¼0.97], demonstrating that almost the same
effects of substituents are involved in the TL and PL energies of 32ꢀꢀ*
and 3ꢀ*, respectively. In other words, little interaction occurs be-
3
tween the diarylmethyl and the allyl radical moieties of 2ꢀꢀ*. This
result is understandable from the standpoint of the dihedral angle
3
between the diarylmethyl and the allyl moieties of 2ꢀꢀ, which is
caþlculated to be smaller than that of the corresponding radical cation
2ꢀ (see below). Note that Hammett analysisþon 2ꢀþ and 3þ give the
ratio
r
(2ꢀþ)/
r
(3ꢀþ)¼0.9,11a suggesting that 2ꢀ has a largely-twisted
molecular geometry and strongly-separated electronic structure.
In a previous effort probing substituent effects on the TL energies
of the singlet excited state of the 1,4-diphenylcyclohexane-1,4-diyl
(110ꢀꢀ*, Chart 1),23 we observed that
D
ETL(110ꢀꢀ*) displays a reason-
able correlation with sꢀ (Arnold) but not with sꢀ (Cheng) and sꢀ
3
1
(Creary). The excited biradicals 2ꢀꢀ* and 10ꢀꢀ* share a common
benzyl radical, yet their ET energies are well correlated with differ-
ent constants, sꢀ (Cheng) and
s (Arnold), respectively. The multi-
plicities of biradicals (triplet or singlet), nature of orbital interaction
between the two radical moieties,24 and/or charge-transfer charac-
ter of the excited states may be the cause(s) of these differences.
3
Fig. 3. Hammett plots of TL energies of 2aefꢀꢀ
*
[
D
ETL(32ꢀꢀ*)] obtained under
ꢀ (Cheng), (b) ꢀ (Arnold),
g
-irradiationeannealing conditions (green dots) against (a)
s
s
and (c) sꢀ (Creary) and of PL energies of 3aefꢀ* [
D
EPL(3ꢀ*)] (green circles) in MCH
matrices at 77 K against (d) sꢀ (Cheng), (e) sꢀ (Arnold), and (f) sꢀ (Creary). Solid fitted
3
lines are obtained by inclusion of data for 2aefꢀꢀ* and 3aefꢀ* while dotted lines are
3
obtained by exclusion of data for 2cꢀꢀ* and 3cꢀ*.
It is interesting that D s
ETL(32ꢀꢀ*) correlates with ꢀ (Cheng) but not
Chart 1. The singlet excited state of the 1,4-diphenylcyclohexane-1,4-diyl (110ꢀꢀ*).
with sꢀ (Arnold) even though these substituent constants originate
from analysis of the same benzylic radical 5ꢀ. A possible reason why
the TL energies of 32ꢀꢀ* do not correlate with ꢀ (Creary) could lie in
s
the fact that s
ꢀ (Creary) comes from a kinetic (or dynamic) analysis,
3.3. DFT calculations of the molecular geometries and
3
electronic structures of 2ꢀꢀ
which pertains to the transition states 7z and not biradical in-
termediates 8ꢀꢀ involved in the rate determining step of the meth-
ylenecyclopropane rearrangement of 6 to give 9 (Scheme 3c).
PL spectra of 3aefꢀ* in MCH matrices at 77 K are shown in
Fig. 2gel. The radicals 3aꢀ*, 3bꢀ*, 3cꢀ*, 3dꢀ*, 3eꢀ*, and 3fꢀ* exhibited
PL bands at lPL¼522, 545, 551, 536, 549, and 538 nm, respectively
(Table 1). The relative PL energies of 3befꢀ* based on 3aꢀ* display
a reasonable but not good correlation (R2¼0.80) with sꢀ (Cheng)
[see Fig. 3d (solid line) and Eq. 4a]. A decrease in the correlation
coefficients R2 occurs when the data for the CH3O-substituted de-
rivative 3cꢀ* is included (Fig. 3d). When the data for 3cꢀ* is excluded
from the analysis, a satisfactory correlation coefficients (R2¼0.97)
Quantum chemical studies of the molecular geometries and
3
electronic structures of 2ꢀꢀ* are essential to develop an un-
derstanding of TL. In general, however, it is difficult to evaluate
accurately the properties of excited states of open-shell species,
such as 32ꢀꢀ*. Thus, DFT and TDDFT calculations were performed on
3
32aefꢀꢀ and 3aefꢀ to gain information about the TL of 2ꢀꢀ*.
3
The optimized molecular geometry of the parent 2aꢀꢀ is dis-
played in Fig. 4. The bond lengths, dihedral angles, spin (r) and
3
charge (q) densities of 2aefꢀꢀ, calculated with UB3LYP/cc-pVDZ,
are listed in Table 2. All C1eC2 bond lengths (l, Table 2) of
for
D
EPL(3ꢀ*) against
s
ꢀ (Cheng) is obtained [see Fig. 3d (dotted line)
ETL(32ꢀꢀ*), plots of EPL(3ꢀ*)
and Eq. 4b]. Similar to the analysis of
D
D
versus sꢀ (Arnold) and sꢀ (Creary) are scattered and poor correla-
tions (R2¼0.47 and 0.67) are observed (see Fig. 3e and f and Eqs. 5
and 6). These results showed that PL characteristics of 3aefꢀ* are
3
similar to TLs of 2aefꢀꢀ*.
ꢄ
D
EPLð3ꢀ Þ ¼ ꢁ10:5sꢀðChengÞ ꢁ 0:37; R2 ¼ 0:80
(4a)
(4b)
3
Fig. 4. (a) Front and (b) side views of optimized molecular geometry of 2aꢀꢀ at
ꢄ
D
EPLð3ꢀ Þ ¼ ꢁ14:9sꢀðChengÞ ꢁ 0:10; R2 ¼ 0:97
UB3LYP/cc-pVDZ level.10c,25

7436
Y. Matsui et al. / Tetrahedron 67 (2011) 7431e7439
Table 2
Bond lengths (l), dihedral angles (
q,
u), and spins (S
r
) and charge distributions (Sq) in the diarylmethyl (subunit I) and allyl (subunit V) moieties of 32aefꢀꢀ (UB3LYP/cc-pVDZ)26
ꢀ
Biradicals
X
Subunit I
Subunit II
l A
q
deg
u deg
S
r
Sq
S
r
Sq
32aꢀꢀ
32bꢀꢀ
32cꢀꢀ
32dꢀꢀ
32eꢀꢀ
32fꢀꢀ
H
þ0.892
þ0.895
þ0.891
þ0.892
þ0.899
þ0.901
þ0.018
þ0.028
þ0.048
þ0.019
þ0.019
ꢁ0.005
þ1.108
þ1.105
þ1.109
þ1.108
þ1.101
þ1.099
ꢁ0.018
ꢁ0.028
ꢁ0.048
ꢁ0.019
ꢁ0.019
þ0.005
1.468
1.469
1.468
1.468
1.470
1.471
32.2
33.0
32.3
32.5
33.7
34.0
36.0
35.2
35.5
35.9
35.0
34.8
CH3
CH3O
F
Cl
CF3
3
ꢀ
2aefꢀꢀ were found to be in the range of 1.468e1.471 A. Dihedral
angles between the diarylmethyl and allyl moieties
(
(
q
¼:C5eC1eC2eC3) and between the TMM and phenyl moieties
u
¼:C1eC2eC3eC4) of 32aefꢀꢀ are in the range of
q
¼32.2e34.0ꢂ
and
u
¼34.8e35.9ꢂ, respectively. Thus, no clear correlation exists
between geometries (l, q, u) and substituents. On the other hand,
the CH3- and CH3O-derivatives 2bꢀꢀ and 2cꢀꢀ possess greater
3
3
positive charge densities in the diarylmethyl moieties (subunit I,
3
3
3
Sq¼þ0.028 for 2bꢀꢀ, þ0.048 for 2cꢀꢀ) as compared with 2aꢀꢀ
3
(Sq¼þ0.018). Conversely, the CF3-derivative 2fꢀꢀ has negative
3
charge density (Sq¼ꢁ0.005 for 2fꢀꢀ) in the subunit I in contrast to
3
Fig. 5. Hammett plots of the calculated ETs of (a) 2aefꢀꢀ and (b) 3aefꢀ calculated at
32aꢀꢀ. Interestingly, as compared to 32aꢀꢀ, the F- and Cl-derivatives 32dꢀꢀ
and 32eꢀꢀ have slightly greater positive charges (Sq¼þ0.019 for 32dꢀꢀ
the TD-UB3LYP/cc-pVDZ level versus sꢀ (Cheng). Solid fitted lines are obtained by in-
3
clusion of data for 2aefꢀꢀ* and 3aefꢀ* while dotted lines are obtained by exclusion of
3
data for 2cꢀꢀ* and 3cꢀ*.
3
and 2eꢀꢀ) in the subunit I, suggesting that the halogens participate
as weak EDGs. Therefore, the observed changes seen in the charge
distributions in the subunit I are reasonable from the standpoint of
the electron-donating and -withdrawing abilities of the substituents.
In contrast to the observations summarized above, the spin
even when the ET of 3cꢀ is excluded from the analysis (R2¼0.92)
[see Fig. 5b (solid and dotted lines) and Eqs. 8a and 8b].
3
densities in subunit I of 2aefꢀꢀ fall in the narrow range of
3
D
D
D
D
EETð 2ꢀꢀÞ ¼ ꢁ17:3sꢀðChengÞ ꢁ 0:22; R2 ¼ 0:80
(7a)
(7b)
(8a)
S
r
¼0.891e0.901, showing that no remarkable substituent effects
take place. In addition, these values show that one of the spins of
biradical is localized in subunit I and the other is localized in the
allyl moieties (subunit II). Thus, the electronic structures contain
isolated subunits I and II. Consequently, introduction of EDGs and
EWGs into the phenyl groups of 32aꢀꢀ does not significantly change
spin densities and, therefore, the major electronic properties of
3
EETð 2ꢀꢀÞ ¼ ꢁ22:2sꢀðChengÞ þ 0:06; R2 ¼ 0:80
EETð3ꢀÞ ¼ ꢁ19:6sꢀðChengÞ ꢁ 0:19; R2 ¼ 0:95
EETð3ꢀÞ ¼ ꢁ22:6sꢀðChengÞ ꢁ 0:06; R2 ¼ 0:92
(8b)
32aꢀꢀ are retained in 2befꢀꢀ.
3
Analyses of the ratio of the slopes of Eqs. 7a and 8a [
r
(32ꢀꢀ)/
r
(3ꢀ)¼0.87] suggest that substituents effect the ETs of 32ꢀꢀ and 3ꢀ to
nearly the same extent. This conclusion is strongly supported by the
ratio
r
(32ꢀ)/
r
(3ꢀ)¼0.98, obtained from Eqs. 7b and 8b.
3.4. Hammett plots of TDDFT calculated ET energies
3
The ETs of 2aefꢀꢀ, calculated with TD-UB3LYP/cc-pVDZ, are
3
3
3
shown in Table 1. The long wavelengths ETs of 2aꢀꢀ, 2bꢀꢀ, 2cꢀꢀ,
3.5. Rate constants of CR processes evaluated by using Miller’s
equation
32dꢀꢀ, 2eꢀꢀ, and 2fꢀꢀ at lET¼412, 423, 435, 413, 431, and 429 nm,
respectively, correspond to transitions of triplet ground states to
lowest triplet excited states (32ꢀꢀ/32ꢀꢀ*). Thus, the transitions are
associated with the TeT fluorescence (32ꢀꢀ*/32ꢀꢀ). Similarly, the
wavelengths for the ETs of 3aꢀ, 3bꢀ, 3cꢀ, 3dꢀ, 3eꢀ, and 3fꢀ at lET¼429,
444, 459, 432, 449, and 446 nm, respectively, corresponds to
transitions of doublet ground states to doublet lowest excited states
(3ꢀ/3ꢀ*) and associated with fluorescence (3ꢀ*/3ꢀ). As previously
3
3
As described above, differences exist in the TL intensities of
32aefꢀꢀ*,26 which should be proportional to the product of the
generation rate constant and the fluorescence quantum yields (FF
)
for 32aefꢀꢀ*. The FF values for 32aefꢀꢀ* cannot be determined since
temperature is continuously changing through out the entire
annealing process in the TL experiments.27 Therefore, the experi-
mental results can only provide indirect information about thþe rate
3
reported, ETs for 2aꢀꢀ and 3aꢀ at 412 and 429 nm, respectively,
originate from SOMOeLUMO transitions.10c
constants (kCR) for generation of 2aefꢀꢀ* via CR between 2ꢀ and
3
3
D
EET values for 2ꢀꢀ were observed to display a reasonable cor-
1ꢀꢁ, which can be discussed in the context of the schematic energy
relation (R2¼0.80) with sꢀ (Cheng) even when the ET of 2cꢀꢀ is
excluded from the analysis (R2¼0.80) [see Fig. 5a (solid and dotted
lines) and Eqs. 7a and b]. In a similar manner, the calculated
diagram for the TL (Fig. 6). The TL mechanism proceeds through
3
þ
seven different states (ievii) and CR between 2ꢀ and 1ꢀꢁ can occur
via four possible exergonic paths (A, B, C, and D in Fig. 6). Only path
D
EET(3ꢀ) values satisfactory correlate (R2¼0.95) with sꢀ (Cheng)
B, which gives rise to state (iv) 32ꢀꢀ*þ1, is involved in the TL of 32ꢀꢀ*.

Y. Matsui et al. / Tetrahedron 67 (2011) 7431e7439
7437
A
DGCR, associated with paths A and B (
D
GCR and
D
GCRB) and
determined by analysis of the energy diagram and exemplified in
Fig. 6 for the parent system 1a/32aꢀꢀ*, were found to be ꢁ2.72 and
A
B
ꢁ0.25 eV, respectively.
DGCR and DGCR for the other derivatives
are summarized in Table 3.2þ6
ꢁ
If kCR associated with 2ꢀ and 1ꢀ at 135 K is well described by
a properly parameterized Eq. 9, a bell-shaped relationship between
log kCR and
D
GCR should exist (see Fig. 7)þ. The resuꢁlts obtained by
3
using Eq. 9 suggest that CR between 2aꢀ and 1aꢀ that involves
generation of 2aꢀꢀ* (paths B) is the minor process, as compared
with that of 32aꢀ (paths A). The respective kCRA(a) and kCRB(a) values
are 1.8ꢃ107 and 1.4ꢃ105
s
ꢁ1, showing that the generation ratio of
32aꢀꢀ* and 32aꢀꢀ [kCRB(a)/kCRA(a)] is 7.7ꢃ10ꢁ3 (Table 3).
On the other hand, the corresponding values of the couple
1c/32cꢀꢀ* are kCRA(c)¼6.8ꢃ107 sꢁ1 and kCRB(c)¼1.3ꢃ103 sꢁ1, showing
that generation ratio þof 32cꢀꢀ* and 32cꢀꢀ [kCRB(c)/kCRA(c)] is 1.9ꢃ10ꢁ5
.
If the amount of 2ꢀ and 1ꢀ formed in systems 1a/32aꢀꢀ* and
ꢁ
1c/32cꢀꢀ* and the FF of 2aꢀꢀ* and 2cꢀꢀ* are nearly the same, the
3
3
comparison of the kCRB/kCR ratio between the 1a/32aꢀꢀ* and
A
1c/32cꢀꢀ* systems means that the TL intensity of 2cꢀꢀ* is 400
3
(w7.7ꢃ10ꢁ3/1.9ꢃ10ꢁ5) times smaller than that of 2aꢀꢀ*. These
3
findings are not inconsistent with the results depicted in Fig. 2.
3
Thus, the differences in TL intensities among 2ꢀꢀ derivatives are
A
B
seems to be governed by kCR and kCR
.
However, similar kinetic analyses of the systems comprised of
the couples 1def/32defꢀꢀ* do not explain the finding that the TLs of
32defꢀꢀ* are more intense than that of 32cꢀꢀ*. Errors associated with
Fig. 6. A schematic energy diagram consisting seven states (ievii) for four possible
paths AeD of CR between 2ꢀþ and 1ꢀꢁ. The energy values in the diagram are derived
from analysis of the parent system (1a/32aꢀꢀ*) and are given in eV.
Table 3
Free energy changes of CR paths, emission energies, and rate constants of CR at 135 K
E1/2OX(1)a
V
E1/2RED(1)b
V
E1/2 (3ꢀ) V
EEM(32ꢀꢀ*)c eV
D
E(32ꢀꢀ) eV
D
GCR eV
D
GCR eV
kCRA sꢁ1
kCR sꢁ1
B
kCRB/kCR
OX
A
B
A
Systems
X
1a/32aꢀꢀ*
H
þ1.82
þ1.65
þ1.35
þ1.80
þ1.93
þ2.18
ꢁ2.49
ꢁ2.46
ꢁ2.51
ꢁ2.27
ꢁ2.11
ꢁ2.01
þ0.23d
2.47
2.40
2.34
2.46
2.36
2.40
þ0.50
þ0.48
þ0.46
þ0.49
þ0.48
þ0.50
ꢁ2.72
(ꢁ2.58)
ꢁ2.45
ꢁ0.25
(ꢁ0.18)
ꢁ0.11
1.8ꢃ107
(2.1ꢃ107)
6.8ꢃ107
1.4ꢃ105
(1.7ꢃ104)
1.3ꢃ103
7.7ꢃ10ꢁ3
(8.1ꢃ10ꢁ4
1.9ꢃ10ꢁ5
(4.2ꢃ10ꢁ7
(2.8ꢃ10ꢁ7
(4.9ꢃ10ꢁ6
1b/32bꢀꢀ*
CH3
CH3O
F
Cl
CF3
(þ0.12)e
ꢁ0.06d
)
1c/32cꢀꢀ
*
1d/32dꢀꢀ*
(þ0.20)e
(þ0.27)e
(þ0.46)e
(ꢁ2.47)
(ꢁ2.38)
(ꢁ2.47)
(ꢁ0.01)
(ꢁ0.02)
(ꢁ0.07)
(5.7ꢃ107)
(1.3ꢃ108)
(5.7ꢃ107)
(2.4ꢃ101)
(3.6ꢃ101)
(2.8ꢃ102)
)
)
)
1e/32eꢀꢀ
*
1f/32fꢀꢀ*
a
OX
Determined with E1/2 (¼Ecpꢁ0.03 V, vs SCE in CH3CN) obtained by cyclic voltammetry.
b
c
OX
Determined with calibration line [EAB¼1.067ꢃ(E1/2 ꢁE1/2RED)ꢁ0.024].35
Obtained by the g-irradiation method.
d
e
Quoted from Ref. 37.
Determined by using the Hammett equation, E1/2 (3ꢀ)¼0.363sꢀþþ0.23.26,36
D
GCR , D
GCR kCRA, kCRB, and kCRB/kCRA determined by using these E1/2 (3ꢀ) values are given in
OX
A
B
OX
parentheses.
On the other hand, non-radiative paths A, C, and D give rise to states
(iii) 32ꢀꢀþ1, (vi) 12ꢀꢀþ1, and (vii) 12ꢀꢀ*þ1, respectively. Although these
non-radiative paths have been disregarded in many earlier efforts,
they have been considered in this work. Unfortunately, it is difficult
to estimate accurately the energies of the states (vi) and (vii), which
contain two semi-stable species of 32ꢀꢀ, i.e., 12ꢀꢀ and 12ꢀꢀ*. As a result,
in the following discussion, we have only given attention to the
generation ratio of 32ꢀꢀ* and 32ꢀꢀ by using CR rate constants of paths
A and B (kCRA and kCRB), whose sum statistically accounts for 75% of
all CR opportunities.30
From the free energy change of each CR process (DGCR), it is
possible to determine kCR29 by using Miller’s equation31a,32 (Eq. 9),
which comes from Marcus theory.31bee Note that Eq. 9 does not
take into account multiplicities and intersystem crossing and only
34
considers substituents effects on
DGCR
.
!
!
(
)
1=2
hvÞ2
N
X
4
p3
h2lskbT
e
ꢁsSu
ð
ls
þ
D
4
GCR
þ
u
kCR
¼
jVj2
exp
ꢁ
u
!
l
skbT
u
¼0
Fig. 7. A bell-shaped relationship of log kCR versus
D
GCR in CR between 2ꢀþ and 1ꢀꢁ
S ¼
l
v=hv
using a parameterized Eq. 9 given Ref. 32 (135 K). Red and blue dots refer to 1a/32aꢀꢀ
and 1c/32cꢀꢀ* systems, respectively.
*
(9)

7438
Y. Matsui et al. / Tetrahedron 67 (2011) 7431e7439
E
RED(1) determined with calibration line, EOX(3defꢀ) determined by
References and notes
using the Hammett equation, parameters in MCH areþpossible
ꢁ
1. (a) Hammett, L. P. J. Am. Chem. Soc. 1937, 59, 96e103; (b) Hammett, L. P. Physical
Organic Chemistry, 2nd ed; McGraw-Hill: New York, NY, 1970.
reasons for this deviation, as well as decompositions of 2ꢀ and 1ꢀ
.
3
As a matter of fact, FF of 2ꢀꢀ* is also an important factor in de-
2. (a) Samori, S.; Tojo, S.; Fujitsuka, M.; Spitler, E. L.; Haley, M. M.; Majima, T. J. Org.
Chem. 2008, 73, 3551e3558; (b) Imoto, M.; Ikeda, H.; Ohashi, M.; Takeda, M.;
Tamaki, A.; Taniguchi, H.; Mizuno, K. Tetrahedron Lett. 2010, 51, 5877e5880; (c)
Kitamura, C.; Tsukuda, H.; Yoneda, A.; Kawase, T.; Kobayashi, T.; Naito, H. Eur. J.
Org. Chem. 2010, 3033e3040; (d) Maeda, H.; Ishida, H.; Inoue, Y.; Merpuge, A.;
Maeda, T.; Mizuno, K. Res. Chem. Intermed. 2009, 35, 939e948; (e) Liddle, B. J.;
Silva, R. M.; Morin, T. J.; Macedo, F. P.; Shukla, R.; Lindeman, S. V.; Gardinier, J. R.
J. Org. Chem. 2007, 72, 5637e5646.
3. (a) Fossey, J.; Lefort, D.; Sorba, J. Free Radicals for Organic Chemistry; John: West
Sussex, 1995; (b) Parsons, A. F. An Introduction to Free Radical Chemistry; Wiley-
Blackwell: Hoboken, NJ, 2000.
4. Wen, Z.; Li, Z.; Shang, Z.; Cheng, J.-P. J. Org. Chem. 2001, 66, 1466e1472.
5. Dust, J. M.; Arnold, D. R. J. Am. Chem. Soc. 1983, 105, 1221e1227.
6. (a) Creary, X. Acc. Chem. Res. 2006, 39, 761e771; (b) Creary, X.; Mehrsheikh-
Mohammadi, M. E.; McDonald, S. J. Org. Chem. 1987, 52, 3254e3263.
7. Hicks, R. Stable Radicals: Fundamental and Applied Aspects of Odd-Electron
Compounds; John: West Sussex, 2010.
termining TL intensities.
4. Conclusion
In the study described above, we experimentally determined
3
the effects of substituents on TL and PL energies of 2aefꢀꢀ* and
3
3aefꢀ* and calculated the ETs of 2aefꢀꢀ and 3aefꢀ. Hammett
analyses using sꢀ (Cheng) constants showed that substituents
have nearly the same effects on the TL and PL energies of
3
32aefꢀꢀ* and 3aefꢀ*. The results suggest that 2ꢀꢀ* has a largely
twisted molecular geometry and a considerably localized elec-
tronic structure containing the diarylmethyl radical (subunit I)
fluorophore. The results of DFT calculations also provide support
for this interpretation. As compared with the TL wavelengths of
gem-methylphenyltrimethylenemethane (lTL¼461 nm) and gem-
(2-naphthyl)phenyltrimethylenemethane (lTL¼607 nm),10a rela-
tively small but finite changes in TL wavelengths take place
when EDGs and/or EWGs are introduced into 32aꢀꢀ*
8. Lv, X.; Mao, J.; Liu, Y.; Huang, Y.; Ma, Y.; Yu, A.; Yin, S.; Chen, Y. Macromolecules
2008, 41, 501e503.
9. (a) Nishide, H.; Iwasa, S.; Pu, Y.-J.; Suga, T.; Nakahara, K.; Satoh, M. Electrochim.
Acta 2004, 50, 827e831; (b) Nishide, H.; Oyaizu, K. Science 2008, 319, 737e738.
10. (a) Namai, H.; Ikeda, H.; Hoshi, Y.; Kato, N.; Morishita, Y.; Mizuno, K. J. Am.
Chem. Soc. 2007, 129, 9032e9036; (b) Ikeda, H. J. Photopolym. Sci. Technol. 2008,
21, 327e332; (c) Ikeda, H.; Matsui, Y.; Akimoto, I.; Kan’no, K.-i; Mizuno, K. Aust.
J. Chem. 2010, 63, 1342e1347.
11. (a) Namai, H.; Ikeda, H.; Kato, N.; Mizuno, K. J. Phys. Chem.
A 2007, 111,
(
lTL¼501 nm). Therefore, EDGs and/or EWGs can be employed to
4436e4442; (b) Ikeda, H.; Namai, H.; Kato, N.; Ikeda, T. Tetrahedron Lett. 2006,
47, 1857e1860.
12. X-irradiation under conditions that are similar to those employed in
ation provided very weak TL upon annealing.
13. Dunning, T. H. J. Chem. Phys. 1989, 90, 1007e1023.
14. Becke, A. D. J. Chem. Phys. 1993, 98, 5648e5652.
15. Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785e789.
fine tune of TL wavelengths. Although induced by either - or X-
g
3
irradiation of 1, some derivatives of 2ꢀꢀ* exhibit very weak or no
TL under the X-irradiation conditions. X-irradiation is a conve-
nient method for generating transients for TL measurements,
g-irradi-
while
severe conditions.
g-irradiation is also a powerful method but it requires
16. Although the wavelengths of ETs calculated with TDDFT do not always exactly
match with values obtained experimentally owing to errors of the excitation
energies involved with ground states and excitedstates that havecharge-transfer
character, they are sufficiently accurate to enable qualitative analyses. See Ref.17.
17. Dreuw, A.; Head-Gordon, D. J. Am. Chem. Soc. 2004, 126, 4007e4016.
18. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, R. E.;
Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.;
Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.;
Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Mo-
rokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.;
Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.; Replogle,
E. S.; Pople, J. A. Gaussian 98, revision A.11.4; Gaussian: Pittsburgh, PA, 1998.
19. Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994,
98, 11623e11627.
3
Substituents on benzene rings of 2aꢀꢀ* not only affect TL
wavelengths but also their intensities, which are controlled mainly
by the rate constant ratio kCRB/kCR and FF. Thermodynamic and
A
kinetic analyses employing Miller’s equation revealed that the ra-
diative path B is inefficient as compared with non-radiative path A
(kCRB/kCRA<10ꢁ2).
In future efforts, we plan to determine FF of 32ꢀꢀ* to complete the
analysis of the CR process and gain a fundamental insight into the
TL phenomena and its application to ORLED.
Acknowledgements
This study was supported by the Cooperation for Innovative
Technology and Advanced Research in Evolutional Area (CITY
AREA) program by the Ministry of Education, Culture, Sports,
Science and Technology (MEXT), Japan. H.I. gratefully acknowl-
edges financial support in the form of a Grant-in-Aid for Scientific
Research on Priority Areas ‘New Frontiers in Photochromism’
(Nos. 20044027 and 21021025 in the Area No. 471) and In-
20. (a) Ikeda, H.; Hoshi, Y.; Namai, H.; Tanaka, F.; Goodman, J. L.; Mizuno, K. Chem.
dEur. J. 2007, 13, 9207e9215; (b) Ikeda, H.; Takasaki, T.; Takahashi, Y.; Konno,
A.; Matsumoto, M.; Hoshi, Y.; Aoki, T.; Suzuki, T.; Goodman, J. L.; Miyashi, T. J.
Org. Chem. 1999, 64, 1640e1649; (c) Ikeda, H.; Minegishi, T.; Abe, H.; Konno, A.;
Goodman, J. L.; Miyashi, T. J. Am. Chem. Soc. 1998, 120, 87e95.
21. Murov, L. S.; Carmichael, I.; Hug, G. L. Handbook of Photochemisty, 2nd ed.;
Marcel Dekker: New York, NY, 1993.
22. The broad TL spectrum of 32cꢀꢀ* (Fig. 2c) was deconvoluted into four bands (blue,
purple, red, and yellow). The shortest band maximum is listed in the Table 1.
23. Namai, H.; Ikeda, H.; Hoshi, Y.; Mizuno, K. Angew. Chem., Int. Ed. 2007, 46,
7396e7398.
novative Areas ‘p-Space’ (Nos. 21108520 and 23108718 in the
Area No. 2007), the Scientific Research (B) (Nos. 20044027 and
23350023), and the Challenging Exploratory Research (No.
21655016) from the MEXT of Japan. K.M. also acknowledges fi-
nancial support in the form of a Grant in-Aid for the Scientific
Research (C) (No. 23550058). Y.M. gratefully acknowledges fi-
nancial support by the Sasakawa Scientific Research Grant from
the Japan Science Society.
1
24. Note that intramolecular orbital interactions likely take place in 10ꢀꢀ* but that
intramolecular charge-transfer interactions are not expected owing to the fact
1
that 10ꢀꢀ* contains essentially the same two radical moieties.
25. Fig. 4 was drawn with WinMOPAC 3.9; Fujitsu Ltd.: Tokyo, Japan, 2004.
26. For the detail, see the Supplementary data.
27. Determinations of FF for short-lived intermediates at room temperature have
been carried out by Majima and co-workers,28 by using multi-laser flash
photolysis spectroscopy. Studies leading to the determinations of FF for
32aefꢀꢀ* employing a similar methodology will be reported elsewhere.
28. (a) Oseki, Y.; Fujitsuka, M.; Sakamoto, M.; Majima, T. J. Phys. Chem. A 2007, 111,
9781e9788; (b) Sakamoto, M.; Cai, X.; Kim, S. S.; Fujitsuka, M.; Majima, T. J.
Phys. Chem. A 2007, 111, 223e229; (c) Cai, X.; Tojo, S.; Fujitsuka, M.; Majima, T. J.
Phys. Chem. A 2006, 110, 9319e9324; (d) Ouchi, A.; Li, Z.; Sakuragi, M.; Majima,
T. J. Am. Chem. Soc. 2003, 125, 1104e1108.
Supplementary data
The details of the methods used to prepare substrates, details
3
of energy diagrams for TL of 2befꢀꢀ*, and computational results
for the ground states of 1aef, 2aefꢀꢀ, and 3aefꢀare given in the
Supplementary data. Supplementary data associated with this
article can be found, in the online version, at doi:10.1016/
j.tet.2011.06.052.
29. We have not considered the dissociations of [2ꢀþ/1ꢀꢁ] to the corresponding
free ions.
30. Helfrich, W.; Schneider, W. G. Phys. Rev. Lett. 1965, 14, 229e232.
31. (a) Miller, J. R.; Beitz, J. V.; Huddleston, R. K. J. Am. Chem. Soc. 1984, 106,
5057e5068; (b) Siders, P.; Marcus, R. A. J. Am. Chem. Soc. 1981, 103, 741e747; (c)
Siders, P.; Marcus, R. A. J. Am. Chem. Soc. 1981, 103, 748e752; (d) VanDuyne, R.
3

Y. Matsui et al. / Tetrahedron 67 (2011) 7431e7439
7439
P.; Fischer, S. F. Chem. Phys. 1974, 5, 183e197; (e) Ulstrup, J.; Jortner, J. J. Chem.
Phys. 1975, 63, 4358e4368.
34. Actually, introduction of the substituents will not only affect
D
GCR but also ls
,
lv, and other parameters.
32. In Eq. 9, V, ls
,
lv
,
n
, and
D
GCR are an electronic coupling matrix element
35. The ERED values of 1aef were estimated with the calibration line [EAB
¼
(V¼18 cmꢁ1), solvent reorganization energy
(
ls¼1.00 eV), vibration re-
1.067ꢃ(E1/2OXꢁE1/2RED)ꢁ0.024] obtained by using EOX and absorption energy
organization energy (lv¼0.25 eV), single average frequency (
n
¼1500 cmꢁ1),
(EAB) of 1aef.10a
and free energy change for CR process (
and
D
GCR), respectively. In addition, h, kb,
36. (a) Wayner, D. D. M.; McPhee, D. J.; Griller, D. J. Am. Chem. Soc. 1988, 110,
132e137; (b) Sim, B. A.; Milne, P. H.; Griller, D.; Wayner, D. D. M. J. Am. Chem.
Soc. 1990, 112, 6635e6638; (c) Workentin, M. S.; Wayner, D. D. M. Res. Chem.
Intermed. 1993, 19, 777e785.
37. Ikeda, H.; Akiyama, K.; Takahashi, Y.; Nakamura, T.; Ishizaki, S.; Shiratori, Y.;
Ohaku, H.; Goodman, J. L.; Houmam, A.; Wayner, D. D. M.; Tero-Kubota, S.;
Miyashi, T. J. Am. Chem. Soc. 2003, 125, 9147e9157.
T
are Planck’s constant, Boltzmann’s constant, and temperature, re-
spectively. The experimental parameters were determined by Kikuchi and
co-workers33 for a variety of benzene derivatives in dichloromethane, and were
used.
33. Niwa, T.; Kikuchi, K.; Matsushita, N.; Hayashi, M.; Katagiri, T.; Takahashi, Y.;
Miyashi, T. J. Phys. Chem. 1993, 97, 11960e11964.