Latent photochromism (pseudothermochromism) and photofatigue of crystalline 2-(2′,4′-dinitrobenzyl)pyridine

Article abstract of DOI:10.1002/poc.738

Along with the metastable 2-(2′,4′-dinitrophenylmethylidene)-1,2-dihydropyridine (NH) and the unstable 6-aci-nitro-2-nitro-5-(2′-pyridylmethylene)-1,3-cyclohexadiene (OH), the stable form of 2-(2′,4′-dinitrobenzyl)pyridine (DNBP), CH, is photochemically converted into small amounts of 1,2-bis(2′,4′-dinitrophenyl)-1,2-bis(2′-pyridyl)ethane, trans-bis[5-nitro-2-(pyridine-2-carbonyl)phenyl]diazene N-oxide, 6-nitro-3-(2′-pyridyl)-2,1-benzisoxazole and 3-nitropyrido[1,2-b]quinolin-6-ium-11-olate. The latent photochromism of DNBP, as shown by x-ray analysis of the structures of the side-products and ESR/IR measurements, is attributed to open-shell reactions that are initiated by hydrogen photoabstraction and subsequent creation of two monoradicals, NH and OH. Large amounts of the radicals (ca 50% NH and 70% OH) confined in the crystalline interior are persistent under ambient conditions. Through quasi-periodic reactions, the remaining radicals partially recover the ground-state isomers CH, NH and OH, or decay to the side-products, which results in crystalline photofatigue. Together with proton tunneling from the excited CH, the radical reactions represent dominant mechanism for the creation of NH and OH in the low-temperature regimes, but are successfully competed by the closed-shell reactions at higher temperatures. The precursor state, whose existence was assumed previously from transient absorption spectroscopy, may be identified as the radical OH. The present work represents the first study of the photofatigue of a 2-(2′,4′-dinitrobenzyl)pyridine compound and extends the 'classical' mechanism of the photochromic reactions of nitrobenzylpyridines with a set of open-shell radical reaction routes. Copyright

Full text of DOI:10.1002/poc.738

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2004; 17: 865–875
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.738
Latent photochromism (pseudothermochromism) and
photofatigue of crystalline 2-(20,40-dinitrobenzyl)pyridine
y
ˇ
Pance Naumov and Yuji Ohashi*
Department of Chemistry and Materials Science, Tokyo Institute of Technology, O-okayama, Meguro-ku, Tokyo 152-8551, Japan
Received 30 June 2003; revised 10 December 2003; accepted 15 December 2003
ABSTRACT: Along with the metastable 2-(20,40-dinitrophenylmethylidene)-1,2-dihydropyridine (NH) and the
epoc
unstable 6-aci-nitro-2-nitro-5-(20-pyridylmethylene)-1,3-cyclohexadiene (OH), the stable form of 2-(20,40-dinitro-
benzyl)pyridine (DNBP), CH, is photochemically converted into small amounts of 1,2-bis(20,40-dinitrophenyl)-1,2-
bis(20-pyridyl)ethane, trans-bis[5-nitro-2-(pyridine-2-carbonyl)phenyl]diazene N-oxide, 6-nitro-3-(20-pyridyl)-2,1-
benzisoxazole and 3-nitropyrido[1,2-b]quinolin-6-ium-11-olate. The latent photochromism of DNBP, as shown by
x-ray analysis of the structures of the side-products and ESR/IR measurements, is attributed to open-shell reactions
that are initiated by hydrogen photoabstraction and subsequent creation of two monoradicals, NH. and OH.. Large
amounts of the radicals (ca 50% NH. and 70% OH.) confined in the crystalline interior are persistent under ambient
conditions. Through quasi-periodic reactions, the remaining radicals partially recover the ground-state isomers CH,
NH and OH, or decay to the side-products, which results in crystalline photofatigue. Together with proton tunneling
from the excited CH, the radical reactions represent dominant mechanism for the creation of NH and OH in the low-
temperature regimes, but are successfully competed by the closed-shell reactions at higher temperatures. The
precursor state, whose existence was assumed previously from transient absorption spectroscopy, may be identified as
the radical OH.. The present work represents the first study of the photofatigue of a 2-(20,40-dinitrobenzyl)pyridine
compound and extends the ‘classical’ mechanism of the photochromic reactions of nitrobenzylpyridines with a set of
open-shell radical reaction routes. Copyright # 2004 John Wiley & Sons, Ltd.
Additional material for this paper is available in Wiley Interscience
KEYWORDS: 2-(20,40-Dinitrobenzyl)pyridine; nitrobenzylpyridines; photochromic ortho-nitrobenzylpyridine; photochro-
mism; photofatigue; radicals
INTRODUCTION
all Ref. 2) and have also been proposed as efficient
second-harmonic generation photomodulators.3
The scientific interest in photochromic compounds is
driven by their potential application as phototriggers for
electrical nano-circuits, photomodulators for non-linear
optical materials and materials for photo-optical memory
storage and switching devices.1 Necessary requirements
for the compounds in such applications are small struc-
tural perturbation during the photoreactions, preservation
of the crystal integrity, stability of the colored forms,
tunability of the yields and high resistance to undesirable
side-reactions (photofatigue). A promising class of com-
pounds that meet the first four of the above conditions are
those similar to the photochromic 2-(20,40-dinitrobenzyl)-
pyridines (DNBPs). The DNBPs have been used as
phototrigger units in the caged compounds (for reviews,
The collection of earlier studies has generally agreed
upon a nitro-assisted proton transfer mechanism of the
photochromic reactions, outlined in Scheme 1. When
photoexcited, the otherwise stable CH forms isomerize
to metastable blue–violet quinones/azamerocyanines
NH, mainly via unstable aci-nitro tautomers OH.4,5
Surprisingly, apart from the application relevance of
this photochromic system, and a few reports concerning
residual yellow coloration,4,6 very little is known about
its photofatigue. Long-term and/or intense irradiation of
powdered or crystalline DNBP with UV–visible light at
any temperature results in residual, unbleachable yellow–
brownish coloration. The cyclability (the number of
cycles needed for the blue absorbance to drop to 50%
of the initial value) of powdered DNBP switched between
the colorless and the blue form is Z50 ꢀ 20 (Fig. 1). The
fatigue is accompanied by a decrease in several absorp-
tions in the 1300–600 cmꢁ1 region, disappearance of a
weak band at 980 cmꢁ1 and evolution of a single, un-
bleachable, weak band at 1020 cmꢁ1. A similar decrease
in the photochromic efficiency is known for alkylnitro-
benzenes in solution.7
*Correspondence to: Y. Ohashi, Department of Chemistry and
Materials Science, Tokyo Institute of Technology, O-okayama,
Meguro-ku, Tokyo 152-8551, Japan.
E-mail: yohashi@cms.titech.ac.jp
yOn leave from the Institute of Chemistry, Faculty of Science, ‘Sv.
Kiril i Metodij’ University, Skopje, Macedonia.
Contract/grant sponsors: MEXT; Japan Science and Technology
Corporation (JST).
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

866
P. NAUMOV AND Y. OHASHI
in a polycrystalline matrix, absorbs at 335 nm and decays
either to OH and then to NH and CH, or directly to CH.
Our motives for studying the photofatigue of DNBP
were mainly based on the following two facts: (1)
although much is known about the main photochromic
reactions of DNBP, the side-reactions of this system have
hardly been given any attention in the past, and (2)
the several spectroscopic observations pointed out above
could not be satisfactorily explained with Scheme 1. For a
better understanding of the photochemistry of DNBP,
therefore, one needs to look for information obtainable
with other techniques, such as, in this case, structural and
product analyses combined with ESR spectroscopy. In
this work we characterized four side-products responsible
for the photofatigue of DNBP and we show that all of the
above experimental observations can be accounted for by
two photochemically created radical intermediates. The
present study, however, was not intended to give a
complete and final decision on the reaction mechanism;
rather, by obtaining proof for the existence of additional
species, it was aimed to provide a direction in which the
answers for the complicated photochemistry of DNBP
should be sought.
Scheme 1. The three-species mechanism of the photochro-
mic proton transfer reactions in dinitrobenzylpyridines
Spectroscopic investigations have indicated that the
actual mechanism of the photoreactions of DNBP is more
complex than that shown in Scheme 1. First, at least two
temperature regimes for the formation of both NH and
OH exist: at higher temperatures (e.g. above ꢂ160 K for
OH and ꢂ230 K for NH in the case of the parent
compound DNBP) the photoproducts are created through
thermally activated proton transfer. At lower tempera-
tures, where the thermal proton transfer is forbidden, low-
yield and nearly temperature-independent processes
(probably proton tunneling from excited state) prevail.8,9
Second, the OH isomer in the solid DNBP evolves (on a
microsecond time-scale) with delay relative to NH, which
indicates a long-lived triplet state or an additional inter-
mediate.8 Third, if DNBP dispersed in polymer matrices is
flashed with UV radiation at cryogenic temperatures, it
remains colorless, but develops the characteristic 600 nm
NH band on warming in the dark. To explain the phenom-
enon, for which we propose here one of the terms latent
photochromism or pseudothermochromism, Sousa and
Loconti10 assumed the existence of a colorless ‘dark
intermediate’ that decays to NH. Recently, Casalegno
and co-workers11,12 characterized this precursor by tran-
sient absorption spectroscopy. It was found that the
species, with a lifetime of about 10ms at room temperature
RESULTS AND DISCUSSION
Photolysis of DNBP
The HPLC analysis of DNBP photolyzed with visible
light from Xe lamp (400–800 nm) showed 13 photopro-
ducts (Fig. 2). The composition of the photolyzates
obtained from powdered bulk DNBP samples under
various conditions was analyzed by preparative-scale
TLC (for details on the irradiation conditions, see Ex-
perimental). From DNBP photolyzed at ambient tem-
perature and at 275–278 K in air, nine and seven products
were isolated, respectively, in addition to a dark, highly
intractable material. Cryogenically irradiated DNBP
samples, warmed to ambient temperature and then kept
in the dark for several days, showed the same composi-
tion as the samples irradiated at ambient temperature.
DNBP irradiated anaerobically at 275 K yielded smaller
amounts of the yellow side-products and a higher con-
centration of the NH isomer than at other temperatures,
which resulted in unusually dark blue coloration. Except
for two components, DNBP irradiated in acetonitrile and
in the solid state at 275 K afforded the same products, but
greater amounts decomposed in the solution (73.2 vs
55.3%). In the presence of a triplet-quenching diene in
acetonitrile, DNBP was largely consumed and decayed to
as many as 18 products.
Figure 1. Time profile of the integrated absorbance of the
600 nm NH band upon switching between the colorless and
the colored DNBP (coloration, 0.1 s flashes from Hg lamp,
l < 440 nm; bleaching, 1 min irradiation with Y-51-filtered
output from Xe arc lamp, l ¼ 500--800 nm)
Photochemical effects on the crystal structure
In order to investigate the structural changes during
the prolonged irradiation more directly, the effects of
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

PHOTOCHROMISM AND PHOTOFATIGUE OF 2-(20,40-DINITROBENZYL)PYRIDINE
867
Figure 3. Change of the cell parameters (a) and the aniso-
tropic displacement parameters (b) of a single crystal of
DNBP irradiated in air. Before t ¼ 0, the colorless crystal
was irradiated 0.25 h with l < 440 nm (Hg lamp, filter UV-
D33S) and the data were collected; the respective data are
plotted at t ¼ 0. From t ¼ 0 the blue crystal was colored
yellow by irradiation with 600 nm (l3) pulses for 0.67 h while
collecting the data for 4 h; the resulting parameters are
plotted at t ¼ 4 h. From t ¼ 4 h the yellow crystal was
bleached with 400--440 nm radiation (l2) (Hg lamp, O-56
filter and Xe lamp, UV-D33S filter) and left in the dark. The
data collected from the colorless crystal are plotted at
t ¼ 13.2 h. Further irradiation was carried out with unfiltered
output from an Xe lamp at l ¼ 540--800 nm. Each point
corresponds to the end of the data collection relative to t ¼ 0
Figure 2. HPLC traces of DNBP (a) before and (b) after 296 h
of irradiation with Xe arc lamp (400--800 nm). The retention
times (in minutes) are given in parentheses
photoirradiation on the structure of DNBP were followed
in situ by single-crystal x-ray structure analysis of the
lattice constants and the thermal parameters of the o-nitro
group at 250K (Fig. 3). As expected from their conforma-
tional similarity,4,5 switching among CH, OH and NH
resulted in small, albeit significant, changes but did not
affect the long-range structural order of DNBP. However,
prolonged irradiation of about 40 h resulted in notable
structural changes: initial abrupt contraction of the cell
(the cell volume decreased by 2.3% and the residual factor
increased by 0.55%) and large expansion with increased
mosaicisity afterwards. Interestingly, standing in the dark
led to partial recovery of the structural order of the brown
crystal (the residual factor decreased by 3.99%). Similar
changes were observed under argon. With the use of
single-photon excitation, no secondary structure could be
clearly discerned during the single-crystal measurements.
Fig. 4. The plot shows an induction period, which was
estimated from the change of the concentrations to about
25–50 h (ꢂ93% DNBP remained), in agreement with the
value obtained from the single-crystal study (ꢂ40 h).
After the induction period, the concentrations of the
products increased and decreased twice until the end of
the irradiation, first within about 50–140 h and then
within 140–240 h from the onset of irradiation. This
indicated quasi-periodic reactions that repeat after about
each 90–100 h of exposure. During each cycle DNBP
decayed and was partially recovered from one product
with a weight ratio of ꢂ12% (peak 5 in Fig. 4) and four
products with weight ratios varying between 0 and 4%
Characterization and crystal structures
of the side-products
The time profile of the products of DNBP that was
irradiated constantly during 10 days is presented in
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

868
P. NAUMOV AND Y. OHASHI
Figure 4. Time dependence of the composition of solid DNBP photolyzed with Xe lamp (400--800 nm) and analyzed by HPLC.
The inset shows the variation of the remaining DNBP (the axes units are the same on both plots). Peaks identified with the
standard addition method: peak 2, side product 2; peak 4, side product 1; peak 6, DNBP; peak 7, side product 3; peak 12, side
product 4
(peaks 1–4), while a small amount was converted
into several photoinactive products with relatively
constant ratios of less than 1% (peaks 7–14). A
thorough physico-chemical characterization of the pro-
ducts proved to be difficult owing to the small
isolable quantities and their photoreactivity. Repeated
recrystallizations afforded crystals from four stable
products, namely 1,2-bis(20,40-dinitrophenyl)-1,2-bis
(20-pyridyl)ethane (1), trans-bis[5-nitro-2-(pyridine-
2-carbonyl)phenyl]diazene N-oxide (2), 6-nitro-3-(20-
pyridyl)-2,1-benzisoxazole (3) and 3-nitropyrido[1,2-b]
quinolin-6-ium-11-olate (4). An additional amine product,
2-(40-amino-20-nitrobenzyl)pyridine, was obtained in the
presence of triplet-quenching diene; the structure was
identical with that of the same compound prepared
chemically.13 The molecular structures of 1–4 as deter-
mined by x-ray diffraction are shown in Fig. 5; the
relevant crystallographic data are listed in Table 1.
tural Database.14 The molecule of 3 is nearly planar. The
angle between the pyridyl and benzisoxazole functions is
4.2(1)ꢃ and the nitro group rotates by 8.7(2)ꢃ around the
C—N bond. The three condensed rings of 4 are coplanar,
with slight deviation of the nitro group by 6.1(6)ꢃ from
—
the common plane. The C—O and N N bond lengths,
—
˚
1.246(4) and 1.360(4) A, are very similar to the values of
˚
the bromo derivative, 1.244(5) and 1.361(5) A, respec-
tively.15 The long C—O distance indicates single-bond
character of the carbonyl bond and thus a significant
contribution of the enolate form to the resonant structure
of the molecule.
Possible intermediates and pathways to produce
the side-products
The presence of the induction period (Fig. 4), the large
number of side-products and their structures imply the
existence of photoinduced radical species. The structure
of the stable dimer 1, the only characterized side-product
with unoxidized benzylic carbons, clearly suggests the
existence of a precursor monomer radical denoted OH. in
Scheme 2. The most probable way for creation of OH. is
abstraction of a single hydrogen atom from DNBP. A
second radical denoted NH., whose existence was con-
firmed by ESR spectroscopy (see below), could be
obtained either by transfer of the remaining benzylic
proton in OH. to the pyridyl nitrogen or by hydrogen
abstraction from NH (Scheme 3). The present results thus
support the assumptions of Kuindzhi et al.18 based on
ESR measurements.
The molecule of 1 exists in the solid state in a gauche
staggered conformation around the central C6—C18
bond with anti-oriented pyridyl and dinitrophenyl rings,
which minimizes the steric repulsion. The central C6—
˚
C18 bond of 1.603(5) A is very long to avoid the short
contact within the molecule. However, the C17—C18—
C19 [109.4(3)ꢃ] and C5—C6—C7 [109.8(3)ꢃ] angles are
normal. The torsion angles of the two dinitrophenyl rings
(C12—C7—C6—C18 and C24—C19—C18—C6) are
59.1(4)ꢃ and 62.0(5)ꢃ, respectively, whereas those of
the two pyridyl rings (N1—C5—C6—C18 and N4—
C17—C18—C6) are 42.1(5)ꢃ and 41.8(4)ꢃ, respectively.
The molecule of the dimer 2 is centrosymmetric with
—
respect to the center of the azoxy bridge. The central N
˚
—
N distance [1.241(5) A] compares well with the mean
Products 1 and 2 indicate dimerization of OH. and
NH., while 3 and 4 suggest intramolecular cyclization.
The quasi-periodic course of the photofatigue and the
˚
value (1.24 A) for uncoordinated dimeric N-unsubstituted
trans-azobenzenes retrieved from the Cambridge Struc-
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

PHOTOCHROMISM AND PHOTOFATIGUE OF 2-(20,40-DINITROBENZYL)PYRIDINE
869
Figure 5. ORTEP diagrams (50% probability level) of the molecular structures of the photoproducts: (1) 1,2-bis(20,40-dinitro-
phenyl)-1,2-bis(20-pyridyl)ethane; (2) trans-bis[5-nitro-2-(pyridine-2-carbonyl)phenyl]diazene N-oxide (1:1 disordered over the
0
---
center of the N N bond); (3) 6-nitro-3-(2 -pyridyl)-2,1-benzisoxazole; (4) 3-nitropyrido[1,2-b]quinolin-6-ium-11-olate
---
anaerobic photolysis experiments indicate that eventual
peroxide-mediated reactions are not the main source of
the oxidized side-products 2–4. It seems unlikely that the
product 4 is created from 3 pyrolytically, since such
reaction requires very high temperatures.15 From the
structures of the photoproducts and the literature data
on photodecay of other nitro compounds,16,17 various
mechanisms for the photofatigue of DNBP can be
assumed (e.g. including diradicals or charged radical
species). Elucidation of the exact reaction mechanism,
however, would need further experimental and theoretical
work. Extensive theoretical investigation of the potential
energy surface for the main and the side reactions in
DNBP is in progress and will be published elsewhere.
even in the non-irradiated sample. The additional broad,
asymmetric signal that initially appears at 312.834 mT
(g ¼ 2.0784) remains unaffected by further irradiation
and is probably due to a small amount of thermally
unstable radicals of very short relaxation time produced
at defects. Accordingly, it disappears immediately upon
annealing at 20 K and can not be recovered by irradiation.
Prolonged irradiation increased the resonance at
324.351 mT revealing overlapped singlet (fitted:
324.454 mT, g ¼ 2.0040, ÁHpp ¼ 1.9 mT) and doublet
(322.596 and 326.697 mT, gav ¼ 2.0029, D ¼ 2.2 mT)
signals. The two signals intensified simultaneously with
irradiation time and were stable indefinitely at 5 K. Upon
warming to 300 K, 30% of the singlet and 50% of the
doublet decayed with different rates (Fig. 7). The remain-
ing radicals were stable in the dark for weeks, even in the
presence of air.
ESR spectra
Scattered reports on photoreactions of nitro aromatics
were summarized by Ramamurthy and Venkatesan.19 The
products, 2-nitrosobenzyl alcohols, were assumed to
form through a diradical obtained by abstraction of the
ꢀ-hydrogen by the o-nitro group. Nitrosobenzyl alcohol
is also obtained by partial decay of 2-nitrotoluene irra-
diated in solution.7 Instead, a structure of benzyl alcohol
with an o-N(H)O. group was assigned to the radical
species formed from several photoirradiated o-nitroben-
zyl compounds.20 In basic solutions, nitro aromatics
To confirm the existence of the radical intermediates and
to supplement the existing scarce data,18 the ESR spectra
of DNBP were investigated. Powdered DNBP irradiated
at 5 K (Fig. 6) developed weak, sharp resonance
(324.351 mT, g ¼ 2.0046) immediately after the irradia-
tion with visible light from Xe lamp. As a result of the
very small amount of long-lived radicals produced by the
diffuse room light during sample preparation, the reso-
nance at 324.351 mT is observable as a very weak signal
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

870
P. NAUMOV AND Y. OHASHI
Table 1. Crystallographic data and experimental details for four side-products produced by the photolysis of 2-(20,40-
dinitrobenzyl)pyridine
Parameter
1
2a
3
4
Empirical formula
Color
Formula weight
Temperature (K)
˚
Wavelength (A)
Crystal system
Space group
Diffractometer
Unit cell dimensions
C24H16N6O8
Colorless
516.43
C24H14N6O7
Orange–red
498.42
C12H7N3O3
Yellow
241.21
C12H9N3O3
Red
259.22
296(2)
95(2)
0.71073
Triclinic
95(2)
223(2)
0.71073
Triclinic
Pꢁ 1
0.71073
Monoclinic
P21/n
0.71073
Monoclinic
P21/n
Pꢁ 1
SMART-CCD
SMART-CCD
SMART-CCD
R-AXIS RAPID
˚
˚
˚
˚
˚
a ¼ 9.1260(4) A
a ¼ 6.1113(4) A
a ¼ 10.7816(2) A
a ¼ 4.2493(1) A
˚
˚
˚
˚
˚
b ¼ 10.9272(4) A
b ¼ 9.8060(4) A
b ¼ 3.7660(4) A
b ¼ 13.6055(7) A
˚
˚
c ¼ 13.8930(8) A
c ¼ 10.1754(7) A
ꢁ ¼ 63.007(2)ꢃ
ꢂ ¼ 87.333(2)ꢃ
ꢀ ¼ 74.221(1)ꢃ
520.77(5)
c ¼ 25.2394(6) A
c ¼ 17.5911(1) A
ꢁ ¼ 101.437(2)ꢃ
ꢂ ¼ 107.156(3)ꢃ
ꢀ ¼ 106.639(2)ꢃ
1206.35(10)
2
ꢂ ¼ 98.728(1)ꢃ
ꢂ ¼ 95.267(1)ꢃ
3
˚
Volume (A )
1013.02(3)
4
1.582
0.118
496
1012.71(6)
4
1.596
0.119
504
Z
2
1.589
0.118
256
Density (calc.) (Mg mꢁ3
)
1.422
0.110
Absorption coefficient (mmꢁ1
F(000)
)
532
0.55 ꢄ 0.45 ꢄ 0.03
Crystal size (mm3)
Theta range for data
collection (ꢃ)
0.20 ꢄ 0.05 ꢄ 0.05
0.20 ꢄ 0.10 ꢄ 0.10
0.50 ꢄ 0.10 ꢄ 0.10
1.61 to 25.00
2.26 to 27.57
1.63 to 27.52
2.33 to 25.00
Index ranges
ꢁ7 h 10,
ꢁ12 k 12
ꢁ16 l 16
6909
ꢁ7 h 7
ꢁ12 k 12
ꢁ13 h 13
5219
ꢁ13 l 13
ꢁ4 k 4
ꢁ32 l 32
9255
ꢁ4 h 5
ꢁ15 k 16
ꢁ20 l 20
7367
Reflections collected
Independent reflections
4155
2359
2303
1751
[R(int) ¼ 0.0486]
Full-matrix least-
squares on F2
4155/0/345
1.004
R1 ¼ 0.0681
wR2 ¼ 0.1462
R1 ¼ 0.1420,
wR2 ¼ 0.1844
0.412/ꢁ0.249
[R(int) ¼ 0.0501]
Full-matrix least-
squares on F2
2359/0/173
1.269
R1 ¼ 0.0630
wR2 ¼ 0.1115
R1 ¼ 0.1446
wR2 ¼ 0.1610
0.579/ꢁ0.547
[R(int) ¼ 0.0633]
Full-matrix least-
squares on F2
2303/0/163
1.022
R1 ¼ 0.0467
wR2 ¼ 0.1102
R1 ¼ 0.0755
wR2 ¼ 0.1248
0.322/ꢁ0.317
[R(int) ¼ 0.0825]
Full-matrix least-
squares on F2
1751/0/159
1.196
R1 ¼ 0.0728
wR2 ¼ 0.1768
R1 ¼ 0.1123
wR2 ¼ 0.2014
0.315/ꢁ0.249
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices
R indices (all data)
Largest dꢁif3f. peak and
˚
hole (e A
)
a
—
The molecule is centrosymmetric with respect to the center of the N N bond.
—
spontaneously produce radical anions.21 Consistent with
the former assignment18 and the above ESR data on nitro
aromatic compounds, the ESR spectra of photoirradiated
DNBP confirm the creation of the OH. and NH. radicals
that were assumed from the product analyses. The con-
formational similarity of the optimized radical structures
[UB3LYP/6–31G(d) level, supplementary material, Table
S1] with the molecular structure of CH in the crystal
suggests that they can be produced and preserved in the
crystalline lattice of CH without significant structural
perturbation (Fig. 8). The spin isodensity surfaces calcu-
lated on the theoretical radical structures show that in
both radicals the unpaired electrons are localized on the
benzylic carbon atom.
Scheme 2. Dimerization of OH. to the side-product 1,2-
bis(20,40-dinitrophenyl)-1,2-bis(20-pyridyl)ethane (1)
Implications for the reaction mechanism
Scheme 3. Possible routes for formation of NH. from OH. or
NH
The above results confirm that the photofatigue of DNBP
in the solid state (and probably in solution) is triggered by
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

PHOTOCHROMISM AND PHOTOFATIGUE OF 2-(20,40-DINITROBENZYL)PYRIDINE
871
that was flashed with UV radiation at 80 K and then
heated to room temperature in the dark (Fig. 9) (similar
latent photochromism was also observed in the UV–
visible–NIR spectra of single crystals of the p-azoxy
dimer of DNBP). The temperature effects on the IR
spectra were correlated with the respective ESR spectra
(Fig. 7).
The difference IR bands, empirically assigned to the
n(CH) (weak negative bands around 3010 cmꢁ1), n(NH)
(complex band with submaxima at ca 3385, 3238 and
3175 cmꢁ1) and n(OH) (3450 cmꢁ1) modes, respectively,
were used as probes. The n(NH) mode of small amounts
of NH in crystalline DNBP is usually observed as a very
weak band at 3385 cmꢁ1 and the n(ND) and n(OD) bands
appear at 2507 and 2606 cmꢁ1, respectively. The appear-
ance of two more components (3238 and 3175 cmꢁ1
)
besides the 3385 cmꢁ1 n(NH) band (some of which may
be due to NH.) remains unclear. A possible explanation
may be that different conformer of NH is obtained by
the radical thermal decay from that resulted by the
proton transfer. The effects on the band intensities in
Fig. 7 of concentration variations of NH and OH due to
their decay during the experiment, the temperature
effects and the ‘negative’ band at 3213 cmꢁ1 were not
accounted for.
After the sample had been flashed at 80 K with UV
radiation, n(CH) modes appeared as weak negative bands,
while only a very weak n(NH) band and no n(OH) bands
could be observed. This indicated photodecay of CH,
creation of a very small amount of NH and practically no
significant amounts of OH. The ESR spectra, on the
other hand, clearly suggest the creation of the radicals at
any temperature within the range 5–300 K. It can be
concluded, therefore, that at low temperature, photoex-
cited CH decays mainly to the radicals and small
amounts of NH, probably through reaction(s) from its
excited state.
Figure 6. ESR spectra recorded from powdered DNBP. (a)
Unirradiated, 5 K; (b) irradiated for 5 min with Xe lamp
(400--800 nm), (c) irradiated for 15 min; (d) dark, heated to
20 K; (e) dark, cooled to 5 K; (f) irradiated for 60 min at 5 K;
(g) irradiated for 75--195 min at 5 K; (h) dark, retained for 1 h
at 5 K; (i) dark, heated from 5 to 250 K; (j) dark, room
temperature; the spectrum remains unchanged for a long
time in air
the creation of two radicals, ascribed to the structures
NH. and OH.. To investigate the reaction pathways that
include the radicals, we monitored the evolution of NH
and OH through the rapid-scan FTIR spectra of DNBP
The FTIR spectra indicate that upon warming in
the dark the amount of NH initially increases and
Figure 7. Thermal decay of the fitted intensities of the two ESR signals (left) and of the characteristic difference n(NH) and
n(OH) bands (right) of photoirradiated DNBP. Irradiation conditions: Xe lamp through quartz windows (ESR) and high-pressure
Hg lamp through quartz windows (IR). The intensity of the lowest points in the right-hand plot equals zero. The band intensities
were not corrected for temperature effects
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

872
P. NAUMOV AND Y. OHASHI
Figure 8. Theoretical [UB3LYP/6--31G(d)] structures of the radicals OH. and NH. and the respective spin isodensity surfaces co-
plotted with the experimental cavities of CH molecules in crystalline DNBP. The cavities are represented as intersected surfaces
˚
˚
of overlapping spheres at 1.2 A from the respective van der Waals radii (C, 1.700; H, 1.200; N, 1.550; O, 1.520 A). The contour
˚
lines represent steps of 0.2 A. (For definition of the reaction cavity, see Refs 28 and 29) (This figure is available in colour online)
Figure 9. Difference rapid-scan FTIR spectra of DNBP irradiated (1 min) through the quartz windows of a low-temperature cell
at 80 K and then warmed in the dark to ambient temperature. The spectra were recorded by heating at 10 K intervals; the
temperature increases from the top spectrum downwards. The difference spectra were obtained by subtraction of the spectrum
of the unirradiated sample from the respective spectra after irradiation at each temperature
subsequently decreases (Fig. 7, right plot). The tempera-
ture of the maximum integrated absorption of the NH
band (190 K) corresponds well with the range 150–200 K
where the doublet signal (OH.) reaches a local maximum
and the singlet signal (NH.) is at a local minimum (Fig. 7,
left plot). This supports the assumption that the NH form
is created from the radical species. At 220 K, most of the
NH produced thermally from the radicals decayed to CH,
which is reflected in the disappearance of the n(NH) band
(Fig. 7). Upon warming above 240 K, the n(OH) band
appears, increases and subsequently decreases. This in-
dicates that OH is thermally produced from the remain-
ing radicals. The OH isomer reaches maximum
concentration at 260 K. At higher temperatures it decays
thermally to the metastable isomer NH (presence of an
overlapped shoulder around 3380 cmꢁ1) and the stable
isomer CH (disappearance of the weak negative bands
around 3010 cmꢁ1). On the basis of the above discussion,
the NH. and OH. radicals can be identified as the long-
lived (>16 h at 77 K) intermediate whose existence was
assumed previously10 and the extended reaction mechan-
ism in Scheme 4 for the photoreactions in DNBP can be
proposed.
Are the radicals identical with the
precursor state?
The recent spectroscopic measurements of Casalegno and
co-workers on DNBP in the crystalline state11 and
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

PHOTOCHROMISM AND PHOTOFATIGUE OF 2-(20,40-DINITROBENZYL)PYRIDINE
873
Scheme 4. Proposed extended reaction mechanism of the proton and hydrogen transfer reactions in photoirradiated solid
DNBP including the closed-shell species characterized previously and the radicals NH. and OH. characterized in this study. The
reactions including the open-shell species are based on the present data and are written in tentative form. The reactions shown
with broken arrows are characteristic for fluid state. The conversion of CH into NH. (homolytic N---H and heterolytic C---H
cleavage) probably proceeds in two consecutive steps through NH or OH..
dispersed in polymer matrices12 showed the creation of a
colorless precursor (lmax ¼ 335 nm in benzene) that is
stable for several hours at temperatures lower than 50 K
in polymers. Since the precursor is considerably stable
and, within the experimental uncertainties, could not be
quenched with oxygen and naphthalene, it was assigned
to an additional ground-state OH conformer rather than
to a diradical or a long-lived triplet state.
The present results confirm undoubtedly that radical
species are created in photoirradiated DNBP. Moreover,
the structure of the dimer 1 made it clear that one of the
radicals is OH., which is only weakly conjugated over the
methylene bridge and should not absorb in the visible
region. Since the behavior (the stability and the reactions)
of OH. conforms with the characteristics reported for the
precursor state, we assume that they might be identical or,
at least, very similar. However, it should be borne in mind
that the previous data have indicated that the reaction
course in DNBP depends strongly on experimental fac-
tors such as the physical state of the sample and the
irradiation conditions. Any generalizations of the reac-
tion mechanism over different reaction conditions, there-
fore, would need further experimental evidence.
OH and subsequently to CH, which accounts for the
latent photochromism (pseudothermochromism). Since
the radical reactions may have small energy barriers,
they represent the dominant mechanism for creation of
NH and OH in the low-temperature regimes. The set of
closed-shell proton transfer pathways is thermally acti-
vated and successfully competes at higher temperatures.
Prolonged irradiation causes quasi-periodic reactions
resulting in partial recovery of DNBP, creation of
cyclized/dimerized side-products and decrease of the
overall photochromic efficiency.
EXPERIMENTAL
Synthesis. Commercial DNBP was purified by column
chromatography (Al2O3, CH2Cl2–hexane) and recrystal-
lized twice from ethanol.
Aerobic solid-state photolyses and product analyses.
DNBP photolyzed in solid state/acetonitrile solution, at
room/low temperature, in air/under argon during various
irradiation times was thoroughly analyzed. Equal por-
tions of multiply recrystallized and finely powdered
DNBP crystals were irradiated with Pyrex-filtered Xe
lamp output (400–800 nm) for 10 days at 275–278 K in
air. The sample gradually turned yellow, brown and
finally into a dark semi-solid mass. At 1-day intervals,
samples from the photolysate were quickly dissolved in
acetone–EtOH in equal concentrations and analyzed by
HPLC. The HPLC analysis of irradiated DNBP was
performed with Hitachi LaChrom HPLC systems, on
analytical [pump, L-7100; UV–visible detector, L-7420,
l ¼ 250 nm; column, Chiralpak AD, 0.46 cm ꢄ 25 cm;
sample, 5 ml; elution, EtOH–hexane (7:3), 0.2 mL/
minꢁ1 (0–38 min), EtOH–hexane (7:3), 1 mL/minꢁ1
CONCLUSIONS
This work revealed that the mechanism of the nitro-
assisted proton transfer in dinitrobenzylpyridine (and
probably in similar compounds) is more complex than
the three-species reaction scheme adopted previously and
involves a set of open-shell reactions initiated by homo-
lytic cleavage of the benzylic C—H bonds. Two radical
intermediates, simultaneously created in the temperature
range 5–300 K, are persistent in the crystalline interior of
DNBP for weeks in air at ambient temperature. Upon
warming, the radicals partially decay initially to NH and
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

874
P. NAUMOV AND Y. OHASHI
(38–70 min), 100% EtOH, 1 mL/minꢁ1 (70–80 min)] and
preparative [pump, L-6250; column, Chiralpak AD (Dai-
sel Chem), sample, 5 mL] scales. After 296 h of irradia-
tion, a total of 14 peaks were observed (denoted peaks 1–
14 from lower retention times upwards in Fig. 2). The
following peaks were identified with the products isolated
by TLC: peak 2, B5 (side-product 2); peak 3, B7; peak 4,
side-product 1; peak 7, A9 (side-product 3); peak 11, A3;
peak 12, A4 (side-product 4). Bulk samples were irra-
diated under a similar regime at 275–278 and 298 K and
analyzed by preparative HPLC and TLC. In addition to
the highly intractable dark brown material at Rf ¼ 0.0
(A0, 8.7%), nine components were isolated by TLC, A1–
A9 (Rf, w): A1 (0.02, 12.3%); A2 (0.05, 3.3%); A3 (0.08,
5.4%, red on TLC); A4 (0.11, 6.4%, yellow on TLC);
A5 þ A6, not separated (0.23, 8.4%); A7 (0.35, 5.1%);
A8 (0.47, 44.7%, DNBP); A9 (0.60, 5.6%, yellow on
TLC). A7 was identified from the analytical data as
unreacted DNBP. A5 and A6 could not be separated
satisfactorily; the structure of the azoxybenzene dimer
was determined from a crystal obtained from their
mixture. These two products gave the only result that
was not readily reproducible: in several TLC experi-
ments, either one or two components with very close Rf
formed owing to secondary chemical reactions. The
colorless A5, isolated by preparative HPLC, always
yielded complex mixtures, which also contained A6
and A3. The dimer 1,2-bis(20,40-dinitrophenyl)-1,2-
bis(20-pyridyl)ethane (side product 1) was isolated as
the first fraction with column chromatography (Al2O3,
gradual elution with CH2Cl2 in hexane) from the DNBP
photolyzate that was kept in air. Secondary reactions
proceeded in several other photoproducts. For the experi-
ment at 275–278 K, DNBP powder (0.134 g) was placed
on a cold plate and irradiated with filtered (Y-43) output
from an Xe lamp for 11 days. At least seven components
could be identified, with Rf values of 0.00, 0.01, 0.03
(yellowish), 0.06 (yellow), 0.40 (DNBP), 0.62 and 0.77.
The products were extracted as described above.
[Al2O3/hexaneEtOAc (7:3)] as described for the solid
DNBP. In addition to the dark residue at Rf ¼ 0.00 (9.3%),
nine components were isolated from the mixture (Rf, w):
B1 (0.01, 6.2%);, B2 (0.05, 8.2%); B3 (0.11, 11.3%,
yellow); B4 (0.18, 9.3%);, B5 (0.22, 8.2%); B6 (0.33,
4.1%); B7 (0.37, 6.2%, yellow); B8 (0.50, 26.8%,
DNBP); B9 (0.91, 10.3%).
Aerobic solution-state photolysis in presence of a triplet
quencher. DNBP (7.4 mg) in anhydrous acetonitrile
(6 mL, 99.8%; Aldrich SureSeal) with five drops of 2,5-
dimethyl-2,4-hexadiene was placed in a quartz cuvette on
a cold plate at 275 K. The solution when irradiated with
unfiltered Xe lamp light turned yellow and then dark. The
solvent was then evaporated and the residue was analyzed
by TLC as described above. As many as 18 components,
C0–C17, were isolated by preparative TLC with Rf
values [Al2O3/hexane–EtOAc (7:3)] of 0.00 (dark),
0.03, 0.06, 0.11 (yellow), 0.15, 0.21, 0.24 (red), 0.30
(yellow), 0.37 (yellowish), 0.41, 0.44 (yellow), 0.48
(yellow), 0.56 (yellow), 0.65 (yellow), 0.75, 0.78, 0.84
and 0.92.
Single-crystal photocrystallographic study. A Quanta-Ray
MOPO-SL optical parametric oscillator pumped with a
pulsed Quanta-Ray Nd:YAG laser was used as mono-
chromatic source. High-pressure Hg and Xe arc lamps
were used as polychromatic sources. The combined
optical outputs were all focused on a single crystal of
DNBP affixed on the goniometer head of the CCD
diffractometer and data were collected at each stage of
irradiation. The colorless crystal was irradiated during
0.25 h with l < 440 nm (Hg lamp, UV-D33S filter) (l1)
and the data were collected from the blue crystal. The
respective data are represented as the point t ¼ 0 in Fig. 3.
At t ¼ 0 the data collection was started together with the
irradiation with 600 nm (l3) laser pulses for 0.67 h,
whereupon the crystal turned yellow. After about 4 h
the data collection was over; the respective data are
plotted as the point t ¼ 4 h. The yellow crystal was then
bleached with 400–440 nm light (l2) (Hg lamp, O-56
filter and Xe lamp, UV-D33S filter), left in the dark for
some time and the data were collected again; the respec-
tive parameters are plotted as the point t ¼ 13.2 h. Further
irradiation was carried out with unfiltered output from an
Xe lamp at l ¼ 540–800 nm. Each remaining point in Fig.
3 corresponds to the end of each data collection in respect
to t ¼ 0.
Anaerobic solid-state photolysis. DNBP (44 mg) was
sealed in a glass ampule under argon, attached to a cold
plate (275 K) and irradiated with unfiltered output from
Xe lamp for 10 days. Whereas on the upper side the
substrate turned dark, the lower side turned intense blue.
After exposure to air, the residue afforded at least seven
products that were separated by TLC [Al2O3/hexane–
EtOAc (7:3)] with Rf ¼ 0.00 (dark), 0.05 (red), 0.09
(yellow), 0.20, 0.41 (DNBP) and 0.64.
X-ray diffraction. X-ray data were collected either with a
SMART-CCD diffractometer,22 equipped with the Ri-
gaku liquid nitrogen cooling system, or with a Rigaku
R-AXIS RAPID IP diffractometer.23 The choice of the
crystal size was limited by the small amount of the
products available. The structures were solved by direct
methods24 and refined25 on F2 (Table 1). The non-H
atoms were refined anisotropically and the aromatic
Aerobic solution-state photolysis. DNBP (29 mg) dis-
solved in anhydrous undegassed acetonitrile (6 mL,
99.8%; Aldrich SureSeal) was placed in quartz cuvettes
on a cold plate at 275 K. During irradiation for 9 days,
4.8 mL of acetonitrile were added to refill the evaporated
solvent. The acetonitrile was then evaporated at room
temperature and the residue was analyzed by TLC
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875

PHOTOCHROMISM AND PHOTOFATIGUE OF 2-(20,40-DINITROBENZYL)PYRIDINE
875
Photochromism: Molecules and Systems. Elsevier: Amsterdam,
2003.
hydrogen atoms were included as riding bodies.
The crystallographic data have been deposited with the
Cambridge Crystallographic Data Centre (CCDC) as
supplementary publication No. CCDC 213691–213694.
2. (a) Corrie JET, Trentham DR. In Biological Applications of
Photochemical Switches, Morrison H (ed). Wiley: New York,
1994; 243–316; (b) Corrie JET, Trentham DR. In Methods in
Enzymology, vol. 291, Marriot G (ed). Academic Press: New
York, 1998; (c) Toscano JP. In Advances in Photochemistry, vol.
26, Neckers DC, von Bu¨nau G, Jenks WS (eds). Wiley: New York,
2001; 79.
3. Houbrechts S, Clays K, Persoons A, Pikramenou Z, Lehn J-M.
Chem. Phys. Lett. 1996; 258: 485–489.
4. Naumov P, Sekine A, Uekusa H, Ohashi Y. J. Am. Chem. Soc.
2002; 124: 8540–8541.
ESR spectra. The ESR spectra from powdered DNBP
were recorded on a JES-TE2000 spectrometer (Jeol)
equipped with liquid He cryostat. The sample was irra-
diated with an Xe arc lamp through the quartz window.
5. Naumov P, Sekine A, Uekusa H, Ohashi Y. Acta Crystallogr.,
Sect. B. Submitted.
IR spectra. The FTIR spectra were recorded in the rapid-
scan mode from KBr pellets with an Excalibur FTS 3000
instrument (Bio-Rad). The samples were irradiated for
1 min with an unfiltered high-pressure Hg lamp through
the double quartz windows of a temperature-controlled
cell (Oxford Instruments) with double KSR5 windows for
the probe beam. Conditions for spectral recording were
sample/background 32/16, resolution 4 cmꢁ1, and inter-
6. (a) Hardwick R, Mosher HS, Passailaigue P. Trans. Faraday Soc.
1960; 56: 44–50; (b) Wettermark G, Sousa J. J. Phys. Chem. 1963;
57: 874–877; (c) Wettermark G. Nature (London) 1962; 194: 677.
7. (a) Wettermark G. J. Phys. Chem. 1962; 66: 2560–2562; (b)
Wettermark G, Ricci R. J. Chem. Phys. 1963; 39: 1218–1223; (c)
Wettermark G, Black E, Dogliotti L. Photochem. Photobiol.
¨
1965; 4: 229–239; (d) Schworer M, Wirz J. Helv. Chim. Acta
2001; 84: 1441–1458.
8. Casalegno R, Corval A, Kuldova K, Ziane O, Trommsdorff HP.
´
val 4000–400 cmꢁ1
.
J. Lumin. 1997; 72–74: 78–80.
9. Scherl M, Haarer D, Fischer J, DeCian A, Lehn J-M, Eichen Y.
J. Phys. Chem. 1996; 100: 16175–16186.
Theoretical calculations. All theoretical calculations were
performed with the Gaussian 98 program suite26 running
on an SGI Origin2000 supercomputer. Ground-state
models of isolated monomers geometrically close to the
crystal CH structure were analytically optimized at the
UB3LYP/6–31G(d) level.27 The Cartesian coordinates of
the optimized radical structures are deposited (Table S1).
10. Sousa JA, Loconti JD. Science 1964; 146: 397–398.
11. Ziane O, Casalegno R, Corval A. Chem. Phys. 1999; 250:
199–206.
12. Corval A, Casalegno R, Ziane O, Burrows HD. J. Phys. Chem. A
2002; 106: 4272–4276.
13. Fujii I, Tanaka K, Deguchi T, Hirayama N. Anal. Sci. 1999; 15:
405–406.
14. Allen FH, Kennard O, Taylor R. Acc. Chem. Res. 1983; 16:
146–153.
15. Ning RY, Yean Chen W, Sternbach LH. J. Heterocycl. Chem.
1974; 11: 125–134.
16. White RC, Selvam T, Ihmels H, Adam W. J. Photochem. Photo-
biol. A 1999; 122: 7–10.
17. Bowman DF, Brokenshire JL, Gillian T, Ingold KU. J. Am. Chem.
Soc. 1971; 93: 6551–6555.
Acknowledgments
P.N. acknowledges a student scholarship granted by the
Ministry of Education, Science, Sports and Culture of
Japan (MEXT). The authors thank Professor T. Enoki
(TIT) for providing the ESR equipment and Dr J. W.
Bode for his comments on the manuscript. Special thanks
are due to Dr A. Miyazaki (TIT) for his kind help with the
ESR spectroscopic measurements. This work was partly
supported by a Grant-in-Aid from MEXT and a CREST
fund from Japan Science and Technology Corporation
(JST).
18. Kuindzhi BM, Igonin LA, Gribova ZP, Shabadash AN. Opt.
Spectrosc 1962; 12: 118–120.
19. Ramamurthy V, Venkatesan K. Chem. Rev. 1987; 87: 433–481.
20. Strom TE, Weinstein J. J. Org. Chem. 1967; 32: 3705–3706.
21. Russel GA, Janzen EG. J. Am. Chem. Soc. 1962; 84: 4153–4154.
22. SMART and SAINT, Area-detector Control and Integration Soft-
ware. Siemens Analytical X-ray Instruments: Madison, WI, 1995.
23. PROCESS-AUTO, Software for Data Collection. Rigaku Corpora-
tion: Tokyo, 1998.
24. Altomare A, Cascarano G, Giacovazzo C, Guagliardi A, Burla MC,
Polidori G, Camalli M. J. Appl. Crystallogr. 1994; 27: 435–436.
25. Sheldrick GME. SHELXL-97, Program for Crystal Structure
¨
¨
Refinement. University of Gottingen: Gottingen, 1997.
26. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,
Cheeseman JR, Zakrzewski VG, Montgomery JA, Stratmann RE,
Burant JC, Dapprich S, Millam JM, Daniels AD, Kudin KN,
Strain MC, Farkas O, Tomasi J, Barone J, Cossi M, Cammi R,
Mennucci B, Pomelli C, Adamo C, Clifford S, Ochterski J,
Petersson GA, Ayala PY, Cui Q, Morokuma K, Malik DK, Rabuck
AD, Raghavachari K, Foresman JB, Cioslowski J, Ortiz JV,
Baboul AG, Stefanov BB, Liu G, Liashenko A, Piskorz P,
Komaromi I, Gomperts R, Martin RL, Fox DJ, Keith T, Al-Laham
MA, Peng CY, Nanayakkara A, Gonzalez C, Challacombe M, Gill
PMW, Johnson B, Chen W, Wong MW, Andres JL, Gonzalez C,
Head-Gordon M, Replogle ES, Pople JA. GAUSSIAN98 (rev.
A11), Gaussian, Inc., Pittsburgh PA, 1998.
Supplementary material
The Cartesian coordinates of the theoretical structures of
the two radicals (Table S1) are available on Wiley
Interscience.
REFERENCES
27. Becke AD. Phys. Rev A. 1988; 38: 3098–3100; (b) Lee C, Yang
W, Parr RG. Phys. Rev. B 1988; 37: 785–789.
1. (a) Irie M. (ed). Photochromism: Memories and Switches. Chem
Rev. 2000; 100: issue 5; (b) Feringa BL. Molecular Switches.
Wiley: New York, 2001; (c) Durr H, Bouas-Laurent H (eds).
28. Ohashi Y, Uchida A, Sasada Y. Acta Crystallogr., Sect. B 1983;
39: 54–61.
29. Ohashi Y. Acc. Chem. Res. 1988; 21: 268–274.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 865–875