Proton-Transfer Processes in Well-Defined Media: Experimental Investigation of Photoinduced and Thermal Proton-Transfer Processes in Single Crystals of 2-(2,4-Dinitrobenzyl)pyridine Derivatives

Article abstract of DOI:10.1021/jp9609242

A detailed spectroscopic study of photoinduced and thermally activated proton-transfer processes for a series of different crystals of 2-(2,4-dinitrobenzyl)pyridine derivatives has been performed.The quantitative analysis of ground- and excited-state activation barriers and preexponential factors in deuterated and nondeuterated crystals shows clearly that the observed photochromism is linked to a proton-transfer process.Furthermore, it is clearly seen that the supramolecular environment of the transferred proton participates in the proton-transfer process.These supramolecular effects control the relative rates and efficiencies of the observed proton-transfer processes in both the ground and excited state, yielding, at room temperature, photoproducts having lifetimes ranging between hours and weeks.At least two proton-accepting groups may be active in the abstraction of the proton from its relatively stable benzylic position.Additionally, low-temperature measurements of proton-transfer processes show that tunneling processes prevail at temperatures below 100 K only in the excited state.No evidence for tunneling could be found for ground-state processes.

Full text of DOI:10.1021/jp9609242

J. Phys. Chem. 1996, 100, 16175-16186
16175
Proton-Transfer Processes in Well-Defined Media: Experimental Investigation of
Photoinduced and Thermal Proton-Transfer Processes in Single Crystals of
2-(2,4-Dinitrobenzyl)pyridine Derivatives
Michael Scherl,† Dietrich Haarer,*,† Jean Fischer,‡ Andre´ DeCian,‡ Jean-Marie Lehn,*,§ and
Yoav Eichen*,§,
Lehrstuhl fu¨r Experimentalphysik IV, UniVersita¨t Bayreuth, 95440 Bayreuth, Germany;
Laboratoire de Chimie Supramole´culaire, UniVersite´ Louis Pasteur, 4, Rue Blaise Pascal,
67000 Strasbourg, France; and Laboratoire de Cristallochimie et de Chimie Structurale,
4, Rue Blaise Pascal, 67000 Strasbourg, France
ReceiVed: March 27, 1996; In Final Form: July 31, 1996X
A detailed spectroscopic study of photoinduced and thermally activated proton-transfer processes for a series
of different crystals of 2-(2,4-dinitrobenzyl)pyridine derivatives has been performed. The quantitative analysis
of ground- and excited-state activation barriers and preexponential factors in deuterated and nondeuterated
crystals shows clearly that the observed photochromism is linked to a proton-transfer process. Furthermore,
it is clearly seen that the supramolecular environment of the transferred proton participates in the proton-
transfer process. These supramolecular effects control the relative rates and efficiencies of the observed
proton-transfer processes in both the ground and excited state, yielding, at room temperature, photoproducts
having lifetimes ranging between hours and weeks. At least two proton-accepting groups may be active in
the abstraction of the proton from its relatively stable benzylic position. Additionally, low-temperature
measurements of proton-transfer processes show that tunneling processes prevail at temperatures below 100
K only in the excited state. No evidence for tunneling could be found for ground-state processes.
Introduction
electro and thermo to piezo processes. In most biological light-
sensory systems, the primary vision step involves a photoinduced
Thermo- and photoinduced inter- and intramolecular trans-
formations provide means to modify and “switch” the physical
properties of solid phases1 such as polymeric films, glasses, and
crystals. The development of such tunable materials is an
important step toward the elaboration of diverse new solid-state
applications such as optical data processing and storage devices.
Proton-transfer (PT) and proton-mediated electron-transfer
(PMET) processes are of particular interest since they present
a number of unique properties.2 (a) PT reactions are, in most
cases, reversible and can be used for many cycles. (b) PT
processes do not necessarily require large structural changes
and therefore can take place in the solid state, be it amorphous
or crystalline, even at very low temperatures; furthermore, due
to the small mass of the proton, quantum mechanical tunneling
may contribute to the PT process. (c) PT processes are, in
general, electronically well shielded, and one does not have to
consider the Coulombic interactions far beyond the reactive site.
(d) The lifetime of the proton-transferred state ranges from the
picosecond regime to seconds, and even hours and days (see
below). (e) PT reactions can be thermo-, photo-, electro-, and
even piezoinduced, ensuring a wide scope of potential applica-
tions for such protonic systems. (f) Protonic systems may be
designed to self-assemble into large and well-defined supramo-
lecular assemblies where the monomers are interconnected via
hydrogen bonds and other noncovalent interactions.
isomerization process of a protonated Schiff base that induces
a subsequent proton cascade leading eventually to a neural
signal,3 which can be further processed. Isolated bacterio-
rhodopsin systems, natural and mutants, have been applied as
photochromic molecules in technical prototypes of data process-
ing and storage model devices.4 Dye molecules having two or
more tautomeric forms such as porphyrins and phthalocyanines
have been shown to undergo both thermo- and photoinduced
tautomerization processes.5 Other suitable chromophoric sys-
tems lacking labile protons were shown to induce proton
reorganization in their solvation shell.6 Both kinds of systems
have been utilized as chromophores in ultrahigh-density memory
devices based on the energy selective spectral hole-burning
technique.7 PT is involved in many photochromic processes
since the proton affinity of excited molecules and groups is often
different from the comparable ground-state values.8 Photo- and
thermoinduced PT processes are responsible for both the
photochromism and thermochromism of N-salicylideneanilides
in different matrixes.9 In many R-alkyl ketones the keto oxygen
was shown to abstract a proton from the alkyl group in the
excited state.10 The existence of low-lying thermally accessible
states enable the thermo- and piezoinduction of proton trans-
locations.11 A thermoinduced PT is associated with the revers-
ible phase transition in 4,4′-bipyridinium-squaric acid crystals,12
which results in a significant color change. Piezoinduced PT
processes were found to be coupled to paramagnetic and optical
changes in benzoquinone-hydroquinone crystals.13 PT is also
the key step in the thermochromism of iodoanilinium picrate.14
Nevertheless, many rather simple PT reactions are not well
understood due to the complex interrelation between electronic,
ionic, and environmental effects. They have, in the past, often
Induced proton translocations have been shown to take place
in many diverse molecular and supramolecular structures
following different stimulations ranging from photo through
† Universita¨t Bayreuth.
‡ Universite´ Louis Pasteur.
§ Laboratoire de Cristallochimie et de Chimie Structurale.
Current address: Department of Chemistry, TechnionsIsrael Institute
of Technology, Technion City, Haifa 32000 Israel.
X Abstract published in AdVance ACS Abstracts, September 15, 1996.
S0022-3654(96)00924-0 CCC: $12.00 © 1996 American Chemical Society

16176 J. Phys. Chem., Vol. 100, No. 40, 1996
Scherl et al.
CHART 1
SCHEME 1
reduced pressure. The crude product was recrystallized from
ethanol-d1. NMR characterization showed 92-95% deuteration,
exclusively at the benzylic position.
6-Benzyl-2,2′-bipyridine: A suspension of zinc (1.1 g, 17
mmol, fine dust) in dry THF, under argon, was activated using
0.1 mL of a dry solution of HCl in 1,4-dioxan (4 M, Aldrich).
When hydrogen evolution ceased, benzyl bromide (1.1 mL, 4.4
mmol) was added (drop by drop during 15 min). The solution
was stirred at room temperature for another 1 h and then filtered.
A solution of 6-bromo-2,2′-bipyridine (1 g, 4.25 mmol) and
(Ph3P)4Pd (0.2 mmol) in 15 mL of THF was then added to the
benzyl zinc bromide, and the solution was heated to 50 °C for
2 h. The THF was then removed under reduced pressure, and
the product extracted to an acidic aqueous solution. The product
was reextracted from the basified aqueous solution using
dichloromethane. Chromatography (alumina/2% ethyl acetate
in hexane, rf ) 0.7) afforded 0.98 g (95%) of pure 6-benzyl-
2,2′-bipyridine, mp (ether) 61-62 °C.
6-(2,4-Dinitrobenzyl)-2,2′-bipyridine (2): 6-Benzyl-2,2′-bi-
pyridine (0.9 g, 3.6 mmol) was dissolved in 11 mL of cold
sulfuric acid (97%). The solution temperature was set to -10
°C, and then 1.5 mL of nitric acid (100%) was added drop by
drop while stirring vigorously. The solution was stirred for
another 3 h at room temperature, then poured on ice, neutralized
with NH4OH, and extracted with three portions of 100 mL of
dichloromethane. The combined organic layers were concen-
trated under reduced pressure and chromatographed (alumina/
ether-hexane 1:1, rf ) 0.7) yielding 0.9 g (75%) of 6-(2,4-
dinitrobenzyl)-2,2′-bipyridine (2), mp (ether-hexane) 110-
110.5 °C, (ethanol) 118-118.5 °C.
6-(2,4-Dinitrobenzyl)-2,2′-bipyridine-d2 (2-d2): 6-(2,4-Dini-
trobenzyl)-2,2′-bipyridine (0.5 g, 1.5 mmol) was dissolved in 5
mL of dry THF under argon. To this solution were added two
drops of dry triethylamine and 10 mL of ethanol-d1. The
solution was heated to 40 °C during 1 week, and then THF
was removed under reduced pressure. The white crystals that
separated from solution were filtered and dried. NMR char-
acterization showed 92-95% deuteration, exclusively at the
benzylic position.
2-Benzyl-1,10-phenanthroline: 1.53 g of (5.07 mmol) ben-
zyltriphenylphosphonium chloride was suspended in 40 mL of
DME (sodium dried), under argon, and cooled to -35 °C.
n-Butyllithium (2.9 mL, 1.6 M solution in hexane) was added
to the suspension and stirring continued for 1 h at -35 °C. A
solution of 0.5 g (2.33 mmol) of 2-chloro-1,10-phenanthroline
in 10 mL of DME was added drop by drop to the stirred
solution, and the mixture was left to reach room temperature
(about 1 h). The suspension was refluxed over 17.5 h, and then
300 mg of sodium carbonate and 10 mL of water were added
to the solution, and reflux was continued for another 3 h. The
solvents were then removed under reduced pressure, and the
product was extracted with three portions of dichloromethane.
The combined organic layers were concentrated and chromato-
been interpreted within schemes of small molecular systems
underestimating supramolecular structures and effects, and thus
often no clear and conclusive picture could be developed.
One of the most interesting solid-state photoactive PT
processes is the photoinduced proton transfer (PIPT) taking place
in 2-(2,4-dinitrobenzyl)pyridine (R-DNBP, 1, Chart 1) crystals.
Scheme 1 shows the proposed mechanism for the reaction in
the crystalline state.15a The phototautomerism of 1 from its
“CH” to its “NH” tautomeric form, 1a, may, in principle, take
place directly or via its metastable “OH” tautomer. The
photochromic properties were rationalized according to the
molecular structure and properties, ignoring the fact that some
basic atoms of neighboring molecules are within interaction
distance from the active protons and the proton path during the
PT process.15a Earlier reports referred to the question of the
reaction path and had concluded that the transfer is intramo-
lecular on the basis of the crystal structure of 1.15b Although
the photochromism of R-DNBP, (1) was reported as early as
1925,16 only little is known about this process in solid matrixes
and about the correlation between the structure (molecular and
crystalline) and its PT properties. Furthermore, none of the other
2,4-dinitrobenzyl derivatives prepared up to date was reported
to be photochromic in the solid state.
In view of preparing new PT systems exhibiting longer
lifetimes for the tautomers and to better understand the
structure-PT dynamics relationship, new 2,4-dinitrobenzyl
derivatives have been prepared and characterized. In this study
we put the main emphasis on the crystalline state.
Experimental Section
Materials. 2-(2,4-Dinitrobenzyl)-pyridine (1) was prepared
according to a literature procedure.16
2-(2,4-Dinitrobenzyl)-pyridine (1-d2): 2-(2,4-Dinitrobenzyl)-
pyridine (1 g, 3.86 mmol) was dissolved in 10 mL of dry THF
under argon. To this solution were added two drops of dry
triethylamine and 10 mL of ethanol-d1. The solution was heated
to 40 °C during 1 week, and then THF was removed under

Proton Transfer in Well-Defined Media
J. Phys. Chem., Vol. 100, No. 40, 1996 16177
TABLE 1: Selected Crystallographic Information on the Different Crystal Structuresa
crystal
formula
MW
1
2I
2II
3
4
5
C12H9N3O4
259.2
C34H24N8O8
672.6
C17H12N4O4
336.3
C13H11N3O4
273.3
C13H11N3O4
273.3
C19H12N4O4
360.3
space group
a, Å
b, Å
c, Å
â, deg
P21/c
Pna21
14.051(4)
5.159()
42.017(2)
90.00
P21/c
P21/c
P21/c
C2/c
10.100(3)
15.178(4)
7.685(2)
101.40(2)
4
1.491
9.304
1558
1168
0.034
8.340(3)
12.740(4)
14.408(4)
92.13(2)
4
1.460
8.563
1903
1381
0.037
7.677(2)
19.750(6)
8.487(3)
96.06(2)
4
1.422
8.684
1571
1115
0.043
11.338(3)
15.390(4)
7.639(2)
106.04(2)
4
1.417
8.654
1607
1058
0.050
27.494(8)
8.662(2)
15.581(4)
3153.4
8
1.518
8.735
3376
1421
0.091
Z
F
4
calc, g cm-1
1.467
8.602
2053
1270
µ(Cu KR), cm-1
no. of reflns
with I > 3σ(I)
R(F)
0.060
Rw(F)
GOF
0.057
1.298
0.090
1.723
0.053
1.256
0.068
1.516
0.075
1.678
0.137
1.915
a Data were collected at -100 °C for 1, 2I, 2II, and 5 and at room temperature for 3 and 4.
SCHEME 2
graphed (alumina/dichloromethane-hexane 1:1) to afford 0.56
g (88%) of 2-benzyl-1,10-phenanthroline.
Crystal Structures. Crystals of 1, 2-II, 3, 4, and 5 were
grown by slow evaporation of their ethanolic solutions in the
dark. 2-I crystals were grown by the slow evaporation of a
solution of 1:1 (v/v) ether-hexane in the dark. Crystallographic
data were collected using a Philips PW1100/16 diffractometer,
The structures were solved by direct methods and refined on
|Fo|. Detailed crystallographic information on the different
crystal structures is presented in Table 1. Further details on
the crystal structures are available on request from the Director
of Cambridge Crystallographic Data Center, 12 Union Road,
GB-Cambridge CB2 1EZ, UK, on quoting the full journal
citation.
2-(2,4-Dinitrobenzyl)-1,10-phenanthroline (5): 2-Benzyl-1,10-
phenanthroline (0.5 g, 1.85 mmol) was dissolved in 11 mL of
cold concentrated sulfuric acid. The solution was cooled to -10
°C, and then 1.5 mL concentrated nitric acid (100%) was added
drop by drop while stirring vigorously the solution. The solution
was stirred for another 3 h at room temperature, then poured
on ice, neutralized with NH4OH, and extracted with three
portions of 100 mL of dichloromethane. The combined organic
layers were concentrated under reduced pressure and chromato-
graphed (alumina/dichloromethane-hexane 1:1) yielding 0.47
g (70%) of 2-(2,4-dinitrobenzyl)-1,10-phenanthroline.
2-Benzyl-6-methylpyridine was prepared similarly to the
method of preparation of 6-benzyl-2,2′-bipyridine; 98% yield,
liquid. Used without further purification.
2-(2,4-Dinitrobenzyl)-6-methylpyridine, (3) 2-Benzyl-6-meth-
ylpyridine was nitrated according to the procedure for the
preparation of 6-(2,4-dinitrobenzyl)-2,2′-bipyridine, 3; 70%
yield, mp 112-113 °C.
2-Benzyl-3-methylpyridine was prepared similarly to the
method of preparation of 6-benzyl-2,2′-bipyridine; 95% yield,
liquid. Used without further purification.
The deuterated crystals 1-d2, 2-d2-I, and 2-d2-II were found
to have identical diffraction patterns and unit-cell dimensions
as their nondeuterated analogues.
Apparatus. Proton-transfer processes were followed by
UV-vis absorption spectroscopy using a setup outlined in
Scheme 2. A current stabilized tungsten lamp, a, coupled to a
1 m monochromator (Spex 1704), b, was used as the analytical
light source. Two photomultipliers, c and d (Hamamatsu R649)
were used to collect the probe and reference beams respectively.
A helium flow cryostat, e (Oxford Instruments CF 204),
equipped with a temperature controller, f (Cryovac TIC 304-
M) was used to stabilize the samples temperature between 4
and 350 K within 0.1 K. The sample temperature was
monitored using a calibrated silicon diode sensor (Lake Shore
DT-470) mounted on the sample holder in close proximity to
the sample. A power stabilized high pressure mercury lamp
(150 W), g, equipped with narrow-band mercury line filters
2-(2,4-Dinitrobenzyl)-3-methylpyridine (4): 2-Benzyl-6-meth-
ylpyridine was nitrated according to the procedure for the
preparation of 6-(2,4-dinitrobenzyl)-2,2′-bipyridine, 2; 75%
yield, mp 121-122 °C.
All compounds were characterized using NMR, FAB-MS,
and elemental analysis.

16178 J. Phys. Chem., Vol. 100, No. 40, 1996
Scherl et al.
(Andover) was used as the exciting light source. Polarizers, h
and i, were used for taking the linearly polarized light spectra.
Time-dependent measurements were performed either by
taking several spectra at different time intervals or by following
the absorption peak as a function of time at a fixed wavelength.
For the relative quantum yield measurements the setup was
provided with appropriate filters (Andover bandpass filters) in
order to separate the excitation light from the probe light. In
this setup, visible absorption spectra could be recorded during
UV irradiation. Initial linear slopes of the photo coloration were
used to evaluate the relative quantum yield and the crystals were
thermally bleached after each data point.
Results and Discussion
Crystal Structures and Photochromic Properties. All the
newly prepared 2-(2,4-dinitrobenzyl)pyridine derivatives gave
crystalline phases. Some of the molecules gave different
polymorphs under different crystallization conditions. The
crystal structures of 2-I (the photochromic phase of 6-(2,4-
dinitrobenzyl)-2,2′-bipyridine, 2), 2-II (the nonphotochromic
phase of 6-(2,4-dinitrobenzyl)-2,2′-bipyridine, 2), 3, 4, and 5
are presented in Figures 1-5, respectively. The revised
structure of 1, which was solved using modern techniques is
presented in Figure 6. Important intramolecular and inter-
molecular parameters for the different structures are sum-
marized in Tables 2 and 3, respectively. 1, 2-II, 3, 4, and 5
crystallized in a monoclinic unit cell while the crystal of 2-I
was orthorhombic. The crystals of 1, 2-II, 3, and 4 contained
four molecules in the unit cell, while those of 2-I, 5 contained
eight molecules. The pyridine-benzene interplanar angles vary
somewhat along the crystal series: ω1 ) 65.1(1)°; ω2-I ) 56.6-
(1)° and 57.2(1)°; ω2-II ) 65.14(5)°; ω3 ) 55.0(1)°; ω4 ) 55.8-
(1)°; ω5 ) 89.3(1)°. In all systems, the plane of the o-nitro
group deviates from the plane of the benzene ring,
R1 ) 31.7(5)°; R2-I ) 36.4(6)° and 37.3(4)°; R2-II ) 23.3(2)°;
R3 ) 40.1(2)°; R4 ) 41.1(4)° and R5 ) 25.5(1)°. In all crystals,
two types of basic atoms interact with the benzylic hydro-
gens. One oxygen atom of the o-nitro group is within close
contact to one of the benzylic protons, d1 ) 2.350(2) Å; d2-I
)
2.409(3) Å and 2.306(3) Å; d2-II ) 2.342(2) Å; d3 ) 2.432(3)
Å; d4 ) 2.428(3) Å and d5 ) 2.303(4) Å. The nitrogen of the
pyridine is also situated in close contact to one of the benzylic
hydrogens, d1 ) 3.015(3) Å; d2-I ) 2.936(6) Å and 2.873(6)
Å; d2-II ) 2.679(1) Å; d3 ) 2.960(2) Å; d4 ) 3.071(3) Å and
d5 ) 2.512(4) Å. Additionally, different basic groups belonging
to neighboring molecules interact with the benzylic protons,
forming supramolecular arrays.
Despite their crystalline and molecular structural similarities
only 1-d2, 2-I, 2-d2-I, 3, and 4 showed photochromic properties
similar to the previously reported 1 crystals whereas 2-II, 2-d2-
II, and 5 did not change their spectra upon illumination (the
crystal did not show any photochromic properties in the 400-
800 nm spectral range at 4-300 K for 2-II and at 77-300 K
for 5). In contrast, all compounds were photochromic in
solutions. The similarities in the intramolecular coordinates
around the part of the molecule where the PT reaction takes
place and the negligible and nonconsistent differences in the
crystal densities (see Table 1) tend to rule out the possibility
that the differences in the reactivity originate from the differ-
ences in the crystal free volume, as was previously suggested
for salicylideneanilides.17 Two different alternative/comple-
mentary arguments for the different photochromic properties
in the crystalline state could be suggested based upon the
differences between the photochromic and nonphotochromic
crystal structures.
Figure 1. Arrangement of molecules in the unit cell of 2-I according
to the crystal structure.
(a) The different photochromic properties may arise from the
differences in the packing of the molecules in the crystal. In
most of the photochromic crystal structures, namely, 2-I, 3, and
4, the molecules are arranged in stacks due to short-range π-π
interactions between the pyridine rings, d(2-I) ) 3.772 Å and
3.839 Å, d(3) ) 3.795 Å, and d(4) ) 3.931 Å24a (1 crystals lack
any close inter-ring contacts). This leaves the 2,4-dinitrobenzyl
chromophore relatively isolated and with no close inter-ring
contacts with other aromatic rings, d(1 ) ) 4.762 Å, d(2-I ) )
min
min
5.159 and 5.159 Å, d(3 ) ) 5.188 Å and d(4 ) ) 4.681 Å.24a
min
min
On the other hand, in the nonphotochromic crystals 2-II and 5
the 2,4-dinitrobenzyl chromophore of one molecule is strongly
interacting with other rings of neighboring molecules. In the
2-II crystal it is tightly sandwiched between one 2-pyridine ring
1
and one 2′-pyridine ring of two neighboring molecules, d(2-II)
) 3.730 Å, d(2-I) ) 3.789 Å.24a In the 5 crystal it is stacked
2
with the rings of the phenanthroline moiety, d(5) ) 3.663 Å.24a
Figure 7 presents the intermolecular stacking found in 4 and 5
crystals as examples for the intermolecular arrangement in

Proton Transfer in Well-Defined Media
J. Phys. Chem., Vol. 100, No. 40, 1996 16179
Figure 2. Arrangement of molecules in the unit cell of 2-II according
to the crystal structure.
Figure 4. Arrangement of molecules in the unit cell of 4 according to
the crystal structure.
(b) The proton abstracted in this process is bound to a carbon
atom, implying a relatively high activation barrier. Therefore,
this process may require a considerable activation involving both
the oxygen of the o-nitro group and the pyridine nitrogen in
the abstraction of the proton from the benzylic carbon. Ac-
cording to the crystal structures presented in Figures 1-6, it is
evident that all the photochromic structures have both the oxygen
of the o-nitro group and the pyridine nitrogen bound to the same
benzylic proton. Furthermore, in all of the photochromic
structures these abstracting atoms are positioned at the same
side of the molecules and are within close distance. Presumably,
this arrangement allows the lowering of the barrier for the
excited-state proton abstraction by forming some kind of a low-
energy path to deprotonation (see Figure 8b). In the two
nonphotochromic structures shown above, at least one of two
requirements is not fulfilled. Figure 8a-c presents a schematic
representation of the different excited-state energy profiles
expected in systems having only one abstracting group, pho-
tochromic structures, and nonphotochromic structures, respec-
tively. It is evident that additional abstracting groups at an
appropriate position may contribute to the lowering of the energy
barrier of the process. Additionally, minute changes in the
relative orientation of the reactive groups may induce an increase
in the proton-transfer barrier and thus inhibit it. These trends
for a correlation of the proton dislocation with the position of
both the oxygen and nitrogen atoms is also backed by quantum
mechanical calculations.23
Figure 3. Arrangement of molecules in the unit cell of 3 according to
the crystal structure.
photoactive and photoinert crystals. The relevant interplanar
distances and angles for the different crystals are presented in
Table 3. In structures where the 2,4-dinitrobenzyl chromophore
is tightly stacked with an adjacent ring of a neigh-
boring molecule, deactivation pathways for the electronic energy
faster than proton transfer could be envisaged due to stronger
electronic coupling between the excited chromophore and the
π-systems of neighboring molecules, thus leading to a nonpho-
tochromic crystal. This argument may be backed by many
similar cases reported for the N-salycilideneanilide series.17,18
Photoinduced Processes. The new photochromic com-
pounds formed transparent to pale yellow crystals having an

16180 J. Phys. Chem., Vol. 100, No. 40, 1996
Scherl et al.
Figure 6. Arrangement of molecules in the unit cell of 1 according to
the revised crystal structure.
TABLE 2: Selected Intramolecular Parameters for the
Different Crystal Structures
crystal
1
2I
2I
2II
3
4
5
Figure 5. Arrangement of molecules in the unit cell of 5 according to
the crystal structure.
pyridine phenyl 65.1 56.6 57.2 65.1 55.0 55.8 89.3
(deg)
o-nitrophenyl
(deg)
p-nitrophenyl
(deg)
N2-H1 (Å)
N2-H2 (Å)
O1-H1 (Å)
O1-H2 (Å)
N1-H1 (Å)
31.7 36.4 37.3 23.3 40.1 41.1 25.5
absorption onset at ca. 400 nm. Similarly to the 1 crystals, the
crystals of 1-d2, 2-I, 2-d2-I, 3, and 4 turned deep blue when
exposed to near-UV light. Upon irradiation, the crystals
developed a new broad absorption band in the visible spectrum
with a maximum intensity at ca. 600 nm. Monitoring this new
absorption peak using polarized light showed that it is composed
of at least two partially overlapping bands with different relative
intensities at different light polarizations. Figure 9 presents the
absorption spectra of 1 at different light polarizations, a similar
effect was found in the visible spectra of all other irradiated
crystals. The two spectra, obtained at the different polarizations,
decayed homogeneously with identical decay rates, indicating
that all of the transitions composing this band originate from
one species. Unlike earlier reports stating that only a very small
fraction of the molecules could be converted into the “NH” blue
tautomer,19 we have found that by using thin crystalline layers,
a high degree of conversion could be attained. Figure 10 shows
the absorption spectra of a ca. 0.5 µm thick layer of polycrys-
talline 1 before and after irradiation. From the relative
intensities of the UV band (λ ) 252 nm) and the “NH” blue
band as well as from complementary IR measurements,24b it
results that at least 16%24c of the molecules were converted to
the “NH” tautomer.
12.30 21.08 18.26 15.57 12.32 14.23 14.62
3.015 2.936 2.873 2.679 2.960 3.071 2.512
3.061 3.207 3.182 3.191 3.087 3.115 2.736
2.350 2.409 2.306 >3.5 2.432 2.428 2.303
>3.5 >3.5 >3.5 2.342 >3.5 >3.5 >3.5
2.927 2.972 2.788 2.730 2.957 2.990 2.735
The photo coloration efficiencies of the different crystals are
strongly temperature dependent.24d Figure 11 presents the
temperature dependence of the photo coloration process ef-
ficiency for the 1 and 2-I crystals. As was previously reported
for the 1 crystal,15a at least two different temperature regimes
exist in its Arrhenius diagram. Three different processes may
contribute to the formation of the “NH” species. (a) Temper-
ature-independent processes, attributed to proton tunneling in

Proton Transfer in Well-Defined Media
J. Phys. Chem., Vol. 100, No. 40, 1996 16181
Figure 7. (a, left) Intermolecular stacking found in the crystals of 4 as an example for the intermolecular arrangement found in photoactive
crystals. (b, right) Intermolecular stacking found in the crystal of 5 as an example for the intermolecular arrangement in photoinert crystals.
TABLE 3: Selected Intermolecular Parameters for the Different Crystal Structures
crystal
1
2I
2I
2II
3
4
5
phenyl-phenyl dist (Å)
ang (deg)
phenyl-pyridyl dist (Å)
ang (deg)
pyridil-pyridyl dist (Å)
ang (deg)
N2-H2′ (Å)
N1-H2′ (Å)
4.762
15.897
5.897
65.147
>6
5.159
0.040
>6
5.159
0.063
>6
5.629
0.000
3.789
7.383
5.765
61.571
>3.5
>3.5
>3.5
2.874
2.960
>3.5
>3.5
2.526
>3.5
>3.5
>3.5
5.188
51.440
5.506
55.150
3.795
0.000
>3.5
5.352
0.000
4.681
55.847
3.931
0.000
>3.5
>6
3.663
8.122
3.717
0.000
2.655
>3.5
>3.5
>3.5
>3.272
>3.5
>3.5
>3.5
>3.5
>3.5
>3.5
3.772
14.910
>3.5
3.337
>3.5
>3.5
>3.5
>3.5
>3.5
>3.5
2.492
>3.5
>3.5
3.839
15.149
>3.5
>3.5
>3.5
>3.5
>3.5
>3.5
>3.5
>3.5
>3.5
>3.5
>3.5
3.174
>3.5
>3.5
>3.5
>3.5
N(p-NO )-H2′ (Å)
N(py′)-2H1′ (Å)
N(py′)-H2′ (Å)
O1-H1′ (Å)
3.392
3.242
3.113
3.490
>3.5
>3.5
2.982
2.664
2.630
>3.5
>3.5
>3.5
2.604
3.441
2.837
>3.5
>3.5
>3.5
2.619
3.242
O1-H2′ (Å)
O2-H1′ (Å)
O2-H2′ (Å)
O3-H2′ (Å)
O4-H2′ (Å)
the different excited states, dominate in the low-temperature
regime (T < 100 K and T < 200 K for 1 and 2-I, respectively);
the contribution is more important in 2-I than in 1. The ratio
of the temperature-independent quantum yield between 2-I and
1 is roughly 8; this process may be attributed to tunneling from
the excited “CH” state to one of the “NH” states (excited or
ground state). (b) At higher temperatures, a temperature-
dependent process gradually increases the overall quantum yield.
It may be attributed to the direct thermally activated transfer
from an excited “CH” form to the proton-transferred “NH” form.
The corresponding barriers were found to be 11 and 4 kJ mol-1
for 1 and 2-I, respectively. (c) At even higher temperatures, a
second thermally activated process originating from the thermal
decay of the “OH” form to the “NH” form may contribute to
the overall “NH” tautomer formation, as can be seen from the
deviation of the data points in the high-temperature regime in
Figure 11. In contrast to 1, where the relatively stable “OH”
form contributes to the “NH” formation only at the high-
temperature regime and therefore allows the deconvolution of
the two processes, the “OH” form of the 2-I crystal is relatively
unstable on the experimental time scale and therefore cannot
be separated from the other thermally activated process. The
barrier height reported for the 2-I crystal is therefore the lowest
barrier of these two processes. The photo coloration quantum

16182 J. Phys. Chem., Vol. 100, No. 40, 1996
Scherl et al.
Figure 8. Schematic representation of the different excited-state energy profiles expected in systems having (a) only one abstracting group, (b)
photochromic structures and (c) nonphotochromic structures.
Figure 9. Absorption spectra of the "NH" form of 1 at different light
polarizations (298 K): (a) 0°; (b) 90°.
Figure 11. Temperature dependence of the photo coloration process
efficiency: (a) in the 1 crystal; (b) in the 2-I crystal. Data points were
taken down to a lowest temperature of 10 K showing no deviations
from the constant low-temperature values.
Upon irradiating the blue band of the “NH” form, this band
weakened, forming a new band (λmax ≈ 435 nm), previously
attributed to the “OH” form15a (Figure 12). A similar effect
with different quantum efficiencies could be observed in all other
photochromic crystals. The marked difference between the
relative intensities of the new “OH” forms in the different
crystals suggest that another route for the photobleaching of
the “NH” form (presumably “NH” f “CH”) is operative and
that its relative efficiency is varying from one system to the
other.
Thermally Activated Processes. The “NH” blue band of
all of the photochromic crystals decays slowly and homoge-
neously in the dark, thus reverting into their “CH” forms.24c
Figure 13 shows the time dependence of the spontaneous decay
of the “NH” absorption band of 2-I at 308 K. The crystals of
1, 2-I, and 3 underwent a similar decoloration process following
a perfect monoexponential decay. In the case of the 4 crystal,
only a biexponential curve could be fitted to the “NH” band
decay; it yields two reaction rates that are different by about 1
order of magnitude. Figure 14 shows the visible absorption
spectra of 4 at different time intervals at 300 K; it is seen that
the band decays homogeneously along the whole spectral range.
It seems that in this case two species having identical absorption
spectra are decaying with two different reaction rates. This may
originate from two slightly different environments occupied by
the “NH” form.
Figure 10. Absorption spectra of a ca. 0.5 µm thick layer of
polycrystalline 1 (298 K): (a) before; (b) after UV irradiation.
yield of 3 crystals was found to be similar to that of 1. In
contrast, the quantum yield of 4 is markedly lower.24e The
relative quantum yields of formation of the “NH” form in the
1d2, 2d2-I are significantly lower than their nondeuterated
analogues, 1 and 2-I. The relative quantum yield for the
formation of the blue form at 300 K is φ“NH”:φ“ND” ) 20:1 (
30% for 1, according to IR measurements using mixed crystal
powders. A similar effect could be observed in the case of
2-d2-I as compared to 2-I. This implies that the rate-determining
step of the photoinduced coloration process is indeed a proton-
transfer reaction. Important structural reorganization processes,
if any, may occur only after the proton abstraction.
Figure 15 shows the time dependence of the “NH” band
intensity at 340 K for the different crystals. Different decay

Proton Transfer in Well-Defined Media
J. Phys. Chem., Vol. 100, No. 40, 1996 16183
Figure 12. Photoinduced “NH” to “OH” transformation in (a) 1, (b) 2-I, (c) 3, and (d) 4 crystals. After UV irradiation using a Schott UG11 filter
for (a) 15 s, (b) 30 s, (c) 15 s, and (d) 180 s. The solid-line spectra were taken at a temperature of 200 K (150 K for (b) and show the absorption
bands of the “NH” form around 600 nm as well as the “OH” form below 500 nm. Additional irradiation into the “NH” band using a Schott OG550
cutoff filter induces a photoreaction of the “NH” form. The dashed line spectra were taken at the same temperatures after (a) 60 s, (b) 1020 s, (c)
60 s, and (d) 300 min.
Figure 13. Time-dependent spontaneous decay of the “NH” absorption
band of 2-I at 308 K.
Figure 14. Time-dependent spontaneous decay of the “NH” absorption
band of 4 at 300 K.
rates were measured, even though all of the molecules have
similar or identical (in the case of deuterated and nondeuterated
analogues) internal coordinates around the reaction center. For
instance, the lifetime of the blue band decay in 2-I crystal is
1230 s, which is about 7 times longer than that measured for
the decay of the same band in the parent 1 crystal (177 s24f).
The decay curves of both deuterated crystals 1-d2 and 2-d2-I
are significantly longer than those of their nondeuterated
analogues. Since in both cases the crystal structures were found
to be independent of the isotope composition, it is evident that
the difference in the lifetimes of the thermal decoloration process
must be attributed to a kinetic isotope effect. The strong isotope
effects, kH/kD ) 4.1 for 1 and kH/kD ) 5.6 for 2-I at 298 K,
suggest that the rate-limiting step of the observed process
involves mainly a proton translocation. These findings are in
marked contrast to many other proton-transfer systems such as
N-salicylideneanilides,20 where no isotope effect could be
observed. In such systems, the long lifetime of the proton-
transferred state was attributed to a skeletal cis-trans reorga-
nization associated with the proton motion.21 Furthermore, it
was previously postulated that long-lived proton-transferred
states must be stabilized by (and therefore associated with) some

16184 J. Phys. Chem., Vol. 100, No. 40, 1996
Scherl et al.
analogues, 97 ( 2.1 kJ mol-1 vs 90.4 ( 1.724 kJ mol-1 for
1-d2 vs 1 and 90.0 ( 3 kJ mol-1 vs 87.4 ( 0.8 kJ mol-1 for
2-d2-I vs 2-I, respectively. No significant difference could be
observed in the preexponential factor, as is expected for a purely
thermally activated proton-transfer reaction. The main contribu-
tion for a primary isotope effect originates from the difference
in the zero-point energies in the educt and transition state, which
appears as a difference in the energies of activation between
the deuterated and nondeuterated compound. A calculation of
the Arrhenius parameters for the ground-state reaction “NH”
f “CH”, applying the RRKM theory for unimolecular reac-
tions22 was carried out using a normal-coordinate analysis
(MOPAC 6.0 PM3 FORCE calculation), based on the minimized
geometries obtained from a PM3-SCF calculation.23 The results
of the calculations and the experimental values are compared
in Table 4. Despite of the differences in the absolute values of
the calculated activation energies, (discussed in ref 23), the
difference between isotopically different molecules fits the
experimental values within the experimental error. The calcu-
lated preexponential factors for the nondeuterated and deuterated
crystals are in good agreement with the experimental results
implying that the ground-state reaction is dominated by thermal
activation.
Figure 15. Time dependence of the “NH” band intensity at 340 K
(342.5 for e) for the different crystals: (a) 1, (b) 1-d2, (c) 2-I, (d) 2-d2-
I, and (e) 3.
The kinetic parameters for 3 resemble those measured for
the parent compound 1, while the kinetic behavior of its isomer
4 is very different. The decay curves of 4 could be correlated
to two identical processes having different barriers and preex-
ponential factors. These present the two extreme examples in
this series of the lowest barrier (77.8 ( 2.5 kJ mol-1) and the
lowest preexponential factor (1010.1(0.3 s-1) as well as the highest
barrier (108.4 ( 8 kJ mol-1) with the highest preexponential
factor (1014.2(1.4 s-1). The two sets of very different thermo-
dynamic parameters originate probably from the same species
occupying different environments. No direct correlation be-
tween the activation energy and observed lifetime could be
drawn. For instance, at ambient temperatures (325 K), the
longest lived species 2-d2-II shows a lifetime of 29 000 s (∼8
h) with a barrier of 90.0 ( 3.0 kJ mol-1, while the shortest
lived species 4 shows a lifetime of 270 s (4.5 min) with a barrier
height of 77.8 ( 1.7 kJ mol-1. In these systems, the differences
in the lifetime of the proton-transferred states clearly originate
from both the changes in the barrier and the preexponential
factor. Further development of improved systems having
controllable properties will require the uncovering and mastering
of new schemes to the selective control of these two parameters.
Figure 16. Arrhenius type diagram of the “NH” form decay constants
as a function of temperature in the different crystals: (a) 1, (b) 1-d2,
(c) 2-I, (d) 2-d2-I, (e) 3, and (f) 4 (plotted separately for reasons of
clarity).
TABLE 4: Kinetic Parameters Measured for the “NH”
Decay Process
crystal τ [s] (T ) 325 K) Ea [kJ/mol] calc log(preexp) calc
1
1-d2
2I
2-d2-I
3
4
750
3080
5200
29000
910
270
400
90.4 ( 1.7 165 11.7 ( 0.3 11.1
97.0 ( 2.1 171 12.1 ( 0.3 11.1
87.4 ( 0.8 164 10.3 ( 0.2 12.4
90.0 ( 3.0 169 10.0 ( 0.5 12.3
The plot of the observed activation energies vs the distance
which the proton must cross during the “NH” f “CH”
conversion, assuming no structural reorganization, shows no
straightforward correlation (Figure 17). The differences in the
“NH” to “CH” span only a very short range (∼0.3 Å) and cannot
explain the large differences in the observed barriers and
lifetimes. Our preliminary simulation of the “NH” f “CH”
conversion using an SCF calculation (MOPAC 6.0 PM3
calculation) based on the crystal structure geometry24g as well
as IRC (intrinsic reaction coordinates) calculations on the free
molecule23 indicated that this process may involve the o-nitro
group acting as a chaperone and escorting the proton along its
path. In addition, there are some other proton accepting groups
of neighboring molecules in close proximity (see Table 3), which
could, in principle, be involved in the complex mechanism of
the proton transfer. Hence the overall transfer mechanism is
not solely dependent on the distance that the proton must cross
but also on the relative orientation and position of the neighbor-
ing electronegative atoms. Thus, the stability of the “NH”
photoproduct in the crystalline state is mainly governed by the
95.0 ( 0.8
77.8 ( 2.5
108.4 ( 8.0
12.3 ( 0.2
10.1 ( 0.3
14.2 ( 1.4
kind of a structural reorganization.17a These systems appear to
be the first long-lived photochromic systems based on a pure
proton-transfer process.
Figure 16 presents an Arrhenius type diagram for the decay
constants of the “NH” form as a function of temperature in the
different crystals. In all cases, the data could be fitted to a linear
function, indicating a pure thermally activated process, as was
previously reported for 1.15a Table 4 summarizes some kinetic
parameters measured for the “NH” f “CH” conversion. The
activation energies found vary from 78 to 108 kJ mol-1
,
depending on the crystal structure and isotope composition. In
all cases the measured barriers were considerably higher than
the kinetic barriers measured in solutions for the same process.15c
The kinetic barriers for the “NH” f “CH” conversion in the
deuterated crystals are slightly higher than in their nondeuterated

Proton Transfer in Well-Defined Media
J. Phys. Chem., Vol. 100, No. 40, 1996 16185
TABLE 5: Kinetic Parameters Measured for the “OH”
Decay Process
crystal
Ea [kJ/mol]
calc
log(preexp)
calc
1
68.6 ( 3.0
72.0 ( 3.0
30.1 ( 1.3
102.5 ( 0.8
62
71
169
11.4 ( 0.6
11.6 ( 0.6
6.8 ( 0.4
13.4 ( 0.6
13.0
13.0
12.6
1-d2
2I
3
that very similar structures exhibit very different kinetic
parameters. For instance, in the 2-I crystal, the energy barrier
for the “OH” decay is only 30.1 kJ mol-1 with a preexponential
factor of 106.8(0.4 s-1, while in the 3 crystal the barrier for the
same process is 102.5 kJ mol-1 with a preexponential factor of
1013.4(0.6 s-1. As in the case of the “NH” state, no argument
based on molecular level processes and properties appears to
allow the rationalization of such dramatic differences in the
barriers and preexponential factors. They must thus originate
from the supramolecular arrangement of the proton-transferred
molecule and its first layer of neighboring molecules. Further
evidence lies in the remarkable stability of the “OH” form in
the 3 crystal. In this system the “OH” form is more stable than
the “NH” form, in marked contrast to all other photochromic
crystals and to all earlier observations in all other noncrystalline
phases.15c This phenomenon could not be rationalized by
intramolecular effects either, as clearly the -NO2H must be
less stable than a protonated pyridine group.
Figure 17. Plot of the observed activation energies vs. the distance
the proton must cross during the “NH” f “CH” process, for crystals
of (a) 1, (b) and (c) 2-I, (d) 3 and the two exponents of 4, (e) and (f)
assuming no structural reorganization. The “NH” form was calculated
using the crystal structure and placing the H at the pyridine N, with a
bond length of 1.00 A
Conclusions
Different photoinduced and thermal processes taking place
in photochromic 2-(2,4-dinitrobenzyl)pyridine were followed
using UV-vis and IR spectroscopy. All characterized processes
were found to have a relatively strong primary isotope effect,
indicating that the rate-determining steps in all of them are pure
proton-transfer processes. All other possible reorganizations
must have lower barriers. These findings together with the high
barriers of ca. 100 kJ mol-1 suggest that it may be possible to
design photochromic systems based on proton-transfer processes
having bistable properties. These systems are the first long-
liVed photochromic systems based on a pure proton-transfer
process.
Figure 18. Arrhenius type diagram of the “OH” form decay constants
as a function of temperature in: a) 1, (b) 1-d2, (c) 2-I, and (d) 3 crystals.
environment. Environmental factors determine not only the
molecular internal coordinates but also the reactivity through
the involvement of electronegative atoms of the cage molecules
surrounding the reaction center.
All investigated dinitrobenzylpyridine derivatives form crys-
talline phases with similar molecular internal coordinates. Not
all phases were found to be photochromic. This phenomenon
cannot be explained on the basis of intramolecular features as
6-(2,4-dinitrobenzyl)-2,2′-bipyridine (2) crystallizes in two
different phases, one photochromic and the other not. The free-
volume argument, previously utilized for explaining the same
phenomenon in the N-salicylideneanilides, cannot hold in this
case since similar crystal densities were measured for the
different crystals. Additionally, the internal coordinates around
the reaction center are very similar and would not explain the
different photochromic properties. Two arguments are sug-
gested for the rationalization of the observed differences in
photochromic properties. (a) The different photochromic
properties may arise from the differences in the packing of the
molecules in the crystal. In nonphotochromic crystals, the 2,4-
dinitrophenyl chromophore is tightly stacked to neighboring
aromatic rings, while in the photochromic crystals this chro-
mophore is isolated. The stacking of the chromophore to other
aromatic rings may offer electronic deactivation pathways faster
then the proton-transfer process. (b) The fact that the proton
abstracted is bound to a carbon atom implies a relatively high
barrier. This may require an important strong activation which
is achieved only if both the oxygen of the o-nitro group and
the nitrogen of the pyridine are involved in abstracting the proton
The “OH” forms of the different crystals decay spontaneously
in the dark, reverting to the “NH” proton-transferred states. In
parallel, some of them partially decay also to the thermody-
namically stable “CH” form. The lifetime of the “OH” state
varies considerably from one system to the other. For instance,
in the 2-I crystals the lifetime of the “OH” form was 6.4 s at
198 K while in 1 crystals the “OH” form is much longer lived
with an estimated lifetime of ca. 55 000 s at the same
temperature. The “OH” form in the 3 crystals is practically
stable under these conditions. The lifetime of the “OH” form
in the 1-d2 crystal is significantly longer than the lifetime of its
nondeuterated analogue 1, indicating a strong isotope effect,
kH/kD ) 3.8 at 250 K. This implies that the rate-limiting step
of the “OH” decay process is also dominated by a proton
translocation. Figure 18 presents an Arrhenius type diagram
of the “OH” form decay constants as a function of temperature
in 1, 1-d2, 2-I, and 3 crystals. Some kinetic parameters of the
different “OH” forms are presented in Table 5. It is clear that
as for the “NH” case, the differences in the stability of the “OH”
forms are governed by a combination of variations in the
preexponential factor and the activation energy. It is evident

16186 J. Phys. Chem., Vol. 100, No. 40, 1996
Scherl et al.
(7) (a) Haarer, D. In Presistent Spectral Hole Burning: Science and
Applications. Top. Curr. Phys. 1988, 44, and references therein. (b)
Ku¨mmerl, R. A.; Scherl, M.; Haarer, D. Opt. Memories Neural Networks
1994, 3, 377. (c) Friedrich, J.; Haarer, D. Angew. Chem., Int. Ed. Engl.
1984, 23, 133. (d) Castro, G.; Haarer, D.; Macfarlane, R. M.; Trommsdorff,
H. P. US Patent 4101976, 1976.
from the benzylic carbon. The arrangement in which both the
pyridine nitrogen and the o-nitro oxygen bind to the same proton
and are located on the same side of the molecule is found only
in the photochromic crystals.
Environmental effects are not restricted to the control of
excited-state properties. The significant changes in the observed
kinetic barriers and preexponential factors for the different
processes (Tables 4 and 5) could not be rationalized upon
utilizing molecular parameters. Preliminary simulations suggest
the involvement of the o-nitro group as a chaperone escorting
the proton along its path. The presence of other neighboring
basic groups in close vicinity implies a more complicated
picture. Clearly, the further development of better photochromic
systems based on proton transfer, having controllable properties
will require the uncovering and mastering of new schemes for
the control of these environmental factors of supramolecular
nature.
(8) Ireland, J. F.; Wyatt, P. A. H. AdV. Phys. Org. Chem. 1976, 12,
131.
(9) Schmidt, G. M. J. In Solid State Photochemistry; Ginsburg, D.,
Ed.; Verlag-Chimie: Weinheim, 1976.
(10) (a) Organic Photochemistry; Padwa, A., Ed.; Marcel Dekker: New
York; Vol. 8 and 11. (b) Huffman, K. R.; Loy, M.; Ullman, E. F. J. Am.
Chem. Soc. 1965, 87, 5417.
(11) (a) Kawato, T.; Koyama, H.; Kanatomi, H.; Isshiki, M. J.
Photochem. 1985, 28, 103. (b) Kawato, T.; Kanatomi, H.; Koyama, H.;
Igrashi, T. J. Photochem. 1986, 33, 199. (c) Koyama, H.; Kawato, T.;
Kanatomi, H.; Matsushita, H.; Yonetani, K. J. Chem. Soc., Chem. Commun.
1994, 579.
(12) Reetz, M. T.; Hoger, S.; Harms, K. Angew. Chem., Int. Ed. Engl.
1994, 33, 181; Agnew. Chem. 1994, 106, 193.
(13) (a) Miati, T.; Saito, G.; Urayama, H. Phys ReV. Lett. 1988, 60,
2299. (b) Miati, T. Mol. Cryst. Liq. Cryst. 1989, 171, 343.
(14) Hadjoudis, E. In ref 1a and references therein.
(15) (a) Sixl, H.; Warta, R. Chem. Phys. 1985, 94, 147. (b) Seff, K.;
Trueblood, K. N. Acta Crystallogr. 1968, B24, 1406.
(16) Chichibabin, A. E.; Kunidzhi, B. M.; Benewolenskaja, S. W. Ber.
Dtsch. Chem. Ges. 1925, 28, 1580.
Further experiments, designed to better understand the
correlation between the structure and reactivity in such systems,
are underway in our laboratories.
(17) (a) ReactiVity of the Photoexcited Organic Molecule; Schmidt, G.
M. J. Interscience: New York, 1967. (b) Bergman, J.; Leiserowitz, L.;
Schmidt, G. M. J. J. Chem. Soc. 1964, 2068. (c) Kawato, T.; Koyama, H.;
Kanatomi, H.; Isshiki, M. J. Photochem. 1985, 28, 103. (d) Kawato, T.;
Kanatomi, H.; Koyama, T.; Igarashi, T. J. Photochem. 1986, 33, 199. (e)
Kawato, T.; Koyama, H.; Kanatomi, H.; Tagawa, H.; Iga, K. J. Photochem.
Photobiol. A: Chem. 1994, 78, 71.
(18) (a) Hoshino, N.; Inabe, T.; Mitani, Maruyama, Y. Bull. Chem. Soc.
Jpn. 1988, 61, 4207. (b) Inabe, T.; Gautier-Luneau, I.; Hoshino, N.;
Okaniwa, K.; Okamoto, H.; Mitani, T.; Nagashima, U.; Maruyama, Y. Bull.
Chem. Soc. Jpn. 1991, 64, 801. (c) Wozniak, K.; He, H.; Klinowski, J.;
Jones, W. J. Chem. Soc., Faraday Trans. 1995, 91, 77. (d) Inabe, T.;
Hoshino, N.; Mitani, T.; Maruyama, Y. Bull. Chem. Soc. Jpn. 1989, 62,
2245.
(19) (a) Clark, W. C.; Lothian, G. F. Trans. Faraday Soc. 1958, 54,
1790. (b) Hardwick, R.; Mosher, H. S.; Assailaigue, P. P. Trans. Faraday
Soc. 1960, 56, 44.
(20) Cohen, M. D.; Schmidt, G. M. J.; Flavian, S. J. Chem. Soc. 1964,
2041.
(21) Anderson, D. G.; Wettermark, G. J. Am. Chem. Soc. 1965, 87,
1433.
(22) The Theory of Unimolecular and Recombination Reactions; Gilbert,
R. G., Smith, S. C., Eds.; Blackwell: Oxford, 1990.
Acknowledgment. This work was supported by the EC
ESPRIT program (No. 7238). The authors have deposited
atomic coordinates for all structures with the Cambridge
Crystallographic Data Center. The coordinates can be obtained,
on request, from the Director, Cambridge Crystallographic Data
Center, 12 Union Road, Cambridge CB2 1EZ, UK.
References and Notes
(1) (a) Photochromism: Molecules and Systems. Studies in Organic
Chemistry 40; Durr, H., Bouas-Laurant, H., Eds.; Elsevier: Amsterdam,
1990; Chapters 23-29, and references therein. (b) Irie, M. Iga, R.
Macromolecules 1986, 19, 2480. (c) Mamada, A.; Tanaka, T.; Kung-
watchakun, D.; Irie, M. Macromolecules 1990, 23, 1517. (c) Shvartsman,
F. P.; Krongauz, V. A. J. Phys. Chem. 1984, 88, 6448. (d) Anzai, J.-C.;
Sakamura, K.; Osa, T. J. Chem. Soc., Chem. Commun. 1992, 888. (e)
Willner, I.; Rubin, S.; Shatzmiller, R.; Zor, T. J. Am. Chem. Soc. 1993,
115, 8690. (f) Optical Processing and Computing; Arsenault, A., Ed.;
Academic Press: New York, 1989. (g) Brige, R. R. Annu. ReV. Phys. Chem.
1990, 41, 683. (h) Friedrich, J.; Haarer, D. Angew. Chem., Int. Ed. Engl.
1984, 23, 133. (i) Ao, R.; Ku¨mmerl, L.; Haarer, D. AdV. Mater. 1995, 7,
495.
(2) (a) The Proton in Chemistry, 2nd ed.; Bell, R. P., Ed.; Cornell
University Press: Ithaca, NY, 1973. (b) Weller, A. Prog. React. Kinet. 1961,
1, 189. (c) ReactiVity of Photo-excited Organic Molecules; Porter, G., Ed.;
Interscience Publishers, Wiley: London, 1967. (d) Kosower, E. M.; Huppert,
D. Annu. ReV. Phys. Chem. 1986, 37, 127. (e) Marcus, R. A. J. Phys. Chem.
1968, 72, 891. (f) Photochromism: Molecules and Systems. Studies in
Organic Chemistry 40; Chapters 16 and 17, Du¨rr, H., Bouas-Laurant, H.,
Eds.; Elsevier: Amsterdam, 1990.
(23) Frank, I.; Grimme, S.; Peyerimhoff, S. D. J. Phys. Chem. 1996,
100, 16187.
(24) (a) Center-to-center distances. (b) Using a repetitive grinding
irradiating cycle until the powders did not change color in grinding. (c)
Corval, A.; Kuldova´, K.; Eichen, Y.; Pikramenou, Z.; Lehn, J. M.;
Trommsdorff, H. P. in preparation. (d) The relative quantum yield is given
by the formation of the phototautomer, assuming that the absolute quantum
yield is low (<0.2, in our experiments, no saturation effect could be observed
in the high-temperature regime) and other nonradiative deactivation
processes are not temperature dependent. (e) Relative quantum yields were
measured by normalizing the light flux using a fulgide dye as actinometer25
and assuming similar extinction coefficients for the different “NH”
tautomers. (f) Our measured barriers and lifetimes are somewhat different
from those reported in ref 15a. (g) Calculation were performed using a
data set where the coordinate of all atoms but the pyridine nitrogen, the
benzylic protons, and the o-nitro group are fixed.
(3) Stoeckenius, W.; Bogomolni, R. A. Annu. ReV. Biochem. 1982,
52, 587.
(4) (a) Osterhelt, D.; Brau¨chle, C.; Hampp, N. Q. ReV. Biophys. 1991,
24, 425. (b) Brau¨chle, C.; Hampp, N.; Osterhelt, D. AdV. Mater. 1991, 3,
420.
(5) Ku¨mmerl, L.; Wo¨hrl, H.; Haarer, D. Chem. Phys. Lett. 1994, 227,
337. (b) Braun, J.; Schlabach, M.; Wehrle, B.; Ko¨her, M.; Voegl, E.;
Limbach, H. H. J. Am. Chem. Soc. 1994, 116, 6593.
(6) Small, G. J. In Spectroscopy and Excitation Dynamics of
Condensed Molecular Systems; Agranovich, V. M., Hochstrasser, R. M.,
Eds.; North-Holland Publishing Co.: Amsterdam, 1983.
(25) Heller, H. G.; Langan, J. R. J. Chem. Soc., Perkin Trans. 2, 1981,
341.
JP9609242