Unidirectional cis-trans photoisomerization of cis-3,3′-bis(diphenylhydroxymethyl)stilbene in inclusion complex crystals

Article abstract of DOI:10.1002/poc.693

When a 1:1 complex of cis-3,3′-bis(diphenylhydroxymethyl)stilbene (1) with acetone was irradiated in the solid state at room temperature using a high-pressure mercury lamp with a Pyrex filter for 6 h, a 1:1 complex of the trans-isomer (2) with acetone was

Full text of DOI:10.1002/poc.693

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2003; 16: 905–912
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.693
Unidirectional cis–trans photoisomerization of
cis-3,30-bis(diphenylhydroxymethyl)stilbene in
inclusion complex crystals
Koichi Tanaka,1 Takaichi Hiratsuka,1 Shigeru Ohba,2 M. Reza Naimi-Jamal3 and Gerd Kaupp3*
1Department of Applied Chemistry, Faculty of Engineering, Ehime University, Matsuyama, Ehime 790-8577, Japan
2Department of Chemistry, Keio University, Hiyoshi 4-1-1, Kohoku-ku, Yokohama 223-8521, Japan
3Faculty 5, Organic Chemistry I, University of Oldenburg, D-26111 Oldenburg, Germany
Received 15 April 2003; revised 20 June 2003; accepted 24 June 2003
ABSTRACT: When a 1:1 complex of cis-3,30-bis(diphenylhydroxymethyl)stilbene (1) with acetone was irradiated in
the solid state at room temperature using a high-pressure mercury lamp with a Pyrex filter for 6 h, a 1:1 complex of the
trans-isomer (2) with acetone was obtained without losing the guest acetone molecules. Similar results were obtained
without a guest or in the presence of other guests. Internal rotation around the double bond bearing the very large
triphenylhydroxymethyl substituent is hardly imaginable in the confinement of the crystal. Therefore, the geome-
trically easier ‘hula-twist’ (HT) mechanism with its movements primarily in the molecular plane provides a viable
mechanistic alternative for the cis–trans isomerization. The solid-state mechanism of the reaction was studied by
x-ray and atomic force microscopy (AFM) analyses. Face-selective efflorescence formed a protective cover on (100)
and can be related to the crystal structure. Further molecular migration was not detected with the sensitivity of AFM
and the crystal of 1ꢀacetone did not lose its clarity and microscopic shape. The cis–trans conversion profits from cages
that are present in the crystal lattice. X-ray structural analysis confirmed a strong loss of crystallinity upon the
photochemical conversion of 1ꢀacetone which precludes a definitive conformational proof of HT by x-ray diffraction.
The formation of the amorphous phase excludes a topotactic reaction. Copyright # 2003 John Wiley & Sons, Ltd.
KEYWORDS: cis–trans isomerization; inclusion crystal; x-ray structure; atomic force microscopy; hula-twist; amorphous
phase
INTRODUCTION
benzoate (3) with ethyl 3-ethynylbenzoate (4) in the pre-
sence of a Pd(II) catalyst gave 3,30-bis(carboethoxy)
diphenylacetylene (5), which upon treatment with PhMgBr
in dry THF followed by hydrogenation on Pd–BaSO4 gave
cis-3,30-bis(diphenylhydroxymethyl)stilbene (1). The host
compound (1) formed stable inclusion complex crystals
with various kinds of guest compounds (Table 1). For
example, when 1 was recrystallized from acetone, a 1:1
inclusion crystal was obtained as colorless prisms.
The various host–guest compounds of cis-1 and also
guest-free 1 (see Experimental) underwent solid-state
geometric isomerization to give the trans-isomer 2.
Very interestingly, when a single crystal (3.0 ꢁ 1.0 ꢁ
0.5 mm) or powder of the 1:1 inclusion complex of 1
with acetone (1ꢀacetone) was irradiated with a 400 W
high-pressure mercury lamp using a Pyrex filter from a
distance of 10 cm at room temperature for 12 or 6 h,
respectively, a pale yellow crystal (or powder) of the 1:1
acetone complex of 2 with acetone (2ꢀacetone) was
formed. A 92% conversion of the powder was obtained.
After photolysis, the single crystal was still clear and the
microscopic crystal shape had not changed. It is also
surprising that the guest acetone molecules were mostly
held in the crystal (about 94% retained as detected by
Photochemical cis–trans isomerizations in constrained
media have attracted special interest in recent years in
relation to the photoisomerization of chromophores in
biological systems (e.g. photoactive yellow protein and
retinoid-binding proteins).1 Recently, various cis–trans
—
isomerizations of C C double bonds, e.g. of 2-benzyli-
—
denebutyrolactone and 1,2-dibenzoylethene, in the crystal-
line state have been reported.2,3 We now report the
unidirectional photoisomerization of cis-3,30-bis(diphenyl-
hydroxymethyl)stilbene (1) to its trans-isomer (2), neat
and in inclusion crystals with guest molecules (Scheme 1).
RESULTS AND DISCUSSION
Preparations and solid-state photolyses
The host compound (1) was prepared according to the
method shown in Scheme 2. Treatment of ethyl 3-bromo-
*Correspondence to: G. Kaupp, Faculty 5, Organic Chemistry I, Uni-
versity of Oldenburg, P.O. Box 2503, D-26111 Oldenburg, Germany.
E-mail: kaupp@kaupp.chemie.uni-oldenburg.de
Copyright # 2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 905–912

906
K. TANAKA ET AL.
thermogravimetric analysis) despite the drastic change in
the shape of the host molecule during isomerization.
Similarly, the cis–trans isomerization of 1 and of its
inclusion crystals with cyclopentanone, ꢀ-butyrolactone,
dioxane, DMSO, DMF and pyridine occurred efficiently
(although slowly) in the solid state without loss of the
guests (Table 1). On the other hand, prolonged irradiation
of 2 or of its acetone inclusion complex 2ꢀacetone did not
change the material: no trace of 1 could be detected by 1H
NMR analysis. The unidirectional photoisomerization of
1 to 2 in the inclusion complex with acetone was
investigated by x-ray diffraction and atomic force micro-
scopy (AFM) analysis.
Scheme 1. Solid-state photoisomerization of 1 to give 2
X-ray analysis
The molecular structure in the crystals of 1ꢀacetone is
Scheme 2. Synthesis of the host compound 1
shown in Fig. 1. The acetone molecule is connected to the
Table 1. Host–guest complexes of cis-3,30-bis(diphenylhydroxymethyl)stilbene (1) and their photoisomerization in the solid-
statea
Guest
Host:guest ratio
Z: m.p. (ꢂC)
Conversion (%)
Host:guest ratio
E: m.p. ( ꢂC)
Acetone
Cyclopentanone
ꢀ-Butyrolactone
Dioxane
DMSO
DMF
Pyridine
—
1:1
1:1
1:1
1:1
1:2
1:2
1:1
—
77–82
75–83
85–90
110–115
130–135
78–83
92
93
95
71
68
72
92
93
1:1
1:1
1:1
1:1
1:2
1:2
1:1
—
70–75
65–70
55–60
85–90
120–130
90–95
75–78
68–73
93–95
170–175
a
The photolyses of powdered crystals were carried out for 6 h.
Figure 1. The molecular structure of 1ꢀacetone with displacement ellipsoids at the 50% probability level
Copyright # 2003 John Wiley & Sons, Ltd. J. Phys. Org. Chem. 2003; 16: 905–912

CIS–TRANS ISOMERIZATION OF A STILBENE IN INCLUSION CRYSTALS
907
host 1 by the O2—H2ꢀ ꢀ ꢀO3 hydrogen bond with an
˚
O2ꢀ ꢀ ꢀO3 distance of 2.829(4) A and it is not disordered.
There is also an intramolecular O1—H1ꢀ ꢀ ꢀO2 hydrogen
bond of the host molecule [O1ꢀ ꢀ ꢀO2 distance 2.993(3)
˚
A]. The dihedral angle between the two phenyl rings
of the cis-stilbene moiety is 47.8(1)ꢂ and the central
ꢂ
—
C21—C23 C24—C25 torsion angle is 7.5(8) .
—
Supermicroscopic analysis with AFM
The crystals of 1ꢀacetone keep their guest upon photo-
isomerization. However, at the molecular resolution (in the
Z-direction) of AFM, surface reconstruction or efflores-
cence may occur with loss of some acetone in the surface
region. This seems to be a problem at room temperature,
but less so at ꢃ17 to ꢃ15 ꢂC (Fig. 2). Low-temperature
irradiation is also required because there are signs of local
melting if high-intensity irradiation (700W Hg lamp
through Pyrex cooled to 14ꢂC) is performed at room
temperature: initially sharp crystal edges become
smoother and AFM [Fig. 3(d)] detects softened surface
regions, but not in irradiations at <ꢃ15 ꢂC. Both almost
acetone-free (<6%) crystals of 1 and the crystals of
1ꢀacetone undergo cis–trans photoisomerization also at
ꢃ17 to ꢃ15 ꢂC and (more slowly) at ꢃ70ꢂC, as checked
Figure 2. Micrographs of a crystal of 1ꢀacetone: the white
bar corresponds to 0.1 mm. Left image, (010) face before
irradiation; right image, after 6 h of irradiation with a 700 W
Hg high-pressure lamp through Pyrex at a 15 cm distance and
ꢃ17 to ꢃ15 ꢂC. The irradiated crystal was tilted in order to
visualize both the (110) face and the (010) face. Two small
crystallites on (010) are the result of some remote crystal
breakage that happened on tilting it with tweezers
crystals. There was the expectation that this highly space
demanding cis–trans isomerization had occurred in a
topotactic manner. Therefore, the AFM results in combi-
nation with x-ray diffraction are of particular interest for
the substantiation or exclusion of that assumption. AFM
has been used successfully to prove and disprove several
claims of topotaxy in various molecular crystals very
sensitively.4
The main surface (100) of a flat plate-like crystal of
1ꢀacetone was irradiated at a low light intensity at 20–
25 ꢂC and the surface change was measured at various
times. Figure 3 shows that the initial surface has some
1
by H NMR analysis. If a prismatic crystal of 1ꢀacetone
was irradiated for 6 h at ꢃ17 to ꢃ15 ꢂC through a Pyrex
filter, no microscopic change was observed even though
the crystal was 30% converted to 2ꢀacetone.
Long irradiation (20 h; the irradiation times for high
conversion are rather long owing to internal filtering by
the strong light absorption of the product 2) up to nearly
complete conversion also did not disintegrate the single
Figure 3. 10 mm AFM topographies of the (100) face of 1ꢀacetone: (a) fresh surface of a flat plate specimen; (b) after 4 h; (c)
after 10 h of irradiation with an old 125 W Hg lamp through Pyrex at a 3 cm distance from the lamp and 20–25 ꢂC; (d) after 3 h
of irradiation with a 700 W Hg lamp at room temperature; (e) a fresh surface after 5 and (f) after 15 min irradiation with a 700 W
Hg lamp at ꢃ17 to ꢃ15 ꢂC and a 15 cm distance from the lamp; the Z-scale is 100 nm
Copyright # 2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 905–912

908
K. TANAKA ET AL.
roughness (Rms ¼ 7 nm, excluding the large feature) with
a maximum height of 62 nm (excluding the large feature)
obviously from some efflorescence on resting for 2 weeks
in a bottle without an acetone atmosphere. These values
increase to 12 and 88 nm in (b) and 9 and 85 nm in (c). It
is clearly seen [(a) and (b)] how the original positive
features increase while the craters disappear. That change
comes to a stop in Fig. 3(c). Clearly, a thin cover is
formed on (100) which prevents further efflorescence,
while the conversion to 2ꢀacetone in the bulk exceeds
50%. If a much higher light intensity from a fresh 700 W
Hg lamp is applied, similar features form within minutes
(Rms ¼ 7 nm, maximum height 55 nm), that keep their
shape but collapse after 3 h of irradiation followed by
apparent local melting [Fig. 3(d)].
A very flat surface part of 1ꢀacetone on (100)
(Rms ¼ 3.5 nm, maximum height 48 nm) was irradiated
under argon at ꢃ17 to ꢃ15 ꢂC with the 700 W Hg lamp. It
gave comparable signs of efflorescence with fewer hills
and formation of craters around them on (100)
(Rms ¼ 12 nm, maximum height 116 nm) [Fig. 3(e/f)].
The surface did not change significantly further after
15 min of irradiation (Rms ¼ 12 nm, maximum height
125 nm). Therefore, these features are again interpreted
as the formation of a cover that is the result of initial
efflorescence but not of bulk photoisomerization.
Figure 4. 7.5 mm AFM topographies of 1ꢀacetone on its
(010) face [frequently larger than (100) in its well-formed six-
sided prisms]: (a) fresh; (b) after 4 h, (c) after 6 h and (d) after
9 h of irradiation under argon at ꢃ15 to ꢃ17 ꢂC with a
700 W Hg lamp through Pyrex at a 15 cm distance. We do
not see precisely the same site on the (010) face in (a)–(d), as
the sample holder had to be removed from the AFM stage
for the irradiations. The Z scale is 140 nm
Well-formed six-sided prismatic crystals of 1ꢀacetone
were probed on all of the paired faces along the c-axis.
The now smaller developed (100) face was securely
identified by the features similar to those in Fig. 3(e)
upon irradiation. The (010) surface (Fig. 2) had humps
parallel to the long crystal c-axis and was scanned with
the AFM tip at sites where these were least pronounced.
No change was detectable. Even very long irradiation did
not change the shape of the side-face, as shown in Fig. 4.
The conversion of the crystal in Fig. 4(d) was detected
The AFM results on (100) are in agreement with the
molecular packing: acetone molecules may escape from
the uppermost monolayer and thus start the nucleation for
the efflorescence also from lower layers which finally
creates a thin tight cover with the hills that are seen in Fig.
3. At lower temperature, the nucleation is more selective
and fewer hills can grow to a limiting height. After
buildup of the stable layer, no further distortion due to
efflorescence occurs. The (010) face can only lose the
acetone molecules that are directly at the outer surface.
The next layer acetone molecules are shielded by host
molecules and cannot escape (Fig. 5). The (110) face
exhibits a similar situation. Thus, protecting efflorescence
is only observed on (100). Furthermore, as all molecular
layers interlock, no molecular migrations apart from the
formation of the protecting efflorescence cover can occur
and no photoproduct molecules exit, as shown in Fig. 4.
How, then, might the extremely space-demanding
cis–trans isomerization of 1 to give 2 occur without
destruction of the crystal? The difficulties are evident
from a comparison of the molecular shapes of 1 (in
1ꢀacetone) and 2 (in 2ꢀ0.75 benzene)5 in Fig. 6. The
vinylic substituents are very large. The huge geometric
change is clearly seen in addition to the conformational
difference in the two experimental shapes of these mole-
cules in the crystals. The geometry of the ‘90ꢂ rotated
transition state’ for a hardly imaginable double bond
rotational isomerization cannot be assessed and is there-
fore not modeled and depicted. A closer inspection of the
1
with H NMR to be 56%. Similarly, no change of the
surface was detected on the (110) face on irradiation
under the same conditions. Clearly, no long-range move-
ments of molecules [except for the initial efflorescence on
(100)] can be detected with AFM and this observation
points to either a topotactic reaction or the formation of
an amorphous phase in the crystal of 1ꢀacetone to give
2ꢀacetone. In both cases no new crystalline phase has to
be formed.
Mechanistic interpretation
The crystal structure of 1ꢀacetone (P21/n) exhibits a
layered structure on (010) and (001). The monolayers
are in the (100) and (010) planes. All hydrogen bonds are
within the host–guest pairs (see also Fig. 1). Thus, the
layers are not interlinked by hydrogen bonds, but they
interlock considerably and do not leave good cleavage
planes for easy molecular migrations.2 This is shown in
Fig. 5, which represents a rather large crystal section.
Copyright # 2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 905–912

CIS–TRANS ISOMERIZATION OF A STILBENE IN INCLUSION CRYSTALS
909
Figure 5. Stereoscopic representation of the crystal packing in 1ꢀacetone on (001) with the (100) face on top and (010) to
the right; acetone can escape from the monolayers on (100) but not through the monolayers on (010); hydrogen atoms are
omitted for clarity
lattice in Fig. 5 reveals the presence of cages that appar-
ently can help to accommodate the more extended mole-
cules 2 which avoids the necessity of leaving the lattice by
far-reaching migration.2 Semiempirical PM3 calculations
indicate only minor changes of the space-filling CPK
volumes (0.3%) and surfaces (2.3%) between 1 and 2.
This is not enough to invoke static ‘cavity’ models.6
Dynamic approaches are still qualitative but can be used
in the fixed lattice for judgments in extreme situations. A
rotational mechanism would meet with the situation that it
does not initially profit from the packing as the modeling
arrives at unsurmountable van der Waals interactions
(down to 45% of the standard distance) already at ꢄ10ꢂ
rotation. If that should proceed to a ‘90ꢂ rotated transition
state’ it would require that numerous neighboring mole-
cules in the first, second and third spheres must be pushed
away from their lattice sites during the lifetime of the
electronic excitation while the reacting molecule also
translocates. The vinylic substituents are far too large for
such a process with 1. The ‘hula-twist’ (HT) mechanism of
Liu and co-workers1,7 appears more likely, as it requires
only the out-of-plane translocation of one of the C—H
units and movement of one of the huge substituents
within its molecular plane. Scheme 3 summarizes the
Figure 6. Stereoscopic representation of the space-filling
molecules from x-ray data in their crystals:5 top, 1ꢀacetone;
bottom, 2ꢀ0.75 benzene. O, meridian; double bond C, grid
Scheme 3. Comparison of the rotational and HT mechan-
isms of the cis–trans photoisomerization of 1 showing the
difference in the conformational result
Copyright # 2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 905–912

910
K. TANAKA ET AL.
two mechanisms giving conformers A and B, respectively.
In the HT mechanism, only one of the double bond C—H
undergoes a space saving out-of-plane translocation,
while the m-triphenylhydroxymethyl substituent at the
same C moves essentially within its original plane and
passes a structure that may be drawn similarly to a triangle.
Continuation of the in-plane movement of the aryl moiety
leads to conformer B, which might survive in the confine-
ment of a crystal lattice.3 In the rotational mechanism
either one end (180ꢂ) or both ends (90ꢂ each) of the double
bond with their huge substituents would have to undergo a
space-consuming out-of-plane translocation to form con-
former A.
Figure 5 indicates that HT might occur much more
easily along the vertical layers inasmuch as there is also
some movement possible along the horizontal layer
direction. This mechanism appears much more likely
for geometric reasons, because the doubly layered aniso-
tropic crystal packing is largely in favor (Fig. 5). How-
ever, both mechanisms can still not be quantitatively
modeled. We note that cis–trans photoisomerizations of
stilbenes with even larger dendrimeric substituents have
been interpreted as HT isomerizations on the basis of
quantum yields and fluorescence arguments, although in
solution.8 In that case the ‘confinement’ is the solvent that
would have to be pushed away by the huge substituent if
it were to rotate by ca 90ꢂ during the 10 ns lifetime of the
electronically excited dendrimer.
Most rewarding are the following questions: can the
HT mechanism be proven by x-ray diffraction in the
present case by means of a conformational analysis of
the product 2 (rotation and HT lead to different con-
formers, Scheme 3)3 if the isomerization of 1 were a
topotactic conversion, or is it a Kohlschu¨tter-type topo-
chemical reaction9 leading to colloidal particles within
the shape of the original crystal (not to be confused with
the Schmidt-type topochemical reaction10 which assumes
formation of a crystalline phase with a provision of
minimal atomic and molecular movements)? It turned
out that an amorphous phase was occurring in the
irradiation of 1 and therefore a conformational proof of
HT was not possible with x-ray diffraction on low-
temperature (250 K) irradiation. The observed lattice
changes were very small when the single crystal was
irradiated up to 36% conversion after 3 h. Importantly, the
peak half-width of the x-ray diffraction signals did not
increase, but the peak intensity decreased gradually to
60% after 1 h and to 20% after 3 h of irradiation, while the
crystal remained transparent. The difference electron-
density maps did not show any contribution of the
trans-isomer 2. Clearly, an x-ray amorphous phase was
produced without (sub)microscopic change of the crystal
(Figs 2 and 4). Hence, the present example failed to
provide definitive experimental proof of HT by exclusion
of rotation. However, volume-conserving HT motions,
where the bulky substituents do not leave the planes
occupied by the vinylic benzene rings and only one of
the olefinic C—H units undergoes out-of-plane translo-
cation,1,3 appear most likely, owing to the presence of the
layers and of the cages that serve also for the accom-
modation of the final product 2 (Fig. 5) (the so-called
‘bicycle pedal’ mechanism for 1,3-dienes11 is also vo-
lume conserving, but does not apply here).
CONCLUSIONS
Solid-state reactions without molecular migrations be-
yond geometric relaxation are considered topotactic
processes irrespective of the chemical mechanism, but
they must always be secured at the molecular precision of
the AFM2,4,12 and essentially keep the crystal structure. A
previous claim of a topotactic trans–cis photoisomeriza-
tion of 1,2-dibenzoylethene13 was excluded by AFM as it
exhibited long-range molecular movements along clea-
vage planes.4,14 The initial hopes to realize the first
topotactic cis–trans photoisomerization with 1ꢀacetone
crystals that did not (sub)microscopically change were
not fulfilled, as an amorphous phase occurred by neces-
sarily short-distance molecular movements that could not
be traced by AFM on the rough surfaces. This shows that
AFM is a necessary but not a sufficient condition for the
proof of topotaxy, as an amorphous phase may form
without detectable molecular migrations at the molecular
Z-scale. The unusual present results further challenge
conclusions about topotaxy that are only derived from
dynamic x-ray investigations (without invoking AFM
evidence), if there was a considerable drop in peak
intensity which is frequently not documented or ex-
cluded. It should be noted that further crystal reactions
have been found previously where the macroscopic shape
of single crystals did not change but which were never-
theless not topotactic when the (thermal) reaction oc-
curred exclusively at outer and inner surfaces15 or after
creation of submicroscopic cracks in the bulk16 or by
using the voids of empty space in the lattice.16
It is to be expected that the search for void space in
suitable positions of crystal lattices will reveal topotactic
crystal reactions4,17 without the requirement for absence
of geometric change (<4%) from reactant to product,4,17
even if there is always the possibility of the creation of an
amorphous phase.
EXPERIMENTAL
1H NMR spectra were recorded in CDCl3 on JEOL
Lambda 300 and Bruker WP 500 instruments. IR spectra
were recorded with a JASCO FT-IR 200 spectrometer.
UV spectra were measured on a Shimadzu MPS-2000
spectrometer. All melting-points were determined using a
Yanaco micro melting-point apparatus and were uncor-
rected. The details of AFM measurements on organic
crystals have been described elsewhere.2 A NanoScope II
instrument (Digital Instruments) was used in contact
Copyright # 2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 905–912

CIS–TRANS ISOMERIZATION OF A STILBENE IN INCLUSION CRYSTALS
911
mode at 10–30 nN force with non-scraping (symmetric)
standard Si3N4 tips (spring constant 0.12 N mꢃ1).2 The
crystals were glued to a conducting tab (Plano) on a
magnetic plate. Low-temperature irradiations used a
double-walled cooled beaker with the sample plate in
direct contact with the cold bottom (circulating bath at
ꢃ17 to ꢃ15 ꢂC or ꢃ70 ꢂC) at a 10 cm distance from the
quartz window and an additional 5 cm from the 700 W Hg
lamp that was cooled by tap water (14 ꢂC) running
through a double-walled Pyrex tube. Fixed-lattice mod-
eling was performed with the program Schakal (E. Keller,
Freiburg).
Recrystallization of 1 from acetone gave the complex
1ꢀacetone; m.p. 80–85 ꢂC. IR (Nujol), ꢁ 3475, 3377
(OH), 1695 (C O) cmꢃ1. Anal. Calcd for C43H38O3
—
—
(1ꢀacetone; 542.66): C, 85.68; H, 6.35. Found C 85.92,
H 6.45%.
Low-temperature photolysis of solid 1. Powdered crys-
tals of 1 (30 mg with <6% acetone) were spread on a
cooled glass surface under vacuum and irradiated with a
700 W Hg high-pressure lamp through a Pyrex filter from
a distance of 15 cm for 12 h. The conversion was detected
by 1H NMR. It was >93% (measured after 8 h) at ꢃ15 to
ꢃ17 ꢂC and 42% at ꢃ70 ꢂC. The corresponding experi-
ments with 1ꢀacetone gave conversions of 56 and 10%,
respectively.
3,30-Bis(carbethoxy)diphenylacetylene (5). A mixture of
ethyl 3-bromobenzoate (3) (4.5 g, 26 mmol), ethyl 3-
ethynylbenzoate (4) (5.9 g, 26 mmol), PdCl2(PPh3)2
(0.05 g), PPh3 (0.26 g) and CuI (0.05 g) in Et3N
(200 ml) was heated under reflux for 5 h. After filtration,
the organic layer was evaporated to give, after recrystal-
lization from acetone, 3,30-bis(carbethoxy)diphenylace-
trans-3,30-Bis(diphenylhydroxymethyl)stilbene (2). Solid-
state photolysis of the powdered 1:1 acetone complex of
1 (0.20 g, 0.38 mmol) for 6 h using a 400 W high-pressure
Hg lamp gave crude crystals of the 1:1 acetone complex
of 2 with 92% conversion. Recrystallization of the crude
crystals from acetone gave the pure 1:1 inclusion com-
plex 2 with acetone as colorless needles of 2ꢀacetone
which, upon heating, gave pure 2 as a white powder
(0.19 g, 92% yield, m.p. 68–73 ꢂC). Irradiation of well-
developed crystals of 1ꢀacetone gave similar conversions
at 12 h irradiation without destruction of the still clear
crystals.
tylene (5) as colorless prisms (5.6 g, yield 67%, m.p. 60–
1
68 C). IR (Nujol), ꢁ 1739 (C O) cmꢃ1, H NMR
ꢂ
—
—
(300 MHz, CDCl3), ꢂ 1.42 (t, J ¼ 7 Hz, 3H, CH3), 4.41
(q, J ¼ 7 Hz, 2H, CH2), 7.45–8.22 (m, 8H, Ar).
3,30-Bis(hydroxydiphenylmethyl)diphenylacetylene (6). A
solution of 5 (4.30 g, 13.3 mmol) in dry THF (30 ml) was
added with stirring to a solution of PhMgBr in THF
(200 ml), prepared from Mg (3.30 g, 133.9 mmol) and
bromobenzene (21.0 g, 133.9 mmol), and the mixture was
stirred for 6 h. Conventional workup gave crude crystals
of 3,30-bis(hydroxydiphenylmethyl)diphenylacetylene
(6). Recrystallization of the crude crystals from acetone
gave the 1:1 inclusion complex of 6 with acetone as
colorless prisms, which upon heating gave pure 6 as a
white powder (5.9 g, yield 81%, m.p. 85–90 ꢂC). IR
2: IR (Nujol), ꢁ 3550, 3445 (OH) cmꢃ1; UV(CHCl3),
1
ꢃmax (") 242 (12 500), 304 (12 900), 316 nm (9200); H
NMR (300 MHz, CDCl3), ꢂ 3.05 (s, 2H, OH), 6.98 (s, 2H,
—
CH), 7.24–7.44 (m, 28H, Ar). Anal. Calcd for
—
C40H32O2 (2; 544.68): C, 88.20; H,5.92. Found: C,
88.07; H, 6.20%.
2ꢀacetone: m.p. 70–75 ꢂC. IR (Nujol), ꢁ 3411 (OH),
1698 and 1685 (C O) cmꢃ1. Anal. Calcd for C43H38O3
—
—
(2ꢀacetone; 542.66): C, 85.68; H, 6.35. Found: C, 85.26;
(Nujol), ꢁ 3407 (OH) cmꢃ1
;
1H NMR (300 MHz,
H, 6.50%.
CDCl3), ꢂ 2.77 (s, 2H, OH), 7.25–7.33 (m, 28H, Ar).
Anal. Calcd for C40H30O2 (6; 542.66): C, 88.53; H, 5.57.
Found C, 88.30; H 5.74%.
Irradiation of 2 and 2ꢀacetone. The powdered crystals of
2 or 2ꢀacetone were irradiated for 12 h using a 400 W
high-pressure Hg lamp to give crystals of 2 or 2ꢀacetone
1
cis-3,30-Bis(diphenylhydroxymethyl)stilbene (1). A mix-
ture of 6 (1.00 g, 1.84 mmol) and Pd–BaSO4 (0.2 g) in
THF (100 ml) was stirred under an H2 atmosphere for 6 h.
After filtration, the organic layer was evaporated to give,
after recrystallization from toluene, cis-3,30-bis(diphe-
nylhydroxymethyl) stilbene (1) as colorless prisms
(0.85 g, yield 85%, m.p. 170–175 ꢂC).These could not
be grown larger than 0.05 ꢁ 0.05 ꢁ 0.3–0.4 mm and had
very rough surfaces, unsuitable for x-ray diffraction or
unchanged (determined by H NMR).
Preparation of single crystals of 1ꢀacetone. Single
crystals of 1ꢀacetone were prepared by slow evaporation
of an acetone solution of 1 at room temperature. Both
prismatic (Fig. 2) and plate-like crystals [the latter with
(100) as the main face] could be obtained. For AFM
investigation, the prismatic crystals were taken directly
from a saturated solution and dried on a filter-paper in air.
All of the paired faces along the c-axis could be mounted
horizontally.
even AFM studies. IR (Nujol), ꢁ 3559, 3459 (OH) cmꢃ1
;
UV (CHCl3), ꢃmax (") 243 (25 000), 285 nm (15 000); 1H
NMR (300 MHz, CDCl3), ꢂ 2.86 (s, 2H, OH), 6.52 (s,
—
2H, CH), 7.14–7.27 (m, 28H, Ar). Anal. Calcd for
—
C40H32O2 (1; 542.66): C, 87.92; H 6.05. Found C, 88.20;
H, 5.92%.
Synthesis of the other inclusion complexes 1ꢀguest in
Table 1. The other inclusion complexes of 1ꢀguest were
synthesized by recrystallization of 1 from the neat guest
Copyright # 2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 905–912

912
K. TANAKA ET AL.
compounds. Guest-free crystals of 1 that were suitable for
x-ray diffraction or AFM could not be obtained from
toluene.
ultra-high-pressure Hg lamp through a Pyrex filter. The
peak half-widths of the x-ray diffraction signals did not
increase, but the peak intensity decreased gradually. On
average, the peak intensity decreased to 60% after 1 h and
to 20% after 3 h of irradiation. The crystal structure
analyses before and after photolysis did not show any
change. The transparent crystal was sent from Yokohama,
Japan, to Oldenburg, Germany, and its chemical conver-
General procedure for the photoisomerization of the
other inclusion complexes in Table 1. Amounts of
50 mg of the powdered crystals of the 1ꢀguest inclusion
complexes were irradiated as described above and the
data are summarized in Table 1. The conversion was
1
sion was measured as 36% by 500 MHz H NMR in
CDCl3.
1
determined by H NMR analysis. The guest molecules
were still present in all cases.
X-ray structure analysis of cis-3,30-bis(diphenylhydroxy-
methyl)stilbeneꢀacetone (1ꢀacetone). Crystal data:
C40H32O2 ꢀ C3H6O, Mr ¼ 602.77, crystal dimensions
0.4 ꢁ 0.3 ꢁ 0.2 mm, colorless prism. X-ray intensities
were measured on a Rigaku AFC-7R diffractometer
with Mo Kꢄ radiation at 297 K, monoclinic P21/n: a,
Acknowledgement
K. T. acknowledges financial support from the Asahi
Glass Foundation.
REFERENCES
ꢂ
˚
24.413(3); b, 16.229(3); c, 8.634(1) A; ꢅ, 99.07(1) ; V,
3
3378.0(9) A ; Z, 4; Dx, 1.185 g cmꢃ3; ꢆ, 0.073 mmꢃ1. A
˚
1. Liu RSH, Hammond GS. Chem. Eur. J. 2001; 7: 4537–4544.
2. Kaupp G, Haak M. Angew. Chem., Int. Ed. Engl. 1996; 35: 2774–
2777; Kaupp G. In Comprehensive Supramolecular Chemistry,
vol. 8, Davies JED (ed.). Elsevier: Oxford, 1996; 381–423.
3. Kaupp G. Photochem. Photobiol. 2002; 76: 590–595.
4. Kaupp G. Curr. Opin. Solid State Mater. Sci. 2002; 6: 131–138.
5. Ohba S, Hiratsuka T, Tanaka K. Acta Crystallogr., Sect. E 2002;
58: o1013–o1015.
6. Gavezzotti A. J. Am. Chem. Soc. 1983; 105: 5220–5225; Zimmer-
man HE, Nesterov EE. J. Am. Chem. Soc. 2002; 124: 1169–1171.
7. Liu RSH, Asato AE. Proc. Natl. Acad. Sci. USA 1985; 82: 259–
263.
8. Uda M, Mizutani T, Hayakawa J, Momatake A, Ikegami M,
Nagahata R, Arai T. Photochem. Photobiol. 2002; 76: 596–605.
9. Kohlschu¨tter V. Z. Anorg. Allg. Chem. 1919; 105: 1–25.
10. Schmidt GMJ. J. Chem. Soc. 1964; 2014–2021.
11. Liu RSH, Hammond GS. Proc. Natl. Acad. Sci. USA 2000; 97:
11153–11158; Mu¨ller AM, Lochbrenner S, Schmid WE, Fuss W.
Angew. Chem., Int. Ed. Engl. 1998; 37: 505–507.
12. Kaupp G. Adv. Photochem. 1995; 19: 119–177.
13. Bart JCJ, Schmidt GMJ. Recl. Trav. Chim. Pays-Bas 1978; 97:
231–238.
crystal specimen was sealed in a capillary, and there was
no intensity decay of standard reflections. Data collection
of 9071 reflections (ꢇmax: 27.5ꢂ), 7764 independent
reflections (Rint: 0.022), 3144 observed reflections with
I > 2ꢈ(I), 423 parameters. The hydroxyl H atoms were
located from difference syntheses and refined isotropi-
cally. The other H-atom positional parameters were
calculated geometrically and constrained. R(F) ¼ 0.054
for 3144 observed reflections, wR(F) ¼ 0.0911 for all
5994 reflections. CCDC 187574 contains the supplemen-
tary crystallographic data for this compound. These data
can be obtained free of charge via www.ccdc.cam.ac.uk/
conts/retrieving.html (or from the Cambridge Crystal-
lographic Data Centre).
In the low-temperature irradiation in the x-ray diffract-
ometer a single crystal of 0.6 ꢁ 0.5 ꢁ 0.2 mm was coated
with adhesive, mounted and kept constant at 250 K using
cold nitrogen gas. Lattice constants before photolysis
14. Kaupp G, Schmeyers J. J. Photochem. Photobiol. B: Biol. 2000;
59: 15–19.
15. Kaupp G, Schmeyers J, Kato M, Tanaka K, Harata N, Toda F.
J. Phys. Org. Chem. 2001; 14: 444–452.
16. Kaupp G, Schmeyers J, Kato M, Tanaka K, Toda F. J. Phys. Org.
Chem. 2002; 15: 148–153.
17. Nakanishi H, Jones W, Thomas JM, Hursthouse MB, Motvalli M.
J. Phys. Chem. 1981; 85: 3636–3642.
(monoclinic P21/n), a, 24.383(5); b, 16.233(5); c,
3
8.6096(16) A; ꢅ, 99.241(15) ; V, 3363.3(14) A changed
ꢂ
˚
˚
ꢂ
˚
to a, 24.416(8); b, 16.246(7); c, 8.625(2) A; ꢅ, 99.18(3) ,
3
V, 3378(2) A after photolysis for 3 h with a 250 W
˚
Copyright # 2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 905–912