Supramolecular Control of the Template-Induced Selective Photodimerization of 4-Methyl-7-O-hexylcoumarin

Article abstract of DOI:10.1002/chem.200305268

A symmetric ditopic molecular receptor (3), containing two identical hydrogen-bonding recognition subunits, was designed and synthesized. These subunits are capable of binding substrates with complementary donor and acceptor sites to form a supramolecular complex through hydrogen bonding, Receptor 3 was designed to accept two guest molecules, which are held in close proximity within the supramolecular species. The substrate molecule, 4-methyl-7-O-hexylcoumarin (1c), forms a 2:1 complex with a binding constant of 150M^-1 for the second substrate, passing first through a 1:1 complex with an affinity constant of 510M^-1. The orientation of two molecules of 1c when bound to the template leads to the selective formation of the trans-syn [2+2] photoproduct 2cB upon irradiation. Other photoproducts typically produced in the absence of the template are suppressed.

Full text of DOI:10.1002/chem.200305268

FULL PAPER
Supramolecular Control of the Template-Induced Selective
Photodimerization of 4-Methyl-7-O-hexylcoumarin
[
a, b]
[a, c]
[a]
William G. Skene,
Emilie Couzignÿ,
and Jean-Marie Lehn*
À1
Abstract: A symmetric ditopic molecu-
lar receptor (3), containing two identi-
cal hydrogen-bonding recognition sub-
units, was designed and synthesized.
These subunits are capable of binding
substrates with complementary donor
and acceptor sites to form a supra-
molecular complexthrough hydrogen
bonding. Receptor 3 was designed to
accept two guest molecules, which are
held in close proximity within the
supramolecular species. The substrate
molecule, 4-methyl-ꢀ-O-hexylcoumarin
(1c), forms a 2:1 complexwith a bind-
ing constant of 15ꢁm for the second
substrate, passing first through a 1:1
complexwith an affinity constant of
À1
51ꢁm . The orientation of two mole-
cules of 1c when bound to the tem-
plate leads to the selective formation
of the trans-syn [2+2] photoproduct
2cB upon irradiation. Other photo-
products typically produced in the ab-
sence of the template are suppressed.
Keywords: coumarins
¥
photo-
photo-
chemistry
¥
selective
dimerization
¥
supramolecular
chemistry ¥ template effects
Introduction
Numerous investigations have been carried out on the sub-
[
1±±]
ject of [2+2] photodimerizations,
and uracil, since their products exhibit high biological toxici-
ty.
especially of thymine
[
4±ꢀ]
Other systems giving rise to [2+2] photoproducts,
among them coumarin and its derivatives (1a±c), have been
actively studied to fully understand the mechanistic path-
ways of these photoreversible reactions and to ultimately
[8±1ꢁ]
control product distribution.
Thus, similar to other pho-
todimerization reactions, photoirradiation of 1 can potential-
ly lead to four cycloaddition products; cis-syn, trans-syn, cis-
anti and trans-anti (2A±D, respectively). Under typical ex-
perimental conditions, however, the two syn (2A and 2B)
photoproducts are formed predominately, while only trace
[ꢀ]
amounts of the anti dimers 2C and 2D are observed. The
[
a] Prof. Dr. J.-M. Lehn, Prof. Dr. W. G. Skene, E. Couzignÿ
Laboratoire de Chimie Supramolÿculaire
ISIS -Universitÿ Louis Pasteur
8
, allÿe Gaspard Monge
BP ꢀꢁꢁ28, 6ꢀꢁ8± Strasbourg cedex(France)
Fax: ( +±±)±-9ꢁ24-514ꢁ
E-mail: lehn@isis.u-strasbg.fr
[
b] Prof. Dr. W. G. Skene
Current address: Dÿpartement de Chimie
Universitÿ de Montrÿal
Montrÿal, Quÿbec
C.P. 6128, succ. Centre-ville, H±C ± Jꢀ (Canada)
[
c] E. Couzignÿ
Current address: Structural GenomiX
1
ꢁ5ꢁ5 Roselle Street, San Diego, CA 92121 (USA)
556ꢁ
¹ 2ꢁꢁ± Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200305268
Chem. Eur. J. 2003, 9, 556ꢁ ± 5566

5
56ꢁ ± 5566
ratio of coumarin photoproducts can significantly be altered
through control of the reaction conditions such as solvent
[
9,11,12]
[1±,14]
choice,
ment,
Lewis acid addition,
micellar environ-
modification of the
[
15,16]
[1±,14]
solid-state irradiation,
[1ꢀ,18]
[15]
coumarin precursor,
tethered precursors,
use of inclusion complexes
or recently with pseudo-rotax-
and
[1ꢀ,19±21]
[
22]
anes. Much effort has been focused on crystal engineering
and packing within the solid state to influence the product
[
2±±2ꢀ]
distribution.
These various modifications are under-
stood to change the photochemical reaction manifolds of
the intermediates by perturbing the transition moment ulti-
[8,1ꢁ,28]
mately influencing the product distribution.
In view of
obtaining control over photoproduct selectivity, we investi-
gated whether a molecular receptor, capable of forming a
specifically oriented supramolecular complexin the ground
state with two dimer precursors, could promote such product
amplification through the geometry and steric confinement
of the supramolecular species, similar to known thermal cy-
[29]
cloadditions. The synthesis of such a molecular template
and our photochemical findings are described herein.
Results and Discussion
The binding of two coumarin units to a ditopic molecular
template such as 3 would lead to a symmetric supramolec-
ular structure, whereby the orientation of the two monomers
would be expected to favour the exclusive formation of syn
dimers upon irradiation, as represented in Scheme 1. Photo-
irradiation of free 1a, like that of other [2+2] precursors,
addition to steric hindrance within the template cleft, the
corresponding orientation of the precursors would not be
achieved (see Scheme 2). Coumarin derivatives 1b and 1c
are suitable candidates for such selective photodimerization
as they satisfy the necessary criteria. In particular, 1c pre-
dominately forms anti dimers
[15±1ꢀ,19,2ꢁ]
upon irradiation
and
was selected for syn product
amplification. Substitution at
the 4- and ꢀ-coumarin positions
creates a weak donation that
changes the transition moment,
in turn affecting the product
[8]
distribution.
The transition
moments can further be influ-
[11]
enced by solvent polarity and
the use of cyclodextrin or mi-
Scheme 1. Schematic representation of the reaction pathway responsible for selective syn photodimer forma-
tion within a supramolecular 2:1 species.
[12,15,16]
celles.
With respect to
[
±ꢁ±±4]
previous reports,
we set
forth the design and subsequent
synthesis of a template, which
leads exclusively to syn dimers, so that binding of two units
of 1a to the receptor 3 would not drastically alter the ratio
among the four potential isomers. For comparison we also
prepared three template analogues of 3, namely compounds
4
±6, synthesized from a combination of compounds 7±9; for
details see the Experimental Section. We therefore sought
out a photodimerizable compound that contained sites capa-
ble of forming a supramolecular complexthrough hydrogen
bonding with a template, but whose major photoproducts
under normal irradiation conditions would be anti dimers.
The formation of a 2:1 complexbetween such a precursor
and a template should preferentially yield syn dimers, dem-
onstrating template-assisted selective photodimerization.
The anticipated anti dimers would be suppressed, since, in
Scheme 2. Orientation of the partners within the 2:1 supramolecular
complexrequired for anti photodimer formation.
Chem. Eur. J. 2003, 9, 556ꢁ ± 5566
www.chemeurj.org
¹ 2ꢁꢁ± Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5561

J.-M. Lehn et al.
FULL PAPER
[±9]
would promote the formation of otherwise minor products
upon photodimerization, through template-assisted selective
product amplification.
corresponds to a 2:1 complex. The titration method em-
ployed followed the changes in chemical shift of the NÀH
proton of 3 due to hydrogen bonding to the C=O group of
1
d or 1c. The limitation of this method is that no informa-
tion concerning binding through the pyridine nitrogen can
be obtained. Therefore, it is not surprising that the meas-
ured isotherms for 1d and 1c are similar, since the centre in-
volved in binding is similar for the two monomers. The
slight variances observed reflect the differences in weak
binding of the guest×s heterocyclic site with the receptor, in
line with a recent report concerning weak synthon bind-
Binding of substrate 1c to receptor 3: The synthesis of tem-
plate 3 was easily achieved in two steps (as compared to the
[±5,±6]
previous strategy involving four steps
) from the dialde-
hyde 7 and 2-hydrazinopyridine in good yield. Previous
work has shown the capability of this receptor to efficiently
bind substrate molecules through hydrogen bonds between
the pyridine nitrogen and the adjacent amine with comple-
mentary donor and acceptor sites, making it a suitable tem-
[4ꢁ]
ing. Even though the measured values are for CDCl , it
±
was found that the formation of the supramolecular complex
between a guest and the template are not dramatically af-
fected by the use of acetone as solvent, the isotherms ob-
[±5±±ꢀ]
plate for the present study.
[±5]
Carbostyril 1d was originally investigated as a potential
dimerization candidate, as it possesses two sites capable of
sustaining such hydrogen bonding with the receptor. The
binding isotherm measured by NMR titration allows the de-
termination of the overall association constants for 1:1 and
served for the two solvents being similar.
The mode of binding of 1c to receptor 3 is not as
straightforward as with previous examples for hydrogen
[±1,±2]
bonding or with 1d.
The optimal bonding situation
would be observed with carbostyril 1d, which possesses the
conventional hydrogen-bonding pattern involving two hy-
drogen bonds between its NÀH and C=O groups and the
complementary pyridine nitrogen and NÀH sites of the tem-
plate 3. With 1c, in which the hydrogen bonding donor
ÀNH is not present, the mode of interaction most likely in-
volves a hydrogen-bonding pattern composed of binding be-
tween the NÀH centre of the template and the lactone
oxygen site, together with two weak CÀH¥¥¥X interactions
between the imino =CÀH and the pyridine nitrogen atom of
the template with the carbonyl oxygen atom and the oppo-
sitely located aromatic CÀH of 1c, respectively, as schemati-
cally represented in compound 10. Such interactions have
previously been observed and are well identified by crystal
2
:1 complexes with 3 in CDCl ; these were found to be
±
À1
À1
about 86ꢁm and 6ꢁm , respectively. Unfortunately, the
photodimerization of this lactam leads exclusively to the syn
dimer, even in the absence of the template, regardless of the
[2,±8]
experimental conditions, consistent with previous results.
Also, the photoproducts precipitate upon formation and can
only be solubilized in trifluoroacetic acid, making characteri-
zation and subsequent quantification of the products very
difficult.
Coumarin 1c also satisfied our selection criterion and
proved a better candidate for selective photoproduct ampli-
À1
fication studies. The association constants measured (51ꢁm
À1
and 154m for the 1:1 and 2:1 complexes, respectively)
[4ꢁ,41]
were comparable to those found for 1d, indicating that 1c
binds equally well to the template. These results imply the
initial formation of a moderately stable 1:1 complexbe-
tween the template and 1c, which thereafter leads to a 2:1
complexupon further addition of coumarin, according to
Scheme 1. This stoichiometry was confirmed by Job×s
method through the use NMR spectrometry (Figure 1),
whereby the observed maximum at constant concentration
structures.
NMR data can provide information about an
interaction between the =CH proton and the guest 1c. Moni-
toring the change in chemical shift of the =CH proton signal
of the template on titration with the guest should result in a
deshielding shift, consistent with hydrogen bonding interac-
tion. Titration of 3 with 1c gives rise to such a change in
chemical shift of approximately Dd=ꢁ.4 ppm, indicative of
the binding represented in 10. Moreover, the binding iso-
therm obtained by this titration experiment agrees with the
binding constants obtained from the titration method used to
observe the chemical shift of the NÀH proton, thus further
confirming the formation of a 2:1 supramolecular complex.
Figure 1. Job plot at constant concentration for the binding of the dimer
precursors 1d (&, 1ꢁ mm) and 1c (*, 1ꢁꢁ mm) to the receptor 3 with var-
iation of 9.6±1ꢁ.5 ppm.
5562
¹ 2ꢁꢁ± Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 556ꢁ ± 5566

Selective Photodimerization
556ꢁ ± 5566
Photodimerization of coumarin 1c: The use of authentic
samples was required to compare HPLC retention times of
the template-assisted photoproducts of 1c with those ob-
tained in its absence. Therefore, irradiation of 1c and subse-
quent isolation of its photodimers was required. Only two
photoproducts were observed at conversions below ±ꢁ%;
they were easily separated by column chromatography and
characterized by 2D NMR spectroscopy and were found to
be 6ꢁ% syn and 4ꢁ% anti dimers. No other products were
detected with the analytical methods used. Determination of
due to a higher local concentration of the precursor as a
result of two molecules of 1c being held in close proximity
within the template. A higher probability of encounter is in-
duced by the 2:1 complexformation, leading to an increase
in quantum yield relative to the native dimerization. Since
À1
the measured triplet value of 3 is about 22 kJmol higher
than that of 1c in ethanol, and even greater in apolar sol-
vents, quenching of the coumarin triplet state by the tem-
plate cannot be responsible for the decreased quantum
yield. Even though endergonic triplet quenching is possible,
it is not present to a high degree in this case, owing to low
concentrations of the reagents. Therefore, the lower quan-
tum yield may arise predominately from a screening effect
of the template, as there is a large spectral overlap between
the two partners. Moreover, the screening effect benefits
photoproduct stability by decreasing photodecomposition of
the dimers to regenerate the reagents, which is known to
occur under UV irradiation. The relatively low concentra-
tion used further ensures that dimerization proceeds from
within the supramolecular complexand not between adja-
cent complexes or via aggregates.
1
the dimer configuration was difficult based on simple H
NMR spectroscopy without authentic samples. The differ-
ence between the two cis and trans isomers is relatively
small and the trans cyclobutyl protons give a NMR signal
ꢁ
.1 ppm upfield from their cis counterparts, due to the
[15]
shielding effects of the carbonyl and phenyl groups.
Through the use of a chiral shift reagent and 2D proton
NMR spectroscopy, complete characterization of the isolat-
ed photoproducts was possible and led to the assignment of
the trans form to both the syn and anti dimers (2cB and
2
cD). The configuration of the observed photodimers is
consistent with previous reports for sensitized dimeriza-
[
15,16,19,2ꢁ]
tion.
No appreciable change was observed when
Control irradiation experiments: To confirm the role of hy-
drogen bonding in the irradiation of coumarin 1c leading to
controlled product formation, the phenyl hydrazone ana-
logue of 3 would be a suitable reference. It would lack the
pyridine nitrogen and the eventual formation of the anti
isomer by irradiation in the presence of 1c would confirm, a
contrario, the bonding scheme depicted in 10. However, at-
tempts to prepare this compound were unsuccessful due to
its thermal and photochemical instability. Starting reagents
involving phenyl hydrazine or its salt systematically resulted
in the formation of benzene and nitrogen as decomposition
products in addition to unreacted dialdehyde 7. Alternative-
ly, protonation should disrupt the hydrogen bonding, conse-
quently leading to an anticipated different photoproduct dis-
tribution, confirming the role of the templated supramolec-
ular complex. Indeed, the addition of such trace amounts of
acid to 3 gave rise to an equal amount of the two photodim-
ers upon irradiation (Figure 2), indicating the destruction of
the template effect. As a control, the photoreaction of 1c
alone in the presence of acid gave nearly the same product
distribution.
performing either direct irradiation in chloroform or sensi-
tized irradiation with acetone. Since the experimental condi-
tions employed are relatively dilute, direct excitation of the
precursor undergoes intersystem crossing forming the triplet
state before quenching occurs. This manifold leads exclu-
sively to the trans isomers; conversely the cis isomers arise
from singlet excimer formation at higher concentrations,
whereby two monomers can be in close proximity required
[21]
for promoting such an interaction.
Photodimerization of coumarin 1c in presence of receptor
: Irradiation of 1c in the presence of ꢁ.5 equivalents of the
3
template 3 led exclusively to the head-to-head trans-syn
dimer 2cB. The production of this dimer can be rationalized
through the pathway described in Scheme 1 further support-
ed by the 2:1 binding constant measured. The complete sup-
pression of the head-to-tail anti dimers can result only if the
coumarin precursors are prevented from adopting the corre-
sponding arrangement (Scheme 2), further supporting the
orientation of the two monomer substrates 1c as represent-
ed in 10. The preferential formation of the trans isomer
stems from steric effects and has previously been observed
[1]
for hydrogen-bonded complexes in photodimerization.
Correct molecular overlap for dimerization to proceed
occurs with the two unsaturated groups in a separated yet
stacked arrangement, but not within the same plane. Such
an arrangement can be achieved through a rotation around
the CÀCH= bonds in the side chains of the template and
avoids steric interaction between the 4-methyl groups within
the 2:1 complexof 1c with 3. The results obtained represent
a successful demonstration of light-induced, selective prod-
uct amplification arising from a template-induced, modified
supramolecular complex.
In the presence of the template 3, the observed quantum
yield (F=ꢁ.ꢁ±) of photodimerization of 1c was roughly half
that observed in its absence. A higher value is anticipated
Figure 2. Photoproduct distribution resulting from photodimerization of
1
c under various conditions in acetone: dark grey syn, and light grey anti
[42]
isomers. *Only trace amounts detected.
Chem. Eur. J. 2003, 9, 556ꢁ ± 5566
www.chemeurj.org
¹ 2ꢁꢁ± Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
556±

J.-M. Lehn et al.
FULL PAPER
To further ascertain the importance of supramolecular
and steric effects for controlled photodimerization, the ana-
logues 4 (synthesized from 8 and 9), 5 and 6 of receptor 3
method in CDCl
±
and the isotherm data subsequently fitted with the
[
44]
Chemequi program. NOE and COSY 2D NMR spectra were recorded
on a Bruker ±ꢁꢁ MHz spectrometer, whereby the absolute configuration
of the [2+2] photodimers was determined by using an optically active
chemical shift reagent, europium(iii)-tris-(±-trifluoromethylhydroxy-
methylene-(+)-camphorate).
[4±]
were synthesized.
The single-arm template analogue 4
contains a donor and acceptor site capable of binding only
one coumarin monomer, in contrast to 3. Under stoichio-
metric conditions of one monomer equivalent, no controlled
photodimer production should be observed with this asym-
metric template. Indeed, the irradiation of 4 in acetone led
only to trace amounts of photoproducts with no control
over the product distribution.
Two other template analogues, 5 and 6, were investigat-
ed, as they do not possess the ÀNH donor sites capable of
sustaining the pre-irradiation arrangement necessary for syn
photoproduct formation. Furthermore, they both have a
cleft similar to 3 in which two monomers could potentially
fit. The photoproduct distribution obtained upon irradiation
of the two donor-free templates 5 and 6 in the presence of
7-Hexyloxy-4-methylchromen-2-one (1c): ꢀ-Hydroxy-4-methylcoumarin
(2.ꢁ1 g, 11.4 mmol) was added to dimethylformamide (125 mL), followed
by sodium carbonate (1.45 g, 1±.9 mmol). The slurry was stirred at room
temperature for 2ꢁ minutes, after which 1-bromohexane (1.92 mL, 1±.5
mmol) was then added and the mixture was subsequently refluxed for 16
h after the addition of a catalytic amount of potassium iodide. The sol-
vent was removed under vacuo, whereby the paste was taken up in di-
chloromethane and washed with water. The organic phase was separated
and the solvent removed followed by flash chromatography (SiO
2
) with
1
ꢁ% ethyl acetate/9ꢁ% dichloromethane. Compound 1c was isolated as
1
a colourless oil (2.85 g, 96%). H NMR (2ꢁꢁ MHz, [D]chloroform): d=
.16 (d, J=8.9 Hz, 1H; Aryl), 6.56 (dd, J=2.4 Hz, 1H; Aryl), 6.42 (d,
J=2.4 Hz, 1H; Aryl), 5.ꢀ9 (s, 1H; Aryl), ±.ꢀ2 (t, 2H; CH ), 2.11 (s, 2H;
ꢀ
2
2 2 2 ±
CH ), 1.51 (m, 2H; CH ), 1.12 (m, 4H; CH ), ꢁ.69 ppm (t, ±H; CH );
1
±
C NMR (5ꢁ MHz, [D]chloroform): d=161.9, 16ꢁ.ꢀ, 155.ꢁ, 152.4, 125.4,
1
1±.1, 112.2, 111.4, 1ꢁ1.ꢁ, 68.4, ±1.4, 28.9, 25.5, 22.5, 18.±, 1±.9 ppm; MS
1
c was roughly identical to those observed in the absence of
+
(
ꢀꢁ eV): m/z (%): 26ꢁ.± (±ꢁ) [M] ; elemental analysis calcd (%) for
the templates within experimental error. These results, com-
bined with the photoirradiation in presence of 4, confirm
that a templating effect of 3 with 1c through supramolecular
control is responsible for the selective photoproduct forma-
tion. The orientation of two coumarin monomers within the
template controlled by the complexis such that only the syn
products can be formed. Furthermore, the bulk of the tem-
plate ensures that photodimerization occurs solely between
two molecules of 1c located within the cleft.
C
16
H
2ꢁ
O
±
(26ꢁ.±±): C ꢀ±.82, H ꢀ.ꢀ4; found: C ꢀ±.±ꢁ, H 8.±2.
3
,10-Bishexyloxy-12b,12c-dimethyl-6a,6b,12b,12c-tetrahydro-5,8-dioxadi-
benzo[a,i]biphenylene-6,7-dione (2cB): Coumarin 1c (5ꢁ± mg, 1.9±
mmol) was dissolved in acetone (25 mL) in a Pyrextest tube, which was
then sealed and the oxygen removed through a stream of argon for ±ꢁ
minutes. The solution was irradiated for 12 h with a 4ꢁꢁ W lamp then the
solvent removed. The crude oil was purified by chromatography (SiO )
2
with 1ꢁ% ethyl acetate/9ꢁ% dichloromethane to yield 2cB as a white
1
solid. H NMR (2ꢁꢁ MHz, [D]chloroform): d=ꢀ.ꢁꢀ (d, J=8.ꢀ Hz, 2H;
Aryl), 6.61 (dd, J=6.ꢁ Hz, 2H; Aryl), 6.ꢁ± (d, J=±.ꢁ Hz, 2H; Aryl), ±.ꢀꢀ
(
m, 4H; CH
2
), ±.4ꢁ (s, 2H; CH), 1.6ꢀ (m, 1ꢁH; CH
2
), 1.±ꢁ (m, 12H;
1
±
CH
2
), ꢁ.89 ppm (t, 6H; CH
±
); C NMR (5ꢁ MHz, [D]chloroform): d=
1
64.9, 159.8, 15ꢁ.±, 12ꢀ.4, 11±.6, 112.4, 1ꢁ2.±, 68.±, 55.±, 41.1, ±1.ꢀ, ±1.5,
+
Conclusion
28.9, 25.6, 22.6, 14.ꢁ ppm; MS-FAB: m/z (%): 521.± (1ꢁꢁ) [M+H]
HRMS-FAB: m/z calcd for C±2 : 521.291ꢁ; found: 521.29ꢁ±.
3,9-Bishexyloxy-6b,12b-dimethyl-6a,6b,12a,12b-tetrahydro-5,11-dioxadi-
benzo[a,b]biphenylene-6,12-dione (2cD): Isolated as a white solid after
column chromatography (SiO ) with 1ꢁ% ethyl acetate/9ꢁ% dichlorome-
thane. H NMR (2ꢁꢁ MHz, [D]chloroform): d=ꢀ.ꢁ4 (d, J=8.ꢀ Hz, 2H;
Aryl), 6.ꢀꢀ (dd, J=6.ꢁ Hz, 2H; Aryl), 6.61 (d, J=±.ꢁ Hz, 2H; Aryl), ±.95
(t, 4H; CH ), ±.±6 (s, 2H; CH), 1.ꢀ9 (m, 4H; CH ), 1.64 (m, 18H; CH ),
;
41 6
H O
Irradiation of the 2:1 supramolecular complex 10 formed by
the symmetric ditopic receptor 3 with 4-methyl-ꢀ-O-hexyl-
coumarin 5 displays formation and selective amplification of
the otherwise minor head-to-head syn [2+2] 2cB photo-
product. This outcome may be rationalized in terms of the
pre-irradiation orientation of the two coumarin substrate
molecules within the supramolecular species. Relative dispo-
sition and steric effects within the complexas well as triplet-
based dimerization may be considered responsible for the
preferential formation of the observed trans isomer rather
than the cis form, while light absorption by the receptor
may account for a reduction in the [2+2] dimer quantum
yield. The results obtained illustrate the ability to control
the regio- and/or stereoselectivity of chemical reactions be-
tween molecular species, by proper design of supramolec-
ular entities and represent a process of supramolecular
structure selective catalysis.
2
1
2
2
2
1
±
±
ꢁ.89 ppm (t, 6H; CH ); C NMR (5ꢁ MHz, [D]chloroform): d=166.1,
1
2
59.9, 151.6, 128.ꢁ, 114.8, 112.4, 1ꢁ2.9, 68.5, 46.6, 45.ꢁ, ±1.6, 29.1, 26.4,
5.ꢀ, 22.6, 14.ꢁ ppm.
(
2,7-Di-tert-butyl-9,9-dimethyl)-9H-xanthene-4,5-dicarbaldehyde (7): 4,5-
Dibromo-2,ꢀ-di-tert-butyl-9,9-dimethyl-9H-xanthene (4±8 mg, ꢁ.91 mmol)
was added to anhydrous diethyl ether (5ꢁ mL) and the mixture was
cooled to Àꢀ88C under an argon atmosphere. A solution of n-butyllithi-
um (ꢀ mL, 11.2 mmol) was slowly added. The reaction was stirred at the
low temperature for 6ꢁ minutes, after which dimethylformamide (2 mL,
2
5.8 mmol) was added and the temperature was raised to room tempera-
ture. The mixture was stirred for 6ꢁ minutes then 2m HCl (5 mL) was
added and stirring continued for a further ±ꢁ minutes. Diethyl ether was
removed under vacuum then the residue was extracted with ethyl acetate.
The solid was purified by chromatography (SiO
tate/9ꢁ% hexane to yield the product as a white solid (2ꢁ6 mg, 6ꢁ%).
M.p. 248±249 C; H NMR (2ꢁꢁ MHz, [D]chloroform): d=1ꢁ.68 (s, 2H;
CH), ꢀ.8± (d, J=2.4 Hz, 2H; Aryl), ꢀ.ꢀ2 (d, J=2.4 Hz, 2H; Aryl), 1.ꢀꢁ
(s, 6H; CH
form): d=188.9, 149.6, 146.ꢀ, 1±ꢁ.6, 129.5, 124.1, 12±.5, 62.2, ±4.8, ±2.5,
±
2
) with 1ꢁ% ethyl ace-
o
1
1
±
Experimental Section
± ±
), 1.±ꢀ ppm (s, 18H; CH ); C NMR (5ꢁ MHz, [D]chloro-
+
1.4, 29.ꢀ ppm; MS-FAB: m/z (%): ±ꢀ9.± (1ꢁꢁ) [M+H] ; elemental anal-
All reagents were used as received from Aldrich without further purifica-
tion unless otherwise stated. HPLC analyses were performed on an HP
1ꢁꢁ series HPLC equipped with a diode array detector and an Eclipse
ysis calcd (%) for C25
8.14.
±ꢁ ±
H O (±ꢀ8.5ꢁ): C ꢀ9.±±, H ꢀ.99; found: C ꢀ9.59, H
TM
1
XDB-C18 reverse phase column. A mobile phase of 85% methanol/15%
water was used and the retention times monitored at 21ꢁ nm; the absorp-
tion spectra and retention times were compared to authentic samples
2,7-Di-tert-butyl-9,9-dimethyl-4,5-bis-(pyridin-2-yl-hydrazonomethyl)-9H-
xanthene (3): Compound 7 (ꢁ.±ꢁ g, ꢁ.ꢀ9± mmol) was dissolved in chloro-
form (1ꢁ mL) and 2-hydrazinopyridine (ꢁ.19 g, 1.ꢀ4 mmol) was added as
a solid. The solution was stirred overnight at room temperature. The mix-
(
vide infra). The binding constants were measured by an NMR titration
5564
¹ 2ꢁꢁ± Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 556ꢁ ± 5566

Selective Photodimerization
556ꢁ ± 5566
ture was then evaporated and subjected to chromatography (SiO
2
) with a
[D]chloroform): d=161.2, 14ꢀ.±, 1±ꢀ.ꢁ, 112.8, 1ꢁꢀ.±, 41.ꢁ ppm; MS (ꢀꢁ
+
gradient of 4ꢁ% ethyl acetate/dichloromethane and increased to 6ꢁ%
ethyl acetate/dichloromethane. The product was obtained (ꢁ.41 g, 9±%)
eV): m/z (%): 12±.2 (1ꢁꢁ) [M] ; elemental analysis calcd (%) for C
6
H
9
N
±
(12±.2): C 58.51, H ꢀ.±ꢀ;N ±4.12 found: C 59.ꢁ1, H ꢀ.42, N ±4.21.
1
as a yellowish powder, which decomposed above 1848C. H NMR (2ꢁꢁ
N-(2,7-Di-tert-butyl-9,9-dimethyl-9H-xanthen-4-ylmethylene)-N’-pyridin-
-yl-1,1’-dimethylhydrazine (6): Hydrazine 9 (264 mg, ꢁ.21 mmol) was
MHz, [D]chloroform): d=9.±± (brs, 2H, NH), 8.±6 (s, 2H; CH), 8.14 (d,
J=4.9 Hz, 2H; Aryl), ꢀ.8± (d, J=2.± Hz, 2H; Aryl), ꢀ.5ꢀ (td, J=ꢀ.8 Hz,
J=1.8 Hz, 2H; Aryl), ꢀ.44 (d, J=2.2 Hz, 2H; Aryl), ꢀ.42 (d, J=8.4 Hz,
2
added to dialdehyde 7 (29 mg, ꢁ.ꢁꢀꢀ mmol) in a 9:1 mixture of ethanol/
chloroform (±5 mL) and then refluxed for 24 h. The solvent was removed
under reduced pressure and compound 6 was quantatively isolated as a
2
1
1
1
H; Aryl), 6.ꢀ4 (t, J=ꢀ.8 Hz, 2H; Aryl), 1.68 (s, 6H; CH
±
), 1.4ꢁ ppm (s,
1
±
8H; CH
±
); C NMR (5ꢁ MHz, [D]chloroform): d=15ꢀ.6, 146.5, 146.1,
1
white solid upon drying under vacuum. Decomposition above 2ꢁꢀ8C; H
45.4, 1±8.4, 1±6.2, 1±ꢁ.1, 12±.1, 122.5, 122.4, 114.4, 1ꢁꢀ.2, ±2.4, ±1.5, 22.6,
NMR (2ꢁꢁ MHz, [D]chloroform): d=8.24 (s, 4H; NH), ꢀ.9ꢀ (d, J=2.1
Hz, 2H; Aryl), ꢀ.ꢀ4 (s, 2H; CH), ꢀ.65 (t, J=2.1 Hz, 2H; Aryl), ꢀ.44 (d,
+
4.1 ppm; MS-FAB: m/z (%): 561.2 (1ꢁꢁ) [M+H] ; elemental analysis
calcd (%) for C±5
C ꢀ5.26, H ꢀ.±8, N 15.ꢁ9.
,7-Di-tert-butyl-9,9-dimethyl-9H-xanthene-4-carbaldehyde (8): Trifluoro-
acetic acid (TFA; 5 mL) followed by hexamethylenetetramine (95.5 mg,
.68 mmol) was added all at once to 2,ꢀ-di-tert-butyl-9,9-dimethyl-9H-
4O 6
H N O (56ꢁ.ꢀ±): C ꢀ4.9ꢀ, H ꢀ.19, N 14.99; found:
J=2.6, 2H; Aryl), 6.ꢀ9 (t, J=6.2 Hz, 2H; Aryl), ±.ꢀꢀ (s, 6H; CH
±
), 1.69
1±
(
s, 6H; CH ), 1.±4 ppm (s, 18H; CH ); C NMR (5ꢁ MHz, [D]chloro-
± ±
2
form): d=15ꢀ.9, 14ꢀ.1, 145.ꢀ, 1±ꢀ.ꢀ, 1±ꢁ.±, 129.4, 122.9, 12ꢁ.5, 115.6,
+
1ꢁ9.9, ±4.ꢀ, ±2.ꢁ, ±1.6, 29.± ppm; MS-FAB: m/z (%): 59ꢁ.8 (1ꢁꢁ) [M+H]
ꢁ
.
xanthene (212 mg, ꢁ.66 mmol). The white heterogeneous solution was re-
fluxed for 18 h eventually giving rise to a deep red homogenous solution.
The TFA was removed by distillation from the resulting red solution. The
red oil was taken up in ethyl acetate (5 mL), after which 2m aqueous
HCl (2ꢁ mL) was added and then the solution heated to 8ꢁ8C for 12 h.
The biphasic system was cooled then separated with ethyl acetate and
then concentrated in vacuo. The yellow oil was purified by chromatogra-
Template-mediated photodimerization: A PyrexNMR tube was charged
À2
with the template 3 (6.5ꢀ mg, 2.4î1ꢁ mmol) and coumarin 1c (6.1± mg,
À2
1
.±î1ꢁ mmol), which were then dissolved in acetone (2 mL). The
oxygen from the homogeneous solution was removed by a stream of
argon and the tube subsequently sealed. The sample was exposed to a
4
ꢁꢁ W lamp for 12 h and the product distribution determined by HPLC
˱
analysis. In the case of the asymmetric template, 4 (2.4 mg, 1.2î1ꢁ
mmol) was dissolved in acetone (1 mL) along with 1c (1.5 mg, 2.4î1ꢁ
phy (SiO
2
) with ꢀꢁ% ethyl acetate/±ꢁ% hexane and the product was iso-
˱
1
lated as a white solid (44 mg, 1±%). M.p. 155±16ꢁ8C; H NMR (2ꢁꢁ
MHz, [D]chloroform): d=1ꢁ.ꢀ1 (brs, 1H; NH), ꢀ.ꢀ8 (d, J=2.6 Hz, 1H;
Aryl), ꢀ.ꢀꢁ (d, J=2.1 Hz, 1H; Aryl), ꢀ.45 (d, J=2.6 Hz, 1H; Aryl), ꢀ.29
mmol) in a Young NMR tube, subjected to four cycles of purge-pump-
thaw, and then irradiated in the sealed tube. The same irradiation proce-
dure was adopted for the stoichiometric irradiation of template analogues
(
d, J=2.6 Hz, 1H; Aryl), ꢀ.ꢁꢀ (d, J=8.ꢀ Hz, 1H; Aryl), 1.68 (s, 6H;
5
and 6.
1
±
± ±
CH ), 1.±5 ppm (s, 18H; CH ); C NMR (5ꢁ MHz, [D]chloroform): d=
1
±
89.ꢀ, 14ꢀ.5, 129.ꢀ, 128.9, 124.8, 12±.9, 122.8, 122.ꢀ, 115.9, ±4.ꢀ, ±4.5, ±2.5,
1.6, ±1.4 ppm; MS-FAB: m/z (%): ±51.2 [M+H] ; elemental analysis
+
calcd (%) for C24
Alternatively, compound 8 can be isolated from the reaction mixture of 7
1ꢁ4 mg, ±±%).
N-(2,7-Di-tert-butyl-9,9-dimethyl-9H-xanthen-4-ylmethylene)-N’-pyridin-
-yl-hydrazine (4): Compound 8 (2ꢁ mg, ꢁ.ꢁ6 mmol) was dissolved chloro-
±ꢁ 2
H O (±5ꢁ.2): C 82.24, H 8.6±; found: C 82.ꢁ5, H 9.18.
Acknowledgments
(
We thank Dr. D. Hickman, Dr. P. Baxter and Dr. V. Berl for helpful sug-
gestions. W.G.S. also thanks the Natural Sciences and Engineering Re-
search Council Canada and the Universitÿ Louis Pasteur for financial
support.
2
form (1ꢁ mL) to which was added (15 mg, ꢁ.14 mmol) 2-hydrazinopyri-
dine; the mixture was then stirred at room temperature for 24 h. The sol-
2
vent was removed and then subjected to chromatography (SiO ) with a
gradient of 5% ethyl acetate/95% hexane to 5ꢁ% ethyl acetate/5ꢁ%
hexane. The product was isolated as a white solid (1ꢀ mg, 69%). M.p.
[
[
1] P. Beak, J. M. Zeigler, J. Org. Chem. 1981, 46, 619±624.
2] D. J. Trecker in Organic Photochemistry, Vol. 2: Photodimerizations
1
1
8
55±16ꢁ8C; H NMR (2ꢁꢁ MHz, [D]chloroform): d=8.95 (s, 1H; NH),
.41 (s, 1H; CH), 8.21 (d, J=5.1 Hz, 1H; Aryl), ꢀ.91 (d, J=2.1 Hz, 1H;
(
Ed.: O. L. Chapman), Marcel Dekker, New York, 1969, pp. 6±±116.
Aryl), ꢀ.64 (t, 1H; Aryl), ꢀ.42 (d, J=ꢀ.ꢀ Hz, 1H; Aryl), ꢀ.28 (s, 1H;
CH), ꢀ.22 (d, J=2.6 Hz, 2H; Aryl), ꢀ.ꢁ5 (d, J=8.ꢀ Hz, 1H; Aryl), 6.81
[
±] N. J. Turro, Modern Molecular Photochemistry, University Science
Books, Sausalito, 1991, p. 628.
(
t, 1H; Aryl), 1.66 (s, 6H; CH
±
), 1.±9 (s, 9H; CH
±
), 1.±5 ppm (s, 9H;
[
[
4] J. Eisinger, A. A. Lamola, Mol. Photochem. 1969, 1, 2ꢁ9±22±.
5] D. Elad, I. K. Rosenthal, S. Sasson, J. Chem. Soc. C 1971, 2ꢁ5±±
1
±
CH
±
); C NMR (5ꢁ MHz, [D]chloroform): d=14ꢀ.ꢀ, 146.6, 145.9, 145.±,
1
±
±8.1, 1±5.4, 1±ꢁ.2, 129.4, 124.5, 12±.8, 122,6, 121.4, 12ꢁ.5, 115.ꢀ, 1ꢁꢀ.6,
4.6, ±2.±, ±1.6 ppm; MS-FAB: m/z (%): 441.6 (1ꢁꢁ) [M+H] ; elemental
2
ꢁ5ꢀ.
+
[
[
6] P. J. Wagner, D. J. Bucheck, J. Am. Chem. Soc. 1970, 92, 181±185.
ꢀ] J. G. Burr in Advances in the Photochemistry, Vol. 6: Advances in the
Photochemistry of Nucleic Acid Derivatives (Eds.: W. A. Noyes, G.
S. Hammond, J. N. Pitts), Wiley, London, 1968, pp. 19±±±ꢁꢁ.
analysis calcd (%) for C29
C ꢀ9.ꢁ1, H 8.ꢁ4, N 9.58.
±5 ±
H N O (441.±): C ꢀ8.8ꢀ, H ꢀ.99, N 9.52; found:
N’-[2,7-Di-tert-butyl-5-(dimethylhydrazonomethyl)-9,9-dimethyl-8a,10a-
dihydro-9H-xanthen-4-ylmethylene]-N,N-dimethylhydrazine (5): N,N-Di-
methyl hydrazine (5ꢁ mL, ꢁ.65 mL) was added to a solution of dialdehyde
[
[
8] J. R. Heldt, J. Heldt, M. Ston, H. A. Diehl, Spectrochim. Acta Part A
1
995, 51, 1549±156±.
9] R. Hoffman, P. Wells, H. Morrison, J. Org. Chem. 1971, 36, 1ꢁ2±
ꢁ8.
1ꢁ] G. S. Hammond, C. A. Stout, A. A. Lamola, J. Am. Chem. Soc.
964, 85, ±1ꢁ±±±1ꢁ6.
11] H. Morrison, H. Curtis, T. McDowell, J. Am. Chem. Soc. 1966, 88,
415±5419.
7
(2± mg, ꢁ.ꢁ6 mmol) in ethanol (1ꢁ mL), solubilized by heating; the mix-
ture was then refluxed for 8 h. The solvent was removed under reduced
pressure and the white solid was further dried under vacuum to quanta-
1
[
[
1
tively afford compound 5. H NMR (2ꢁꢁ MHz, [D]chloroform): d=ꢀ.8
1
(
d, J=2.6 Hz, 4H; Aryl), ꢀ.±4 (d, J=2.1 Hz, 2H; CH), ±.ꢁ2 (s, 12H;
1
±
± ± ±
CH ), 1.±6 (s, 6H; CH ), 1.±6 ppm (s, 18H; CH ); C NMR (5ꢁ MHz,
5
[
±
D]chloroform): d=145.8, 1±ꢁ.ꢁ, 128.4, 12±.±, 122.ꢁ, 119.8, 42.9, ±4.ꢀ,
1.9, ±1.6 ppm; MS-FAB: m/z (%): 464.ꢀ (1ꢁꢁ) [M+H] .
[
[
12] K. Muthuramu, V. Ramamurthy, J. Org. Chem. 1982, 47, ±9ꢀ6±±9ꢀ9.
1±] F. D. Lewis, D. K. Howard, J. D. Oxman, J. Am. Chem. Soc. 1983,
105, ±±44±±±45.
+
N-Methyl-N-pyridin-2-ylhydrazine (9): N-methylhydrazine (2ꢁ mL, ±6ꢁ
mmol) was added to 2-bromopyridine (5 mL, 5 mmol) and the mixture
was refluxed under an argon atmosphere for 2 h. The residual solvent
was distilled off and the oil fully dried under vacuum. This residual oil
[14] F. D. Lewis, S. V. Barancyk, J. Am. Chem. Soc. 1989, 105, 865±±
8661.
[15] J. N. Moorthy, K. Venkatesan, R. G. Weiss, J. Org. Chem. 1992, 57,
was purified by chromatography (SiO
2
) in dichloromethane with a gradi-
±292±±29ꢀ.
ent to ±% methanol to yield the product as a slightly yellow oil (4.86 g,
[16] K. Muthuramu, N. Ramnath, V. Ramamurthy, J. Org. Chem. 1983,
48, 18ꢀ2±18ꢀ6.
1
ꢀ
9%). H NMR (2ꢁꢁ MHz, [D]chloroform): d=8.14 (dd, J=±.1 Hz, 2H;
NH), ꢀ.4± (t, 2H; Aryl), 6.95 (d, J=ꢀ.ꢀ Hz, 2H; Aryl), 6.55 (t, 2H;
[1ꢀ] L. H. Leenders, E. Schouteden, F. C. De Schryver, J. Org. Chem.
1973, 38, 95ꢀ±966.
1
±
Aryl), 4.ꢁ9 (brs, 2H; NH), ±.19 ppm (s, 6H; CH
±
)
C NMR (5ꢁ MHz,
Chem. Eur. J. 2003, 9, 556ꢁ ± 5566
www.chemeurj.org
¹ 2ꢁꢁ± Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5565

J.-M. Lehn et al.
FULL PAPER
[18] N. Ramnath, V. Ramamurthy, J. Org. Chem. 1984, 49, 282ꢀ±28±ꢁ.
[19] Y. Chen, C.-S. Jean, J. Appl. Polym. Sci. 1997, 64, 1ꢀ49±1ꢀ58.
[2ꢁ] Y. Chen, K.-H. Chen, J. Polym. Sci. Part A 1997, 35, 61±±624.
[21] F. C. De Schryver, J. Put, L. Leenders, H. Loos, J. Am. Chem. Soc.
[±4] S. Y. Jon, Y. H. Ko, S. H. Park, H.-J. Kim, K. Kim, Chem. Commun.
2001, 19±8±19±9.
[±5] H. Risler, Master×s dissertation, Universitÿ Louis Pasteur (France),
2002.
1
974, 96, 6994±ꢀꢁꢁꢁ.
[±6] W. G. Skene, V. Berl, H. Risler, J.-M. Lehn, 39th IUPAC Congress
and 86th Conference of The Canadian Society for Chemistry, Ottawa
(Canada) 2003.
[
[
22] D. G. Amirsakis, A. M. Elizarov, M. A. Garcia-Garibay, P. T. Glink,
J. F. Stoddart, A. J. P. White, D. J. Williams, Angew. Chem. 2003, 115,
1
158±1164; Angew. Chem. Int. Ed. 2003, 42, 1126±11±2.
[±ꢀ] V. Berl, Ph.D. thesis, Universitÿ Louis Pasteur (France), 2000.
[±8] T. Chiba, M. Koada, T. Kato, J. Heterocycl. Chem. 1982, 19, 1521±
1525.
[±9] K. A. Connors, Binding Constants: The Measurement of Molecular
Complex Stability, Wiley, 1987, p. 411.
[4ꢁ] P. Vishweshwara, R. Thaimattamb, M. JaskÛlski, G. R. Desiraju,
Chem. Commun. 2002, 18±ꢁ±18±1.
[41] G. R. Desiraju, Acc. Chem. Res. 2002, 35, 565±5ꢀ±.
[42] Based on F=ꢁ.ꢁ68 for direct and sensitized irradiation of 5,ꢀ-dime-
thoxy-chromen-2-one in acetonitrile; see: S. C. Shim, K. Y. Choi, P.
S. Song, Photochem. Photobiol. 1978, 27, 25±±1.
[4±] A hydrazinophenyl analogue of 3 rather than the hydrazinopyridine
employed would be a better alternative for model studies; however,
the former is photochemically and thermally very unstable.
[44] V. P. Solov×ev, V. E. Baulin, N. N. Strakhova, V. P. Kazachenko, V. K.
Belsky, A. A. Varnek, T. A. Volkova, G. Wipff, J. Chem. Soc. Perkin
Trans. 2 1998, 1489±1498.
2±] H. Tanaka, F. Toda, E. Mochiwuki, N. Yasui, Y. Kai, I. Miyahara, K.
Hirotsu, Angew. Chem. 1999, 111, ±ꢀ±±±±ꢀ±6; Angew. Chem. Int.
Ed. 1999, 38, ±52±±±525
[
[
[
24] P. Rolf, F. Herzog, A. Schuster (F. Hoffmann-La Roche Ag), US
6
2ꢁ1ꢁ8ꢀ, 2001.
25] K. Gnanaguru, N. Ramasubbu, K. Venkatesan, V. Ramamurthy, J.
Org. Chem. 1985, 50, 2±±ꢀ±2±46.
26] L. R. MacGillivray, J. L. Reid, J. A. Ripmeester J. Am. Chem. Soc.
2
000, 122, ꢀ81ꢀ±ꢀ818.
[
2ꢀ] T. Fri sœ cœ i cœ , L. R. MacGillivray, Chem. Commun. 2003, 1±ꢁ6±1±ꢁꢀ.
[28] P. P. Wells, H. Morrison, J. Am. Chem. Soc. 1975, 97, 154±159.
[29] S. J. Howell, N. Spencer, D. Philp, Tetrahedron 2001, 57, 4945±4954.
[±ꢁ] Y. Kato, S. Nishizawa, N. Teramae, Org. Lett. 2002, 4, 44ꢁꢀ±441ꢁ.
[±1] D. M. Bassani, V. Darcos, S. Mahony, J.-P. Desvergne, J. Am. Chem.
Soc. 2000, 122, 8ꢀ95±8ꢀ96.
[
±2] D. M. Bassani, X. Sallenave, V. Darcos, J.-P. Desvergne, Chem.
Commun. 2001, 1446±144ꢀ.
Received: June 24, 2ꢁꢁ± [F5268]
[
±±] T. K. Park, J. Schroeder, J. Rebek, Jr., J. Am. Chem. Soc. 1991, 113,
5
125±512ꢀ.
5566
¹ 2ꢁꢁ± Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 556ꢁ ± 5566