Solid-state photolysis of anthracene-linked ammonium salts: The search for topochemical anthracene photodimerizations

Article abstract of DOI:10.1016/S0040-4020(00)00508-1

The reaction of 9-N,N-dimethylaminomethylanthracene 1 with aromatic carboxylic acids gave crystalline salts (2a, 2b, 3a and 3b), which were irradiated in the solid state. Whereas salt 3a was selectively transformed to the dimer 4a, the other salts were ph

Full text of DOI:10.1016/S0040-4020(00)00508-1

TETRAHEDRON
Pergamon
Tetrahedron 56 (2000) 6867±6875
Solid-State Photolysis of Anthracene-Linked Ammonium Salts:
The Search for Topochemical Anthracene Photodimerizations
a,*
b
Heiko Ihmels, Dirk Leusser, Matthias Pfeiffer and Dietmar Stalke
b
b
a
Institute of Organic Chemistry, University of W uÈ rzburg, Am Hubland, D-97074 W uÈ rzburg, Germany
Institute of Inorganic Chemistry, University of W uÈ rzburg, Am Hubland, D-97074 W uÈ rzburg, Germany
b
Received 7 January 2000; accepted 20 February 2000
AbstractÐThe reaction of 9-N,N-dimethylaminomethylanthracene 1 with aromatic carboxylic acids gave crystalline salts (2a, 2b, 3a and
b), which were irradiated in the solid state. Whereas salt 3a was selectively transformed to the dimer 4a, the other salts were photoinert. The
3
results were rationalized on the basis of X-ray structure analysis. In the presence of oxygen, solid-state irrradiation of anthracene salt 3b leads
to selective photooxygenation. The selectivity of both solid-state photoreactions was not observed in solution. q 2000 Elsevier Science Ltd.
All rights reserved.
Introduction
actions are considered to be strong and are often accom-
panied by hydrogen bonding. These forces may lead to a
more selective crystallization process which could reduce
the formation of defect sites.
The performance of photoreactions in the crystalline state
and the continuing improvement of this methodology have
provided access to selectivities in solid-state photoreactions
1
that are not observed in solution photochemistry. The
Ammonium salts of the easily available aminomethyl-
substituted anthracene derivative 1 with selected aromatic
carboxylic acids were chosen as photoactive model
investigation of the solid-state structures of the photoactive
compounds and comparison with their reactivity have led to
fundamental knowledge of the geometric requirements and
8
compounds, which may photodimerize on irradiation. We
2
the steric course of well-established photoreactions. Due to
report the investigation of the photochemistry of selected
crystalline salts of amine 1 and the correlation of the solid-
state reactivity with their crystal structure.
the dense packing in the crystal lattice, solid-state reactions
proceed with a minimum of atomic or molecular motion.
Therefore, any reaction in the crystalline state is proposed to
3
take place with the least motion. Such a control of the solid-
state reactivity by the packing arrangement of the crystal is
known as the topochemical postulate. However, the photo-
dimerization of several crystalline anthracene derivatives
3
a,4
has been shown to be non-topochemical,
because the
product formation could not be directly correlated with
the observed solid-state structure. Reasonable explanations
of these striking exceptions from the topochemical postulate
are based on observations, that the reactions of the anthra-
5
6
cene derivatives take place at defect sites or on the surface
of the crystals. Since one of the factors which cause non-
topochemical behaviour may be the relatively high degree
of mobility of the chromophores in the crystal lattice, the
introduction of attractive ionic interactions by salt forma-
7
tion may be useful to achieve a topochemical reaction.
Thus, a positively or negatively charged chromophore
must remain in the close proximity of its counter ion,
which should restrict its mobility. Moreover, ionic inter-
Results
Synthesis
9-Dimethylaminomethylanthracene (1) and its hydro-
chloride salt 1´HCl were synthesized according to literature
Keywords: solid-state photolysis; topochemical anthracene photodimeriza-
tion.
*
Corresponding author. E-mail: ihmels@chemie.uni-wuerzburg.de
0040±4020/00/$ - see front matter q 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(00)00508-1

6
868
H. Ihmels et al. / Tetrahedron 56 (2000) 6867±6875
Figure 1. Structure of salt 2a in the solid state.
9
procedure. The salts of amine 1 were obtained by addition
Figure 3. Structure of salt 3b in the solid state.
of the corresponding aromatic carboxylic acid to the amine
1
and subsequent crystallization from methanol/ethyl
acetate, ethyl acetate or ethyl acetate/n-hexane. Whereas
the salts 2a and 2b crystallized in an amine-acid ratio of
showed an amine-acid ratio of 1:2. However, this salt could
not be obtained analytically pure.
1
:1, crystallization of a mixture of amine 1 with 2-furan-
Solid-State Structures
The ammonium salt 2a crystallized in the space group P1
Å
carboxylic acid or 9-anthracenecarboxylic acid yielded
solid samples of ammonium salts 3a and 3b with an
amine-acid ratio of 1:2. This ratio was observed on
crystallization from all the employed solvents and solvent
mixtures; even when a 1:1 mixture of the two components
was used. With 4-nitrobenzoic acid, the intentional use of an
amine-acid ratio of 1:2 for the salt formation resulted in
crystallization of the 1:1 salt 1b. Attempts to prepare
crystalline salts by the reaction of the amine 1 with anisic
acid, 4-bromobenzoic acid or naphthylacetic acid failed,
because these salts were obtained as viscous oils at room
temperature. With phenylacetic acid, an amorphous solid
was obtained by slow solidi®cation of the initially precipi-
(from methanol/ethyl acetate). The aromatic carboxylate
ions form an almost planar layer in the crystal lattice,
and the anthracene moieties are arranged perpendicular
to this plane (Fig. 1). Moreover, the ammonium func-
tionality is embedded in the anionic aromatic layer by
hydrogen bonding to the carboxylate group, whereas
two anthracene moieties arrange as closest pairs with
a head-to-tail orientation. The molecular planes of the
two anthracene molecules are almost coplanar.
Single crystals of salt 3a were obtained by crystallization
from ethyl acetate. Interestingly, the achiral compound crys-
tallizes in the chiral space group P1. In the solid state, one
furancarboxylate moiety and one free acid molecule form
hydrogen-bonded, nearly planar complexes, which build up
a distorted layer within the crystal lattice (Fig. 2). Like in
the solid-state structure of salt 2a, the ammonium function-
alities are hydrogen bonded to the carboxylate ions. The
connection of the ammonium substituent with the carboxyl-
ate layer results in the formation of stacks of anthracene
chromophores which are located between two furan layers.
The anthracene moieties are arranged as pairs of closest
molecules in a head-to-tail orientation.
1
tated viscous oil at ±48C. H NMR-spectroscopic analysis
The ammonium salt 3b also crystallizes in an amine-acid
ratio of 1:2 (Fig. 3). Resembling the solid-state arrangement
of salt 3a, three molecules are connected by hydrogen bond-
ing: One free carboxylic acid moiety is hydrogen bonded to
the carbonyl functionality of the carboxylate, whereas the
negatively charged oxygen atom of the latter is connected to
the ammonium hydrogen atom. But compared to salt 3a, the
molecular planes of the aromatic substituents of the acid and
the carboxylate are not coplanar but perpendicular towards
Figure 2. Structure of salt 3a in the solid state.

H. Ihmels et al. / Tetrahedron 56 (2000) 6867±6875
6869
±
5
Figure 4. Emission spectra of the anthracene derivatives 1, 2b, 3a and 3b. 1: 3a in MeOH (10 M); 2: crystalline 1; 3: crystalline 2b; 4: crystalline 3b; 5:
crystalline 3a.
each other. Moreover, there is no head-to-tail oriented pair
of two ammoniummethyl-substituted anthracene moieties in
the solid-state. Instead, the closest coplanar arrangement of
two anthracene chromophores consists of one ammonium-
methyl-substituted anthracene and one anthracene carboxyl-
ate molecule; however, these molecules are signi®cantly twisted
towards each other and separated by relatively large distances.
Solid-State Emission Spectra
Fig. 4 shows the solid-state emission spectra of the amine 1,
and its salts 2b, 3a and 3b compared with the ¯uorescence
spectrum of salt 3a in methanol. The solution spectrum of
salt 3a resembles the one observed for unsubstituted anthra-
cene. The solid-state emission of the amine 1 also exhibits a
structured ¯uorescence pattern; however, the band maxima
at longer wavelength are signi®cantly broadened. The solid-
state emission spectra of the salts 2b and 3b exhibit one
broad, red-shifted emission band at lˆ462 nm and
lˆ470 nm, respectively, whereas salt 3a emits with a
very broad band at lˆ513 nm. The salt 2a does not emit
in the solid state.
The salts 2a, 2b and 3b did not react on solid-state irradia-
tion (l.320 nm); however, in the case of the anthracene
salt 3b, small amounts of anthraquinone 5 were detected by
1
H NMR spectroscopy. When the irradiation of a solid
sample of anthracene salt 3b was conducted in an oxygen-
gas atmosphere, enhanced formation of anthraquinone 5
was observed. No other product could be detected under
these conditions. When the salts 2b and 3a were irradiated
under oxygen-gas atmosphere, only trace amounts of
1
Solid-State Photolyses
anthraquinone 5 were detected by H NMR spectroscopy;
The free amine 1 was crystallized from n-hexane or metha-
nol, and irradiation of both solid samples did not result in a
photoreaction. The solid hydrochloride salt 1´HCl, which
was crystallized from methanol/ethyl acetate, was also
photoinert. However, irradiation of a thoroughly ground
crystalline sample of salt 3a (l.320 nm) gave the head-
to-tail dimer 4a in almost quantitative yield as was deter-
even after prolongued irradiation times. In the case of
compound 3a, the formation of the dimer 4a was observed
under these conditions, but the reaction rate is signi®cantly
slower than under inert-gas atmosphere. Irradiation of
anthracene salt 2a in the presence of oxygen did not result
in anthraquinone formation.
1
mined by H NMR spectroscopy. Small amounts (,5%) of
The head-to-tail structure of the dimer 4a was deduced from
the characteristic signal of the bridgehead protons in the H
1
non-isolable, unidenti®ed byproducts were also detected.
Bulk single crystals exhibited a very slow reaction under
the same conditions. For example, after 8 h of irradiation the
ground sample showed a monomer/dimer ratio of ca. 1:1,
whereas on photolysis of the bulk single crystal under the
same conditions, a 10:1 ratio was observed. Several ®lters
were employed to investigate the wavelength-dependency
of the photolysis of salt 3a: With every ®lter used (Schott
GG 395, WG 360, WG 320, WG 295), the result of the
photoreaction did not change.
NMR spectrum, because the observed shift of dˆ3.88 is
comparable to the reported one for the closely related
aminomethyl-substituted anthracene dimer 4b (dˆ3.70).
8
The bridgehead proton of the head-to-head dimer is
expected give a H NMR signal at higher ®eld, as was
observed for the head-to-head isomer of dimer 4b
1
8
(dˆ4.15). Moreover, the head-to-tail connection was
con®rmed by a NOE effect between the bridgehead proton
and the opposite methylene group. The characteristic signals

6
870
H. Ihmels et al. / Tetrahedron 56 (2000) 6867±6875
Figure 5. De®nition of the geometric parameters d, x, y and a.
of the bridgehead carbon atoms (C-9: dˆ57.9; C-10: 62.7)
and from the amine-acid ratio which is present in solution.
Thus, the crystallization process of the salts 3a and 3b
results from a selective self-assembly in the solid state. It
may be assumed that this self-organization depends on the
acidity of the carboxylic acid; however, the salt of amine 1
with 4-nitrobenzoic acid (pK ˆ3.41), whose pK value is
further con®rmed the dimeric structure.
For comparison, the ammonium salts 1´HCl, 3a and 3b were
also irradiated in solution. After 2 h of irradiation of salt 3a
in methanol, the dimer 4a was the main product (.90%;
a
a
1
conversion 40%). However, H NMR-spectroscopic analy-
comparable with the one of 2-furancarboxylic acid
sis of the crude reaction mixture revealed that several by-
products were also formed in small amounts. At full
conversion, these byproducts accumulate; and additional
byproducts were formed. The dimer 4b could not be isolated
from the reaction mixture. Irradiation of anthracene salt 3b
(pK ˆ3.23), does not crystallize in an amine-acid ratio of
a
1:2 ± even when such a ratio was initially used for crystal-
lization.
In contrast to the salt 3a, the anthracene salt derivatives
1´HCl, 2a, 2b and 3b did not dimerize on irradiation in
the solid state. These observations may be explained by
the arrangement of the chromophores in the crystalline
state. Thus, the photoinertness of the solid samples may
be caused by a large separation between the reaction centers
of two anthracene moieties, whereas the photodimerization
of salt 3a should result from a favorable arrangement of the
chromophores in the crystalline medium. The solid-state
structure analysis of salts 2a and 3a reveals that both salts
form crystal lattices in which two anthracene molecules
arrange as a close coplanar pair in a distorted head-to-tail
orientation. According to structure-reactivity correlation
studies on the solid-state photodimerization of anthracene
in DMSO-d under inert-gas atmosphere also yielded a
6
1
complex reaction mixture. It was shown by H NMR-
spectroscopic analysis of the photolysate, that the head-to-
tail dimer 4c was formed as the main product, which was
indicated by a singlet at dˆ5.70 whose shift is comparable
to the one of the head-to-tail dimer of 9-anthracene-
1
0
carboxylic acid (dˆ5.63). Irradiation of salt 3b in
oxygen-saturated DMSO-d also yielded the dimer 4c
6
among other unidenti®ed byproducts. Anthraquinone 5
was only formed in trace amounts (.5%) under these
conditions. Irradiation of the hydrochloride salt 1´HCl in
methanol gave the head-to-tail dimer as main product; but
as observed in the case of salt 3a, several byproducts were
also formed in small amounts.
4
and its derivatives, we de®ned the parameters d, Dx, Dy
and a to obtain a quantitative analysis of the solid-state
0
arrangement (Fig. 5). The distances d and d between the
Discussion
opposite reactive meso positions (i.e. C-9 and C-10), which
may form the new C±C bonds in the photodimerization,
were taken as structural parameters, whose ideal value
The ammonium salt 3a represents one of the few achiral
molecules that crystallize in chiral space groups. Such a
3
a,4
should be d,420 pm. The deviation from an ideal super-
posed position of two anthracene molecules was quanti®ed
by the shift (Dx and Dy) of one molecule with respect to
another along the long and the short molecule axes, x and y.
In the case of a perfect superposition, the ideal values are
DxˆDyˆ0 pm. Moreover, the parameter a de®nes the
twisting angle between two anthracene molecules, i.e. the
deviation from a parallel orientation of the long molecule
1
c,11
behaviour is rarely observed,
and in most cases, there
is no correlation between the tendency of a molecule to
crystallize in a chiral space group and a distinct structural
feature. Since the other ammonium salts 2a and 3b, crystal-
lize in achiral space groups, the ammoniummethyl-
substituted anthracene moiety is unlikely to be a general
structural element that forces crystallization in a chiral
one. However, it should be noted, that other achiral,
axes x and x (Fig. 5). Thus, a is the angle between the long
1 2
9
-substituted anthracene derivatives are known, that crystal-
molecule axes of two anthracene molecules which are
projected to one molecule plane. The ideal angle a should
1
2
14
lize in chiral space groups. Another remarkable property
of salts 3a and 3b is their crystallization in an amine-acid
ratio of 1:2. The solid-state structure analysis of salts 3a and
be 08. The observed structural features of compounds 2a and
3a (Table 1) reveal that in both salts, the distances d
between the meso positions of the anthracene moieties are
smaller than 400 pm. These distances are generally con-
sidered to be suf®ently short for a photodimerization in
3
b reveals that the association of a carboxylate functionality
with one carboxylic acid molecule is the basic structural
element of this arrangement, which has also been observed
with other carboxylates. Interestingly, the formation of
this solid-state structure is independent from the solvent
1
3
4
the solid state. However, in both cases, the anthracene
frameworks have to move signi®cantly in order to achieve

H. Ihmels et al. / Tetrahedron 56 (2000) 6867±6875
Table 1. Geometric parameters d, d , Dx, Dy and a of the salts 2a, 3a and 3b derived from X-ray analysis
6871
0
d [pm]
0
386.4
386.4
1.44
1.10
0
354.8
393.5
1.20
0.29
501.7
397.7
d [pm]
a
Dx [pm]
Dy [pm]
a [deg]
a
23.40
77.5
a
Since the molecule axes are nearly perpendicular towards each other, the determination of Dx and Dy is not useful.
a perfect superposition of the p systems, which is a neces-
sary prerequisite for a photodimerization. To achieve a
better overlap, one anthracene moiety of salt 2a has to
shift along both molecule axes for more than 100 pm (Dx
and Dy), whereas the aromatic rings of salt 3a only have to
shift signi®cantly along one molecular axis (Dx) along with
a twist (aˆ23.48) of the two molecules towards each other.
Although, such deviations from an ideal arrangement of the
two chromophores may hinder the photodimerization in the
solid state, the photoreaction must not necessarily be
suppressed. Wang and Jones showed for example, that
such a deviation from the ideal superposition does not
inhibit a photodimerization, because shifts along the
molecular axes larger that 100 pm still allow the solid-
state photodimerization of acridizinium salts, whose
tion. In the case of salt 3a, the twisting of one anthracene
moiety may lead to a more favorable orientation of the two
chromophores towards each other, because the values for Dx
and a become smaller during the twisting movement and
should thus facilitate the photodimerization.
The observed relatively slow solid-state photoreaction of
salt 3a shows that the geometric arrangement of the chromo-
phores is far from being ideal and may be close to the limit
beyond which a photoreaction is no longer possible. It was
observed, that the reaction is getting slower with prolongued
reaction times and that the bulk single crystal has a signi®-
cantly lower conversion compared to carefully ground solid
samples. Since the photodimer does not absorb at wave-
lengths higher that lˆ300 nm, this photoproduct does not
act as an internal light ®lter which may slow down further
photoreactions. These observations lead to the conclusion,
that the reaction does not take place at a defect side, since
such reactions were shown to get faster after an induction
1
5
structure is closely related to anthracene.
To rationalize the photoreactivity of the salts 2a and 3a, it is
helpful to further investigate the interaction between the
ionic centers of the molecules. The packing diagram reveals
that the ammonium moiety is embedded in the dense layer
which is formed by the aromatic carboxylate counterions.
This arrangement is likely to be governed by attractive ionic
interactions and by hydrogen bonding between the proto-
nated amine and the carboxylate oxygen atom. Thus, in
addition to the steric constraints of the crystal packing, the
ammonium functionality is `anchored' within the counter-
ion layer, which should further restrict the freedom of move-
ment of the anthracene moiety in the crystal lattice. Consid-
ering this `®xed' ammonium functionality, the least
hindered motion of the anthracene moiety may result from
a rotation around the CH ±C bond or around the N±CH
4
period. However, it cannot be excluded, that the photoreac-
tion proceeds mostly on the surface of the solid, a phenom-
enon which is frequently observed in solid-state
photochemistry.
1
b,1e
In contrast to the observed solid-state structures of salts 2a
and 3a, X-ray structure analysis of salt 3b shows that the
anthracene moieties of this compound do not form close
coplanar pairs with a parallel alignment of the p systems.
The closest anthracene pair observed consists of one ammo-
niummethyl-substituted and one 9-anthracenecarboxylate.
However, the separation between the meso positions is
rather large (dˆ397.7 and 501.7 pm). Furthermore, the
molecules are severely twisted towards each other by
77.58. Consequently, a solid-state photodimerization is not
expected due to a highly unfavorable orientation of the
chromophores.
2
ar
2
bond. The latter leads to a twisting of the anthracene
moieties towards each other rather than to a translatory
shift. In the case of salt 1´HCl, such a motion results in a
less favorable arrangement of the two chromophores,
because the improvement of the Dx value will be accom-
panied by a signi®cant increase of the angle a. Thus, the
combination of the signi®cant deviation from an ideal super-
position of two anthracene molecules with their restricted
mobility should result in the suppression of the dimeriza-
A reasonable correlation between the solid-state reactivity
and the solid-state emission spectrum of the chromophores
was observed. The photoinert salts 2b and 3b exhibit a
broad emission band at lˆ460±470 nm, which is suggested

6
872
H. Ihmels et al. / Tetrahedron 56 (2000) 6867±6875
to be a red-shifted monomer emission. Such a bathochromic
shift compared to the emission in solution may result from
partial overlap of the chromophore with the p system of
another anthracene moiety. Thus, by comparison with the
salts 2b and 3b, it may be concluded that the more
pronounced red-shift of the solid-state emission of salt 3a
results from a stronger overlap between the opposing p
systems of the chromophores. In fact, this assumption was
con®rmed by X-ray structure analysis results of salts 3a and
9-substituted anthracene peroxides to anthraquinone with
2
2
the loss of the substituent has been obsereved; however,
the yields are relatively low and additional rearrangement
products are also formed in solution. Since the formation of
such rearrangement products requires signi®cant molecular
motion, we assume that in the case of endoperoxide 6, the
corresponding reaction pathways are suppressed by the
steric constraints in the crystal lattice. Thus, the formation
of anthraquinone 5 from solid-state irradiation of salt 3b
under oxygen-gas atmosphere represents a highly selective
reaction. This selectivity is even more remarkable consider-
ing the observation that photolysis of salt 3b in an oxygen-
saturated solution only gave trace amounts of anthraquinone
5. Analysis of the reaction mixture showed that both the
carboxy- and the ammoniummethyl-substituted anthracene
were consumed during anthraquinone formation. Thus, all
three anthracene moities per salt unit are photooxygenated
in the solid state. At this stage, the distinct mechanism of
this reaction is not quite clear and further experiments to
isolate the intermediate peroxide 6 and to identify subse-
quent reaction intermediates are necessary.
3b. The signi®cant shift of the emission band of salt 3a may
be caused by excimer formation, since the observed band
shift maximum at lˆ513 is comparable to the known
1
6c
excimer emission of anthracene derivatives. Such a corre-
lation of the red-shift of the solid-state ¯uorescence band
with the degree of intermolecular p ±p interaction has
already been observed for several alkenes and polycyclic
1
6
aromatic molecules, and may serve as an ef®cient tool to
evaluate the orientation of two p systems towards each
other in the solid state. From these results, it may be
deduced that two anthracene moieties of the solid salt 2b
are overlapping to almost the same extend than the ones of
salt 3b. The emission spectrum of the amine 1 represents an
ensemble of a nearly `undisturbed' anthracene emission and
a band which may result from ¯uorescence of one
anthracene chromophore which partially overlaps with a
neighboring aromatic p system. Moreover, the latter
broad band may also result from exciplex formation of the
amine functionality and the anthracene chromophore.
Irradiation of salt 3a in methanol gave dimer 4a as main
reaction product only at low conversion. At full conversion,
a considerable number of byproducts was formed. This may
be due to the formation of the head-to-head dimer and addi-
tional reaction pathways which result from a photo-induced
The formation of the peroxide 6 requires the presence of
singlet oxygen, since the reaction between singlet-state
anthracene and triplet oxygen is spin forbidden. The singlet
oxygen may result from a photosensitization of ground-state
triplet oxygen by a singlet-excited anthracene chromophore.
The ef®ciency of this process is likely to be caused by the
presence of three anthracene chromophores per salt
molecule. That such a high density of anthracene molecules
may be required for an ef®cient photooxidation is shown by
the observation that the other anthracene salts 2a, 2b and 3a
only yielded traces of anthraquinone 5 under the same
conditions.
cleavage of the CH ±N bond. Such photocleavage reactions
2
of benzylammonium salts are known to yield a variety of
photoproducts. On solid-state photolysis of salt 3a, this
1
7
1
8
reaction pathway is obviously suppressed. However, it is
not clear whether the initially formed radicals quickly
recombine or whether these reactive intermediates are not
formed at all in the solid state. In order to trap the radicals
with an alcohol functionality, the 4-hydroxymethyl-
substituted benzoate was used as counter ion. It was
proposed that salt 2b arranges in a similar crystal lattice
than 2a, so that the hydroxy fuctionality may be positioned
in the near vicinity of the formed radicals. However, no
trapping product was observed on irradiation of salt 2b, so
that it may be concluded that the radicals which result
from a photo-induced C±N-bond cleavage are not
formed.
In conclusion, most derivatives of salts of 9-N,N-dimethyl-
aminomethylanthracene 1 with selected aromatic carboxylic
acids are photoinert in the solid state. In the case of salt 2a, it
was shown that the suppressed reactivity is mainly caused
by attractive ionic interactions and hydrogen bonding which
restricts the freedom of movement of the anthracene
chromophores, which need to move signi®cantly for a
photodimerization. However, the photoactive salt 3a repre-
sents one example for a selective solid-state photoreaction,
because irradiation of crystalline samples gave almost
exclusively the head-to-tail dimer 4a, whereas solution
photolysis yielded a large number of products. The trans-
formation of anthracene salt 3b to anthraquinone 5 under
oxygen-gas atmosphere also represents a highly selective
solid-state photoreaction. Since it was shown that the
anthracene molecules may act as photosensitizers in the
solid state and no disturbing photodimerization takes
place, this structural feature may offer new opportunities
Although the anthracene salt 3a does not photodimerize in
the solid state, a remarkably selective reaction to give
anthraquinone
atmosphere. It is already known that autoxidation of solid
5
was observed under oxygen-gas
1
9
anthracene yields anthraquinone 5 as reaction product, and
,10-dicyanoanthracene is photooxygenated to anthra-
quinone 5 in solution. Moreover, it has also been reported
9
that anthracene peroxide rearranges to anthraquinone on
X-ray irradiation in the solid state. Thus, we assume that
the initially formed photoproduct of salt 3b may be the
2
0
2
1
endoperoxide 6 which subsequently rearranges to the
anthraquinone 5. The thermally-induced transformation of
2
3
for selective solid-gas-phase photoreactions.

H. Ihmels et al. / Tetrahedron 56 (2000) 6867±6875
Table 2. Crystallographic data of compounds 2a, 3a and 3b
3a 3b
6873
shock-cooled crystals on an Enraf±Nonius CAD4 four
circle diffractometer (graphite±monochromated Mo-Ka
radiation, lˆ71.073 pm) equipped with a low temperature
2
a
2
4
empirical formula
formula weight
crystal size [mm]
crystal system
space group
a [pm]
b [pm]
c [pm]
a [8]
b [8]
C
402.44
24
H
22
N
2
O
4
C
459.48
0.5£0.4£0.4
triclinic
P1
835.73(14)
1116.39(21)
1374.19(23)
76.659(14)
73.967(14)
69.180(15)
1.1392(3)
2
1.340
0.095
484
3.12 to 22.51
5898
5898
3
622
0.0492
0.0347
0.0820
27
H
25NO
6
C
679.78
47
H
37NO
4
device at 193(2) K.
No absorption correction was
employed. The structures were solved by direct methods
(
ods against F (Shelxl-97 ).
0.5£0.5£0.4
0.3£0.3£0.1
25
Shelxs-97 ) and re®ned by full-matrix least squares meth-
triclinic
monoclinic
2
26
Å
P1
1
P2 /c
760.44(9)
1077.32(13)
1223.88(15)
97.595(7)
97.465(7)
92.959(7)
0.9830(2)
2
1.360
0.093
424
3.20 to 22.45
2905
2563
1421.75(22)
1667.67(25)
1569.39(24)
90
108.966(21)
90
3.5190(9)
4
1.283
0.08
1432
3.00 to 22.51
4987
4585
0
480
0.1305
0.0544
0.1218
All non-hydrogen atoms were re®ned with anisotropic
displacement parameters. The following hydrogen atoms
were located by difference Fourier syntheses and re®ned
isotropically: H1N bonded to N1 in 2b. H1O and H2O
bonded to O1 and O2, and H1N and H2N bonded to N1
and N2, respectively, in 3a. H1N bonded to N1 and H1O
bonded to O4 in 3b. All other hydrogen atoms of the
molecules were assigned ideal positions and re®ned iso-
tropically using a riding model with Uiso constrained to
1.2 or 1.5 times the Ueq of the parent atom. Additional
crystallographic data of the re®nements and the data collec-
tions of compound 2a, 3a and 3b are given in Table 2.
Crystallographic data (excluding structure factors) for the
structures reported in this paper has been deposited with the
Cambridge Crystallographic Data Centre as supplementary
publication no. CCDC-138799 (2a), CCDC-138800 (3a)
and CCDC-138801 (3b). Copies of the data can be obtained
free of charge on application to CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK [fax: (internat.) 144(1223)336-
033; e-mail: deposit@ccdc.cam.ac.uk].
g [8]
V [nm
Z
23
]
2
3
r
m
c
[Mgm
]
±
1
]
c
[mm
F(000)
u range [8]
no. re¯ns. measd.
no. unique re¯ns.
no of restraints
0
278
0.0386
0.0321
0.0913
no of parameters
R1 [I.2s(I)]
a
a
R1 [I.4s(I)]
b
wR2 (all data)
c
2
1
g ; g
0.0536; 0.2048 0.0473; 0.0040 0.0486; 0
0.0100(13) 0.00219(42)
largest diff peak and 139 and 2154 146 and 2132 172 and 2168
extinction coef®cient 0.029(3)
23
hole (e nm
)
Pꢀꢀ
ꢀ
ꢀ
ꢀ
ꢀꢀ
ꢀꢀ
ꢀ
ꢀ
ꢀ
F
2 F
ꢀ
a
o
c
R1 ˆ
Pꢀ
ꢀ
ꢀ
Fo
s
P
2
2
2
c
wꢀF 2 F †
Photolysis in solution and in the solid state. The solution
photolyses were performed in deoxygenated solvents with a
150 W mercury high-pressure lamp (Heraeus TQ 150).
Solid-state irradiations were carried out with a 500 W
mercury high-pressure lamp (Heraeus TQ) with the crystal-
line samples crushed between Pyrex slides in argon-¯ushed
polyethylene bags or by distributing the ground solid on the
inner walls of a test tube (DURAN glassware). If not stated
otherwise, the samples were placed without further cooling
ca. 5 cm in front of the cooling mantle of the lamp. The
b
c
o
wR2 ˆ
w ˆ
P
2
2
wꢀF †
o
2
o
2
c
1
ꢀmaxꢀF ; 0† 1 2F †
;
P ˆ
2
2
o
2
s ꢀF † 1 ꢀg P† 1 g P
3
1
2
Most notably, no example for a non-topochemical solid-
state reaction was observed, because the photochemical
behaviour of the crystalline salts 2a and 3a,b could be
successfully correlated with their solid-state structures.
1
photolysate was analyzed by means of H NMR. In prepara-
tive runs the solid samples (20±40 mg) were placed in a
¯ask (DURAN glassware) and ¯ushed with argon. During
photolysis the ¯ask was frequently removed and placed in a
sonic bath for 1±2 minutes.
Experimental
1
13
General Methods. H NMR and C NMR: Bruker AC 200
1
13
H NMR: 200 MHz; C NMR: 50.3 MHz); H NMR
1
(
1
3
chemical shifts refer to dTMSˆ0.0. C NMR chemical shifts
N-9-Anthracenylmethyl-N,N-dimethylammonium 4-nitro-
benzoate (2a). To a mixture of 75 mg (0.32 mmol) of amine
1 in 3 mL of ethyl actetate were added 58 mg (0.32 mmol)
of 4-nitrobenzoic acid in 2 mL of methanol. The solution
was stirred at 508C for 10 min, concentrated to ca. 2±3 mL
and cooled slowly to 248C to give 99 mg (0.25 mmol, 74%)
of salt 2a as bright yellow prisms, mp 159±1608C. UV
(MeOH): lmax (log e)203 (4.23), 215 (4.25), 250 (4.57),
refer to solvent signals (DMSO-d : dˆ39.5; CD OD:
6
3
dˆ49.2, CDCl ˆ77.0). UV: Hitachi U-3200 spectro-
3
photometer. Fluorescence: Perkin±Elmer LS50 spectro-
meter. Solid-state ¯uorescence spectra were recorded with
a Front Surface Accessory, Perkin±Elmer. Emission spectra
were recorded with an excitation wavelength of
l ˆ380 nm (with salt 3a in methanol: l ˆ370 nm)
ex
ex
1
Elemental analyses were performed by Mr. C.-P. Kneis at
the University of W uÈ rzburg, Institute of Inorganic
Chemistry. Melting points were determined with a B uÈ chi
B-545 and are uncorrected.
331 (3.50), 347 (3.75), 365 (3.92), 384 (3.89). H NMR
(CD
7.51±7.72 (m, 4 H, ar-H), 8.01±8.18 (m, 6 H, ar-H), 8.40
OD) d 2.89 (s, 6 H, NCH ), 5.31 (s, 2 H, CH N),
3 3 2
1
3
(dd, Jˆ9 Hz, 1 Hz, 2 H, ar-H), 8.68 (s, 1 H, ar-H). C NMR
(
CD OD) d 44.1 (NCH ), 54.1 (CH N), 122.9 (ar-C ), 124.1
3
3
2
q
9-N-Dimethylaminomethylanthracene 1 (mp 65±668C)
and its hydrochloride 1´HCl (mp 252±2548C, dec.) were
synthesized according to Marcus and Fitzpatrick.
(ar-CH), 124.6 (ar-CH), 126.8 (ar-CH), 129.1 (ar-CH),
130.9 (ar-CH), 131.4 (ar-CH), 132.2 (ar-CH), 133.0 (ar-
9
C
not detected. Anal. calcd for C H N O (459.50) C: 71.63,
), 144.6, (ar-C
q
), 150.6 (ar-C
), 172.1 (CˆO), one ar-C
q
q
q
2
4
24
2
4
X-Ray Crystallography. All data were collected from
H: 5.51, N: 6.96; found C: 71.25, H: 5.36, N: 6.59.

6
874
H. Ihmels et al. / Tetrahedron 56 (2000) 6867±6875
1
N-9-Anthracenylmethyl-N,N-dimethylammonium
hydroxymethyl benzoate (2b). A mixture of 70 mg
0.46 mmol) of amine 1 and 108 mg (0.46 mmol) of
-hydroxymethylbenzoic acid was dissolved in 5 mL of
4-
H NMR signals in the aromatic region and at d2.5±4.5 (in
CD OD). On concentration of the solution by evaporation of
3
(
4
solvent, a white solid precipitated, which was washed three
times with ca. 5 mL of ethyl acetate to give 31 mg of dimer
4a as white solid (0.06 mmol, 46%). mp 248±2508C (dec.)
methanol/ethyl acetate (1:1). Slow evaporation of the
solvent at room temp. gave 121 mg (0.33 mmol, 68%) of
UV (MeOH): lmax (log e)ˆ2.03 (4.32), 216 sh (4.11), 254
1
salt 2b as yellow cubes, mp 150±1518C. H NMR (CD OD)
1
(3.11), 296 (zero-onset). H NMR (CD OD) d 2.63 (s, 12 H,
3
3
d 2.71 (s, 6 H, NCH ), 4.63 (s, 2 H, CH OH), 5.04 (s, 2 H,
3
NCH ), 4.03 (s, 2 H, 9-H), 4.58 (s, 4 H, CH N), 6.94±7.17
2
(m, 18 H, ar-H). C NMR (D O/CD OD, 5:1) d46.2
2 3
(NCH ), 57.3 (C-9), 59.1 (CH N), 61.6 (C-10), 127.6 (ar-
3 2
2
3
1
3
CH N), 7.36 (d, Jˆ8 Hz, 2 H, ar-H), 7.49±7.68 (m, 4 H,
2
ar-H), 7.91 (d, Jˆ8 Hz, 2 H, ar-H), 8.10 (dd, Jˆ9 Hz, 1 Hz,
2
H, ar-H), 8.41 (d, Jˆ9 Hz, 2 H, ar-H), 8.63 (s, 1 H, ar-H).).
CH), 127.8 (ar-CH), 128.5 (ar-CH), 129.7 (ar-CH), 140.7
(ar-C ), 142.5 (ar-C ).
1
3
C NMR (CD OD) d 44.7 (NCH ), 54.6 (CH N), 65.0
3
3
2
q
q
(
CH OH), 125.0 (CH), 125.3 (C ), 126.6 (CH), 127.5
2 q
(
1
CH), 128.5 (CH), 130.8 (CH), 130.9 (CH), 131.2 (CH),
32.9 (C ), 133.1 (C ), 135.4 (C ), 146.5 (C ), 173.4
Basi®cation of the dimer with aqueous NaOH and subse-
quent extraction with dichloromethane gave the free amine
whose H NMR spectrum was identical with the one
q
q
q
q
1
(
6
CˆO). Anal. calcd for C H NO (387.9) C: 77.49, H:
2
5
25
3
.50, N: 3.61; found C: 77.25, H: 6.40, N: 3.67.
obtained from the photodimer of amine 1.
N-9-Anthracenylmethyl-N,N-dimethylammonium
furancarboxylate´2-furancarboxylic acid complex (3a).
2-
Photodimerization of amine 1. A solution of 45 mg
(0.19 mmol) of amine 1 was irradiated for 6 h in 8 mL of
n-hexane (l.320 nm). During photolysis, a yellow solid
precipitated, which was washed twice with 4 mL of
n-hexane. The remaining solid was crystallized from
A mixture of 121 mg (0.51 mmol) of amine 1 and 115 mg
1.03 mmol) of 2-furancarboxylic acid was dissolved in
(
3
1
mL of ethyl acetate. The solution was re¯uxed for
0 min and then cooled slowly to 248C to give 185 mg
methanol/dichloromethane to give 26 mg (0.05 mmol,
58%) of the head-to-tail photodimer. H NMR (CDCl ) d
1
(
0.41 mmol, 78%) of salt 3a as pale yellow prisms, mp
3
9
(
7±988C. UV (MeOH): lmax (log e)ˆ336 (3.45), 350
2.21 (s, 12 H, NCH ), 3.69 (s, 4 H, CH N), 3.81 (s, 2 H, 10-
3
H), 6.70±7.00 (m, 16 H, ar-H).
2
1
3.75), 368 (3.89), 386 (3.81). H NMR (CD OD) d 2.93
3
(
s, 6 H, NCH ), 5.38 (s, 2 H, CH N), 6.48 (dd, Jˆ2 Hz,
3
2
3
7
8
Hz, 2 H, ar-H), 7.02 (dd, Jˆ3 Hz, 1 Hz, 2 H, ar-H),
Photodimerization of salt 3a in the solid state. The salt 3a
was crystallized from ethyl acetate, and the crystals were
thoroughly ground. A portion of 20 mg (0.04 mmol) was
placed on the inner wall of a DURAN test tube and
irradiated at 108C (l.320 nm) for 32 hours. Analysis of
.49±7.69 (m, 6 H, ar-H), 8.09 (d, Jˆ9 Hz, 2 H, ar-H),
1
3
.39 (d, Jˆ9 Hz, 2 H, ar-H), 8.65 (s, 1 H, ar-H). C NMR
(
CD OD) d 43.9 (NCH ), 53.8 (CH N), 112.7 (ar-CH),
3 3 2
117.0 (ar-CH), 121.9 (ar-C ), 124.5 (ar-CH), 126.8 (ar-
CH), 129.2 (ar-CH), 130.9 (ar-CH), 132.5 (ar-CH), 132.9
q
1
the crude reaction mixture by N-NMR spectroscopy
showed that the photodimer has forme1d almost
(
ar-C ), 133.0 (ar-C ), 146.6 (ar-CH), 149.5 (ar-C ), 164.5
q
q
q
(
5
CˆO). Anal. calcd for C H NO (459.50) C: 70.58, H:
quantitatively (.95%, conversion 78%).
H
NMR
2
7
25
6
.48, N: 3.05; found C: 70.30, H: 5.24, N: 3.05.
(CDCl ) d 2.53 (s, 12 H, NCH ), 3.88 (s, 2 H, 9-H),
3 3
4
6.84±6.97 (m, 14 H, ar-H), 7.22±7.28 (m, 6 H, ar-H),
.50 (s, 4 H, CH N), 6.52 (dd, Jˆ2 Hz, 3 Hz, 4 H, ar-H),
2
N-9-Anthracenylmethyl-N,N-dimethylammonium
anthracenecarboxylate´9-anthracenecarboxylic
complex (3b). To a mixture of 65 mg (0.28 mmol) of
amine 1 in 1 mL of ethyl acetate were added 122 mg
9-
acid
1
3
7.58 (dd, Jˆ2 Hz, 1 Hz, 6 H, ar-H). C NMR (CD OD)
3
d46.0 (NCH ), 57.9 (C-9), 59.4 (CH N), 62.7 (C-10),
3
2
112.8 (ar-CH), 117.1 (ar-CH), 127.7 (ar-CH), 128.0
(ar-CH), 128.3 (ar-CH), 129.8 (ar-CH), 141.6 (ar-CH),
143.1 (ar-C ), 146.7 (ar-C ), 149.5 (ar-C ), 164.1
(
0.56 mmol) of 9-anthracenecarboxylic acid in 3 mL of
methanol. Slow evaporation of the solvent at room temp.
yielded 131 mg (0.19 mmol, 69%) of salt 3b as pale yellow
plates, mp 218±2208C (dec.). UV (MeOH): lmax (log e) 218
q
q
q
(CˆO).
(
(
2
4.44), 248 (4.55), 330 (3.78), 346 (4.08), 364 (4.25), 384
4.20). H NMR (DMSO-d ) d 2.45 (s, 6 H, NCH ), 4.62 (s,
H, CH N), 7.58±7.73 (m, 12 H, ar-H), 8.10±8.24 (m, 10
2
The identity of the photodimer was con®rmed by basi-
®cation with aqueous 0.1N NaOH, extraction with
deuterochloroform and comparison of the H NMR spec-
1
6
3
1
H, ar-H), 8.59 (d, Jˆ8 Hz, 2 H, ar-H), 8.70 (s, 1 H, ar-H),
trum with an authentic sample of the head-to-tail dimer of
amine 1.
1
.76 (s, 2 H, ar-H). C NMR (DMSO-d ) d 44.4 (NCH ),
3
8
5
1
1
6
3
3.5 (CH N), 124.9, 125.1, 125.3, 125.6, 126.1, 126.6,
2
26.8, 127.6, 128.0, 128.2, 128.5, 128.9, 130.6, 130.8,
30.9, 131.4, 170.6 (CˆO). Anal. calcd for C H NO
Photolysis of salt 3b in the solid state under oxygen-gas
atmosphere. A ground sample of salt 3b (15 mg) was
placed on the inner walls of a DURAN test tube, which
was subsequently ®tted with a septum. Syringe needles
were used as inlet and outlet to maintain a permanent ¯ow
of oxygen-gas over the sample. The solid was irradiated
l.320 nm for 8 hours. During that period the solid sample
turned deep yellow. Analysis of the crude reaction mixture
4
7
37
4
(
679.8) C: 83.04, H: 5.49, N: 2.06; found C: 82.60, H:
6.32, N: 2.45.
Photolysis of salt 1´HCl in methanol. A solution of 68 mg
0.25 mmol) of hydrochloride 1´HCl in 10 mL of methanol
(
1
was irradiated for 6 h at l.320 nm. H NMR-spectroscopic
analysis showed that all starting material was consumed.
The photodimer 4a was formed as main product among
several byproducts as indicated by a multitude of additional
1
by N NMR spectroscopy showed that anthraquinone 6 was
formed almost exclusively (.95%, conversion 30%).
Anthraquinone 6 was identi®ed by comparison with an

H. Ihmels et al. / Tetrahedron 56 (2000) 6867±6875
6875
1
authentic sample: H NMR (DMSO-d ) d 8.04, 8.32
Olovsson, G.; Scheffer, J. R.; Trotter, J. J. Am. Chem. Soc. 1998,
120 12755.
6
(
AA'BB' system, 8 H, ar-H).
8
. For the photodimerization of aminomethyl-substituted anthra-
cenes in solution, see: (a) Brimage, D. R. G.; Davidson, R. S.
J. Chem. Soc., Chem. Commun. 1971, 1385. (b) Mori, Y.;
Maeda, K. Bull. Chem. Soc. Jpn. 1997, 70 869.
Acknowledgements
This work was generously ®nanced by the Bundes-
ministerium f uÈ r Bildung und Forschung, the Deutsche
Forschungsgemeinschaft, and the Fonds der Chemischen
Industrie. Constant encouragement and generous support
by Prof. W. Adam is gratefully appreciated. We thank A.
Ancev for assistance with the synthetic work (IAESTE
exchange student from the University of Skopje,
Macedonia).
9
1
1
2
1
1
. Marcus, E.; Fitzpatrick, J. T. J. Org. Chem. Jpn. 1959, 24, 1031.
0. Ito, Y.; Fujita, H. J. Org. Chem. 1996, 61, 5677.
1. Koshima, H.; Matsuura, T. J. Synth. Org. Chem. Jpn. 1998,
68.
2. Some examples are given in Ref. 3a.
3. Jaulmes, S.; Cassanas, G.; Laruelle, P. Acta Crystallogr. Sect.
B 1982, 38, 279.
4. This de®nition neglects the fact that the molecular planes of
1
two anthracenes are not perfectly parallel; however, this geometric
0
References
feature is taken into account with the parameters d and d .
15. Wang, W. N.; Jones, W. Mol. Cryst. Liq. Cryst. 1994, 242,
227.
1. For reviews, see: (a) Photochemistry in Organized and
Constrained Media; Ramamurthy, V., Ed.; VCH-Publishers:
New York, 1991. (b) Kaupp, G. In Comprehensive Supramolecular
Chemistry, Davies, J. E. D., Ripmeester, J. A., Eds.; Elsevier:
Oxford, 1996; Vol. 8, p 381. (c) Sakamoto, M. Chem. Eur. J.
16. (a) Sandman, D.; Lopez, G. J. Mol. Cryst. Liq. Cryst. 1996,
277, 28. (b) Ortmann, I.; Werner, S.; Kr uÈ ger, C.; Mohr, S.;
Schaffner, K. J. Am. Chem. Soc. 1992, 114, 5048. (c) Yakot, Y.;
Cohen, M. D.; Ludmer, Z. Adv. Photochem. 1979, 11, 241.
1997, 3, 684. (d) Ito, Y. Synthesis 1998, 1. (e) Keating, A. E.,
Garcia-Garibay, M. A. In Organic and Inorganic Photochemistry;
Ramamurthy, V., Schanze, K. S. Eds.; Marcel Dekker: New York,
(
d) Stevens, B. Spectrochim. Acta 1962, 18, 439.
7. (a) Appleton, D. C.; Bull, D. C.; Givens, R. S.; Lillis, V.;
1
McKenna, J.; McKenna, J. M.; Thackery, S. J. Chem. Soc., Perkin
Trans. 2 1980, 77. (b) Walley, A. R.; Breslin, D. T.; Saeva, F. D.
J. Org. Chem. 1988, 53, 713.
1
2
998; p 195.
. [212] Photocycloaddition: (a) Cohen, M. D.; Schmidt, G. M. J.
J. Am. Chem. Soc. 1964, 86, 1996. (b) Schmidt, G. M. J. J. Chem.
Soc. 1964, 2014. Di-p-methane rearrangement: (c) Scheffer, J. R.,
Trotter, J., Gudmundsdottir; A. In CRC Handbook of Organic
Photochemistry and Photobiology; Horspool, W. M., Song P.-S.,
Eds.; CRC: Boca Raton, 1995; p 204. Norrish Type II reaction:
1
8. Preliminary experiments with N,N-dimethylbenzylammonium
salts did not show evidence for photolytic C±N-bond cleavage on
solid-state photolysis. Ihmels, H., unpublished results.
1
9. (a) Kaupp, G.; Plagmann, M. J. Photochem. Photobiol. A
994, 80, 399. (b) Mizuno, K,; Tamai, T.; Nakashini, I.; Ichinose,
1
(
d) Ihmels, H., Scheffer, J. R. Tetrahedron 1999, 55, 885. Type B
N.; Otsuji, Y. Chem. Lett. 1985, 2065.
0. (a) Londsdale, K.; Nave, E.; Stephens, J. F. Phil. Trans. Roy.
reaction of enones: (e) Zimmermann, H. E.; Sebek, P. J. Am. Chem.
Soc. 1997, 119, 3677.
2
Soc. London 1966, A261, 1. (b) See also: Buckland, S. J.;
Davidson, R. S. Phosphorous and Sulfur 1983, 18, 225.
3
. (a) Schmidt, G. M. J. Pure Appl. Chem. 1971, 27, 647.
b) Weiss, R. G.; Ramamurthy, V.; Hammond, G. S. Acc. Chem.
Res. 1993, 26, 530.
. Ramamurthy, V. Photochemistry in Organized and Constrained
Media; Ramamurthy, V., Ed.; VCH: New York, 1991; p 133.
. (a) Craig, D. P., Sarti-Fantoni, P. J. Chem. Soc., Chem.
(
(
c) Clennan, E. L.; Foote, C. S. In Organic Peroxides, Ando, W.,
Ed., Wiley and Sons, Chichester, 1992, 255.
1. For photooxidation of rubrene in the solid state, see:
Hochstrasser, R. M. Can. J. Chem. 1959, 37 1123
2. Rigaudy, J.; Baranne-Lafont, J.; Defoin, A.; Cuong, N. K.
Tetrahedron 1978, 34, 73.
3. For ef®cient solid-gas-phase reactions, see: Kaupp, G.; Kuse,
A. Mol. Cryst. Liq. Cryst. 1998, 313 361.
4
2
5
2
Commun. 1966, 742. (b) Cohen, M. D.; Ludmer, Z.; Thomas,
J. M.; Williams, J. O. J. Chem. Soc., Chem. Commun. 1969, 1172.
2
6
. Kaupp, G. Angew. Chem. 1992, 104, 609; Angew. Chem. Int.
Ed. Engl. 1992, 31, 595.
. Solid-state photoreactions of crystalline salts have already been
24. Stalke, D. Chem. Soc. Rev. 1998, 27, 171.
25. Sheldrick, G. M. Acta Crystallogr. Sect. A 1990, 46, 467.
26. Sheldrick, G. M. Shelxl-97, Program for Crystal Structure
7
used with remarkable success: (a) Gamlin J. G.; Jones, R.;
Leibovitch, M.; Patrick, B.; Scheffer, J. R.; Trotter, J. Acc.
Chem. Res. 1996, 29, 203. (b) Ref. 1d. (c) Leibovitch, M.;
Re®nement, University of G oÈ ttingen, 1997.