Structural study of the solid-state photoaddition reaction of arylidenoxindoles

Article abstract of DOI:10.1021/jo991873i

The photochemistry of isomeric 2-furyliden- and benzylidenoxindoles (2H- indol-2-ones) is examined. In solution E-Z isomerization is the only process via the excited singlet state (which fluoresces in glassy solution at 77 K and not at room temperature). In the crystalline state, the two (Z) derivatives are photostable, in accordance with the prediction based on the structural determination of the furylidene derivative, which adopts the unreactive Schmidt's γ type arrangement. The (E) furylidene derivative (1a) gives efficiently (Φ = 0.3) the head-to-tail dimer, as indicated by the crystal structure, which is of the reactive α type, in full accord with the topochemical principles. In contrast, the corresponding benzylidene (1b) derivative reacts sluggishly (Φ < 0.01) and mainly gives polymers, despite the fact that crystal structure determination shows that it likewise pertains to the α type and complies with the topochemical rules. The difference in reactivity is explained on the basis of (i) the twist of the phenyl ring with respect to the indole plane, and (ii) the higher overall cohesion energy and the lower interaction energy between facing molecules, as found from the charge density analysis for the crystals of 1b in comparison to those of 1a. This evidences a further stringent requirement for the occurrence of topochemical photodimerizations.

Full text of DOI:10.1021/jo991873i

3416
J . Org. Chem. 2000, 65, 3416-3425
Str u ctu r a l Stu d y of th e Solid -Sta te P h otoa d d ition Rea ction of
Ar ylid en oxin d oles
Marco Milanesio
Dipartimento di Chimica IFM, Universita`, Via P. Giuria 7, I-10125 Torino, Italy
Davide Viterbo*
Dipartimento di Scienze e Tecnologie Avanzate, Universita` del Piemonte Orientale “A. Avogadro”,
Corso T. Borsalino 54, I-15100 Alessandria, and Dipartimento di Chimica IFM, Universita`,
Via P. Giuria 7, I-10125 Torino, Italy
Angelo Albini* and Elisa Fasani
Dipartimento di Chimica Organica, Universita`, Via Taramelli 10, I-27100 Pavia, Italy
Riccardo Bianchi and Mario Barzaghi
Centro di Studio per le Relazioni tra Struttura e Reattivita` Chimica, CNR, Via C. Golgi 19,
I-20133 Milano, Italy
Received December 6, 1999
The photochemistry of isomeric 2-furyliden- and benzylidenoxindoles (2H-indol-2-ones) is examined.
In solution E-Z isomerization is the only process via the excited singlet state (which fluoresces in
glassy solution at 77 K and not at room temperature). In the crystalline state, the two (Z) derivatives
are photostable, in accordance with the prediction based on the structural determination of the
furylidene derivative, which adopts the unreactive Schmidt’s γ type arrangement. The (E) furylidene
derivative (1a ) gives efficiently (Φ ) 0.3) the head-to-tail dimer, as indicated by the crystal structure,
which is of the reactive R type, in full accord with the topochemical principles. In contrast, the
corresponding benzylidene (1b) derivative reacts sluggishly (Φ < 0.01) and mainly gives polymers,
despite the fact that crystal structure determination shows that it likewise pertains to the R type
and complies with the topochemical rules. The difference in reactivity is explained on the basis of
(i) the twist of the phenyl ring with respect to the indole plane, and (ii) the higher overall cohesion
energy and the lower interaction energy between facing molecules, as found from the charge density
analysis for the crystals of 1b in comparison to those of 1a . This evidences a further stringent
requirement for the occurrence of topochemical photodimerizations.
In tr od u ction
years, exceptions to the rules were also found, i.e.,
products expected to react did not react at all or gave
the “wrong” cycloadduct, while molecules in unsuitable
topochemical positions in the crystal reacted. In an
attempt at explaining these exceptions Kaupp6 inter-
preted his AFM experiments in terms of significant mass
displacements at the crystal surface contradicting the
topochemical postulate of minimal motion. Heteroge-
neous surface transformations may well occur with mass
transport and phase separation, and this type of mech-
anism can possibly explain some of the cases when the
reaction takes place in absence of topochemical prereq-
uisites, but it is unable to account for the more frequent
cases of nonreactive crystals complying with the rules.
Organic reactions in the crystalline state have been
known for more than a hundred years, and among them
photodimerizations have received considerable attention.
A major contribution to the understanding of the require-
ments needed to realize these reactions in the solid state
came from the work of Schmidt and co-workers,1 who
established the fundamental topochemical rules. The
basic assumption of these rules is that the arrangement
of the molecules in the crystal lattice determines the
result, in the sense that only molecules arranged in such
a way that the formation of the new bonds requires a
minimal motion react and the regio- and sterochemistry
of the cycloadducts reproduces the preexistent arrange-
ment in the crystal. The typical example is that of the
different crystalline forms of cinnamic acids giving dif-
ferent [2 + 2] photocycloadducts. Schmidt’s geometrical
rules were validated by a rather large number of studies
on photocycloaddition reactions.2 The most convincing
evidence of the validity of the topochemical rules came
from the studies of a small number of single crystal to
single crystal reactions.3-5 At the same time, over the
(2) Ramamurthy, V.; Venkatesan, K. Chem. Rev. 1987, 87, 433-
481. Garcia-Garibay, M. A.; Constable, A. E.; J ernelius, J .; Choi, Y.;
Cizmeciyan, D.; Shin, S. H. In Physical Supramolecular Chemistry;
Echegoyen, L., Kaifer, A. E., Eds.; Kluwer Academic Press: Dordercht,
1996; pp 289-312. Kaupp, G. In Organic Photochemistry and Photo-
biology; Horspool, W. M., Song, P. S., Eds.; CRC: Boca Raton, 1995;
pp 50-63.
(3) Novak, K.; Enkelmann, V.; Wegner, G.; Wagener, K. B. Angew.
Chem., Int. Ed. Engl. 1993, 32, 1614-1616.
(4) Enkelmann, V.; Wegner, G.; Novak, K.; Wagener, K. B. J . Am.
Chem. Soc. 1993, 115, 10390-10391.
(5) Ko¨hler, W.; Novak, K.; Enkelmann, V. J . Chem. Phys. 1994, 101,
10474-10480.
(1) Cohen, M. D.; Schmidt, G. M. J .; Sonntag, F. I. J . Chem. Soc.
1964, 2001-2013. Schmidt, G. M. J . J . Chem. Soc. 1964, 2014-2021.
(6) Kaupp, G. Angew. Chem., Int. Ed. Engl. 1992, 31, 592-598.
10.1021/jo991873i CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/11/2000

Solid-State Photoaddition of Arylidenoxindoles
J . Org. Chem., Vol. 65, No. 11, 2000 3417
Sch em e 1
We report here some results on the solid-state reactiv-
ity of some (hetero)arylidenoxindoles. These can be
considered as derivatives both of cinnamic acid and of
stilbene, i.e., of the two main classes of compounds known
to photodimerize, with the so far scarcely investigated
additional feature of a further conjugated substituent7,8
at the supposedly reactive CdC double bond (cf. Scheme
3). This crowding around the reactive center was expected
to be informative of the requisites for photoreaction. As
it will appear in the following, the photochemical behav-
ior was in fact quite structure-dependent, and we feel
that our results offer further insight into the mechanism
of solid-state photodimerization.
Resu lts
(Hetero)arylidenoxindoles are conveniently obtained by
condensation of oxindole with the appropriate aldehyde
(see the Experimental Section). A mixture of geometric
isomers is obtained and can be separated. The compounds
were acetylated to avoid a directing effect that the N-H
bond had on the molecular packing. In this work, we
considered the (Z) and (E) isomers of N-acetyl 2-furyliden-
and 2-benzylidenoxindoles 1 and 2 (see Scheme 1).
P h otoch em ica l Stu d ies in Solu tion . Irradiation of
the above oxindoles in solution (benzene, acetonitrile, or
chloroform) led in every case to rapid E-Z interconver-
sion (1 f 2 and vice versa). In every case a close to 1:1
steady-state mixture was obtained by irradiation with
Pyrex-filtered light. The quantum yield for the 1a f 2a
isomerization was 0.4, and the other isomerizations
occurred at a similar rate. Prolonged irradiation led to
the formation of small amounts of different products,
some of which corresponded to those obtained in the solid
state and identified as dimers (see below).
F igu r e 1. Emission spectra of arylidenoxindoles 1 and 2 by
irradiation at 360 nm in EPA glass at 77 K (dotted line) and
in the crystalline state (continuous line): (a) 1a , (b) 1b, (c)
2a , (d) 2b, and absorption spectra (e) of arylidenoxindole 1a
(1 × 10-4 M, dotted line) and of dimer 3a (5 × 10-5 M,
continuous line) in MeCN at room temperature.
readily observed in a glassy matrix at 77 K (Figure 1,
dotted line) and was recognized as fluorescence, since it
was not detected when a phosphoroscope was used.
P h otoch em ica l Stu d ies of 1a in th e Solid Sta te.
In the solid state the (E)-furylidene derivative 1a was
highly photoreactive. This compound formed large yellow
needles that, when exposed to a lamp emitting in the
near UV or to solar light, reacted in a few hours. The
large crystals “burst”, scattering around fragments for
several centimeters. Observation under the microscope
showed that the crystal cracked perpendicularly to the
elongation axis, corresponding to the short b axis (cf.
Table 1). The reaction could be conveniently carried out
by grinding the crystals and shaking them from time to
time to change the exposed surface. Under these condi-
tions gram samples of 1a could be easily brought to g85%
conversion in a few hours by means of a mercury arc.
The material was purified by recrystallization and rec-
ognized to be the head-to-tail cyclobutane dimer 3a on
the basis of the spectroscopic features and the single-
None of these oxindoles showed a detectable fluores-
cence in solution. On the other hand, emission was
(7) Intramolecular crystal state cycloadditions involving trisubti-
tuted CdC double bonds, with the third one being a nonconjugating,
usually an alkyl group, are largely known and include cases where
the double bond is exo to a ring, showing that purely steric crowding
is not per se a limitation. See, e.g., refs 7 and 8: Nakanishi, H.; J ones,
W.; Thomas. J . M.; Hursthouse, M. B.; Motevalli, M. J . Phys. Chem.
1981, 85, 3636-3642. J ones, W.; Ramdas, S.; Theocaris, C. R.; Thomas,
J . M.; Thomas, N. W. J . Phys. Chem. 1981, 85, 2594-2597. Kearsley,
S. K.; Desiraju, G. R. Proc. R. Soc. London, Ser. A, 1985, 397, 157-
181. Rabinovitch, D.; Schmidt, G. M. J . J . Chem. Soc. B 1967, 144-
149. Kaupp, G.; J ostkleigrewe, E.; Hermann, H. J . Angew. Chem., Int.
Ed. Engl. 1982, 21, 435-436. Chase, D. B.; Amey, R. L.; Holtje, W. G.
Appl. Spectrosc. 1982, 36, 155-158. Waschen, E.; Mantsch, R.,
Krampity, D, Hartke, K. Liebigs Ann. Chem. 1976, 2137-2144.
(8) Theocaris, C. R.; J ones, W.; Thomas, J . M.; Motevalli, M.;
Horsthouse, M. B. J . Chem. Soc., Perkin Trans. 2 1984, 71-76.

3418 J . Org. Chem., Vol. 65, No. 11, 2000
Ta ble 1. Cr ysta l Da ta a n d Str u ctu r e Refin em en t for Com p ou n d s 1a , 1b, 2a , a n d 3a
Milanesio et al.
1a
1b
2a
3a
Crystal Data
empirical formula
rormula weight
crystal system
C
15H11NO3
C17H13NO2
263.28
monoclinic
P21/c
C15H11NO3
253.25
monoclinic
P21/c
C30H22N2O6
506.50
monoclinic
C2/c
253.25
monoclinic
P2/n
space group
unit cell dimensions
a ) 12.006(1) Å
b ) 5.523(1) Å
c ) 18.407(2) Å
â ) 104.30(1)°
1182.7(3) Å3
4
a ) 8.099(1) Å
b ) 9.299(1) Å
c ) 17.511(2) Å
â ) 97.95(1)°
1306.1(3) Å3
4
a ) 4.585(3) Å
b ) 17.699(12) Å
c ) 14.904(7) Å
â ) 91.02(5)°
1209.3(13) Å3
4
a ) 14.961(2) Å
b ) 12.563(2) Å
c ) 12.779(2) Å
â ) 94.18(1)°
2395.5(6) Å3
4
volume
Z
density (calculated)
absorption coefficient
F(000)
1.422 Mg/m3
0.100 mm-1
528
1.339 Mg/m3
0.088 mm-1
552
1.391 Mg/m3
0.098 mm-1
528
1.404 Mg/m3
0.099 mm-1
1056
crystal size (mm)
0.30 × 0.35 × 0.45
0.40 × 0.58 × 0.86
Data Collection
2.49° to 35.03°.
0.60 × 0.21 × 0.18
0.22 × 0.38 × 0.64
θ range for
1.84° to 27.57°.
1.79° to 22.51°.
2.12° to 25.00°.
data collection
index ranges
0 e h e 15,
0 e k e 7,
-23 e l e 23
2833
2692 with
R(int) ) 0.0212
98.5%
-3 e h e 13,
-15 e k e 14,
-28 e l e 28
6101
5677 with
R(int) ) 0.0219
99.0%
-4 e h e 4,
0 e k e 19,
0 e l e 15
2419
1555 with
R(int) ) 0.0907
98.9%
0 e h e 17,
0 e k e 14,
-15 e l e 15
2175
2094 with
R(int) ) 0.0498
98.8%
reflections collected
independent
reflections
completeness
Refinement
5677/233
1.006
data/parameters
goodness-of-fit on F2
final R indices
[I > 2σ(I)]
2692/216
1.022
R1 ) 0.0404,
1555/172
1.043
R1 ) 0.0894,
2094/216
1.003
R1 ) 0.0587,
R1 ) 0.0517,
wR2 ) 0.0943
wR2 ) 0.1283
wR2 ) 0.2130
wR2 ) 0.1293
R indices (all data)
R1 ) 0.0688,
R1 ) 0.0837,
R1 ) 0.1776,
R1 ) 0.1151,
wR2 ) 0.1085
wR2 ) 0.1469
wR2 ) 0.2686
wR2 ) 0.1584
largest diff. peak
and hole
0.214 and -0.188 e Å-3
0.459 and -0.234 e Å-3
0.243 and -0.470 e Å-3
0.239 and -0.287 e Å-3
crystal structure determination. NMR analysis showed
that no other photoproduct was present in the reaction
mixture in a significant amount.
developed (see Figure 3). The position and intensity of
these reflections differed from those calculated from the
crystal structure of recrystallized 3a (Figure 7). The
difference in the diffraction patterns, together with the
above-mentioned minor differences in the IR spectrum,
indicated that the “as formed” substance was a poly-
morph of the recrystallized sample (NMR spectra of the
dissolved samples confirmed the conversion into 3a ).
Indeed, after washing the resulting powder with a little
chloroform to take away unreacted 1a , partial recrystal-
lization of 3a took place and was revealed by the
appearance of the expected reflections in the diffraction
pattern.
The reaction could be monitored by various techniques.
Irradiation of a sample in a KBr pellet could be followed
until the IR absorption bands typical of the starting
material could no longer be detected, indicating that the
conversion was g85-90%. The shift of the CdO absorp-
tion bands from 1695 and 1740 to 1710 and 1760 cm-1
(compare Figure 2a and 2b) were a clear indication that
reaction had taken place at the conjugated CdC double
bond. In fact, the spectrum obtained (Figure 2b) was
identical to that of recrystallized 3a (Figure 2c) in the
1800-1300 cm-1 but showed some small differences in
the C-H stretching band at ca. 3100 cm-1 and in the
region 1300-500 cm-1, although the general pattern was
the same.
Under this condition the rate of the reaction was 60%
of that of the photorearrangement of o-nitrobenzaldehyde
to o-nitrosobenzoic acid.9 Since a quantum yield of 0.5
has been reported for the latter reaction in the solid state
and the KBr pellets were prepared in the same way so
that the fraction of scattered light was reasonably the
same, we inferred that the photodimerization of 1a occurs
with a quantum yield of 0.3.
P h otoch em ica l Stu d ies in th e Solid Sta te of th e
Oth er Oxin d oles. The crystalline (Z) furylidene deriva-
tive 2a was much more photostable. After 7 days of
irradiation only tiny amounts of new products were
present. The NMR spectrum of the mixture was compat-
ible with the formation of the four possible cyclobutane
dimers (two of them in a larger proportion), although the
overall conversion remained low (about 10%).
The (E) benzylidene derivative 1b was also rather
stable. Prolonged irradiation gave a highly insoluble
material. A part of the irradiated solid could be extracted
with DMSO and was shown to consist of unchanged 1b
along with 25% of different products, reasonably a
mixture of the four possible dimers. As for the non-
extracted material, the displacement of the CdO IR
absorption bands was consistent with saturation of the
double bond, and therefore we assumed that this was a
polymer resulting from addition at that bond (tentative
Powder diffraction experiments on ground crystalline
samples that had been exposed for different irradiation
times (with occasional shaking during the prolonged
irradiation) showed that the starting material was
consumed up to about 85%, while a new set of reflections
(9) Kuhn, H. J .; Braslavsky, S. E.; Schmidt, R. Pure Appl. Chem.
1989, 61, 187.

Solid-State Photoaddition of Arylidenoxindoles
J . Org. Chem., Vol. 65, No. 11, 2000 3419
F igu r e 3. Comparison of the powder diffraction patterns of
1a and of its photoreaction products obtained under different
conditions, using Cu KR X-ray radiation. Only the 2θ positions
of the diffraction lines are shown. SL indicates the sample
exposed to two 15 W, 360 nm phosphor-coated lamps for 1 h;
HL is the sample exposed for 10 h (with occasional shaking);
and HLW is the same HL sample but after washing with
chloroform.
F igu r e 2. (a) IR spectrum of a dispersion of furylidenoxindole
1a in KBr. (b) IR spectrum of the same KBr pellet irradiated
until ca. 85% conversion of 1a had been reached (see Experi-
mental Section). (c) IR spectrum of a recrystallized sample of
the dimer 3a (see Experimental Section).
Sch em e 2
structure, 4 in Scheme 2). The corresponding (Z) isomer
2b was essentially photostable in the solid state.
Comparison with o-nitrobenzaldehyde in KBr showed
that with all of the above oxindoles the quantum yield
F igu r e 4. (a) Drawing of the molecular structure of compound
1a , showing the adopted labeling scheme, with displacement
ellipsoids drawn at the 30% probability level. (b) Drawing of
the two facing and potentially reactive molecules in the 1a
crystal.
for photoreaction was ,10-2
.
In the crystalline state, both (Z) derivatives 2a and 2b
exhibited an intense yellow emission, which was 30-40
nm red-shifted with respect to the luminescence observed
in ether/pentane/ethyl alcohol 5/5/2 (EPA) glass (see
Figure 1). With the (E) derivative 1b the luminescence
was weaker and practically unshifted, while no emission
was detected from crystalline 1a .
Cr ysta l Str u ctu r es. We have solved the crystal
structures of the two (E) furylidene and benzylidene
derivatives 1a and 1b, of the (Z) furylidene derivative
2a [the (Z) benzylidene derivative 2b did not give suitable

3420 J . Org. Chem., Vol. 65, No. 11, 2000
Milanesio et al.
F igu r e 6. (a) Drawing of the molecular structure of compound
2a , showing the adopted labeling scheme, with displacement
ellipsoids drawn at the 30% probability level. (b) Drawing of
two facing and potentially reactive molecules in the 2a crystal.
F igu r e 5. (a) Drawing of the molecular structure of compound
1b, showing the adopted labeling scheme, with displacement
ellipsoids drawn at the 30% probability level. (b) Drawing of
two facing and potentially reactive molecules in the 1b crystal.
crystals], and of the furyl dimer 3a . These structures are
shown in Figures 4a, 5a, 6a, and 7, while Figures 4b, 5b,
and 6b show the facing of pairs of potentially reactive
monomers. The most relevant bond distances are listed
in Table 3 for all derivatives.
The molecule of the (E) furylidene derivative 1a was
almost planar (dihedral angle between the furyl and the
oxindole planes, 7.6°) in keeping with a π delocalization
of the stylbene type involving both ring systems through
the C3dC9 double bond. This type of conjugation was
confirmed by the trend of bond distances given in Table
3. The small twist angle indicated that the five-membered
ring was well accommodated in the molecule without
relevant steric hindrance. In the crystal the molecules
were held together by van der Waals forces and by
weak C-H‚‚‚O intermolecular interactions (C2′-
H‚‚‚O8: C2′‚‚‚O8 ) 3.227 Å, H‚‚‚O8 ) 2.34 Å; C6-
H‚‚‚O11: C6‚‚‚O11 ) 3.431 Å, H‚‚‚O11 ) 2.50 Å).
Instead, the molecules of the (E) benzylidene derivative
1b (Figure 5a) were not planar, and the phenyl ring was
rotated by 38.9° with respect to the oxindole moiety. The
stylbene-type π delocalization was thus reduced (but not
wholly canceled) as shown by the significant lengthening
of the C9-C1′ bond distance (cf. Table 3) with respect to
1a . As in 1a the molecules in the 1b crystal were held
F igu r e 7. Drawing of the molecular structure of compound
3a , showing the adopted labeling scheme, with displacement
ellipsoids drawn at the 30% probability level.
together only by van der Waals and weak C-H‚‚‚O
interactions (C6-H‚‚‚O11: C6‚‚‚O11 ) 3.392 Å, H‚‚‚O11
) 2.41 Å).
The molecules of the (Z) furylidene derivative 2a
(Figure 6a) were also planar (dihedral angle between the

Solid-State Photoaddition of Arylidenoxindoles
J . Org. Chem., Vol. 65, No. 11, 2000 3421
Ta ble 2: Va lu es of th e Most Releva n t Geom etr ica l
P a r a m eter s F or eseen by th e Top och em ica l Ru les for 1a
a n d 1ba
Sch em e 3
1a
1b
ideal
θ1 (deg)
0.0
92.0
115.0
3.498
0.175
0.0
95.1
111.2
3.877
0.384
0.0
90.0
90.0
<4.2
0.0
θ2 (deg)
θ3 (deg)
center-to-center distance (Å)
D1 (Å)
clobutane ring was forced by symmetry to be planar but
was not a regular square. The two sides were very similar
[C3-C9 ) 1.575(6) Å, C3-C9b ) 1.579(6) Å], but the
angle at C3 of 88.4° was significantly smaller than that
at C9 of 91.6°. The furyl and the oxindole moieties stuck
out of the cyclobutane plane, forming with it dihedral
angles of 46.5° and 85.6° respectively. The dihedral angle
between the furyl and the oxindole planes was 49.4°. The
weak C-H‚ ‚ ‚O intermolecular interaction (C4′-H‚ ‚ ‚
O8: C4′‚ ‚ ‚O8 ) 3.278 Å, H‚ ‚ ‚O8 ) 2.41 Å) was retained
in the dimer.
Ch a r ge Den sity Ca lcu la tion s. On the low-temper-
ature data of 1a and 1b we also carried out a charge
density study with the aim also of evaluating the packing
energies. A calculation of the molecular interaction
energies, derived from the charge density analysis,
showed that the cohesion energy was smaller (-3.0 kcal/
mol) for 1a than for 1b (-5.0 kcal/mol), while the value
of the interaction energy between pairs of facing (and
potentially reactive) molecules (Figures 4b and 5b) was
significantly greater (-4.0 kcal/mol) for 1a than for 1b
(-1.0 kcal/mol).
a
As illustrated in ref 2 (Figure 5, page 440) by Ramamurthy
and Venkatesan.
Ta ble 3. Selected Bon d Len gth s (Å), An gles (d eg), a n d
Tor sion An gles (d eg) for 1a , 1b, 2a , a n d 3a
1a
1b
2a
3a
Bond Lengths
N(1)-C(10)
N(1)-C(2)
N(1)-C(7a)
C(7)-C(7a)
C(3a)-C(7a)
C(3a)-C(4)
C(3)-C(3a)
C(4)-C(5)
C(5)-C(6)
C(6)-C(7)
C(3)-C(9)
C(2)-C(3)
C(2)-O(8)
C(9)-C(1′)
C(10)-O(11)
C(10)-C(12)
C(3)-C(9b)
1.405(2) 1.403(1) 1.417(9)
1.422(2) 1.422(1) 1.405(9)
1.427(2) 1.427(1) 1.422(8)
1.406(5)
1.407(4)
1.441(4)
1.383(2) 1.391(1) 1.381(10) 1.387(5)
1.412(2) 1.402(1) 1.385(10) 1.391(5)
1.394(2) 1.393(1) 1.381(10) 1.384(5)
1.460(2) 1.459(1) 1.466(10) 1.495(4)
1.391(2) 1.394(1) 1.382(10) 1.394(5)
1.386(2) 1.391(2) 1.378(12) 1.373(6)
1.392(2) 1.396(2) 1.382(11) 1.391(5)
1.352(2) 1.348(1) 1.337(10) 1.580(4)
1.490(2) 1.490(1) 1.510(10) 1.522(4)
1.209(2) 1.210(1) 1.204(8)
1.424(2) 1.462(1) 1.431(10) 1.483(4)
1.212(2) 1.212(1) 1.205(9) 1.220(4)
1.201(4)
1.495(3) 1.495(2) 1.466(10) 1.484(6)
1.572(4)
Bond Angles
C(10)-N(1)-C(2)
C(10)-N(1)-C(7a)
C(2)-N(1)-C(7a)
C(7)-C(7a)-C(3a)
C(7)-C(7a)-N(1)
C(3a)-C(7a)-N(1)
C(4)-C(3a)-C(7a)
C(4)-C(3a)-C(3)
C(7a)-C(3a)-C(3)
C(5)-C(4)-C(3a)
C(6)-C(5)-C(4)
C(5)-C(6)-C(7)
C(7a)-C(7)-C(6)
C(9)-C(3)-C(3a)
C(9)-C(3)-C(2)
C(3a)-C(3)-C(2)
O(8)-C(2)-N(1)
O(8)-C(2)-C(3)
N(1)-C(2)-C(3)
C(3)-C(9)-C(1′)
O(11)-C(10)-N(1)
126.0(1) 126.2(1) 125.7(6)
124.6(1) 124.7(1) 123.9(7)
109.2(1) 109.0(1) 110.3(6)
121.6(1) 121.5(1) 120.6(7)
128.9(1) 128.8(1) 130.2(7)
109.5(1) 109.6(1) 109.1(6)
119.2(1) 119.7(1) 121.1(7)
132.9(1) 131.9(1) 129.7(7)
107.9(1) 108.1(1) 109.2(6)
119.1(1) 119.3(1) 118.3(8)
120.8(2) 120.1(1) 120.4(8)
121.1(2) 121.5(1) 121.6(8)
118.1(1) 117.7(1) 118.0(8)
136.8(1) 135.7(1) 127.2(7)
116.4(1) 117.9(1) 127.7(7)
106.7(1) 106.4(1) 105.1(6)
125.8(1) 125.7(1) 127.2(7)
127.6(1) 127.7(1) 126.3(7)
106.7(1) 106.6(1) 106.3(6)
132.9(1) 130.4(1) 132.0(7)
119.6(2) 119.5(1) 119.4(7)
125.5(3)
109.3(3)
124.8(3)
122.1(3)
128.5(3)
109.3(3)
120.3(3)
129.8(3)
109.8(3)
118.0(3)
120.9(4)
122.1(3)
116.5(3)
116.0(3)
114.0(3)
103.3(3)
126.4(3)
125.4(3)
108.2(3)
117.7(3)
119.2(4)
123.2(4)
117.6(3)
91.5(2)
Discu ssion
The photochemistry of oxindoles 1 and 2 in solution
requires little comment, being just another case of
photoinduced geometrical isomerization. The high rate
of rotation around the exocyclic double bond precludes
other excited-state processes such as emission and ad-
dition. The quantum yield of isomerization is 0.5 ( 0.1,
as often observed for this kind of process,10 and is
independent from the presence of oxygen. When rotation
is inhibited by viscosity in the glass at 77 K an intense
fluorescence appears, while there is no phosphorescence.
Thus, there is no evidence for the occurrence of a
significant intersystem crossing to the triplet, and the
isomerization appears to proceed from the singlet state.
Compound 1a can be conveniently photodimerized in
the solid state. As it appears from Figure 1e, the dimer
is transparent in the wavelength range used, and thus
the reaction proceeds until up to 85% of the starting
material is consumed. As mentioned above, a rough
evaluation of the quantum yield (taking o-nitrobenzal-
dehyde as an actinometer) shows that the dimerization
is efficient (Φ = 0.3). The different behavior of the four
arylidenoxidoles examined deserves some discussion in
the frame of the current rationalization on the topic.
The reported examples fall into three classes: fumaric
acid derivatives (A), cinnamic acid derivatives (B), stil-
benes (C). However, to our knowledge, no example where
all three substituents at the CdC double bond have a
conjugated multiple bond (type D) had been previously
investigated.7,8 The present study of a type D case shows
a diversity of behaviors for the derivatives considered,
from an efficient and selective dimerization to no reaction
at all.
O(11)-C(10)-C(12) 122.1(1) 122.1(1) 121.5(8)
N(1)-C(10)-C(12)
C(3)-C(9)-C(3b)
C(9)-C(3)-C(9b)
C(2′)-C(9)-C(3b)
C(3a)-C(3)-C(9b)
C(2)-C(3)-C(9b)
118.3(1) 118.4(1) 119.1(8)
88.5(2)
120.6(3)
123.9(3)
111.3(3)
Torsion Angles
179.7(2) 152.3(1) 2.0(2)
C(3)-C(9)-
C(1′)-C(2′)
C(3a)- C(3)-
C(9)-C(1′)
-133.1(4)
-106.6(3)
2.9(3)
-7.7(2)
177.4(8)
furyl and the oxindole planes 1.4°) with the same degree
of π delocalization as in 1a .
The molecules of the furyl head-tail dimer 3a (Figure
7) sat on a crystallographic inversion center coinciding
with their center of mass in the middle of the cyclobutane
ring. The four carbon atoms C3, C9, C3b, C9b of the ring,
having now an sp3 hybridization, could no longer con-
tribute to the π delocalization, and the furyl and oxindole
rings became independent aromatic systems. The cy-
(10) Saltiel, G.; D’Agostino, J .; Megarity, E. D.; Metts, L.; Neuberger,
K. R.; Wrighton, M.; Zafiriou, O. C. Org. Photochem. 1973, 3, 1.

3422 J . Org. Chem., Vol. 65, No. 11, 2000
Milanesio et al.
The first case is represented by the (E) furylidenoxin-
dole 1a. Following the classification proposed by Schmidt1,2
for photochemical dimerizations, the crystal packing of
this compound is of the R type: (i) The distance between
parallel molecules related by a lattice translation is quite
long (5.523 Å), because of the large angle æ ) 52.5°
between the translation vector and the normal to the
molecular planes. (ii) Instead, the pairs of molecules
related by an inversion center face each other in a head-
tail arrangement (Figure 4b) favorable to photoreaction,
with a distance between the double bonds of only 3.498
Å (cf. Table 2 for the other geometrical parameters
foreseen by the topochemical rules). Only van der Waals
forces and weak C-H‚ ‚ ‚O interactions held together the
molecules. Acetylation of the indole nitrogen prevents the
formation of stronger interactions via the N-H group.11
Indeed, crystals of this substance are highly photore-
active, and the observed cracking of the prismatic crystals
in a direction normal to their elongation axis is in keeping
with the observed crystal packing. The planes of flat
molecules stack one on top of the other along the direction
of the short b axis, which therefore corresponds to the
easiest crystal growth direction (elongation axis). The
reacting molecules, approaching each other, move away
from the parallel translationally related molecules up to
the point of breaking the crystal lattice in the direction
perpendicular to the translation. The details of the
process have not been fully clarified. The powder diffrac-
tion analysis of the product of the cyclophotoaddition of
1a shows that the “as formed” crystalline dimer has a
polymorphic form different from that of the recrystallized
sample (the crystal structure of which has been deter-
mined). To determine the crystal structure of the “as
formed” dimer, which may be considered as an interme-
diate step of the process, we collected the high resolution
powder diffraction pattern at the ESRF synchrotron
source. So far, all our attempts at indexing the pattern
have failed, probably because the sample contains more
than one crystalline phase.
inversion center (Figure 5b) are much closer and in a
topochemically favorable arrangement (Table 2), even
though the distance between the two double bonds (3.877
Å) is slightly larger than in 1a . The twist of the phenyl
ring can be due to steric hindrance with the neighboring
ring and/or energy gain in crystal packing. Indeed, the
twist allows a favorable interaction between one of the
phenyl protons and the π cloud of the six-membered ring
of the oxindole system of the facing molecule, as shown
in Figure 5b.
Despite the favorable molecular arrangement, no to-
pochemical photocycloaddition takes place with 1b. The
quantum yield for conversion is ,10-2, and the products
formed (a mixture of dimers, some polymeric material)
may arise from reactions at defects rather than being
lattice-determined. The reason for the difference between
the two (E) arylidene derivatives 1a and 1b is not
apparent at first sight, because both compounds display
the same R crystal packing and satisfy the topochemical
requirements as far as distance and parallelism of the
relevant CdC bond are concerned, even though the
distance is slightly (by 0.378 Å) larger for the unreactive
benzylidene derivative. The most relevant difference
found in the crystal structures is the different planarity
of the molecules and thus the lesser degree of “stilbene-
type” conjugation in the phenyl derivative.
Molecular packing in the crystal may play a role. From
a qualitative point of view, one may notice that, according
to the rules proposed by Kitaigorodski,13 the packing of
the benzylidene derivative 1b, crystallizing in the ubiq-
uitous P21/c space group, should be more compact than
that of the furylidene derivative 1a , crystallizing in the
rarer P2/n space group with less translational symmetry.
The values of the calculated densities seem to point in
the opposite direction, but the difference is small. A more
quantitative estimate of the packing effects was attained
from the charge density estimation of the molecular
interaction energies. Indeed, the cohesion energy was 2
kcal/mol smaller for 1a (-3 kcal/mol) than for 1b (-5
kcal/mol), and the value of the interaction energy between
potentially reactive molecules was 3 kcal/mol larger for
1a (-4 kcal/mol) than for 1b (-1 kcal/mol).
We have tried to confirm these results by carrying out
ab initio theoretical calculations both on the periodic
crystals and on the isolated molecules of 1a and 1b.
Unfortunately the results on the crystals did not yield
interpretable results.14 The molecular calculations were
carried out both on single molecules and on pairs of facing
molecules, as in Figures 4b and 5b, to evaluate the
interaction energy. Even though these calculations were
performed with a variety of methods, the values of the
interaction energy turned out to be close to zero, as
expected for a purely dispersive interaction. This result
is in keeping with the small values of the interaction
energy obtained from the charge density analysis. The
most significant outcome was the clear indication that
the rotation of the phenyl ring in 1b is mainly due to
the steric repulsion between the protons at C6′ and C4,
as shown by the value of the twist angle (51.2° in the
The fact that dibenzylidenecyclopentanone, a molecule
that bears some resemblance to the present derivatives,
yields a photodimer as an “amorphous” phase rather than
undergoing a crystal-to-crystal transformation has been
previously rationalized8 as being due to the fact that the
intrinsic rigidity of this large molecule precludes that the
dimer fits into the monomer lattice, as is possible with
smaller or more flexible derivatives, and at any rate that
the formation of new products disrupts the arrangement
of the crystals.12 The similar rigidity of 1a may be the
cause of our failure to carry out single crystal to single
crystal reaction upon irradiation into the outermost
absorption flank.3-5
As for the (E) phenylidene derivative 1b, the arrange-
ment of the molecules in the crystal is again of the R type,
with æ ) 62.0° and translation distance of 8.10 Å, while
the two head-to-tail faced molecules related by the
(11) These hydrogen bond interactions were indeed found in the
crystal structure of 2-benzylidenoxindole, the crystals of which had
been obtained in one of our attempts at growing suitable crystals of
2b. The crystal structure will not be described here, but the data will
be deposited with the Cambridge Crystallographic Data Centre, with
deposition number CCDC-137079.
(13) Kitaigorodski, A. I. Organic Chemical Crystallography; Consult-
ants Bureau, 1961,
(12) This is another case where, despite the fact that the dimer does
not fit easily within the structure of the monomer, an important
condition according to Kaupp’s scheme, the dimerization occurs ef-
ficiently (ca 90%) as predicted topochemically, though not in a crystal
to crystal way.
(14) The calculated lattice energies for 1a and 1b tend to zero or
even to positive values. This may be attributed to the known inability
of the Hartree-Fock and Density Functional methods to handle weak
dispersive interactions. We intend to verify this hypothesis by carrying
out calculations on smaller molecular crystals.

Solid-State Photoaddition of Arylidenoxindoles
J . Org. Chem., Vol. 65, No. 11, 2000 3423
isolated molecule and 40.5° in the facing pair). On the
other hand, the optimized geometry of the molecule of
1a becomes completely flat.
ciple1 alone, applied to the analysis of the crystal
structures, it has been possible to explain both the
absence of photocycloaddition in the (Z) derivative 2a ,
which crystallizes in the unreactive γ form, and the high
reactivity of the (E) furylidene derivative 1a , which
adopts the R form with a very favorable topochemical
head-tail facing of the molecules. To explain the unex-
pected lack of reactivity of the (E) benzylidene derivative
1b, which also adopts the R form with a favorable
topochemical head-tail facing of the molecules, a more
detailed analysis of the molecular geometry, the weak
intermolecular interactions, and the crystal packing had
to be carried out. The calculation of the interaction
energies from the charge density study on the low-
temperature diffraction data proved to be a valuable tool
for analyzing these factors. The twist of the phenyl ring
with respect to the oxindole moiety, the interaction
between a phenyl proton and the π cloud of the six-
membered ring in the oxindole nucleus of a facing
molecule, the relatively high cohesion energy, and the
low interaction energy between facing molecules in the
crystal were found to be the main factors preventing
photoreaction in the case of 1b. This points to a further
stringent requisite for topochemical dimerizations and
suggests that cohesion energy should be taken into
account when rationalizing negative results.
The lack of reactivity in the (E) benzylidene derivative
1b can thus be attributed to two connected factors: the
twist of the phenyl ring and the greater cohesion of the
packing. As for the first factor, apparently a planar 1,2-
disubstitued ethylene moiety is required for the reaction
(indeed there are no examples of reactive monosubsti-
tuted or 1,1-disubstituted ethylenes). Thus, the (E)
furylidene derivative 1a undergoes a model photocy-
cloaddition under topochemical control,2,15 and the excited
molecule finds no barrier toward the funnel leading to
addition,16a as shown by the high quantum yield and
absence of luminescence. However, in 1b a slight torsion
of the stilbene moiety introduces a barrier into the path
leading to cycloaddition and hinders the reaction. In this
case the excited state decays either radiatively (notice
that the fluorescence from the solid is similar in shape
to that in glassy solution, indicating that the conforma-
tion is similar in the two cases) or via a nonradiative
internal conversion (as suggested by the fact that the
luminescence is weak), which presumably involves a
molecular motion initially not dissimilar to that leading
to cycloaddition. The small increase in energy in 1b is
counterbalanced by the closer packing, revealed by the
larger cohesion energy and mostly due to the phenyl
C-H-oxindole ring interaction shown in Figure 5b.
Ariel et al.16b have explained the similar lack of
photoreactivity observed in 4,4,8R-trimethyl-8aâ-car-
bomethoxy-4aâ,5,8,8a-tetrahydro-1(4H)-naphthalen-1-
one as caused by a “steric compression effect” due to a
repulsive interaction between methyl hydrogens in dif-
ferent molecules, which, as the reacting molecules ap-
proach each other, get so close one to the other as to
hinder the possible reaction. In our case the more
effective packing is attained also through a favorable
C-H‚ ‚ ‚π interaction, and the “steric compression” may
be interpreted as the lack of sufficient space for rotation
and reaction.
A number of problems still remain open. Following the
suggestions of a number of recent crystal engineering
studies on fluorinated derivatives,17,18 we also intend to
carry out the photoreaction of fluorinated analogues of
1b and test the capability of fluorine to steer reactivity.
Finally, we think that the mechanisms that are at the
basis of these types of solid-state reactions need further
attention, especially in the cases in which the topochemi-
cal principle is unable to explain the reactivity. Progress
in this direction may come from some recent Raman
phonon studies,14,19-22 evidencing the relation between
the occurrence of topochemical reactions and lattice-
phonon interactions, as well as from the appropriate
considerations of the limits imposed by the tightness of
crystal packing to the molecular movements necessary
for the reaction to take place.23-27 A very recent study of
a solid-state photochromic reaction by first principles
molecular dynamics techniques28 might also open a new
way to tackle this kind of problem.
In contrast, the crystal packing for the (Z) furylide-
noxindole 2a (and by inference also for 2b, for which we
were unable to grow suitable crystals) corresponds to the
unreactive Schmidt’s γ form, with the closest molecules
separated by the shortest a ) 4.596 Å lattice translation
(Figure 6b), which is too large for the photoreaction to
occur. This explains the lack of photocycloaddition for the
two (Z) derivatives in terms of unsuitable γ crystal
packing. Noteworthy, the crystals of these materials
exhibit an intense luminescence, which is red-shifted
with respect to that observed in a glassy solution at 77
K, possibly due to a more strict planarity of these
molecules in the crystalline state (see Figure 6a for 2a ).
(17) Vishnumurthy, K.; Guru Row, T. N.; Venkatesan, K. J . Chem.
Soc., Perkin Trans. 2 1996, 1475-1478. Vishnumurthy, K.; Guru Row,
T. N.; Venkatesan, K. J . Chem. Soc., Perkin Trans. 2 1997, 615-619.
(18) Coates, G. W.; Dunn, A. R.; Henling, L. M.; Ziller, J . W.;
Lobkovsky, E. B.; Grubbs, R. H. J . Am. Chem. Soc. 1998, 120, 3641-
3649.
(19) Misra, T. N.; Prasad, P. N. Chem. Phys. Lett. 1982, 85, 381-
386.
(20) Ghosh, U.; Misra, T. N. Bull. Chem. Soc. J pn. 1985, 58, 2403-
2406.
Con clu sion s
(21) Chakrabarti, S.; Gantait, M.; Misra, T. N. Proc. Ind. Acad. Sci.
(Chem. Sci.) 1990, 102, 165-172.
With the present study we have been able to rationalize
the photochemical behavior of a family of four arylide-
noxindole derivatives by comparing the results of a series
of crystal structure analyses with those of the photo-
chemical study. On the basis of the topochemical prin-
(22) Ghosh, M.; Chakrabarti, S.; Misra, T. N. J . Phys. Chem. Solids
1997, 59, 753-757.
(23) Cohen, M. D. Angew. Chem., Int. Ed. Engl. 1975, 14, 386-393.
(24) Gavezzotti, A. J . Am. Chem. Soc. 1983, 105, 5220-5225.
(25) Murthy, G. S.; Arjunan, P.; Venkatesan, K.; Ramamurthy, V.
Tetrahedron 1987, 43, 1225-1240.
(26) Marubayashi, N.; Ogawa, T.; Hamasaki, T.; Hirayama, N. J .
Chem. Soc., Perkin Trans. 2 1997, 1309-1314.
(27) Marubayashi, N.; Ogawa, T.; Hirayama, N. Bull. Chem. Soc.
J pn. 1998, 71, 321-327.
(28) Irmgard, F.; Marx, D.; Parrinello, M. J . Phys. Chem. 1999, 103,
7341-7344.
(15) Ghosh, M.; Chakrabarti, S.; Misra; T. N. J . Phys. Chem. Solids
1996, 57, 1891-1895.
(16) (a) Caldwell, R. A. J . Am. Chem. Soc. 1980, 102, 4004-4007.
(b) Ariel, S.; Askari, S.; Scheffer, J . R.; Trotter, J .; Walsh, L. J . Am.
Chem. Soc. 1984, 106, 5726-5728.

3424 J . Org. Chem., Vol. 65, No. 11, 2000
Milanesio et al.
in water and extracted with ethyl acetate, and the organic
phase was analyzed by HPLC. A similar reaction was carried
out on identically prepared KBr pellets containing 2 mg of
2-nitrobenzaldehyde, which was taken as an actinometer (Φ
) 0.5).9
Exp er im en ta l Section
Ma ter ia ls. The arylidenoxindoles 1 and 2 were prepared
according to literature methods,29 purified by chromatography,
and recrystallization from ethanol. The main spectroscopic
characteristics are reported below.
The photochemistry of the other oxindoles was similarly
investigated. Irradiation of 1b (7 days) gave a highly insoluble
material. Part of this could be dissolved in DMSO and analyzed
by NMR. Besides unreacted 1b, it contained only traces (ca.
10% overall) that were suggestive of dimers, since separate
signals for the acetyl group and the cyclobutyl protons (in the
δ 5-5.5 region) were detected. The insoluble part showed a
(E)-1-Acetyl-2,3-dih ydr o-3-(2-fu r ylm eth ylen e)-2H-in dol-
1
2-on e (1a ): H NMR [(CD3)3SO] δ 2.65 (s, 3H), 6.88 (dd, J )
2, 4 Hz), 7.30 (dt, J ) 2, 7 Hz), 7.42 (dd, J ) 2, 7 Hz), 7.45 (d,
J ) 4 Hz), 7.55 (s, 1H), 8.25 (dd, J ) 2, 7 Hz), 8.27 (d, J ) 2
Hz), 8.60 (dd, J ) 2, 7 Hz); IR (KBr) 1740, 1695, 1180 cm-1
.
(E)-1-Acetyl-2,3-d ih yd r o-3-(2-p h en ylm eth ylen e)-2H-in -
d ol-2-on e (1b): 1H NMR [(CD3)3SO] δ 2.65 (s, 3H), 7.12 (dt, J
) 2, 7 Hz), 7.4 (dt, J ) 2, 7 Hz), 7.5-7.75 (m, 5H), 7.85 (s,
IR band at 1760 cm-1
.
Similar irradiation of 2a (7 days) also showed formation of
bands at 1760 cm-1. The photolyzed crystals could be dissolved
for the main part in DMSO and analyzed by NMR. Signals
attributed to two dimers could be detected in the spectrum of
the mixture. The signals attributed to the acetyl, cyclobutyl,
and two of the furyl ring signals were as follows: main product
δ 2.62, 4.87, 5.52, 6.35; minor 2.58, 4.87, 5.33, 6.20. Two other
isomers were present in trace, as deduced from similar but
less intense signals. Overall the dimers amounted to ca. 25%
of the starting oxindole. Attempted chromatographic separa-
tion was unsuccessful.
1H), 8.2-8.25 (m, 2H); IR (KBr) 1740, 1700, 1170 cm-1
.
(Z)-1-Acetyl-2,3-dih ydr o-3-(2-fu r ylm eth ylen e)-2H-in dol-
1
2-on e (2a ): H NMR [(CD3)3SO] δ 2.68 (s, 3H), 6.88 (dd, J )
2, 4 Hz), 7.25 (dt, J ) 2, 7 Hz), 7.35 (dt, J ) 2, 7 Hz), 7.92 (s,
1H), 7.95 (dd, J ) 2, 7 Hz), 8.08 (d, J ) 2 Hz), 8.15 (d, J ) 7
Hz), 8.22 (d, J ) 4 Hz); IR (KBr) 1730, 1695, 1160 cm-1
.
(Z)-1-Acetyl-2,3-d ih yd r o-3-(2-p h en ylm eth ylen e)-2H-in -
d ol-2-on e (2b): 1H NMR [(CD3)3SO] δ 2.62 (s, 3H), 7.28 (dt, J
) 2, 7 Hz), 7.38 (dt J ) 2, 7 Hz), 7.5-7.6 (m, 5H), 7.92 (dd, J
) 2, 7 Hz), 8.05 (s, 1H), 8.15 (dd, J ) 2, 7 Hz); IR (KBr) 1725,
1695, 1155 cm-1
.
X-r a y An a lyses. Crystals of 1a , 1b, 2a , and 3a suitable
for X-ray analysis were obtained by slow evaporation of ethanol
solutions. The diffraction experiments were carried out at 150
K for all derivatives except 2a , which did not give good enough
crystals. Crystallographic data and details of data collections
and refinements are given in Table 1. Data reduction was
carried out by the Siemens P3/PC program,30 and Lorentz and
polarization corrections were performed. No absorption cor-
rection was applied. The structures were solved by direct
methods using the SIR97 program31 and refined using the
SHELXTL/IRIS32 and the SHELXL-9733 programs. The refine-
ment was by least-squares full matrix method, with anisotropic
displacement parameters for all non-hydrogen atoms. The
hydrogen atoms were located in calculated positions and
treated as riding atoms during the refinement.34
P h otoch em ica l Rea ction s in Solu tion . Solutions (100
mL, 1 × 10-2 M) of the oxindoles 1 and 2 in MeCN were flushed
with argon and irradiated by means of a 125 W Pyrex-filtered
mercury arc (medium pressure) in an immersion well ap-
paratus at 17 °C. After 1-2 h the steady state E/Z composition
was reached (6/4 to 1/1 mixtures in all cases) as determined
by HPLC (on the basis of appropriate calibration curves, by
using a Supelcosil LC-SI column and cyclohexane ethyl
acetate mixtures as eluants) and NMR of the residue after
solvent distillation. Irradiation for up to 10 h produced small
amounts of further products (colorless spots in TLC).
Quantum yield measurements were carried out on similar
solutions in quartz tubes. These were inserted in a merry-go-
round apparatus fitted with eight 15 W phosphor-coated lamps
(center of emission, 366 nm). The product formation was
followed by HPLC, and the light flux was determined by
ferrioxalate actinometry.
P h otoch em ica l Rea ction s in th e Solid Sta te. A finely
ground sample of compound 1a (200 mg) was evenly spread
on a 5 cm × 20 cm glass plate and exposed to the light from
two 15 W phosphor-coated lamps (366 nm) at 10 cm distance.
The powder was occasionally mixed. After 10 h, the powder
was washed with a little chloroform, dissolving some unreacted
starting material. The colorless residue (140 mg) was practi-
cally pure dimer 3a , as shown by the spectra. This could be
recrystallized from nitroethane. When large crystals were
used, these burst into smaller fragments that were scattered
around for several centimeters. The photodecomposition was
readily observed by simply leaving the crystals on the bench
exposed to natural or artificial light.
Powder diffraction experiments were carried out on
a
Siemens D-5000 diffractometer using Cu KR radiation (λ )
1.54178 Å). The powder pattern of the “as obtained” dimer 3a
was collected at the beamline BM-16 of the European Syn-
chrotron Radiation Facility in Grenoble (France).
Ch a r ge Den sity An a lyses. Since the diffraction data of
1a and 1b extended to sin θ/λ ) 0.6513 and 0.8076 Å-1
respectively, the analyses were carried out using only the
monopole model, within the aspherical-atom formalism devel-
oped by Stewart35 and implemented in the VALRAY set of
programs.36 The two monopoles of each C, N, and O atom were
derived from the canonical SCF s- and p-orbitals.37 The outer
monopole was shaped with a variable scaling parameter k,
describing the contraction/expansion of the spherical compo-
2,9-Dia cetyl-2,9-d ia za -3,4:10,11-d iben zo-6,12-bis-(2-fu -
r yl)-d isp ir o[4.1.4.1]d od eca n -1,8-d ion e (3a ): mp 225 °C
(nitroethane, with decomposition; monomeric 1a was a prod-
(30) P3/ PC Diffractometer Program, Version 3.13; Bruker Analytical
X-ray Instruments Inc.: Madison, WI; 1989.
(31) Altomare, A.; Burla, M. C.; Camalli, M.; Cascarano, G.; Giaco-
vazzo, C.; Guagliardi, A.; Moliterni, A. G.; Polidori, G.; Spagna, R. J .
Appl. Crystallogr. 1999, 32, 115-119.
(32) Sheldrick, G. M. SHELXTL/ IRIS; Bruker Analytical X-ray
Instruments Inc.: Madison, WI, 1990.
(33) Sheldrick, G. M. SHELXL-97; University of Go¨ttingen: Got-
tingen Germany, 1997.
(34) Supporting Information (atomic coordinates, anisotropic ther-
mal parameters, bond distances and angles) has been deposited with
the Cambridge Crystallographic Data Centre, with deposition numbers
CCDC-137080 for 1a , CCDC-137077 for 1b, CCDC-137076 for 2a , and
CCDC-137078 for 3a .
(35) Stewart, R. F. Electrostatic properties of molecules from dif-
fraction data. In The Application of Charge Density Research to
Chemistry and Drug Design; J effrey, A., Piniella, J . F., Eds.; Plenum
Press: New York, 1991; p 63 and references therein.
(36) Stewart, R. F.; Spackman M. A., VALRAY users’s manual;
Department of Chemistry, Carnegie-Mellon University: Pittsburgh,
1983.
uct). Anal. Found: C, 71.1; H, 4.2; N, 5.4. Calcd for C15H11
-
NO3: C, 71.14; H, 4.37, N, 5.53. 1H NMR [(CD3)3SO] δ 2.52 (s,
3H), 4.78 (s, 1H), 5.72 (d, J ) 4 Hz), 6.28 (dd, J ) 2, 4 Hz),
7.42 (dd, J ) 2, 7 Hz), 7.47 (d, J ) 2 Hz), 7.48 (dt, J ) 2, 7
Hz), 8.02 (dd, J ) 2, 7 Hz), 8.18 (dd, J ) 2.7 Hz). IR (KBr)
1755, 1710, 1275, 1170 cm-1
.
The quantum yield of reaction was measured by mixing and
grinding 2.5 mg of 1a with 250 mg of KBr and exposing the
synthesized pellet to the light from a focused 150 W high-
pressure mercury arc fitted with an interference filter (trans-
mission peak at 366 nm). Under these conditions, light was
completely absorbed by the sample. IR spectra were registered.
After 2, 4, and 6 min of irradiation the pellets were dissolved
(29) Tacconi, G.; Marinone, F. Ric. Sci. 1968, 38, 1239-1244. Stanek,
J .; Ryber, D. Chem. Listy 1946, 40, 173-190. Tacconi, G.; Marinone,
F.; Desimoni, G. Gazz. Chim. Ital. 1971, 101, 173-182.
(37) Clementi, E.; Roetti, C. At. Data Nucl. Data Tables 14 1974,
177-148.

Solid-State Photoaddition of Arylidenoxindoles
J . Org. Chem., Vol. 65, No. 11, 2000 3425
nent of the valence electron density.38 The H atom monopole
was expressed by the radial function exp(-2.48r). The calcula-
tion of the molecular interaction energies was carried out using
the model proposed by Spackman et al.39-40
Ack n ow led gm en t. We wish to thank Dr. Paolo
Redana for his valuable contribution to this work in the
initial stages and Prof. Vincenzo Massarotti and Dr.
Doretta Capsoni for carrying out some of the powder
diffraction experiments.
Th eor etica l Ca lcu la tion s. For our attempt at calculating
the lattice energy of the crystals by the ab initio Hartree-
Fock (HF) method, we have employed the CRYSTAL9841
program. Ab initio Hartree-Fock (HF) and Density Fuctional
(DFT) SCF-MO calculations were performed using the Gauss-
ian9842 program. The initial geometry was the one obtained
from the X-ray analysis, and the geometry optimization was
carried out, at HF level, using 6-31G(d,p) and/or 6-31G(d)43
basis sets. Single point Becke3LYP DFT calculations were
performed on the above optimized geometries to obtain more
accurate estimates of their relative energy.
J O991873I
(42) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.; Baboul, A. G.; Stefanov,
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.;
Martin, R. L.; Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
J ohnson, B.; Chen, W.; Wong, M. W.; Andres, J . L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J . A. Gaussian 98, Revision A.7;
Gaussian, Inc.: Pittsburgh, PA, 1998.
(43) Hehre, W. J .; Radom, L.; Schleyer, P.v.R.; Pople, J . A. Ab initio
molecular orbital theory, 1st ed.; J ohn Wiley & Sons: New York, 1986;
Chapter 4.
(44) Ugliengo, P.; Viterbo, D.; Chiari, G. Z. Kristallogr. 1993, 207,
9-23. Available at http://www.ch.unito.it/ifm/fisica/moldraw/mold-
raw.html.
The graphic analysis was mainly performed using the
MOLDRAW program.44
(38) van der Wal, R. J .; Stewart, R. F. Acta Crystallogr. 1984, A40,
587-593.
(39) Spackman, M. A. J . Chem. Phys. 1986, 85, 6579-6586.
(40) Spackman, M. A.; Weber H. P.; Craven B. M. J . Am. Chem.
Soc, 1988, 110, 775-782.
(41) Saunders, V. R.; Dovesi, R.; Roetti, C.; Causa`, M.; Harrison, N.
M.; Orlando, R. Zicovich-Wilson C. M. CRYSTAL98, Version 1.0, User’s
Manual; 1999.