Stoichiometrically sensitized decarboxylation occurring in the two-component molecular crystals of aza aromatic compounds and aralkyl carboxylic acids

Article abstract of DOI:10.1021/ja971156a

Highly selective photodecarboxylation could be achieved by utilizing a series of two-component molecular crystals of aralkyl carboxylic acids such as 3-indolepropionic acid (a) and 1-naphthylacetic acid (b) combined with acridine (1) or phenanthridine (2) as an electron acceptor. The 1:1 hydrogen bonded crystals were prepared by recrystallization from the solutions. Irradiation of the crystals at -70°C caused highly selective decarboxylation to give corresponding decarboxylated compounds alone in nearly quantitative yields, due to a smaller thermal motion of the radical species in the crystal lattice. Upon irradiating a crystal, a carboxylate radical and hydroacridine or hydrophenanthridine radical are produced via electron transfer from the acid to 1 or 2 and subsequent proton transfer followed by decarboxylation. Next hydrogen abstraction by an active aralkyl radical occurs in highest priority over the shortest distance of 3.2-3.5 A? resulting in the formation of a corresponding decarboxylated product and the regeneration of 1 or 2. Occurrence of radical coupling is low due to the longer coupling distance of 4.5-6.5 A? estimated from the crystallographic data of the starting two-component molecular crystals. The hydrogen abstraction path in the crystal lattice could be confirmed by the reaction of a deuterated crystal 1 · bD in which the CO₂H group was replaced by CO₂D, giving a deuterated 1-methyl(CH₂D)naphthalene as a product. Regeneration of 1 or 2 means that the acceptor plays the role of a stoichiometrical sensitizer, which can act in only one cycle, retaining the initial crystal structure. Such a stoichiometrical sensitization is a novel photochemical process, which occurs specifically in the solid state.

Full text of DOI:10.1021/ja971156a

J. Am. Chem. Soc. 1997, 119, 10317-10324
10317
Stoichiometrically Sensitized Decarboxylation Occurring in the
Two-Component Molecular Crystals of Aza Aromatic
Compounds and Aralkyl Carboxylic Acids
Hideko Koshima,*,† Kuiling Ding,† Yosuke Chisaka,† Teruo Matsuura,†
Ikuko Miyahara,‡ and Ken Hirotsu‡
Contribution from the Department of Materials Chemistry, Faculty of Science and Technology,
Ryukoku UniVersity, Seta, Otsu 520-21, Japan, and Department of Chemistry, Faculty of Science,
Osaka City UniVersity, Sugimoto, Sumiyoshi, Osaka 558, Japan
ReceiVed April 11, 1997X
Abstract: Highly selective photodecarboxylation could be achieved by utilizing a series of two-component molecular
crystals of aralkyl carboxylic acids such as 3-indolepropionic acid (a) and 1-naphthylacetic acid (b) combined with
acridine (1) or phenanthridine (2) as an electron acceptor. The 1:1 hydrogen bonded crystals were prepared by
recrystallization from the solutions. Irradiation of the crystals at -70 °C caused highly selective decarboxylation to
give corresponding decarboxylated compounds alone in nearly quantitative yields, due to a smaller thermal motion
of the radical species in the crystal lattice. Upon irradiating a crystal, a carboxylate radical and hydroacridine or
hydrophenanthridine radical are produced via electron transfer from the acid to 1 or 2 and subsequent proton transfer
followed by decarboxylation. Next hydrogen abstraction by an active aralkyl radical occurs in highest priority over
the shortest distance of 3.2-3.5 Å resulting in the formation of a corresponding decarboxylated product and the
regeneration of 1 or 2. Occurrence of radical coupling is low due to the longer coupling distance of 4.5-6.5 Å
estimated from the crystallographic data of the starting two-component molecular crystals. The hydrogen abstraction
path in the crystal lattice could be confirmed by the reaction of a deuterated crystal 1‚bD in which the CO2H group
was replaced by CO2D, giving a deuterated 1-methyl(CH2D)naphthalene as a product. Regeneration of 1 or 2 means
that the acceptor plays the role of a stoichiometrical sensitizer, which can act in only one cycle, retaining the initial
crystal structure. Such a stoichiometrical sensitization is a novel photochemical process, which occurs specifically
in the solid state.
Introduction
The photodecarboxylation of organic carboxylic acids in
solution is known to occur via PET mechanisms,6 using various
acceptors such as aza aromatic compounds,7-9 polycyanoaro-
matics,10,11 or dye sensitizers.8 However, the product selectivi-
ties are not necessarily high due to the co-occurrence of
subsequent radical coupling in solutions. We have reported in
a preliminary communication that a highly selective and efficient
photodecarboxylation occurs in a two-component molecular
crystal of phenanthridine and 3-indoleacetic acid at low tem-
perature.12 Furthermore, photoirradiation of two-component
molecular crystals of an aralkyl carboxylic acid combined with
an electron acceptor caused decarboxylative condensation
between the two components with remarkable selectivities in
comparison to those observed in solution.12-16 Thus, chirality
of S-(+)-2-(6-methoxy-2-naphthyl)propanoic acid in a CT
One of the features of solid state photoreactions is the higher
product selectivity due to the restricted motion of molecules in
the crystal lattice than in solution photoreactions. Since the
rigorous analysis of topochemical [2 + 2] photocycloadditions
in the solid state by Cohen and Schmidt,1 a large number of
highly selective photoreactions of one-component molecular
crystals have been reported. However, the reaction types
documented most frequently are generally limited to [2 + 2]
cycloaddition and intramolecular reactions.2,3 In the field of
solution photochemistry, a large number of photoinduced
electron transfer (PET) reactions are already known.4 Recently
we have shown that various types of solid state bimolecular
reactions via PET also can be caused by irradiating two-
component molecular crystals of which electron donor and
acceptor species are combined.5
(6) Budac, C.; Wan, P. J. Photochem. Photobiol. A: Chem. 1994, 67,
135-166.
† Ryukoku University.
(7) Noyori, R.; Kato, M.; Kawanishi, M.; Nozaki, H. Tetrahedron 1969,
25, 1125-1136.
‡ Osaka City University.
X Abstract published in AdVance ACS Abstracts, October 1, 1997.
(1) (a) Cohen, M. D.; Schmidt, G. M. J. Chem. Soc. 1964, 1996-2000.
(b) Schmidt, G. M. J. Pure Appl. Chem. 1971, 27, 647-678.
(2) Desiraju, G. R., Ed. Organic Solid State Chemistry; Elsevier:
Amsterdam, 1987.
(8) Brimage, D. R. G.; Davidson, R. S.; Steiner, P. R. J. Chem. Soc.,
Perkin Trans. 1 1973, 526-529.
(9) (a) Okada, K.; Okubo, K.; Oda, M. Tetrahedron Lett. 1989, 30, 6733-
6736. (b) Okada, K.; Okubo, K.; Oda, M. J. Photochem. Photobiol. A:
Chem. 1991, 57, 265-277.
(3) Ohashi, Y., Ed. ReactiVity in Molecular Crystals; VCH: Weinheim,
1993.
(4) Fox, M. A., Chanon, M., Eds. Photoinduced Electron Transfer;
Elsevier: Amsterdam, 1988; Part C.
(5) (a) Koshima, H.; Matsuura, T. Kokagaku 1995, 19, 10-20. (b)
Koshima, H.; Matsuura, T. J. Photochem. Photobiol. A: Chem. 1996, 100,
85-91. (c) Koshima, H.; Wang, Y.; Matsuura, T.; Miyahara, I.; Mizutani,
H.; Hirotsu, K.; Asahi, T.; Masuhara, H. J. Chem. Soc., Perkin Trans. 2
1997, 2033-2037.
(10) Tsujimoto, K.; Nakao, N.; Ohashi, M. J. Chem. Soc., Chem.
Commun. 1992, 366-367.
(11) Libman, J. J. Am. Chem. Soc. 1975, 97, 4139-4141.
(12) Koshima, H.; Ding, K.; Matsuura, T. J. Chem. Soc., Chem. Commun.
1994, 2053-2054.
(13) Koshima, H.; Ding, K.; Chisaka, Y.; Matsuura, T. Tetrahedron:
Asymmetry 1995, 6, 101-104.
(14) Koshima, H.; Ding, K.; Chisaka, Y.; Matsuura, T. J. Am. Chem.
Soc. 1996, 118, 12059-12065.
S0002-7863(97)01156-6 CCC: $14.00 © 1997 American Chemical Society

10318 J. Am. Chem. Soc., Vol. 119, No. 43, 1997
Koshima et al.
Chart 1
crystallized in a chiral space group (P21). In the crystal lattice,
two-component molecules are connected through the O-H‚‚‚N
hydrogen bonding between the hydroxyl group of a-e and the
N atom of 1 or 2. In the case of the crystals containing a or e,
additional N-H‚‚‚OdC hydrogen bonding is formed between
the N-H group of the indole ring and the carbonyl group of
the adjacent molecule of a or e. Table 3 summarizes the length,
angle, and dihedral angle data of hydrogen bonding pairs and
their head-to-tail or head-to-head stacking manner in the crystal
lattice.
Figure 1 shows the representative molecular packing dia-
grams. In the case of 1‚b, hydrogen bonding pairs of 1 and b
are stacked head-to-tail and four pairs are packed in a unit cell.
Surprisingly, the crystal 1‚bD of which CO2H was replaced by
CO2D resulted in a change of crystal system from monoclinic
(1‚b) to triclinic (1‚bD) (Figure 1 and Table 2). The H‚‚‚N
length of hydrogen bonding in 1‚bD (1.77 Å) is slightly longer
than that in 1‚bD (1.42 Å) due to the weaker hydrogen bonding
ability of deuterium than hydrogen. The conformations of
hydrogen bonding pairs between 1‚b and 1‚bD are different;
the dihedral angles of the carboxyl plane and the acridine plane
are considerably different (Table 3). However, the molecular
packing arrangements in head-to-tail pairs seem to be rather
similar.
In the crystal lattice of 2‚a (Figure 1), a linear chain structure
is formed through the N-H‚‚‚OdC hydrogen bondings among
the molecules of a, which is also connected with the molecules
of 2 by another O-H‚‚‚N hydrogen bonding. Further, the
molecules of 2 are stacked head-to-tail and parallel, with a
centrosymmetry in the crystal lattice. In the 2‚c case, the
hydrogen bonding pairs arrange head-to-head to form a column
structure in the crystal lattice.
Solution and Solid State Photoreaction. Irradiation of
solutions of aza aromatic compounds and carboxylic acids in
acetonitrile or benzene caused decarboxylation followed by
radical coupling to give typically four products. The results
are shown in Scheme 1 and Table 4. They are decarboxylated
product 3, condensation product 4, the dimer of 3 (7), and the
dimer of 1 (8) or 2 (9). In some cases, dehydrogenated
condensation product 5 or 6 was obtained. Namely, all the
possible radical coupling products formed in low selectivities
due to the high mobility of intermediate radical species in the
solution.
Solid state irradiation of the two-component molecular
crystals caused selective decarboxylation except for 2‚e to give
the corresponding decarboxylated product 3 in high yields as
the main product (Scheme 1). Table 5 summarizes the results
at various irradiation temperatures in the range -70 to 15 °C.
The condensation product 4 and dehydrogenated condensation
products 5 or 6 were obtained as the minor products. The
dimers 7, 8, and 9 were not produced at all. The high product
selectivities are due to the restricted motion of radical species
in the crystal lattice, very different from those in solution (Table
4). Lowering the irradiation temperature led to an increase in
the selectivities due to a smaller thermal motion of the radical
species. In particular, irradiation at -70 °C of 1‚c, 1‚d, 2‚a,
and 2‚c caused completely selective decarboxylation to give the
corresponding decarboxylated product alone, not being ac-
companied by any consumption of 1 and 2.
crystal with 1,2,4,5-tetracyanobenzene was retained in the
decarboxylative condensation product,13 and an absolute asym-
metric synthesis using a chiral two-component molecular crystal
formed from achiral diphenylacetic acid and acridine was
achieved by the enantioselective decarboxylative condensation.14
We describe here how the highly selective photodecarboxy-
lation can be extended to a series of two-component molecular
crystals of aralkyl carboxylic acids combined with aza aromatic
compounds. On the basis of the correlation between crystal
structure and selectivity, we have established the new concept12
of stoichiometric sensitization.
Results and Discussion
Two-Component Molecular Crystals. Acridine (1) and
phenanthridine (2) were chosen as aza aromatic compounds and
3-indoleacetic acid, (a) 1-naphthylacetic acid (b), 9-fluoren-
ecarboxylic acid (c), 9-fluoreneacetic acid (d), and 3-indolepro-
pionic acid (e) as aralkyl carboxylic acids (Chart 1). Deuterated
1-naphthylacetic acid (bD), of which the CO2H carboxylic acid
group was exchanged to CO2D, was also used. Nine two-
component molecular crystals were prepared by recrystallization
from the equimolar solutions of both components. The solvents,
melting points, and elemental analyses of the two-component-
molecular crystals are summarized in Table 1. The molar ratio
of all the crystals is 1:1. The melting points of 1‚a, 1‚c, 2‚c,
and 2‚e are between those of the component crystals, while the
others are lower than those of the two components. The melting
point of deuterated 1‚bD (74-76 °C) is slightly lower than that
of 1‚b (77-79 °C).
Eight two-component molecular crystals were submitted to
X-ray crystallographic analysis (Table 2). A high-quality single
crystal of 2‚b for X-ray analysis could not be obtained. Seven
crystals belong to achiral space groups, except 2‚e which
However, the crystal 2‚e had no photoreactivity to result in
complete recovery of the starting materials. The reason is
probably due to the very fast back electron transfer of the
indolepropionic acid radical cation.
Solid State Reaction Mechanism. How does the photore-
action occur with high product selectivities in the crystal lattice?
(15) Koshima, H.; Ding, K.; Chisaka, Y.; Matsuura, T.; Ohashi, Y.;
Mukasa, M. J. Org. Chem. 1996, 61, 2352-2357.
(16) Koshima, H.; Ding, K.; Miura, T.; Matsuura, T. J. Photochem.
Photobiol. A: Chem. 1997, 104, 105-112.
(17) Koshima, H.; Hayashi, E.; Matsuura, T.; Tanaka, K.; Toda, F.; Kato,
M.; Kiguchi, M. Tetrahedron Lett. 1997, 38, 5009-5017.

Aza Aromatic Compounds and Aralkyl Carboxylic Acids
J. Am. Chem. Soc., Vol. 119, No. 43, 1997 10319
Table 1. Characterization of the Two-Component Molecular Crystals of Aza Aromatic Compounds and Aralkyl Carboxylic Acids
elemental anal
calcd (%)
H
found (%)
crystal (color)
solvent
melting point (°C)
formula (molar ratio)
C
N
C
H
N
1‚a (yellow)
1‚b (yellow)
1‚bD (yellow)
1‚c (yellow)
1‚d (yellow)
2‚a (colorless)
2‚b (colorless)
2‚c (colorless)
2‚e (yellow)
MeCN
127-129
77-79
C23H18N2O2 (1:1)
C25H19NO2 (1:1)
C25H18DNO2 (1:1)
C27H19NO2 (1:1)
C28H21NO2 (1:1)
C23H18N2O2 (1:1)
C25H19NO2 (1:1)
C27H19NO2 (1:1)
C24H20N2O2 (1:1)
77.95
82.17
81.94
83.27
85.35
77.95
82.17
83.27
78.23
5.12
5.24
4.96
4.92
5.25
5.12
5.24
4.92
5.47
7.90
3.83
3.82
3.60
3.47
7.90
3.83
3.60
7.61
77.80
82.09
82.11
83.40
83.40
78.06
82.42
83.48
78.06
5.38
5.41
5.25
5.10
5.40
5.27
5.44
5.11
5.58
7.89
3.79
3.71
3.60
3.39
7.92
3.72
3.48
7.63
benzene-cyclohexane (1:1)
benzene-cyclohexane (1:1)
74-76
MeOH
MeCN
EtOAc
MeOH
EtOAc
MeCN
135-138
100-102
96-98
83-85
128-130
112-114
Table 2. X-ray Crystallographic Data for the Two-Component Molecular Crystals and the Products
1‚a 1‚b 1‚bD 1‚c 1‚d
2‚aa
2‚c
2‚eb
4c-2
5a‚C6H12
formula
fw
space group
a (Å)
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
C23H18N2O2 C25H19NO2 C25H18DNO2 C27H19NO2 C28H21NO2 C23H18N2O2 C27H19NO2 C24H20N2O2 C26H16N
C28H28N2
392.54
Pnma
353.41
Pbca
365.45
P21/c
366.43
P1h
389.45
P21/n
24.22(1)
14.883(9) 8.945(6)
4.936(3)
90.0
106.01
90.0
1940(2)
4
403.48
P21/c
354.41
P1h
389.45
P21/n
368.43
P21
342.42
P21/n
19.496(4) 10.001(1) 9.354(1)
24.864(5) 23.584(6) 13.333(2)
7.560(3)
90.0
90.0
90.0
12.165(7) 8.5082(5) 4.71(3)
13.251(1) 12.760(1) 17.014(3)
5.327(1) 14.789(1) 14.239(8)
19.244(7) 8.4683(9) 25.111(7) 13.6091(7) 19.2786(8) 9.537(3)
13.407(1) 16.85(1)
8.115(2)
90.0
101.654
90.0
1874.7
4
8.244(1)
106.85(1)
104.44(1)
81.86(1)
950.2(3)
2
90.0
91.33
90.0
2093(1)
4
103.018(7) 90.0
106.269(6) 91.9(1)
81.260(6) 90.0
899.4(1)
2
90.0
104.779(7) 92.642(5) 90.0
90.0 90.0 90.0
1990(11) 928.9(2) 3634.2(3) 2310(2)
90.0
90.0
3644(1)
8
4
2
8
4
F
calcd (g cm-3
radiation
)
1.281
Mo KR
1.295
Mo KR
0.82
1.277
Mo KR
0.81
50.0
3418
2371
257
1.33
Mo KR
0.98
1.280
Mo KR
0.80
1.309
Mo KR
0.84
1.299
Mo KR
0.82
50.0
4118
1954
348
1.317
Cu KR
6.74
120.0
3244
2708
334
1.252
Cu KR
5.55
120.1
5938
4512
602
1.128
Mo KR
0.65
50.2
2404
1461
140
absorp coeff (cm-1) 0.83
2θmax (deg)
no. of reflns measd 3680
no. of observations 2298
50.0
50.0
50.0
50.0
55.0
3749
1567
330
3425
1999
348
4141
2429
281
4403
3023
251
no. of variables
317
R
Rw
0.041
0.028
0.050
0.062
0.046
0.073
0.054
0.079
0.052
0.048
0.051
0.081
0.040
0.031
0.033
0.052
0.046
0.075
0.099
0.108
a Reference 12. b Reference 17.
Table 3. Detail Data of Hydrogen Bonding Pairs
1‚a
1‚b
1‚bD
1‚c
1‚d
2‚a
2‚c
2‚e
H-bonding (O-H‚‚‚N)
H‚‚‚N (Å)
1.53
2.60
179
1.42
2.64
175
1.77a
2.69
154a
1.69
2.72
167
1.64
2.67
168
1.68
2.64
168
1.50
2.65
169
1.51
2.64
171
O‚‚‚N (Å)
O-H‚‚‚N (deg)
H-bonding (N-H‚‚‚OdC)
H‚‚‚O (Å)
2.11
2.92
151
2.01
2.91
154
2.09
2.93
161
N‚‚‚O (Å)
N-H‚‚‚O (deg)
dihedral angle of H-bonding pair (deg)
R/COO
NAr/COO
NAr/R
64.0
93.1
102.7
68.6
65.2
89.8
78.9
15.5
85.1
64.5
33.5
83.7
152.2
51.3
113.2
89.9
30.0
84.7
59.6
25.7
65.1
77.1
2.2
78.3
stacking manner of H-bonding pairsb
h-t
h-t
h-t
h-h
h-t
h-t
h-h
h-h
a H is exchanged to D. b The terms h-t and h-h mean head-to-tail and head-to-head stacking, respectively, where the nitrogen atom side of each
heteroaromatic refers to head and the other side to tail.
Scheme 2 shows the possible reaction mechanism. Irradiation
of a crystal excites 1 or 2 followed by electron transfer from
the acid to 1 or 2 to afford an anion radical (10) and a cation
radical (11). Then proton transfer can give a carboxylate radical
(12) and a hydroacridine radical (13) or hydrophenanthridine
radical (14), which are shown by the most favorable canonical
forms. Next, decarboxylation of 12 produces a ‚CH2- aralkyl
radical (15). Finally, hydrogen abstraction can occur from the
N-H of 13 or 14 by the active radical 15 to result in the
formation of a decarboxylated product 3 and the regeneration
of 1 or 2, while radical coupling between 15 and 13 or 14 gives
a condensation product 4.
lifetimes to lead to hydrogen abstraction and radical coupling.
The reaction paths can be discussed on the basis of crystal-
lographic data of two-component molecular crystals such as the
estimated distances in Table 6. The occurrence of hydrogen
abstraction in higher priority than radical coupling can be
explained from the shorter distances for hydrogen abstraction
(3.2-3.5 Å) than those for radical coupling (4.5-6.5 Å).
Furthermore, the hydrogen atom is very small in contrast to
the large size of 13 or 14, implying its higher mobility in the
solid state. Figure 2 illustrates the reaction paths occurring in
the crystal lattice. In all the crystals except unreactive 2‚e,
hydrogen abstraction occurs between radical species 13 or 14
(H1) and 15 (C1), which are derived from a pair of hydrogen
bonds since the C1- - -H1 distance (3.2-3.5 Å) is shortest in
Although these processes inevitably lead to the alteration of
the crystal lattice, the radical species can move a little in their

10320 J. Am. Chem. Soc., Vol. 119, No. 43, 1997
Koshima et al.
Figure 1. Molecular packing diagrams of the two-component molecular crystals 1‚b, 1‚bD, 2‚a, and 2‚c.
the crystal lattice (open arrow in Figure 2), assuming that the
relative positions of these radical species formed in the crystal
lattice have not changed much from those of the starting crystals.
On the other hand, radical coupling occurs between a
geminate radical pair 15 and 13 or 14, which are derived from
a hydrogen bonding pair (C1- - -C2) or neighboring hydrogen
bonding pairs (C1- - -C3), depending on the molecular arrange-
ment in the crystal lattice. Only in the crystal 2‚a can the radical
coupling occur between C1 and C2 within the geminate radical
pair derived from a hydrogen bonding pair because the C1- -
-C2 distance (5.34 Å) is shorter than the C1- - -C3 (8.04 Å)
distance due to the head-to-tail stacking (Figure 2 and Table
6). In the crystals 1‚a, 1‚b, 1‚bD, and 1‚d, the C1- - -C3
coupling is preferred to the C1- - -C2 coupling because the head-
to-tail stacking of hydrogen bonding pairs results in a shorter
C1- - -C3 distance than the C1- - -C2 distance, and further the
C1- - -C2 coupling should be accompanied by steric hindrance
or half rotation of the acridine ring.
In the crystal lattice of 2‚c, the hydrogen bonding pairs of 2
and c are stacked in a head-to-head manner to make a column
structure. The C1- - -C3 distance between the neighboring pairs
in a single column is slightly shorter (4.75 Å) than the C1- - -
C2 distance (5.23 Å) within a hydrogen bonding pair, resulting
in the radical coupling between C1 and C3. The crystal 1‚c
also has head-to-head stacking of the molecular pairs to form a
column structure; the C1- - -C3 radical coupling between
neighboring columns is caused because of the shorter C1- - -
C3 distance (6.49 Å) than the C1- - -C2 distance (7.71 Å).
The fact that 2‚e showed no photoreactivity is difficult to
explain from the molecular packing arrangement (Figure 2)
because the C1- - -H1 (3.47 Å) and C1- - -C3 (4.53 Å) distances
are short enough to cause hydrogen abstraction and radical
coupling. The complete recovery of starting materials 2 and e
means that any radical species such as 15 and 13 or 14 were
not produced, probably due to the fast back-electron transfer of
the radical anion 10 and the radical cation 11.

Aza Aromatic Compounds and Aralkyl Carboxylic Acids
J. Am. Chem. Soc., Vol. 119, No. 43, 1997 10321
Scheme 1
Table 4. Solution Photoreaction of Aza Aromatic Compounds and
Aralkyl Carboxylic Acids at 15 °C
giving an assignment as a 1-methylnaphthalene (CH2D-) (3bD).
For comparison, the mass fragment data obtained by the
irradiation of 1‚b were 142.15 (100.00), 141.15 (86.56), 115.10
conversion (%) yield based on consumed acid (%)
components
1
8a 9a
(30.19), and 70.40 (30.26). The H NMR chemical shift of
(solvent)
1 or 2 acid
3
4
5
6
7
1-CH2D-Na was δ 2.67 (2H, m); and that of 1-CH3-Na was
2.64 (3H, s). These results confirm that the hydrogen abstraction
path in Figure 2 is correct.
1 + a (MeCN)
1 + a (benzene)
1 + b (MeCN)
1 + b (benzene)
1 + c (MeCN)
1 + d (MeCN)
2 + a (MeCN)
2 + b (MeCN)
2 + c (MeCN)
80
64
81
90
86
75
15
40
11
75
33
87
0
49
18
70
37
4
40
0
0
0
0
0
45
52
0
2
1
10 12
100
3
0
0
52
6
37
Acridine (1) or phenanthridine (2) are regenerated by
hydrogen abstraction from the radical species 13 or 14 in the
solid state. Dimerization of 13 or 14 is not caused by their
small mobility in the crystal lattice, despite the formation of
dimers 8 or 9 in the solution photoreaction. In particular, after
irradiation of 2‚a and 2‚c at -70 °C, 2 was completely recovered
without any consumption. It means that the aza aromatic
compounds behave like a sensitizer, acting only on one cycle,
i.e. a stoichiometrical sensitizer. Such stoichiometrical sensi-
tization is a new concept, which is specific in the solid state
photochemistry.
98 28 43,b 10c
67 21 30
25 14
26
35 35
23
12
0
87 17
99 85
37
10
9
0
a Yield based on consumed aza aromatic compound. b Compound
4c-1. c Compound 4c-2.
We could obtain the evidence that the radical 15 eventually
abstracted the H1 hydrogen atom of the CO2H group by
irradiating the deuterated crystal 1‚bD. As shown in Table 5,
1‚bD similarly caused highly selective decarboxylation. The
irradiated crystal was treated with CH2N2 immediately after the
irradiation to prevent further photoreaction caused by diffused
light, and the products were analyzed by GLC-MS and 1H NMR
spectrometry. The mass fragment data of a GLC peak corre-
sponding to methylnaphthalene were m/e 143.15 (intensity
100.00), 142.14 (87.91), 116.10 (23.53), and 70.80 (33.83),
Conclusions
A series of two-component molecular crystals of aza aromatic
compounds and aralkyl carboxylic acids were prepared by
recrystallization from the solutions. Irradiation of the crystals
at lower temperatures caused more selective and efficient
decarboxylation to give finally the decarboxylated products

10322 J. Am. Chem. Soc., Vol. 119, No. 43, 1997
Koshima et al.
Table 5. Photoreaction of the Two-Component Molecular Crystals
of Aza Aromatic Compounds and Aralkyl Carboxylic Acids
IR spectra were recorded on a JASCO FT/IR-8300 spectrophotometer.
UV spectra were measured on a Shimadzu UV-3100 spectrophotometer.
Mass spectra were taken on a Shimadzu PARVUM QP-5000 GLC-
MS. Differential scanning calorimetry (DSC) was done on a Rigaku
Thermoflex TAS-200 DSC8230D and melting points (mp) were not
corrected. Elemental analysis was carried out with a Yanaco CHN
Corder MT-5. HPLC with a photodiode-array detector were used for
determining the products on a Waters HPLC system. All the reagents
were commercially available.
Preparation of Two-Component Molecular Crystals of Aza
Aromatic Compounds (1, 2) and Aralkyl Carboxylic Acids (a-e).
Nine two-component molecular crystals of aza aromatic compounds
and aralkyl carboxylic acids were prepared by slow evaporation at room
temperature or with gentle heating of the equimolar solutions (Chart
1). Deuterated 1-naphthylacetic acid (bD, mp 125-128 °C) was
prepared by exchanging the CO2H carboxylic acid group with CO2D
in MeOD followed by dryness. A crystal of 1‚bD was obtained by
recrystallization from a 1:1 solution of 1 and bD in benzene-
cyclohexane (1:1) shaken once with D2O. The solvents for crystal-
lization, melting points, and elemental analyses of the crystals are
summarized in Table 1.
conversion
(%)
yield based on consumed
acid (%)
irrad temp
crystal
(°C)
1 or 2 acid
3
4
5
6
1‚a
1‚a
1‚a
1‚a
1‚a
1‚b
1‚b
1‚bD
1‚bD
1‚c
1‚c
1‚c
1‚d
1‚d
1‚d
2‚a
2‚a
2‚a
2‚a
2‚a
2‚a
2‚b
2‚b
2‚c
2‚c
2‚c
-70
-50
-30
-10
15
-70
15
-70
-10
-70
-30
15
-70
-30
15
-70
-50
-30
-10
15
12
12
9
10
12
15
25
86
50
85
30
63
5
13
92
12
46
61
23
38
40
51
62
91
35
80
30
35
100
90
83
73
73
71
81
61
88a
68a
94
94
82
93
73
77
92
89
92
81
69
77
93
68
98
95
97
5
0
7
4
4
1
0
0
0
0
0
0
0
0
0
7
7
12
12
12
4
15
3
6
29
19
36
9
14
4
4
29
4
8
20
0
3
4
5
0
0
10b
0
0
5
0
5
7
Preparative Photoreaction in Solutions. For the common proce-
dure, a solution (100 mL) of an aza aromatic compound (5 mmol) and
a carboxylic acid (5 mmol) was internally irradiated with a 100-W high-
pressure mercury lamp under argon at room temperature. The reaction
mixture was filtered to separate a precipitate of 8 or 9 and the solution
was submitted to preparative TLC (silica gel plate, benzene-ethyl
acetate), and in some cases preparative HPLC (C18 column, MeOH-
H2O) was subsequently carried out.
5
7
14
20
5
30
0
3
2
13
10
4
16
0
3
2
15
-70
15
-70
-30
15
Irradiation of 1 (895 mg) and a (875 mg) in MeCN for 2 h gave
4a‚1/2CCl4 (1270 mg), 5a‚cyclohexane (174 mg), and 8 (95 mg) in
66%, 9%, and 11% yields, respectively. 4a‚1/2CCl4: mp 171-172 °C
a 1-Methyl(CH2D)naphthalene 3bD. b Compound 4c-1.
1
(from 1:3 CH2Cl2-CCl4); H NMR (acetone-d6) δ 9.73 (broad, 1H),
Scheme 2
7.80 (broad, 1H), 6.50-7.50 (m, 13H), 4.30 (t, J ) 7.0 Hz, 1H), 2.95
(d, J ) 7.0 Hz, 2H); IR (KBr) 3419, 3371, 1609, 1454, 1301, 1095,
789 cm-1; UV (MeCN) λmax 208 (log ꢀ 4.85), 222 (4.63), 282 nm (4.34).
Anal. Calcd for C22H18N2‚1/2CCl4: C, 69.78; H, 4.68; N, 7.23.
Found: C, 70.06; H, 5.30; N, 7.45. C, H, and N all had higher found
than calculated values. Possibly it is easy to lose CCl4. 5a‚cyclo-
hexane: mp 228-230 (from 1:3 THF-cyclohexane); 1H NMR (THF-
d8-MeCN-d3) δ 7.00-8.50 (m, 12H), 6.30 (m, 1H), 5.10 (d, J ) 1.5
Hz, 2H), 1.45 (s, 12H); IR (KBr) 3080, 2845, 1611, 1518, 1457, 1354,
1227, 824 cm-1; UV (MeCN) λmax 221 (log ꢀ 4.70), 252 (5.15), 360
(3.96), 383 nm (3.73). Anal. Calcd for C22H16N2‚C6H12: C, 85.67;
H, 7.19; N, 7.14. Found: C, 85.47; H, 7.18; N, 7.27. The molecular
structure of 5a was confirmed by X-ray structure analysis of
5a‚cyclohexane (Scheme 1, Table 2). 8: mp 211-212 °C (from THF).
The IR spectrum and elemental analysis of 8 were consistent with those
of the authentic sample.5c
Irradiation of 1 (895 mg) and b (930 mg) in benzene for 1 h gave
3b (7 mg), 4b (750 mg), 7b (18 mg), and 8 (327 mg) in 1%, 37%, 3%,
and 47%, respectively. 4b: mp 180-181 °C (from benzene); 1H NMR
(acetone-d6) δ 6.53-8.15 (m, 15H), 4.33 (t, J ) 7.0 Hz, 1H), 3.20 (d,
J ) 7.0 Hz, 2H); IR (KBr) 3373, 3040, 2925, 2900, 2860, 1606, 1578,
1472, 790 cm-1; UV (MeCN) λmax 210 (log ꢀ 4.83), 223 (4.83), 223
(4.88), 279 nm (4.35). Anal. Calcd for C24H19N: C, 89.68; H, 5.96;
N, 4.36. Found: C, 89.91; H, 6.28; N, 4.32. 7b: 1H NMR (CDCl3) δ
6.80-8.20 (m, 14H), 3.50 (s, 4H); IR (KBr) 3035, 1586, 1460, 1394,
786 cm-1. The structure of 7b was tentatively given.
Table 6. Estimated Distances for Hydrogen Abstraction and
Radical Coupling
distancea (Å) 1‚a 1‚b 1‚bD 1‚c 1‚d 2‚a 2‚c
2‚e
Hydrogen Abstraction within H-Bonding Pair
C1- - -H1
C1- - -C2
3.33 3.47 3.20b 3.31 3.30 3.18 3.40 3.47
Radical Coupling within H-Bonding Pair
7.40 7.57 7.75 7.71 7.53 5.34 5.23 5.39
Irradiation of 1 (895 mg) and c (930 mg) in MeCN for 2 h gave 3c
(265 mg), 4c-1 (683 mg), 4c-2 (140 mg), and 8 (100 mg) in 32%,
40%, 8%, and 11% yields, respectively. 3c: mp 111-112 °C. The
1H NMR and IR spectral data are consistent with those of an authentic
Radical Coupling between Neighboring H-Bonding Pairs
C1- - -C3
5.95 6.06 5.08 6.49 5.41 8.04 4.75 4.53
1
a See Figure 2 for the numbering. b H is exchanged to D.
alone in high yields. Aza aromatic compounds play the role of
stoichiometrical sensitizer, which can act only in one cycle in
the crystal lattice.
sample. 4c-1: mp 245-246 °C (from 1:1 THF-MeCN); H NMR
(DMSO-d6) δ 8.55 (s, 1H), 6.10-8.00 (m, 16H), 5.00 (d, J ) 3.8 Hz,
1H), 4.17 (d, J ) 3.8 Hz, 1H); IR (KBr) 3390, 3035, 1608, 1488, 1310,
742 cm-1; UV (MeCN) λmax 208 (log ꢀ 4.84), 250 (4.19), 277 nm (4.38).
Anal. Calcd for C26H19N: C, 90.40; H, 5.54; N, 4.05. Found: C,
90.49; H, 5.65; N, 3.98. 4c-2: mp 167-170 °C (from 1:3 THF-
MeCN); 1H NMR (CDCl3) δ 7.00-8.00 (m, 13H), 6.65 (dd, J ) 10.0,
2.0 Hz, 1H), 6.20 (dd, J ) 9.5, 2.0 Hz, 1H), 4.30 (d, J ) 3.8 Hz, 1H),
3.73 (m, 1H), 2.80 (s, 1H), 2.67 (d, J ) 3.8 Hz, 1H); IR (KBr) 3030,
Experimental Section
General Procedures. 1H-NMR spectra were measured on a 60-
MHz JEOL spectrometer with tetramethylsilane as an internal standard.

Aza Aromatic Compounds and Aralkyl Carboxylic Acids
J. Am. Chem. Soc., Vol. 119, No. 43, 1997 10323
Figure 2. Reaction path in the crystal lattice. The atom numbers correspond to those in Table 6. Black and dotted atoms identify N and O atoms,
respectively.
2860, 1610, 1486, 910, 760 cm-1; UV (MeCN) λmax 210 (log ꢀ 4.73),
223 (4.64), 255 (4.70), 291 (4.33), 329 (3.76), 344 (3.76). Anal. Calcd
for C26H19N: C, 90.40; H, 5.54, N, 4.05. Found: 90.68; H, 5.67; N,
4.12. The molecular structure of 4c-2 was determined by X-ray
crystallographic analysis for the verification (Scheme 1, Table 2).
Photoreaction of 1 (895 mg) and d (1121 mg) in MeCN for 14 h
gave 3d (15 mg), 4d (144 mg), 5d (170 mg), and 8 (115 mg) in 3%,
8%, 10%, and 13% yields, respectively. 3d: mp 44-45 °C; 1H NMR
(CDCl3) δ 7.18-7.85 (m, 8H), 3.90 (t, J ) 7.5 Hz, 1H), 1.50 (d, J )
7.67 (m, 10H), 3.10 (s, 4H); IR (KBr) 3400, 3050, 1614, 1454, 1334,
1220, 1090, 744 cm-1; UV (MeCN) λmax 225 (log ꢀ 4.86), 282 (4.11).
Anal. Calcd for C18H16N2: C, 83.04; H, 6.19; N, 10.76. Found: C,
82.63; H, 6.33; N, 10.90. 9: mp 230 °C dec; IR (KBr) 3380, 3340,
3070, 1603, 1465, 1280, 755 cm-1. Anal. Calcd for C26H20N2: C,
86.63; H, 5.59; N, 7.77. Found: 86.16; H, 5.83; N, 8.04. The structure
of 9 was tentatively given.
Irradiation of 2 (895 mg) and b (930 mg) in MeCN for 10 h gave
3b (54 mg), 6b (692 mg), 7b (13 mg), and 9 (250 mg) in 8%, 43%,
2%, and 28% yields, respectively. 6b: 143-145 °C (from 1:1 CH2-
1
7.5 Hz, 3H). 4d: white powder, mp 248-250 °C (from MeOH); H
1
NMR (CDCl3) δ 6.60-8.33 (m, 16H), 6.10 (s, broad, 1H), 3.74-4.15
(m, 2H), 2.10 (q, J ) 6 Hz, 2H); IR (KBr) 3400, 3055, 1606, 1500,
1440, 1308, 740 cm-1; UV (MeCN) λmax 208 (log ꢀ 4.91), 268 nm
(4.46). Anal. Calcd for C27H21N: C, 90.21; H, 5.89; N, 3.90.
Found: C, 90.39; H, 5.95; N, 3.65. 5d: pale yellow crystal, mp 215-
217 °C (from MeCN); 1H NMR (THF-d8) δ 6.60-8.43 (m, 16H), 4.50
(t, J ) 7.8, 1H), 4.00 (d, J ) 7.8 Hz, 2H); IR (KBr) 3060, 1608, 1550,
1513, 1472, 1404 750 cm-1; UV (MeCN) λmax 209 (log ꢀ 4.77), 253
(5.16), 359 (3.99). Anal. Calcd for C27H21N: C, 90.72; H, 5.36; N,
3.92. Found: C, 90.36; H, 5.40; N, 3.65.
Cl2-hexane); H NMR (CDCl3) δ 6.84-8.70 (m, 15H), 5.16 (s, 2H);
IR (KBr) 3045, 1608, 1480, 1164, 792 cm-1; UV (MeCN) λmax 224
(log ꢀ 4.97), 249 (4.65), 272 (4.24), 344 nm (3.23). Anal. Calcd for
C24H17N: C, 90.28; H, 5.33; N, 4.39. Found: C, 90.04; H, 5.67; N,
4.30.
Irradiation of 2 (895 mg) and c (1050 mg) in MeCN for 9 h gave 3c
(750 mg), 6c (134 mg), and 9 (60 mg) in 90%, 8%, and 7% yields,
1
respectively. 3c: mp 113-116 °C. The H NMR and IR spectra of
3c are identical with those of an authentic sample. 6c: mp 225-228
°C (from 1:3 THF-MeCN); 1H NMR (THF-d8) δ 6.50-8.80 (m, 16H),
Photoreaction of 2 (895 mg) and a (875 mg) in MeCN for 10 h
gave 3a (84 mg), 6a (305 mg), 7a (96 mg), and 9 (370 mg) in 13%,
20%, 15%, and 41% yields, respectively. 6a: mp 202-204 °C (from
MeCN); 1H NMR (CDCl3) δ 6.90-8.67 (m, 13H), 6.60-6.73 (m, 1H),
4.80 (d, J ) 2 Hz, 2H); IR (KBr) 3230, 3065, 1608, 1520, 1452, 1366,
1224, 1104, 762 cm-1; UV (MeCN) λmax 221 (log ꢀ 4.76), 248 (4.69),
329 (3.43), 343 nm (3.36). Anal. Calcd for C22H16N2: C, 85.69; H,
5.23; N, 9.08. Found: C, 85.78; H, 5.46; N, 9.08. 7a: mp 263-265
5.90 (s, broad, 1H); IR (KBr) 3075, 1608, 1570, 1478, 1350, 754 cm-1
;
UV (MeCN) λmax 208 (log ꢀ 4.81), 252 (4.80), 301 (4.09). Anal. Calcd
for C26H17N: C, 90.90; H, 4.99; N, 4.08. Found: C, 90.80; H, 5.09;
N, 4.12.
HPLC Study of Photoreaction in MeCN. A MeCN solution (10
mL) of 0.05 M aza aromatic compound and 0.05 M aralkyl carboxylic
acid in a Pyrex test tube was externally irradiated with a 400-W high-
pressure mercury lamp under argon bubbling for 1-15 h at 15 °C.
After the reaction mixture was filtered off to remove 8 or 9, the solution
1
°C (from MeCN); H NMR (THF-d8) δ 9.73 (s, broad, 2H), 6.83-

10324 J. Am. Chem. Soc., Vol. 119, No. 43, 1997
Koshima et al.
was methylated with CH2N2 and analyzed by HPLC. The results are
shown in Table 4.
matrix least-squares method at the final stage. All the calculations were
carried out with use of the teXsan crystallographic software package,
Molecular Structure Corporation.
The crystals 1‚bD and 5a‚cyclohexane were measured in glass
capillaries to prevent the degradation. Detailed crystal data for the
ten two-component molecular crystals and the photoproducts 5a‚cyclo-
hexane and 4c-2 are summarized in Table 2.
Solid State Photoreaction of the Two-Component Molecular
Crystals. A crystal (20 mg) was pulverized in a mortar, placed between
two Pyrex glass plates in a circle 5 cm in diameter, and irradiated with
a 500-W xenon short arc lamp with a UV transparent filter (290-400
nm irradiation) or a 400-W high-pressure mercury lamp (>290 nm
irradiation) under argon at various temperatures in the range -70 to
15 °C for 1-3 h. The irradiated mixture was treated with CH2N2 and
then analyzed by HPLC. The results are listed in Table 5. For the
crystals 1‚b and 1‚bD, the mass spectra were also measured by GC-
MS immediately after the irradiation followed by methylation with
CH2N2.
Acknowledgment. This work was supported by PRESTO,
Japan Science and Technology Corporation, and a Grant-in-
Aid for scientific research from the Ministry of Education,
Science and Culture, Japan. One of the authors (K.D.) is a guest
researcher from Zhengzhou University, China.
X-ray Crystallographic Analyses. Data collections were performed
with a Riguku AFC7R automatic four-circle X-ray diffractometer and
a rotating anode equipped with a graphite monochromated Mo KR (λ
) 0.71069 Å) or Cu KR (λ ) 1.54178 Å) radiation. Three standard
reflections were monitored every 150 measurements for all data
collections, and no degradation was ascertained. Absorption correction
was applied. These structures were solved by direct methods and
expanded by using Fourier techniques. The non-hydrogen atoms were
refined anisotropically. Atomic parameters were refined by the full-
Supporting Information Available: Tables giving full data
collection parameters and further details of refinement, atomic
coordinates, anisotropic displacement parameters, bond lengths
and angles, and torsion angles of 1‚a, 1‚b, 1‚bD, 1‚c, 1‚d, 2‚c,
4c-2, and 5a‚C6H12 (132 pages). See any current masthead page
for ordering and Internet access instructions.
JA971156A