Novel Photocoupling Reaction in Two-Component Crystals of Tetracyanobenzene with Benzyl Cyanide

Full text of DOI:10.1021/ja964451o

5974
J. Am. Chem. Soc. 1997, 119, 5974-5975
Novel Photocoupling Reaction in Two-Component
Crystals of Tetracyanobenzene with Benzyl Cyanide
Yoshikatsu Ito* and Sotaro Endo
Department of Synthetic Chemistry & Biological Chemistry
Faculty of Engineering
Kyoto UniVersity, Kyoto 606, Japan
Shigeru Ohba
Department of Chemistry, Faculty of Science
and Technology, Keio UniVersity, Yokohama 223, Japan
ReceiVed December 30, 1996
We have been studying photoreactions occurring in two-
1
component crystals or in solid mixtures for more than a decade.
Here, we report that a crystalline 1:2 adduct of tetracyanoben-
zene (TCNB) with benzyl cyanide (BzCN) undergoes selectively
an unusual photocoupling reaction.
The crystalline adduct TCNB‚2BzCN (1) was obtained as
colorless plates by recrystallization of TCNB from hot BzCN.
This crystalline adduct 1 showed no clear melting point; solid
TCNB and liquid BzCN separated above 66 °C, and the TCNB
part melted at 273-275 °C (TCNB, mp ) 270-272 °C). The
C,H,N analyses (C, 75.82; H, 3.96; N, 19.68) were consistent
with those calculated for (TCNB)(BzCN)2.2 (i.e., C, 76.05; H,
1
4
.02; N, 19.92). The H NMR spectrum in DMSO-d6 was a
simple sum of the spectra for TCNB and BzCN. The IR
spectrum in a KBr disk exhibited strong absorptions at 3113,
Figure 1. (a) Crystal structure of TCNB‚2BzCN (1). (b) View a is
horizontally rotated by -34°. (c) View a is horizontally rotated by 64°.
For b and c, hydrogens are removed for clarity.
3
033, 2919, 2254, 2243, 1494, 1451, 1406, 1282, 943, 935, 736,
-
1
and 694 cm . This is not a simple sum of the spectra of the
components. For TCNB, IR (KBr) peaks at 3114, 3048, 2245,
-1
1
485, 1277, and 917 cm were found, and for BzCN, IR (KBr)
-
1
Irradiation of 1 in the solid state led to selective formation
of a novel coupling product (2). While 2 was nearly stable
either as a solid or in a DMSO solution, it was unstable in
solvents such as MeOH, EtOH, acetone, and MeCN or in the
presence of BzCN, rearranging into an isoindole derivative 3
readily or in the course of several days at room temperature. It
was sparingly soluble in ether and insoluble in CHCl3.
peaks at 2252, 1497, 1454, 1415, 735, and 696 cm were
-
1
observed. The major peaks at 943 and 935 cm were found
neither for TCNB nor BzCN, indicating that 1 is not a simple
mechanical mixture of TCNB and BzCN.
The crystal structure of 1 was determined by X-ray diffrac-
tion.2 Two molecules of BzCN and one molecule of TCNB
form a unit cell and are arranged like a BzCN‚‚‚TCNB‚‚‚BzCN
sandwich in a parallel fashion (Figure 1a-c). There is a center
of symmetry at TCNB. The interplane distance is 3.6 Å. This
In a typical experiment, the crystals of 1 (40 mg) were spread
between two Pyrex plates and irradiated with a 400 W high-
pressure mercury lamp for 20 h under argon. During the
structure suggests that BzCN and TCNB are interacting through
1
d
weak charge-transfer (CT) forces.3 In fact, a CT transition was
irradiation, the photolysis vessel was cooled from the outside
by circulation of cold water (4 °C). After the irradiation, the
orange photolysate was immediately evaporated in vacuo at 60
observed around 367 nm in the diffuse reflectance spectrum of
the adduct 1 (Figure 2). From absorption spectral studies in
acetonitrile, TCNB was also found to form a weak CT complex
°
C to remove unreacted BzCN. The residue, where 2 was
present as a sole product in a 65% yield on the basis of TCNB
NMR), was fractionally recrystallized with MeCN (0.6 mL)
-
1
with BzCN (K ) 0.07 M ) with a CT transition around 322
(
nm.
to furnish, fortunately, 3 mg of pure 2 as orange-yellow plates:
mp > 300 °C (dec.). The separation of 2 from unreacted TCNB
was not easy, because 2 often cyclized into 3 during the workup.
Although the crystals of 2 obtained were twin crystals, their
X-ray data were collected and the structure could be estimated
*
Author to whom correspondence should be addressed at the follow-
ing: tel 075-753-5670; fax 075-753-5676; e-mail yito@sbchem.kyoto-
u.ac.jp.
(1) For recent publications, see: (a) Ito, Y.; Asaoka, S.; Saito, I.; Ohba,
S. Tetrahedron Lett. 1994, 35, 8193-8196. (b) Ito, Y.; Borecka, B.; Scheffer,
J. R.; Trotter, J. Tetrahedron Lett. 1995, 36, 6083-6086. (c) Ito, Y.;
Borecka, B.; Olovsson, G.; Trotter, J.; Scheffer, J. R. Tetrahedron Lett.
(
R ) 0.270). The conformation of 2 is shown in Scheme 1
4
and is called anti here. When the spectral data as well as the
rearrangement reaction 2 f 3 are considered, the structure of
the primary product 2 is doubtless.
1
995, 36, 6087-6090. (d) Ito, Y. Mol. Cryst. Liq. Cryst. 1996, 277, 247-
2
53. (e) Ito, Y.; Fujita, H. J. Org. Chem. 1996, 61, 5677-5680. (f) Ito, Y.;
Olovsson, G. J. Chem. Soc., Perkin Trans. 1 1997, 127-133.
(
2) Rigaku four-circle diffractometer. Crystal data for 1: C26H16N6 (Mr
In a separate experiment, fractional recrystallization of the
orange photolysate from 1 mL of MeCN, followed by further
)
412.45), triclinic P 1h , a ) 7.767, b ) 10.957, and c ) 7.682 Å, R )
3
9
3.80°, â ) 117.61°, γ ) 103.05°, V ) 553.2 Å , Z ) 1, Dx ) 1.238
3 -1
g/cm , λ (Mo KR) ) 0.710 73 Å, µ ) 0.077 mm , R ) 0.053. Crystal
data for 3: C18H9N5 (Mr ) 295.30), triclinic P 1h , a ) 7.721, b ) 13.393,
(4) For 2: 1H NMR (DMSO-d6, 400 MHz) δ 7.29-7.34 (1 H, m), 7.38
(2 H, s, disappeared on deuteration), 7.45-7.47 (4 H, m), 8.72 (1 H, s),
3
and c ) 7.219 Å, R ) 103.19°, â ) 105.65°, γ ) 78.45°, V ) 692.2 Å ,
3
-1
13
Z ) 2, Dx ) 1.417 g/cm , λ (Mo KR) ) 0.710 73 Å, µ ) 0.090 mm , R
8.98 (1 H, s); C NMR (DMSO-d6, 100 MHz) δ 114.31, 114.55, 114.57,
)
0.053. The details have been deposited at the Cambridge Crystallographic
116.64, 116.66, 118.75, 121.19, 127.12 (CH), 127.77 (CH), 129.16 (CH),
132.67, 135.31 (CH), 138.24 (CH), 144.90, 151.80; IR (KBr) 3460, 3362
(strong), 2241, 2198 (strong), 1622 (strong), 1583, 1529, 1493, 1446, 1411,
Data Centre.
(3) Dahl, T. Acta Chem. Scand. 1994, 48, 93-106. Kobayashi, T.;
-
1
+
Yoshihara, K.; Nagakura, S. Bull. Chem. Soc. Jpn. 1971, 44, 2603-2610.
Fukuzumi, S.; Kochi, J. K. J. Org. Chem. 1981, 46, 4116-4126.
1275, 925, 915, 770, 704 cm ; MS (EI) m/e (rel intensity) 295 (92, M ),
294 (100); UV/vis (DMSO) λ 435 (ꢀ 2500), 292 (14 200) nm.
S0002-7863(96)04451-4 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 25, 1997 5975
Scheme 1
system in solution was very different from that in the solid state.
For example, an MeCN solution containing TCNB (0.5 M,
solubility limit) and BzCN (1.0 M) was irradiated through Pyrex
for 15 h at 20 °C under Ar bubbling. No reaction products
were detected by NMR. On the other hand, similar irradiation
of a solution of TCNB (0.042 M, solubility limit) in BzCN
produced a complex mixture with a trace of recovered TCNB,
from which 4 (44%), 5 (17%), and 6 (3%) were isolated by
preparative TLC. However, 2 and 3 were undetectable in the
reaction mixture. Whether this difference in the reaction course
is due to an effect of crystal packing on the primary photo-
chemical hydrogen abstraction step or on the following radical
reaction step is not clear at the moment.
1
1
recrystallization from 10 mL of acetone afforded 6 mg of 3 as
orange-yellow needles: mp > 300 °C.5 The structure of 3 was
unequivocally determined by X-ray analysis.2 The X-ray quality
crystals were obtained from recrystallization of 3 from MeCN/
EtOH (1:1 v/v).
A rational reaction mechanism is presented in Scheme 1.
Since irradiation of crystals of 1 through either a Pyrex filter
(
>280 nm) or a solution filter (7 g/L of BiCl3 in 10% HCl,
Attempts were made to prepare other weak adduct crystals
of TCNB by means of recrystallization. For example, a variety
of 4-substituted BzCN derivatives such as 4-HOBzCN,
>
355 nm) produced 2 with comparable efficiencies, the primary
step must be CT excitation of 1. The CT excitation ac-
companied by a proton transfer will result in a hydrogen
abstraction by the cyano group of TCNB. Although a similar
process was proposed to occur upon irradiation (at 254 nm) of
4
4
-MeOBzCN, 4-MeBzCN, 4-NCCH2BzCN, 4-ClBzCN, and
-O2NBzCN were tested. In all cases, however, we failed to
obtain adduct crystals. It appears that size and shape of the
donor molecule is an important factor. As displayed in Figure
6
TCNB in an acetonitrile or isobutyronitrile solution, photo-
chemical hydrogen abstraction reactions by the cyano group are
quite rare. Coupling of the photogenerated radical pair and a
1
c, the cyanomethyl group of BzCN is just fit for the space
7
between the o-cyano groups of TCNB. Therefore, the 4-sub-
stituents on BzCN may hamper close packing. So far, other
adduct crystals were successfully obtained only with benzyl
alcohol (BzOH) to afford colorless plates of a 1:1 adduct
TCNB‚BzOH. This adduct also underwent photolysis unique
to the crystalline state, giving a dark brown photolysate, but
subsequent 1,3-proton shift will give rise to 2. The observed
high stability of 2 in the solid state is understandable, because
the amino group and the relevant o-cyano group are fixed in an
anti relationship in the rigid crystalline environment. In MeOH,
EtOH, acetone, MeCN, or BzCN solution, however, the
conformation of 2 will be able to change from anti to syn, where
cyclization to 3 may well occur.8
Two types of the TCNB/BzCN close contact are indicated
in Figure 1a. The distance between one of the cyano nitrogens
of TCNB and the nearest benzylic hydrogen of BzCN is 2.77
Å, and the corresponding C‚‚‚C distance is 3.65 Å. These are
probably short enough for the hydrogen abstraction by the cyano
group as well as the radical coupling of the resultant radical
pair to occur, because the sum of the van der Waals radii for N
and H is 2.75 Å and that of C and C is 3.4 Å. Hydrogen
abstraction is feasible over distances around the van der Waals
sum.
Alternatively, the distance between the cyano nitrogen of
TCNB and the second nearest benzylic hydrogen of BzCN is
3
Although these distances are rather long, we consider that this
arrangement cannot be neglected since it requires less molecular
rotation through a reaction pathway from 1 to 2. Recently, we
have found several photocycloaddition reactions in multicom-
ponent crystals, where the component olefin molecules are
considerably separated or are moving.9
1
2
product isolation has been unsuccessful yet.
In summary, complex photoproducts 4-6 were produced
from irradiation of TCNB in a BzCN solution, while the
cocrystal TCNB•2BzCN (1) underwent selective photocoupling
to give 2, which subsequently cyclized into 3 in solution. The
present results encourage us that though it is still in an
1
,13
embryonic stage,
the research on two- or multicomponent
crystals is an intriguing treasury of novel photoreactions which
are completely different from those in solution. Multicomponent
crystals to be studied are almost limitless in number, and we
believe that this work has opened the door to this promising
field of organic photochemistry.
.42 Å, and the corresponding C‚‚‚C distance is 4.50 Å.
Acknowledgment. This work was supported by the Grant-in-Aid
for Scientific Research on Priority Area from the Japanese Government.
Supporting Information Available: Spectral data and formation
mechanism for 4-6, CT absorption spectra, and X-ray data (for 2 and
4) (16 pages). See any current masthead page for ordering and Internet
access instructions.
As aforementioned, TCNB and BzCN formed a weak CT
complex in MeCN. Photochemistry of the TCNB-BzCN
JA964451O
(
5) For 3: 1H NMR (DMSO-d6, 400 MHz) δ 7.40 (1 H, t, J ) 7.5 Hz),
.49 (2 H, t, J ) 7.8 Hz), 8.20 (2 H, d, J ) 7.8 Hz), 8.65 (1 H, s), 8.81 (1
H, s), 8.88 (1 H, s, disappeared on deuteration), 9.19 (1 H, s, disappeared
(9) For example, â-type photodimerization of trans-o-methoxycinnamic
7
acid in the double salt crystals with ethylenediamine (double-bond separation
1c
>5.35 Å) and â-type photodimerization of trans-cinnamamide in the 2:1
hydrogen-bonded cocrystals with phthalic acid (double-bond separation 7.52
Å or dynamic disorder).10
13
on deuteration); C NMR (DMSO-d6, 100 MHz) δ 115.06, 115.82, 116.08,
1
1
3
1
16.15, 119.55, 126.70 (CH), 127.46 (CH), 128.55 (CH), 128.81 (CH),
29.63 (CH), 133.00, 136.54, 142.85, 155.95, 165.47, 165.56; IR (KBr)
464, 2235, 2205, 1671 (strong), 1528 (strong), 1444, 1392, 1324, 1272,
(10) Ito, Y.; Ohba, S. To be published.
(11) These products suggest that the same type of reaction as the
-
1
+
6
194, 772, 697 cm ; MS (EI) m/e (rel intensity) 295 (87, M ), 294 (100);
previously reported one occurred initially. See Supporting Information.
HRMS calcd for C18H9N5 295.0858, found 295.0845; UV/vis (MeCN) λ
29 (ꢀ 13 300), 278 (9800), 239 (25 500), 206 (19 400) nm; UV/vis (DMSO)
λ 446 (ꢀ 24 100), 285 (19 700) nm.
6) Tsujimoto, K.; Abe, K.; Ohashi, M. J. Chem. Soc., Chem. Commun.
(12) Like the TCNB-BzCN system, irradiation of TCNB and BzOH in
MeCN yielded essentially no products, while irradiation of TCNB in BzOH
gave 4 in a 70% yield.
(13) For intensive investigations on a particular type of two-component
crystals, see: Gamlin, J. N.; Jones, R.; Leibovitch, M.; Patrick, B.; Scheffer,
J. R.; Trotter, J. Acc. Chem. Res. 1996, 29, 203-209 (organic salts). Toda,
F. Acc. Chem. Res. 1995, 28, 480-486 (host-guest crystals). Ito, Y.
Synthesis To be submitted (review).
4
(
1
983, 984-985.
(
7) Cantrell, T. S. J. Chem. Soc., Chem. Commun. 1975, 637-638.
(
8) Meyers, A. I.; Sircar, J. C. The Chemistry of the Cyano Group;
Rappoport, Z., Ed.; Wiley-Interscience: New York, 1970; pp 400-404.