Radical Addition to Isonitriles: A Route to Polyfunctionalized Alkenes through a Novel Three-Component Radical Cascade Reaction

Article abstract of DOI:10.1021/jo991871y

The reaction of aromatic disulfides, alkynes, and isonitriles under photolytic conditions affords polyfunctionalized alkenes - β-arylthio-substituted acrylamides or acrylonitriles - in fair yields through a novel three-component radical cascade reaction. The procedure entails addition of a sulfanyl radical to the alkyne followed by attack of the resulting vinyl radical to the isonitrile. A fast reaction, e.g., scavenging by a nitro derivative or β-fragmentation, is necessary in order to trap the final imidoyl radical, since addition of vinyl radicals to isonitriles seems to be a reversible process. The stereochemistry of the reaction is discussed, particularly with respect to the stereochemical outcome of related hydrogen abstraction reactions by the same vinyl radicals. The lower or even inverted preference for either geometrical isomer observed in our cases with respect to that encountered in hydrogen abstraction reactions is explained in terms of transition-state interactions and/or isomerization of the final imidoyl radical. The latter possibility is supported by semiempirical calculations, which show that the spin distribution in the imidoyl radical can allow rotation of the adjacent carbon - carbon double bond prior to β-fragmentation.

Full text of DOI:10.1021/jo991871y

J . Org. Chem. 2000, 65, 2763-2772
2763
Ra d ica l Ad d ition to Ison itr iles: A Rou te to P olyfu n ction a lized
Alk en es th r ou gh a Novel Th r ee-Com p on en t Ra d ica l Ca sca d e
Rea ction
Rino Leardini, Daniele Nanni,* and Giuseppe Zanardi
Dipartimento di Chimica Organica “A. Mangini”, Universita` di Bologna, Viale Risorgimento 4,
I-40136 Bologna, Italy
Received December 6, 1999
The reaction of aromatic disulfides, alkynes, and isonitriles under photolytic conditions affords
polyfunctionalized alkenessâ-arylthio-substituted acrylamides or acrylonitrilessin fair yields
through a novel three-component radical cascade reaction. The procedure entails addition of a
sulfanyl radical to the alkyne followed by attack of the resulting vinyl radical to the isonitrile. A
fast reaction, e.g., scavenging by a nitro derivative or â-fragmentation, is necessary in order to
trap the final imidoyl radical, since addition of vinyl radicals to isonitriles seems to be a reversible
process. The stereochemistry of the reaction is discussed, particularly with respect to the
stereochemical outcome of related hydrogen abstraction reactions by the same vinyl radicals. The
lower or even inverted preference for either geometrical isomer observed in our cases with respect
to that encountered in hydrogen abstraction reactions is explained in terms of transition-state
interactions and/or isomerization of the final imidoyl radical. The latter possibility is supported by
semiempirical calculations, which show that the spin distribution in the imidoyl radical can allow
rotation of the adjacent carbon-carbon double bond prior to â-fragmentation.
In tr od u ction
cascade reactions involving isonitriles bearing suitable
unsaturated side chains.
Our long interest in the chemistry of imidoyl radicals4
recently led us to develop new synthetic strategies for
the generation of those intermediates via radical addition
to isonitriles. Quinoxaline derivatives have thus been
synthesized by addition of cyano-substituted vinyl,4h
alkyl,4l and sulfanyl4l radicals to aromatic isonitriles. Of
special interest was our vinyl radical-mediated cyclo-
penta-fused quinoxaline production,4h since it provided
the first trimolecular version of the radical addition,
tandem cyclization strategy. This finding seemed worthy
of further investigations since three-component radical
Radical addition to isonitriles dates back to the 1960s,
when Shaw and Saegusa discovered the isonitrile-nitrile
isomerization mediated by methyl1a or stannyl radicals.1b
Until recent years, this reaction has been studied both
from a mechanistic and a synthetic point of view,2
especially as an approach to nitriles2i or a deamination
method.2f,j
Only since 1991 have radical reactions with isonitriles
been successfully employed in the synthesis of heterocy-
clic compounds.3 In his pioneering work,3a Curran carried
out several 4 + 1 annulations using isonitriles as geminal
radical acceptor/radical donor synthons: this strategy
allowed the synthesis of cyclopenta-fused quinolines and
provided easy access to the antitumor agents of the
camptothecin family.3b,e,h,i,l,n Almost in the same years,
Bachi performed studies on the free-radical cyclization
of alkenyl and alkynyl isonitriles, thus accomplishing the
syntheses of pyrroline and pyroglutamate derivatives.3c,f,g
He also achieved the stereo-3j and enantioselective
syntheses3k,m of (()- and (-)-R-kainic acid through radical
(3) (a) Curran, D. P.; Liu, H. J . Am. Chem. Soc. 1991, 113, 2127. (b)
Curran, D. P. Liu, H. J . Am. Chem. Soc. 1992, 114, 5863. (c) Bachi,
M. D.; Balanov, A.; Bar-Ner, N. J . Org. Chem. 1994, 59, 7752. (d)
Fukuyama, T.; Chen, X.; Peng, G. J . Am. Chem. Soc. 1994, 116, 3127.
(e) Curran, D. P.; Ko, S.-B.; J osien, H. Angew. Chem., Int. Ed. Engl.
1995, 34, 2683. (f) Bachi, M. D.; Melman, A. J . Org. Chem. 1995, 60,
6242. (g) Bachi, M. D.; Melman, A. Synlett 1996, 60. (h) Ryu, I.; Sonoda,
N.; Curran, D. P. Chem. Rev. 1996, 96, 177. (i) Curran, D. P.; Liu, H.;
J osien, H.; Ko, S.-B. Tetrahedron 1996, 52, 11385. (j) Bachi, M. D.;
Bar-Ner, N.; Melman, A. J . Org. Chem. 1996, 61, 7116. (k) Bachi, M.
D.; Melman, A. J . Org. Chem. 1997, 62, 1896. (l) J osien, H.; Curran,
D. P. Tetrahedron 1997, 53, 8881. (m) Bachi, M. D.; Melman, A. Pure
Appl. Chem. 1998, 70, 259. (n) J osien, H.; Ko, S.-B.; Bom, D.; Curran,
D. P. Chem. Eur. J . 1998, 4, 67. (o) Kobayashi, Y.; Fukuyama, T. J .
Heterocycl. Chem. 1998, 35, 1043.
(4) (a) Leardini, R.; Pedulli, G. F.; Tundo, A.; Zanardi, G. J . Chem.
Soc., Chem. Commun. 1984, 1320. (b) Leardini, R.; Tundo, A.; Zanardi,
G.; Pedulli, G. F. Synthesis 1985, 107. (c) Leardini, R.; Nanni, D.;
Pedulli, G. F.; Tundo, A.; Zanardi, G. J . Chem. Soc., Perkin Trans. 1
1986, 1591. (d) Leardini, R.; Nanni, D.; Tundo, A.; Zanardi, G. Gazz.
Chim. Ital. 1989, 119, 637. (e) Leardini, R.; Nanni, D.; Tundo, A.;
Zanardi, G. J . Chem. Soc., Chem. Commun. 1989, 757. (f) Leardini,
R.; Nanni, D.; Santori, M.; Zanardi, G. Tetrahedron 1992, 48, 3961.
(g) Guidotti, S.; Leardini, R.; Nanni, D.; Pareschi, P.; Zanardi, G.
Tetrahedron Lett. 1995, 36, 451. (h) Nanni, D.; Pareschi, P.; Rizzoli,
C.; Sgarabotto, P.; Tundo, A. Tetrahedron 1995, 51, 9045. (i) Leardini,
R.; McNab, H.; Nanni, D. Tetrahedron 1995, 51, 12143; (j) Nanni, D.;
Pareschi, P.; Tundo, A. Tetrahedron Lett. 1996, 37, 9337. (k) Leardini,
R.; Nanni, D., Pareschi, P.; Tundo, A.; Zanardi, G. J . Org. Chem. 1997,
62, 8394. (l) Camaggi, C. M.; Leardini, R.; Nanni, D.; Zanardi, G.
Tetrahedron 1998, 54, 5587.
(1) (a) Shaw, D. H.; Pritchard, H. O. Can. J . Chem. 1967, 45, 2749.
(b) Saegusa, T.; Kobayashi, S.; Yoshihiko, I.; Yasuda, N. J . Am. Chem.
Soc. 1968, 90, 4182.
(2) (a) Saegusa, T.; Kobayashi, S.; Yoshihiko, I. J . Org. Chem. 1970,
35, 2118. (b) Banks, R. E.; Haszeldine, R. N.; Stephens, C. W.
Tetrahedron Lett. 1972, 3699. (c) Singer, L. A.; Kim, S. S. Tetrahedron
Lett. 1974, 861. (d) Blum, P. M.; Roberts, B. P. J . Chem. Soc., Chem.
Commun. 1976, 535. (e) Kim, S. S. Tetrahedron Lett. 1977, 2741. (f)
Barton, D. H. R.; Bringmann, G.; Motherwell, W. B. J . Chem. Soc.,
Perkin Trans. 1 1980, 2665. (g) Blum, P. M.; Roberts, B. P. J . Chem.
Soc., Perkin Trans.
2 1978, 1313. (h) Meier, M.; Ru¨chardt, C.
Tetrahedron Lett. 1983, 24, 4671. (i) Stork, G.; Sher, P. M. J . Am.
Chem. Soc. 1983, 105, 6765. (j) Wirth, T.; Ru¨chardt, C. Chimia 1988,
42, 230. (k) Barton, D. H. R.; Ozbalik, N.; Vacher, B. Tetrahedron 1988,
44, 3501. (l) Diart, V.; Roberts, B. P. J . Chem. Soc., Perkin Trans. 2
1992, 1761. For a very recent application resulting in imidoylation of
glycoside derivatives, see: Yamago, S.; Miyazoe, H.; Goto, R.; Yoshida,
J . Tetrahedron Lett. 1999, 40, 2347.
10.1021/jo991871y CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/08/2000

2764 J . Org. Chem., Vol. 65, No. 9, 2000
Leardini et al.
reactions, besides being quite rare,5 are potentially very
powerful synthetic strategies. Indeed, they might provide
access to polyfunctionalized moleculessor polyfused cy-
clic compoundssfrom very simple, easily accessiblesoften
commercially availablesstarting materials.
Sch em e 1
In light of the fair propensity of vinyl radicals to add
to isonitriles,4h we were prompted to exploit these reac-
tions in search of other synthetically useful three-
component procedures. Toward this purpose, the gen-
eration of vinyl radicals in the presence of isonitriles was
carried out through regioselective addition of sulfanyl
radicals to alkynes.6 Sulfanyl radicals in turn were
produced by light-induced homolytic cleavage of the
corresponding disulfides. The usual hydrogen abstraction
from thiols was not considered, since thiols are known
to be highly efficient scavengers of vinyl radicals. Here
we are pleased to report that irradiation of aromatic
disulfides in the presence of suitable alkynes and isoni-
triles can actually result in a novel three-component
radical cascade reaction involving initial addition of a
sulfanyl radical to the alkyne and subsequent addition
of the ensuing vinyl radical to the isonitrile.7
Resu lts a n d Discu ssion
The first experiment was carried out by UV irradiation
of a benzene solution (40 mL) of commercially available
diphenyl disulfide 1a (0.5 mmol) and phenylacetylene 2
(5 mmol) and easily accessible 4-methoxyphenyl isonitrile
3 (10 mmol).8 Benzene proved to be the best reaction
solvent. Reactions carried out in methylene chloride,
methanol, or hexane were characterized by lower yields
and/or longer reaction times. Comparable yields but
longer reaction times were obtained by performing the
experiments in more concentrated solutions.9 Unexpect-
edly, the reaction afforded only products 4a , 5a , and 6,
derived from 1a and 2 with no intervention of the
isonitrile (Scheme 1, reaction a). This curious outcome,
quite unpredictable in light of our previous results,4h led
us to repeat the experiment in the presence of a radical
scavenger to trap some reaction intermediate. Indeed,
when the same reagents were allowed to react under
identical conditions in the presence of m-dinitrobenzene
(MDNB) we obtained again minor amounts of the above
compounds together with amides 7 and 8 in ca. 60%
overall yield (Scheme 1, reaction b). Structure 8 was
assigned on the basis of the X-ray structure determined
for another byproduct obtained from analogous, previ-
ously reported photolyses of disulfides and isonitriles.4l,10
This result can be rationalized according to the reaction
pathway shown in Scheme 2. Fast addition of a sulfanyl
radical to the alkyne gives vinyl radical 9; subsequent
addition of 9 to isonitrile 3 generates imidoyl radical 10.
In the absence of MDNB, radical 10 cannot evolve rapidly
into other intermediates;11 since the addition of 9 to 3 is
likely to be reversible,12 it rather prefers to give back
vinyl radical 9, which is responsible for the formation of
compounds 4a , 5a , and 6.13 On the other hand, in the
presence of MDNB, imidoyl 10 is efficiently trapped by
(5) Ryu, I.; Yamazaki H.; Kusano, K.; Ogawa, A.; Sonoda, N. J . Am.
Chem. Soc. 1991, 113, 8558. Keck, G. E.; Kordik, C. P. Tetrahedron
Lett. 1993, 34, 6875. Russell, G. A.; Kulkarni, S. V. J . Org. Chem. 1993,
58, 2678. Lee, E.; Hur, C. U.; Rhee, Y. H.; Park, Y. C.; Kim, S. Y. J .
Chem. Soc., Chem. Commun. 1993, 1466. For multicomponent radical-
ionic sequences see: Ishiyama, T.; Murata, M.; Suzuki, A.; Miyaura,
N. J . Chem. Soc., Chem. Commun. 1995, 295. Takai, K.; Ueda, T.;
Ikeda, N.; Moriwake, T. J . Org. Chem. 1996, 61, 7990. Tsunoi, S.; Ryu,
I.; Yamasaki, S.; Tanaka, M.; Sonoda, N.; Komatsu, M. J . Chem. Soc.,
Chem. Commun. 1997, 1889. Takai, K.; Matsukawa, N.; Takahashi,
A.; Fujii, T. Angew. Chem., Int. Ed. Engl. 1998, 37, 152. For notable
nonradical multicomponent procedures involving isonitriles, see: Ugi,
I. Angew. Chem., Int. Ed. Engl. 1982, 21, 810. Hanusch-Kompa, C.;
Ugi, I. Tetrahedron Lett. 1998, 39, 2725.
(6) (a) Benati, L.; Montevecchi, P. C.; Spagnolo, P. J . Chem. Soc.,
Perkin Trans. 1 1991, 2103 and references therein. See also: (b) Benati,
L.; Montevecchi, P. C.; Spagnolo, P. J . Chem. Soc., Perkin Trans. 1
1992, 1659. (c) Benati, L.; Capella, L.; Montevecchi, P. C.; Spagnolo,
P. J . Chem. Soc., Perkin Trans. 1 1995, 1035. (d) Benati, L.; Capella,
L.; Montevecchi, P. C.; Spagnolo, P. J . Org. Chem. 1995, 60, 7941. (e)
Montevecchi, P. C.; Navacchia, M. L. J . Org. Chem. 1997, 62, 5600. (f)
Montevecchi, P. C.; Navacchia, M. L. J . Org. Chem. 1998, 63, 537. (g)
Montevecchi, P. C.; Navacchia, M. L.; Spagnolo, P. Tetrahedron 1998,
54, 8207. (h) Montevecchi, P. C.; Navacchia, M. L.; Spagnolo, P. Eur.
J . Org. Chem. 1998, 1219.
(7) Under these conditions, the only side reaction is the isonitrile-
nitrile self-isomerization (see: Casanova, J ., J r.; Werner, N. D.;
Schuster, R. E. J . Org. Chem. 1966, 31, 3473). Since the isonitrile was
present in a large excess, this reaction did not significantly affect the
reaction outcome; however, it occurred quite remarkably only in the
case of very long reaction times.
(8) Isonitriles can be readily synthesized in very good yields by
conversion of the corresponding amines into formamides and subse-
quent dehydration with phosphorus oxychloride (Wolber, E. K. A.;
Schmittel, M.; Ru¨chardt, C. Chem. Ber. 1992, 125, 525).
(10) Leardini, R.; Nanni, D.; Zanardi, G. Eur. J . Org. Chem., 2000,
707. Analogous compounds were always obtained, generally in small
amounts, in all of the subsequent reactions (see the Experimental
Section).
(11) At least one fate was in principle conceivablesand expecteds
for imidoyl radical 10, i.e., six-membered cyclization onto the S-phenyl
ring. In fact, it has been reported that imidoyls can give 1,6-ring closure
onto aromatic rings, at least under oxidative conditions (see ref 4b).
Under the present conditions this process is probably slow and it cannot
compete with the R-fragmentation reaction giving back the isonitrile
and the vinyl radical. Another possibility was the trap of radical 10
by another sulfanyl radical to give an N-bis(arylthio)methylene deriva-
tive. These products were actually obtained in low to moderate yields
by reacting isonitriles and disulfides in the absence of alkynes or
hydrogen donors, but they were never observed in the present study.
(12) Although data on the reversibility of radical addition to
isonitrilessas well as homolytic fragmentations providing vinyl radicalss
are not available in the literature, we found evidence that, at least
with particularly stable radicals, this reaction is likely to be reversible
(see ref 4j). Further support to a reversible radical addition to isonitriles
will be reported below.
(9) This is consistent with the disulfide dissociation equilibrium,
which is shifted towards the reagent by lowering the dilution.
Synthetically useful reactions with photolytically generated sulfanyl
radicals are therefore usually carried out at high dilutions.

Radical Addition to Isonitriles
Sch em e 2
J . Org. Chem., Vol. 65, No. 9, 2000 2765
Sch em e 3
Sch em e 4
the scavenger, giving the intermediate 11;14 fragmenta-
tion of 11 followed by hydrogen abstraction of amidyl
radical 12 affords the final amide 7. An identical mech-
anism can explain the formation of amide 8 starting from
sulfanyl radical 13, which is formed by ortho-selective
photo-Fries rearrangement of the starting disulfide 1a.4l,10
This result was particularly encouraging, since the
reaction affordedsin fair overall yieldscompounds de-
rived from a selective three-component radical cascade
reaction.15 In addition, the reaction provided a one-pot
procedure for the synthesis of potentially interesting
polyfunctionalized alkenes from three simple molecules.
On this basis, we were prompted to explore the synthetic
potential of this reaction by employing a different isoni-
trile, which might lead to an imidoyl radical capable of
undergoing a fast evolution without the intervention of
any additional fourth component (e.g., MDNB). The
subsequent reactions were therefore carried out with
commercially available tert-butyl isonitrile (14). Indeed,
this is known to afford imidoyl radicals that undergo
quite fast â-scission with release of a tert-butyl radical
and formation of a nitrile.1a,b,2a-c,e,h-j,4g,16
When 1a (0.5 mmol), 2 (5 mmol), and 14 (10 mmol)
were allowed to react under the usual conditions, the
unsaturated nitrile 15a was produced in satisfactory
yield (46%),17 in addition to minor amounts of compounds
4a (17%), 5a (2%), and 6 (21%) (Scheme 3). Nitrile 15a
occurred as a mixture of the E- and Z-isomer, whose
relative ratio was originally 4.5:1, but it became 6.7:1
upon chromatographic separation. The reaction mecha-
nism is outlined in Scheme 4. As before, tandem addition
of sulfanyl radical to phenylacetylene and reaction of the
resulting vinyl radical with the isonitrile give imidoyl
radical 16. This intermediate is efficiently “trapped” by
â-fragmentation with loss of tert-butyl radicalsa pathway
that is precluded to imidoyl radical 10sand it is thus
responsible for the formation of nitrile 15. Molecules with
(13) All of these compounds have been commonly observed in the
reactions between sulfanyl radicals and phenylacetylene (see ref 6a,b).
However, since we did not use thiols as a source of sulfanyl radicals,
we were quite astonished to obtain the hydrogen-abstraction product
4a in significant amounts. In our reaction, one of the possible hydrogen-
atom sources could be tentatively identified as the cyclohexadienyl
radicals involved in the formation of compound 5a .
(14) For trapping of radicals by nitro groups resulting in oxygen-
transfer reactions from the nitro moiety to the radical center, see, for
example: Topiwala, U. P.; Luszniak, M. C.; Whiting, D. A. J . Chem.
Soc., Perkin Trans. 1 1998, 1185.
(15) It is worth noting that the literature has reported several
examples of both sulfanyl radical addition to isonitriles (see refs 2a,d,g
and 4l) and imidoyl radical addition to alkynes (see ref 4a,c,h and:
Dan-oh, Y.; Matta, H.; Uemura, J .; Watanabe, H.; Uneyama, K. Bull.
Chem. Soc. J pn. 1995, 68, 1497). Therefore, one could envisage an
alternative reaction pathway including sulfanyl radical addition to 3
followed by reaction of the resulting R-thio-imidoyl radical with 2 to
give a vinyl radical. Cyclization of the latter intermediate onto the
isonitrile aryl ring could lead to a quinoline derivative. Nevertheless,
this kind of product was never observed under any experimental
conditions.
(16) Ohta, H.; Tokumaru, K. J . Chem. Soc., Chem. Commun. 1970,
1601.
(17) A large excess of isonitrile is essential to trap the vinyl radical
efficiently. When we used 5 mmol of 14, the yield dropped to 15%.

2766 J . Org. Chem., Vol. 65, No. 9, 2000
Leardini et al.
Ta ble 1. Yield s of Com p ou n d s 4-6, a n d 15 Obta in ed in
th e Rea ction s of Disu lfid es 1a -d w ith 2 a n d 14a ,b
Sch em e 5
R
t (h)
4 (%)
5 (%)
6 (%)
15 (%, E/ Z)
a
b
c
d
H
72
7
30
24
17
8
9
2
12
9
21
8
7
46, 4.5:1
60, 2.5:1
45, 4:1
CN
OMe
Cl
8
12
9
60, 3.7:1
a
Products 4-6 were usually obtained as mixtures, and yields
b
were determined by GC and/or NMR analysis. Yields of 15 are
for the pure compound obtained after column chromatography; the
E/ Z ratio was determined by GC analysis prior to column
chromatography.
a 15-like structure are potentially very interesting, since
they combine the functionalities of R,â-unsaturated ni-
triles and vinyl sulfides, both of considerable importance
in organic synthesis.18 Moreover, the specific skeleton of
15sa â-thio-substituted acrylonitrileshas been success-
fully employed in the syntheses of fungicides, herbicides,
and insecticides,19 cephalosporin derivatives,20 and push-
pull dienes for nonlinear optics applications,21 and it is
considered a notable building block for a wide range of
natural or biologically active compounds.22
The reaction was repeated with various aromatic
disulfides and alkynes (Table 1). The formation of nitrile
15 was favored by electron-withdrawing disulfide sub-
stituents: indeed, with R ) CN or Cl a good yield of 15
(60%) was obtained. Furthermore, any R group was found
to lower reaction times remarkably. This effect was
particularly notable with disulfide 1b, whose reaction
was about 10 times faster than that of disulfide 1a .
Therefore, reactions with further alkynes were normally
studied with disulfide 1b.
We first examined the use of trimethylsilylacetylene
(17), which is known to react with sulfanyls to give vinyl
radicals structurally analogous to those obtained with
phenylacetylene.23 As shown in Scheme 5, in this case
the formation of nitrile 21 was significantly hindered by
the competing intramolecular ring closure of the vinyl
radical to give benzothiophene 19. This behavior was
somewhat expected, since R-trimethylsilyl-â-(phenylsul-
fanyl)vinyl radicals are reported to be much more prone
to cyclize to benzothiophenes than R-phenyl-â-(phenyl-
sulfanyl)vinyl radicals.6b
Sch em e 6
with a 25% yield of the starting disulfide 1b, which was
initially shown by GC-MS analysis to be totally absent.
Presumably, nitrile 21 was preferentially formed as the
Z-isomer, which, under chromatographic conditions, could
suffer extensive syn-elimination of transient aryl tri-
methylsilyl sulfide 22; this eventually gave disulfide 1b
by subsequent decomposition (Scheme 6). The outcoming
stereochemistry of nitrile 21, as well as that of the above
nitrile 15, can be explained in terms of preferential
trapping of linear 1-trimethylsilyl- and 1-phenyl-2-(aryl-
sulfanyl)vinyl radicals by isonitrile 14 on the side op-
posite to the sulfanyl moiety (see below for further
discussion).
To test the possibility that the low yield of 21 could be
a result of steric hindrance between the trimethylsilyl
group of the vinyl radical and the tert-butyl substituent
of the incoming isonitrile, we carried out the same
reaction in the presence of an isonitrile bearing a linear
side-chain, i.e., n-dodecyl isonitrile (24). Unexpectedly,
after 9 hsthe same time needed for complete disappear-
ance of the starting material in the analogous reaction
with 14sthe reaction mixture contained major amounts
of unreacted disulfide 1b and small quantities of 18 and
19: no trace of 21 was detected. Although not useful to
throw more light on the reason 21 was obtained in low
The acrylonitrile 21 was produced as a 2.5:1 mixture
of the geometrical isomers in 35% overall yield. However,
column chromatography of the crude led to isolate only
a 10% yield of a 1:1 mixture24 of (Z)- and (E)-21, together
(24) In all of the reactions described in this paper, the alkene having
the alkyne-derived group and the sulfanyl moiety on the same side of
the C-C double bond is more stable than the other isomer. This was
suggested by the silica-catalyzed isomerization observed during column
chromatography and it was definitively proved by photolyses of each
of the (supposed) less stable isomers in the presence of the correspond-
ing disulfide. After 24 h under these conditions, they gave almost
complete conversion to the alkene with the opposite configuration, due
to reversible sulfanyl radical addition to the C-C double bond. See:
(a) Kice, J . L. In Free Radicals; Kochi, J . K., Ed.; Wiley: New York,
1973; Vol. II, pp 720-724. (b) Oswald, A. A.; Griesbaum, K.; Hudson,
B. E., J r.; Bregman, J . M. J . Am. Chem. Soc. 1964, 86, 2877. (c)
Kampmeier, J . A.; Chen, G. J . Am. Chem. Soc. 1965, 87, 2608. (d)
Heiba, E. I.; Dessau, R. M. J . Org. Chem. 1967, 32, 3837. It is worth
noting that the most stable configuration is named E for the alkene
obtained from phenylacetylene (and 1-hexyne, see below) but Z for that
arising from trimethylsilylacetylene, due to the presence of the silicon
atom instead of a carbon.
(18) For the use of vinyl sulfides in organic synthesis, see: Cookson,
R. C.; Parson, P. J . J . Chem. Soc., Chem. Commun. 1976, 990. Ager,
D. J . J . Chem. Soc., Perkin Trans. 1 1983, 1131. Oshima, K.; Shimoji,
K.; Takahashi, H.; Yamamoto, H.; Nozaki, H. J . Am. Chem. Soc. 1973,
95, 2694.
(19) Frazza, E. J .; Rappoport, L. U.S. 3,271,408, 1966; Chem. Abstr.
1966, 65, 16975a. Strong, J . G. U.S. 3,828,091, 1974; Chem. Abstr.
1974, 81, 105054w.
(20) Carlo Erba S.p. A. Belg. 849,177, 1977; Chem. Abstr. 1978, 88,
190855v.
(21) Ono, N.; Matsumoto, K.; Ogawa, T.; Tani, H.; Uno, H. J . Chem.
Soc., Perkin Trans. 1 1996, 1905.
(22) Bakuzis, P.; Bakuzis, M. L. F. J . Org. Chem. 1981, 46, 235.
Aurich, H. G.; Quintero, J .-L. R. Tetrahedron 1994, 50, 3929.
(23) Griller, D.; Cooper, J . W.; Ingold, K. U. J . Am. Chem. Soc. 1975,
97, 4269.

Radical Addition to Isonitriles
Sch em e 7
J . Org. Chem., Vol. 65, No. 9, 2000 2767
Sch em e 8
Sch em e 9
yields from 14, this finding interestingly gave additional
support to the above postulated reversible attack of vinyl
radicals to isonitriles. As a matter of fact, there is no
plausible reason to postulate very different reaction rates
between vinyl radical 23 and the two alkyl isonitriles 24
and 14. Even if we do, the fastest reaction should be that
with 24 rather than the other one, since the linear side
chain of 24 should minimize the steric hindrance in the
transition state. Nevertheless, this reaction did not afford
21 at all. The only conceivable, meaningful difference
between the two processes resides in the next step, i.e.,
the â-scission of the imidoyl radical, which furnishes
either a tert-butylswith 14sor a primary alkyl radicals
with 24. Assuming a reversible addition step between
vinyl radical 23 and the two isonitriles (Scheme 7),
imidoyl radical 25 is rapidly consumed by the fast
â-fragmentation with loss of a tert-butyl radical, thus
driving the equilibrium toward nitrile 21. On the con-
trary, the â-scission of imidoyl 26, with loss of a primary
carbon radical, cannot compete efficiently with the
reverse reaction and the process exclusively affordssand
in much longer timessproducts derived from vinyl radi-
cal 23. In our opinion, the overall present and previous4j
results strongly suggest that radical addition to isoni-
triles is a reversible process, at least whenslike in the
case of vinyls 9 and 23sthe attacking radicals possess a
fair stability.
Much better results were obtained by reacting disulfide
1b and isonitrile 14 with 1-hexyne (27). This reaction
afforded small amounts of byproducts 28 and 29 (3%
each)s2,4-di-n-butylthiophene analogous to 6 and 20 was
not detected at allstogether with a good 50% yield of
nitrile 30 (Scheme 8). In this case, a low yield of 29 was
expected on the basis of the reported low propensity of
R-alkyl-â-(phenylsulfanyl)vinyl radicals to cyclize to
benzothiophenes.6b
Having shown that the three-component radical cas-
cade reaction can be efficiently applied to alkynes with
very different properties, in the next section we shall deal
with a detailed mechanistic discussion about the stereo-
chemical outcome of the reaction itself. As a matter of
fact, the stereochemistry of the reactions of â-(arylsul-
fanyl)vinyl radicals with isonitriles differs significantly
from the analogous reactions of those radicals with
hydrogen donors.
First, let us summarize the results obtained with the
three alkynes (Scheme 9). The reactions carried out with
phenyl- (2) and trimethylsilylacetylene (17) afforded
predominantly the alkenes arising from approach of the
isonitrile to the side of the vinyl radical opposite to the
sulfanyl group (E for the former and Z for the latter
alkyne). Confining the discussion to alkyne 2 for sake of
simplicitysthe reaction with 17 suffers from decomposi-
tion problems of the major productsthe two isomers were
identified by mp and spectral data comparison between
each of the two alkenes obtained in the reaction of
disulfide 1d , partially separable by column chromatog-
raphy, and authentic samples of compounds (E)-15d and
(Z)-15d .25 Identification was then extended to products
15a -c on the basis of spectral analogies and assuming
that no substantial change in the stereochemical outcome
could be produced by changing the disulfide substituent.
In addition, an NOE experiment was carried out on the
presumed compound (E)-15c, which was unique to show
sufficiently separated resonances in the aromatic region.
By irradiating the two ortho protons of the R-phenyl ring,
we observed an exclusive positive effect on the signals of
the two vicinal meta protons, whereas no effect at all was
(25) Rappoport, Z.; Topol, A. J . Chem. Soc., Perkin Trans. 2 1975,
982.

2768 J . Org. Chem., Vol. 65, No. 9, 2000
Leardini et al.
noticed for the vinylic proton. The E/ Z ratio increased
slightly during the reaction, as proved by GC-MS
analyses performed at regular times, and, for compounds
15, it showed a remarkable enhancement after column
chromatography (Scheme 9, values in parentheses).
The major production of the E-isomer was consistent
with the results of previous reactions of a hydrogen donor
with â-(arylsulfanyl)vinyl radicals.6a R-Phenylvinyl radi-
cals have been suggested to be linear π-radicals,26 and
to minimize steric hindrance, a hydrogen donor ap-
proaches an R-phenyl-â-(arylsulfanyl)vinyl radical from
the side trans to the sulfanyl moiety, yielding the
corresponding alkenesunder kinetic controlsin a 1:9 E/Z
ratio. This approach seems to be additionally favored by
significant bonding interactions occurring between the
unpaired electron and the sulfur orbitals.6a This behavior
should be substantially imitated also when the scavenger
is an isonitrile and, as the isocyano group is at least as
encumbering as the thiol group, we should expect for
nitriles 15 an E/Z ratio of ca. 9:1 or even a little higher.27
On the contrary, although we could not perform our
reactions under perfect kinetic control, we can definitely
say that in our case the E-isomer was actually still the
major product, but the E/Z ratio was quite low, i.e., only
2.5-4.5:1. In addition, since for nitrile 15 the thermo-
dynamically and kinetically favored isomers are the same
compound (E)-15,24 the little isomerization that occurred
during the reaction and altered the isomer ratio with
respect to the one obtainable under kinetic control caused
the E/Z ratio to increase rather than decrease. Therefore,
we can reasonably conclude that, under absolute kinetic
control, the E/Z ratio should be even lower than that
observed and, thus, far away from what was expected.
Even more interesting were the results obtained from
1-hexyne 27. With this alkyne, the reaction is brought
about by R-alkyl-â-(arylsulfanyl)vinyl radicals, whichs
unlike the linear R-phenyl-substituted analoguessare
bent and rapidly interconverting σ-radicals.26 The choice
of an incoming scavenger is again dictated by steric
factors: if the reaction is dominated by the steric
hindrance between the R- and â-group of the vinyl
radical, then a hydrogen donor enters on the same side
of the sulfanyl group, giving the (E)-alkene; otherwise,
it approaches the opposite side, yielding the Z-isomer. It
is worth pointing out that with 27 the hydrogen-abstrac-
tion reaction follows the latter pathway affording the
reduction product in a 1:9 E/Z ratio, thus perfectly
resembling the result obtained with R-phenyl-substituted
vinyl radicals.6a This means that also with 1-hexyne the
product distribution is governed by steric interactions in
the transition state between the â-substituent and the
incoming scavenger. The negligible steric hindrance
between the R- and â-substituent of the vinyl radical,
compared with that between the â-group and the isoni-
trile moiety, is also supported by the stability of com-
pound (E)-30 that, like in the case of nitriles 15, is
thermodynamically favored with respect to its (Z)-
counterpart. This is an additional point that let us predict
an approach of the isonitrile from the side opposite to
the sulfanyl group.
As far as our reaction with 27 is concerned, the (E)-
30/(Z)-30 ratio was not only low beyond any expectation,
but, additionally, the kinetically favored product was (Z)-
30 instead of the predicted (E)-isomer. As the matter of
fact, when the reaction was stopped after 8 h, i.e., much
before complete conversion of the starting material, the
measured (E)-30/(Z)-30 ratio was 1:2.5. The ratio changed
to 1:1.3 and 1:1 at the end of the reaction and after
column chromatography, respectively, sinceslike in the
case of the other nitrilessthe E-isomer is the more stable
and thence considerable Z/E isomerization occurred
during the reaction and chromatographic separation. The
assignment of the Z-structure to the kinetic product (the
major isomer at early reaction times) was made on the
basis of NOE measurements carried out on a pure sample
of presumed (Z)-30 obtained after chromatographic work-
up. Irradiation on the allylic methylene signals caused a
marked enhancement of the vinyl proton.
Thus, the markedly decreased ratio of the (E)- and (Z)-
nitriles 15 and the inversion of the expected stereochem-
istry for nitrile 30 clearly indicate that the scavenging
of â-(arylsulfanyl)vinyl radicals by isonitriles and/or the
stereochemical fate of the resulting imidoyl radicals are
governed by more complicated factors than those involved
in analogous hydrogen abstraction reactions.
The stereochemical outcome of our reactions might be
explained in terms of possible vinyl radicalsisonitrile
interactions in the transition state. Since carbon radicals
are well-known to attack intermolecularly the sulfur
atom of aryl sulfides to give sulfuranyl radicals28sand
isonitriles can be considered diradical species in their
geminal radical acceptor/radical donor behaviorsit would
not be unreasonable to postulate that in the transition
state the incoming isonitrile might interact not only with
the vinyl-radical carbon but also with the sulfur atom
(Scheme 10). This effect could direct the approach of the
isonitrile to the same side of the sulfanyl group, thus
partially balancing steric hindrance. Due to the signifi-
cant intramolecular bonding interactions between the
unpaired electron and the sulfur orbitals,6a the three-
center interaction could be slightly important in the
reactions of radical 9. On the contrary, this is likely to
be quite significant in the case of radical 31, where the
intramolecular overlap is instead unimportant. This
hypothesis would lead to predict not only a low E/Z ratio
for the alkene products in the reactions with phenyl-
acetylene but also a marked preference for the Z-isomer
with 1-hexyne, which is consistent with our experimental
data.
(26) (a) Singer, L. A.; Chen, J . Tetrahedron Lett. 1969, 4849. (b)
Bennett, J . E.; Howard, J . A. Chem. Phys. Lett. 1971, 9, 460. (c) Ito,
O.; Omori, R.; Matsuda, M. J . Am. Chem. Soc. 1982, 104, 3934. (d)
Giese, B.; Lachhein S. Angew. Chem., Int. Ed. Engl. 1982, 21, 768. (e)
Giese, B.; Gonzales-Gomez, J . A.; Lachhein, S.; Metzger, J . O. Angew.
Chem., Int. Ed. Engl. 1987, 26, 479. (f) Giese, B. Angew. Chem., Int.
Ed. Engl. 1989, 28, 969. (g) Curran, D. P.; Kim, D. Tetrahedron 1991,
47, 6171. (h) J ournet, M.; Malacria, M. Tetrahedron Lett. 1992, 33,
1892. (i) J ournet, M.; Malacria, M. J . Org. Chem. 1992, 57, 3085.
(27) Calculations of both molar volume and parachor for thiol and
isocyano groups gave slightly greater values for the latter moiety. On
this occasion, to avoid confusion, we would like to point out that the
E/ Z convention overturns passing from hydrogen abstraction products
4, 18, and 28 (ref 6a) to our nitriles.
A different approach to rationalize the product distri-
bution is to consider the geometry and electronic struc-
ture of the imidoyl radical arising from addition of the
vinyl radical to the isonitrile (Scheme 11). In the absence
(28) Sulfuranyl radicals are assumed to be true intermediates in
intramolecular cyclizations onto diarylsulfide moieties; see: Leardini,
R.; Pedulli, G. F.; Tundo, A.; Zanardi, G. J . Chem. Soc., Chem.
Commun. 1985, 1390 and references therein. See also: Kampmeier,
J . A.; Evans, T. R. J . Am. Chem. Soc. 1966, 88, 4096. Beckwith, A. L.
J .; Boate, D. R. J . Chem. Soc., Chem. Commun. 1986, 189. Franz, J .
A.; Roberts, D. H.; Ferris, K. F. J . Org. Chem. 1987, 52, 2256.

Radical Addition to Isonitriles
Sch em e 10
J . Org. Chem., Vol. 65, No. 9, 2000 2769
Sch em e 12
Sch em e 11
bent geometry at the radical center,2d,g,31 but recent
studies have suggested thatslike in the case of the
isoelectronic vinyl radicals26b,estheir geometry and de-
localization degree of the unpaired electron may depend
on the substitution pattern, i.e., the R-group and par-
ticularly the nitrogen substituent.32
Both energy and equilibrium geometry of radical (E)-
32a were first optimized. The results shown in Scheme
12 indicate that a bent geometry (120°) at the radical
center was transformed into a quasilinear arrangement
(170.5°) of the NdC-C moiety. In addition, the order of
the CR-Câ and Câ-Cγ bonds was changed to 1.5 from 1
and 2, respectively, with the latter bond slightly longer
than the former, however. Finally, the estimated spin
density showed a maximum on the Cγ atom (0.40) instead
of the starting CR carbon (0.28), with some distribution
also on the nitrogen (0.07) and sulfur (0.15) atoms. These
data are more consistent with structure 33a rather than
the initial imidoyl 32a . The energies and equilibrium
geometries were then calculated for the other radicals
(Z)-32a , (E)-32b, and (Z)-32b, and the rotation barriers
were estimated for both pairs of radical configurations
(6.0 and 5.7 kcal/mol, respectively, Scheme 13). Finally,
from the kinetic data previously reported for â-fragmen-
tations of analogous N-tert-butyl-substituted imidoyls,2g
we estimated a value of ca. 8-10 kcal/mol for the
activation energy concerning release of the tert-butyl
of significant interactions in the transition state other
than steric hindrance, adduct 32 could be preferentially
formed in its predicted E-configuration. Nevertheless,
before undergoing â-fragmentation, (E)-32 might equili-
brate to some extent to the Z-isomer through rotation
around the virtually single Câ-Cγ bond of the contribut-
ing resonance structure 33. Imidoyl radical (Z)-32 could
be the thermodynamically preferred isomer due to some
intramolecular bonding interactions between the un-
paired electron and the sulfur orbitals, which should be
even more effective than in vinyl radical 9 (see above).
Very recently, some isoelectronic R,â-unsaturated acyl
radicals 34 have been described in terms of analogous
mesomeric structures (Scheme 11).29
Since no data have been so far reported concerning the
geometry and electronic structure of R,â-unsaturated
imidoyl radicals, semiempirical MNDO-d calculations30
were performed on radicals 32 (a : R ) H, X ) Ph; b: R
) H, X ) Me). Generally, imidoyls are σ-radicals with a
(29) De Boeck, B.; Herbert, N.; Pattenden, G. Tetrahedron Lett.
1998, 39, 6971. De Boeck, B.; Pattenden, G. Tetrahedron Lett. 1998,
39, 6975.
(30) The choice of MNDO-d parametrization was made to better take
into account the possible involvement of sulfur d-orbitals.
(31) Danen, W. C.; West, C. T. J . Am. Chem. Soc. 1973, 95, 6872.
(32) Pareschi, P. Imidoyl Radicals: Spectroscopic Properties, Reac-
tivity, and Use in the Synthesis of Heterocyclic Compounds. Ph.D.
Thesis, University of Bologna, Italy, 1996, Chapter 1, in collaboration
with J . C. Walton, University of St. Andrews, U.K.

2770 J . Org. Chem., Vol. 65, No. 9, 2000
Leardini et al.
Sch em e 13
radical intermediate 36 could undergo efficient cycliza-
tion on the nitrile moiety despite its kinetically predict-
able (and unfavorable) E-configuration. In light of the
present data calculated for imidoyl radicals 32, ready E/Z
isomerization of 36 through rotation around the Câ-Cγ
bondsand thence approach of the radical center to the
cyano groupsnow becomes plausible.
Con clu sion s
The photolytic reaction of disulfides with alkynes and
isonitriles represents a novel three-component radical
cascade reaction that proceeds by tandem addition of a
sulfanyl radical to the alkyne and attack of the resulting
vinyl radical to the isonitrile carbon atom to give an
imidoyl radical. With aromatic isonitriles, a scavenger
(e.g., MDNB) is needed to trap the produced imidoyl
radical prior to back fragmentation to the vinyl precursor.
With tert-butyl isonitrile, the imidoyl intermediate is
“trapped” by a â-scission process resulting in release of
a tert-butyl radical and production of a nitrile. This one-
pot protocol affords fair yields of synthetically useful
polyfunctionalized alkenes using disulfides, isonitriles,
and alkynes with very different properties. Moreover, our
present results suggest that vinyl radical addition to
isonitriles is a reversible process: this fact is also
supported by the reaction of disulfide 1b and alkyne 17
in the presence of a nonbranched aliphatic isonitrile (24).
Sch em e 14
The stereochemical outcome of the three-component
reactions differs significantly from that previously en-
countered with related hydrogen abstraction reactions of
thio-substituted vinyl radicals. This could be the conse-
quence of vinyl radical-isonitrile interactions in the
transition state and/or fast E/Z isomerization of the
eventual imidoyl radical. Support to the latter possibility
was given by structure and spin-density data obtained
by MNDO-d semiempirical calculations.
Exp er im en ta l Section
Gen er a l P r oced u r es. 1H and 13C NMR spectra were
recorded in deuteriochloroform using tetramethylsilane as an
internal standard. Mass spectra (MS) and high-resolution
mass spectra (HRMS) were performed by electron impact with
a beam energy of 70 eV: relative intensities are given in
parentheses. GC-MS analyses were carried out on a quadru-
polar instrument equipped with a Quadrex capillary column
(007, 25 m × 0.25 mm i.d.). The E/Z ratios for vinylic
compounds were determined by a 60-260 °C temperature
ramp with a rate of 10 °C/min, unless otherwise stated; the
differences in retention times (∆rt) are given in seconds. IR
spectra were recorded in film or chloroform solution. Column
chromatography was carried out on 60-Å silica gel or basic
aluminum oxide (70-230 mesh, activity 3) using light petro-
leum (40-70 °C) and a light petroleum/diethyl ether gradient
(from 0 up to 100% diethyl ether) as eluant. UV photolyses of
disulfides were performed with a Heraeus TQ 150 high-
pressure mercury lamp (150 W). Previously reported reaction
products were identified by spectral comparison and/or mixed
mp determination with authentic specimens. The purity of new
compounds was confirmed by HRMS and elemental analysis;
in the case of inseparable mixtures of compounds, the purity
was proved by the absence of any significant extraneous peak
in the 1H NMR spectra and/or by GC-MS analysis.
radical from 32-like intermediates. On this basis, it can
be reasonably assumed that, under our experimental
conditions, radicals 32 can exhibit fair rotation around
the Câ-Cγ in competition with the â-scission process. The
lower barrier predicted for radical 32b suggests that
isomerization might be even more effective in the reaction
carried out with 1-hexyne, which is consistent with the
experimental data.
Although the above findings cannot afford a full
explanation of the present stereochemical data, they
strongly suggest that the structure of the R,â-unsaturated
imidoyl radical can have a major influence on the
configuration of the reaction products and it can thence
considerably alter the expected, kinetically favored out-
come. However, at this stage, concomitant or even
alternative contribution of transition state interactions
(Scheme 10) cannot be ruled out.
It is worth pointing out that the above theoretical data
also serve to clarify some as yet unexplained behavior of
the related imidoyl radical 36. This was the key inter-
mediate in the smooth formation of the cyclopenta-fused
quinoxaline 38 (50% yield) by tandem 5-exo-dig cycliza-
tion onto the cyano group and subsequent six-membered
ring closure of the resulting iminyl radical 37 onto the
aromatic ring of the isonitrile (Scheme 14).4h At that time,
we had no convincing explanation of the reasons why the
Diphenyl disulfide (1a ), 4-methoxythiophenol, 4-chlorothio-
phenol, 4-hydroxybenzonitrile, p-anisidine, n-dodecylamine,
phenylacetylene (2), tert-butyl isonitrile (14), trimethylsilyl-
acetylene (17), 1-hexyne (27), and m-dinitrobenzene (MDNB)
were commercially available. 4-[(4-Cyanophenyl)disulfanyl]-

Radical Addition to Isonitriles
J . Org. Chem., Vol. 65, No. 9, 2000 2771
benzonitrile (1b),33 bis(4-methoxyphenyl) disulfide (1c),34 and
bis(4-chlorophenyl) disulfide (1d )35 were prepared according
to the literature. N-(4-Methoxyphenyl)formamide36 and N-n-
dodecylformamide37 were synthesized from the corresponding
amines by treatment with boiling formic acid.36 4-Methoxy-
phenyl isonitrile (3)38 and n-dodecyl isonitrile (24)39 were
prepared from the corresponding formamides by reaction with
diisopropylamine and phosphorus oxychloride in dichlo-
romethane.40
Rea ction s of Disu lfid es w ith Alk yn es a n d Ison itr iles.
Gen er a l P r oced u r e. A benzene (40 mL) solution of disulfide
(0.5 mmol), alkyne (5 mmol), and isonitrile (10 mmol) in a
quartz Erlenmeyer flask was kept at rt for 7-72 h under a
nitrogen atmosphere and UV irradiation. Disappearance of the
starting material was monitored by TLC and/or GC-MS
analysis. After removal of the solvent, the residue was chro-
matographed. In each reaction, products are listed according
to the elution order.
in a 2:3 ratio (20% overall yield, ∆rt ) 13). 4b: MS m/z 237
(M+, 100); HRMS calcd for C15H11NS 237.0612, found 237.0628.
5b: MS m/z 235 (M+, 100); HRMS calcd for C15H9NS 235.0456,
found 235.0448. 4b + 5b: 1H NMR (200 MHz) δ 6.46 (d, J )
10.6 Hz, (Z)-4b), 6.80 (d, J ) 10.6 Hz, (Z)-4b), 6.82 (d, J )
15.3 Hz, (E)-4b), 6.96 (d, J ) 15.3 Hz, (E)-4b), 7.24-7.73 (m),
7.99 (d, J ) 8.4 Hz, 5b), 8.22 (dd, J 1 ) 1.3 Hz, J 2 ) 0.8 Hz,
5b).
Further elution gave 4-[[2-cyano-2-phenylethenyl]sulfanyl]-
benzonitrile (15b) as a 4:1 E/Z mixture (2.5:1 before chroma-
tography, 60% overall yield, ∆rt ) 30): mp ) 124-126 °C (from
light petroleum/benzene 70:30 v/v); 1H NMR (200 MHz) δ
7.35-7.75 (m); 13C NMR (50 MHz) δ 112.41 (q), 112.60 (q),
118.35 (q), 118.76 (q), 125.86, 127.47 (q), 128.51, 129.47,
129.74, 129.98, 130.10, 130.52, 130.80, 132.23 (q), 133.53,
133.62, 139.57, 139.91 (q), 141.97; MS m/z 262 (M+, 100);
HRMS calcd for C16H10N2S 262.0565, found 262.0577. Anal.
Calcd for C16H10N2S: C, 73.26; H, 3.84; N, 10.68; S, 12.22.
Found: C, 73.59; H, 3.83; N, 10.61; S, 12.29.
Rea ction s of 1a w ith 2 a n d 3. After 48 h, in the absence
of MDNB, the reaction afforded, besides polymeric products
and tarry material, a mixture (∼40%) of phenyl 2-phenylethe-
nyl sulfide (4a ),6a 3-phenylbenzo[b]thiophene (5a ),41 and 2,4-
diphenylthiophene (6).41 In the presence of MDNB (2 mmol),
after 72 h, we obtained again, by chromatography on alumi-
num oxide, a first fraction containing 4a , 5a , and 6 (∼20%),
followed by N-(4-methoxyphenyl)-2-phenyl-3-(phenylsulfanyl)-
Final elution gave presumably4l,10 4-[[2-cyano-2-phenylethe-
nyl]sulfanyl]-3-[(4-cyanophenyl)sulfanyl]benzonitrile (5%): oil;
MS m/z 395 (M+, 34); HRMS calcd for C23H13N3S2 395.0551,
found 395.0559.
Rea ction of 1c w ith 2 a n d 14. After 30 h, column
chromatography of the reaction mixture afforded a mixture
of 4c,42 5c, and 6 in a 1 (two isomers in a 1:1 ratio):1:0.8 ratio
(25% overall yield). 5c: MS m/z 240 (M+, 100); HRMS calcd
for C15H12OS 240.0608, found 240.0618. 4c + 5c + 6: 1H NMR
(200 MHz) δ 3.79 (s, (Z)-4c), 3.82 (s, (E)-4c), 6.43 (d, J ) 10.6
Hz, (Z)-4c), 6.53 (d, J ) 10.6 Hz, (Z)-4c), 6.53 (d, J ) 15.3 Hz,
(E)-4c), 6.86 (d, J ) 15.3 Hz, (E)-4c), 6.85-6.98 (m), 7.06 (dd,
J 1 ) 8.9 Hz, J 2 ) 2.4 Hz, 5c), 7.20-7.76 (m), 7.78 (d, J ) 8.9
Hz, 5c).
Further elution gave 3-[(4-methoxyphenyl)sulfanyl]-2-phen-
yl-2-propenenitrile (15c) as a 3.3:1 E/ Z mixture (4:1 before
chromatography, 45% overall yield, ∆rt ) 29): oil; 1H NMR
(300 MHz, benzene-d6) δ 3.42 (s, (Z)-15c), 3.44 (s, (E)-15c),
6.75 (A part of AA′BB′, J ) 9.1 Hz, (E)-15c), 6.80 (A part of
AA′BB′, J ) 9.1 Hz, (Z)-15c), 7.14-7.40 (m + B part of AA′BB′,
J ) 9.1 Hz, (E)-15c), 7.29 (s, (E)-15c), 7.87-7.93 (m, (E)-15c).
A NOEDIF experiment was carried out on a fraction contain-
ing 15c in a very high E/ Z ratio by irradiating the two ortho
protons of the unsubstituted R-phenyl ring (7.87-7.93 ppm);
the experiment showed an exclusive positive effect (3.5%) on
the signals of the two vicinal meta protons (7.30-7.37 ppm),
whereas no effect at all was noticed for the singlet ascribable
to the vinylic proton (7.29 ppm): MS m/z 267 (M+, 100); HRMS
calcd for C16H13NOS 267.0718, found 267.0725. Anal. Calcd
for C16H13NOS: C, 71.88; H, 4.90; N, 5.24; S, 11.99. Found:
C, 72.05; H, 4.86; N, 5.26; S, 12.07.
1
2-propenamide 7 (0.1 g, 38%) as a 1:1 E/ Z mixture: oil; H
NMR (300 MHz) δ 4.03 (3 H, s), 4.04 (3 H, s), 7.09 (4 H, AA′BB′,
J ) 6.3 Hz), 7.30 (1 H, bs, NH, removed upon treatment with
D2O), 7.40 (1 H, s), 7.50-7.85 (24 H, m), 8.34 (1 H, s); 13C NMR
(50 MHz) δ 55.95, 114.65, 122.19, 122.27, 128.52, 129.04,
129.70, 129.83, 129.89, 129.93, 130.17, 130.31, 131.12, 131.40
(q), 131.50 (q), 131.60, 131.86 (q), 134.65 (q), 135.32 (q), 138.15
(q), 138.20 (q), 142.38, 145.43, 157.05 (q), 163.22 (q), 164.92
(q); MS m/z 361 (M+, 56); HRMS calcd for C22H19NO2S
361.1137, found 361.1129. Anal. Calcd for C22H19NO2S: C,
73.10; H, 5.30; N, 3.87; S, 8.87. Found: C, 73.55; H, 5.33; N,
3.89; S, 8.84. A subsequent fraction contained N-(4-methoxy-
phenyl)-2-phenyl-3-[[2-(phenylsulfanyl)phenyl]sulfanyl]-2-pro-
penamide 8 (0.05 g, 20%) as a 3:1 mixture of stereoisomers:
oil; 1H NMR (300 MHz) δ 4.03 (s, minor isomer), 4.04 (s, major
isomer), 7.06-7.12 (d + d, J 1 ) 8.2 Hz, J 2 ) 7.6 Hz, A part of
AA′BB′, both isomers), 7.25-7.90 (m); MS m/z 469 (M+, 69);
HRMS calcd for C28H23NO2S2 469.1170, found 469.1191.
Rea ction of 1a w ith 2 a n d 14. After 72 h, column
chromatography of the reaction mixture gave a mixture of 4a ,
5a , and 6 in a 1.7 (two isomers in a 0.6:0.4 ratio):0.1:2 ratio
(38% overall yield), 2-phenyl-3-(phenylsulfanyl)-2-propeneni-
trile (15a ) as a 6.7:1 E/Z mixture (4.5:1 before chromatography,
46% overall yield) [oil; 1H NMR (200 MHz) δ 7.25-7.70 (11 H,
m); MS m/z 237 (M+, 100); HRMS calcd for C15H11NS 237.0612,
found 237.0625], and the presumable4l,10 photo-Fries rear-
rangement compound 2-phenyl-3-[[2-(phenylsulfanyl)phenyl]-
sulfanyl]-2-propenenitrile (6%): oil; [1H NMR (300 MHz) δ
7.05-7.70 (15 H, m); MS m/z 345 (M+, 100); HRMS calcd for
A careful analysis of the final elution fractions showed that
this reaction did not afford the photo-Fries rearrangement
product.
Rea ction of 1d w ith 2 a n d 14. After 24 h, column
chromatography of the reaction mixture afforded a mixture
of 4d ,24b 5d ,43 6, and tert-butyl 4-chlorophenyl sulfide44 in a
1.4 (2 isomers in a 1:0.8 ratio):2:1.4:1 ratio (35% overall yield).
Further elution gave 3-[(4-chlorophenyl)sulfanyl]-2-phenyl-
2-propenenitrile (15d ) as a 6.6:1 E/ Z mixture (3.7:1 before
chromatography, 60% overall yield, ∆rt ) 24), mp ) 99-100
°C (from light petroleum/ethanol 70:30 v/v) (lit.25 mp ) 105-
107 °C); comparison of spectral data of the two isomers of 15d
with those reported in the literature was essential to establish
the absolute configurations of the major and minor isomer.
(E)-15d : IR (film) νmax (cm-1) 3000, 2220 (CN), 1550, 1480,
1450, 1400, 1100, 1010, 865, 810, 765, 695 [lit.25 IR (KBr) νmax
(cm-1) 2210 (CN), 1630, 865; according to the authors, the band
at 865 cm-1 is characteristic of the E-isomer; actually it was
not observed in the IR spectrum of the minor isomer, see
C
21H15NS2 345.0646, found 345.0658. The E/ Z ratio for
compound 15a was determined by GC with a rate of 8 °C/min
(∆rt ) 22).
Rea ction of 1b w ith 2 a n d 14. After 7 h, column chro-
matography of the reaction mixture afforded 6 (8%) and a
mixture of 4-[[2-phenylethenyl]sulfanyl]benzonitrile (4b) (1:1
E/ Z ratio) and 3-phenyl-1-benzothiophene-5-carbonitrile (5b)
(33) Krishnamurthy, S.; Aimino, D. J . Org. Chem. 1989, 54, 4458.
(34) Harpp, D. N.; Ash, D. K.; Smith, R. A. J . Org. Chem. 1980, 45,
5155.
(35) King, K. F.; Bauer, L. J . Org. Chem. 1971, 36, 1641.
(36) Beilsteins Handbuch der Organischen Chemie, 1930, 13, 459.
(37) Katritzky, A. R.; Parris, R. L.; Ignatchenko, E. S.; Allin, S. M.;
Siskin, M. J . Prakt. Chem. 1997, 339, 59.
(38) Ugi, I.; Meyr, R. Chem. Ber. 1960, 93, 239.
(39) Ugi, I.; Fetzer, U.; Eholzer, U.; Knupfer, H.; Offermann, K.
Angew. Chem. 1965, 77, 492.
(40) Wolber, E. K. A.; Schmittel, M.; Ru¨chardt, C. Chem. Ber. 1992,
125, 525.
1
below]; H NMR (300 MHz) δ 7.36-7.51 (8 H, m), 7.60-7.65
(42) Baliah, V.; Rathinasamy, T. K. Indian J . Chem. 1971, 220.
(43) Dickinson, R. P.; Iddon, B. J . Chem. Soc. C 1970, 2592.
(44) Landini, D.; Montanari, F.; Rolla, F. J . Org. Chem. 1983, 48,
604.
(41) Boberg, F. J ustus Liebigs Ann. Chem. 1964, 679, 118.

2772 J . Org. Chem., Vol. 65, No. 9, 2000
Leardini et al.
(2 H, m) [lit.25 1H NMR (60 MHz) δ 7.22-7.61 (m); one end of
the resonance region is significantly shifted to lower fields for
the E-isomer; this feature is clearly shown by the major isomer
of 15d , see also below]; MS m/z 273 (M+ + 2, 25), 271 (M+,
100), 236 (26), 203 (52), 159 (32), 155 (16), 108 (22), 75 (31)
[lit.25 MS m/z 273 (M+ + 2, 36), 271 (M+, 100), 270 (29), 236
(40), 203 (58), 159 (24), 143 (29), 108 (29)]; Z-15d : IR (film)
νmax (cm-1) 3000, 2220 (CN), 1500, 1480, 1450, 1400, 1100,
1015, 810, 750, 700 [lit.25 IR (KBr) νmax (cm-1) 2240 (CN), 1620];
1H NMR (300 MHz) δ 7.35-7.52 (10 H, m) [lit.25 1H NMR (60
MHz) δ 7.25-7.53 (m)]; MS m/z 273 (M+ + 2, 26), 271 (M+,
100), 236 (25), 203 (51), 159 (33), 155 (16), 108 (25), 75 (38)
[lit.25 MS m/z 273 (M+ + 2, 37), 271 (M+, 100), 270 (30), 236
(28), 203 (44), 159 (31), 155 (25), 139 (36), 108 (27)].
uct arising from attack of the vinyl radical to the solvent
benzene] in a 1 (two isomers in a 1:1 ratio, ∆rt ) 8):1:1:1.6 (2
isomers in a 1:0.3 ratio) ratio (14% overall yield). 28 (first
isomer): MS m/z 217 (M+, 37). 28 (second isomer): MS m/z
217 (M+, 36); HRMS calcd for C13H15NS 217.0925, found
217.0919. 29: MS m/z 215 (M+, 22); HRMS calcd for C13H13
-
NS 215.0769, found 215.0764. 40 (major isomer): MS m/z 293
(M+, 57). 40 (minor isomer): MS m/z 293 (M+, 67); HRMS calcd
for C19H19NS 293.1238, found 293.1241. 28 + 29 + 39 + 40:
1H NMR (300 MHz) δ 0.80-1.05 (m), 1.20-1.50 (m), 1.65-
1.80 (m), 2.20-2.35 (m), 2.59 (t, J ) 6.4 Hz), 2.77 (t, J ) 6.5
Hz), 2.84 (t, J ) 6.7 Hz), 6.04-6.22 (m, (E)-28 + (Z)-28), 6.24
(s, 40 [major]), 6.40 (s, 40 [minor]), 7.20-7.64 (m), 7.72
(AA′BB′, J ) 8.2 Hz), 7.92 (d, J ) 8.2 Hz, 29), 8.06 (br s, 29).
Further elution gave 4-[[2-cyano-1-hexenyl]sulfanyl]ben-
zonitrile (30) as a 1.1:1 Z/E mixture (1.3:1 before chromatog-
raphy, 2.5:1 if the reaction is stopped after 8 h, 50% overall
yield, ∆rt ) 5): oil. (Z)-30: 1H NMR (300 MHz) δ 0.95 (3 H, t,
J ) 7.2 Hz), 1.32-1.45 (2 H, m), 1.52-1.67 (2 H, m), 2.37 (2
H, td, J t ) 7.5 Hz, J d ) 1.1 Hz), 6.90 (1 H, t, J ) 1.1 Hz), 7.45
(2 H, A part of AA′BB′, J ) 8.6 Hz), 7.65 (2 H, B part of AA′BB′,
J ) 8.6 Hz); MS m/z 242 (M+, 61). (E)-30: 1H NMR (300 MHz)
δ 0.98 (3 H, t, J ) 6.8 Hz), 1.30-1.50 (2 H, m), 1.50-1.70 (2
H, m), 2.35 (2 H, br t, J ) 6.8 Hz), 7.15 (1 H, t, J ) 0.8 Hz),
7.48 (2 H, A part of AA′BB′, J ) 8.5 Hz), 7.67 (2 H, B part of
AA′BB′, J ) 8.5 Hz); MS m/z 242 (M+, 61); HRMS calcd for
Final elution gave traces of the presumable4l,10 photo-Fries
rearrangement product, 3-([4-chloro-2-[(4-chlorophenyl)sulfa-
nyl]phenyl]sulfanyl)-2-phenyl-2-propenenitrile, MS m/z 413
(M+, 100); HRMS calcd for C21H13Cl2NS2 412.9867, found
412.9875.
Rea ction of 1b w ith 17 a n d 14. After 9 h, column
chromatography of the reaction mixture afforded trimethyl-
[4-(trimethylsilyl)-2-thienyl]silane45 [MS m/z 228 (M+, 13)].
Further elution gave a mixture of 18 and 19 in a 1 (2 isomers
in a 0.9:0.1 ratio):3.5 ratio (26% overall yield, ∆rt ) 18); E-18:
MS m/z 233 (M+, 23); Z-18: MS m/z 233 (M+, 20); HRMS calcd
for C12H15NSSi 233.0695, found 233.0701; 19: MS m/z 231 (M+,
9); HRMS calcd for C12H13NSSi 231.0538, found 231.0531;46
18 + 19: 1H NMR (300 MHz) δ 0.09 (9 H, s, (Z)-18), 0.13 (9
H, s, (E)-18), 0.41 (9 H, s, 19), 6.14 (1 H, d, J ) 12.3 Hz, (Z)-
18), 6.22 (1 H, d, J ) 17.9 Hz, (E)-18), 6.65 (1 H, d, J ) 17.9
Hz, (E)-18), 7.02 (1 H, d, J ) 12.3 Hz, (Z)-18), 7.34-7.42 (4 H,
A part of AA′BB′, (E)-18 + (Z)-18, J E ) 8.3 Hz), 7.55 (1 H, dd,
J 1 ) 8.1 Hz, J 2 ) 0.9 Hz, 19), 7.56-7.62 (4 H, B part of AA′BB′,
(E)-18 + (Z)-18, J E ) 8.3 Hz), 7.66 (1 H, s, 19), 8.02 (1 H, d, J
) 8.1 Hz, 19), 8.18 (1 H, d, J ) 0.9 Hz, 19).
Further elution yielded presumable 7-[(4-cyanophenyl)-
sulfanyl]-3-(trimethylsilyl)-1-benzothiophene-5-carbonitrile [MS
m/z 364 (M+, 14)], 4-[[2-cyano-2-(trimethylsilyl)ethenyl]sulfa-
nyl]benzonitrile (21) as a 1:1 Z/ E mixture (2.5:1 before
chromatography, 10% overall yield, ∆rt ) 32) [1H NMR (300
MHz) δ 0.28 (9 H, s, (Z)-21), 0.40 (9 H, s, (E)-21), 7.50 (2 H, A
part of AA′BB′, J ) 8.3 Hz), 7.68 (2 H, B part of AA′BB′, J )
8.3 Hz), 7.80 (1 H, s, (Z)-21), 7.87 (1 H, s, (E)-21) [peak
assignment was made on the basis of NMR analyses of
fractions with different Z/ E ratios]; MS m/z ((Z)-21) 258 (M+,
23); MS m/z ((E)-21) 258 (M+, 16); HRMS calcd for C13H14N2-
SSi 258.0647, found 258.0653], disulfide 1b (25%), and trace
amounts of the desilylation product 4-[[2-cyanoethenyl]sulfa-
nyl]benzonitrile as a 3:1 isomer mixture [MS m/z 186 (M+,
100); HRMS calcd for C10H6N2S 186.0252, found 186.0250].
Assuming that disulfide 1b arises from decomposition of (Z)-
18, the total yield of 18 is about 35% (25% and 10% for (Z)-18
and (E)18, respectively, see the Results and Discussion).
C
14H14N2S (E/ Z mixture) 242.0878, found 242.0883. Anal.
Calcd for C14H14N2S (E/ Z mixture): C, 69.39; H, 5.82; N, 11.56;
S, 13.23. Found: C, 69.62; H, 5.79; N, 11.60; S, 13.30. A
NOEDIF experiment was carried out on a fraction containing
pure (Z)-30 by irradiating the two allylic protons (2.37 ppm);
the experiment showed an exclusive positive effect (2%) on the
signal of the vinylic proton (6.90 ppm).
Final elution gave traces of the presumable4l,10 photo-Fries
rearrangement product, 4-[[2-cyano-1-hexenyl]sulfanyl]-3-[(4-
cyanophenyl)sulfanyl]benzonitrile: oil; MS m/z 375 (M+, 100);
HRMS calcd for C21H17N3S2 (E/Z mixture) 375.0864, found
375.0879.
Sem iem p ir ica l Ca lcu la tion s. Semiempirical calculations
on radicals (E)-32a ,b and (Z)-32a ,b were carried out with the
CS MOPAC routine included in the CambridgeSoft ChemOffice
Pro 4.0 package. After a careful conformational search, the
geometries of the open-shell intermediates were fully optimized
following the MNDO-d parametrization. Geometries and ener-
gies for radicals (E)-32a ,b and (Z)-32a ,b are shown in Schemes
12 and 13. The rotation barrier for each couple of intermediates
(Scheme 13) was calculated by determining the maximum of
the energy profile for rotation of the Câ-Cγ bond; this was
obtained by calculating single point energies for a series of
intermediates with decreasing CR-Câ-Cγ-S dihedral angles.
Ack n ow led gm en t. We gratefully acknowledge fi-
nancial support from CNR, MURST (1998-1999 Grant
for “Free Radicals and Radical Ions in Chemical and
Biological Processes”), and the University of Bologna
(1997-1999 Funds for Selected Research Topics). We
also thank Dr. Alessandro Mezzetti for the experimental
work.
Rea ction of 1b w ith 17 a n d 24. After 9 h, TLC and GC-
MS analyses of the reaction mixture revealed great amounts
of starting material (at least 70%) together with small quanti-
ties of 18 and 19. Nitriles 21 were absolutely absent.
Rea ction of 1b w ith 27 a n d 14. After 24 h, column
chromatography of the reaction mixture yielded a mixture of
28, 29, 4-(tert-butylsulfanyl)benzonitrile (39),47 and presum-
ably 4-[[2-phenyl-1-hexenyl]sulfanyl]benzonitrile (40) [the prod-
Su p p or tin g In for m a tion Ava ila ble: Full IR and MS data
for compounds 4b, 5b,c, 7, 8, 15a -c, 18, 19, 21, 28-30, and
40 and photo-Fries rearrangement products. This material is
available free of charge via the Internet at http://pubs.acs.org.
(45) O’Donovan, A. R. M.; Sheperd, M. K. Tetrahedron Lett. 1994,
35, 4425.
(46) A GC-MS analysis showed the presence of trace amounts of
another isomer with m/z ) 231, whose structure was not investigated
[MS m/z 231 (M+, 25), 216 (100), 200 (2), 186 (9), 176 (7), 140 (10), 114
(9), 108 (15)].
J O991871Y
(47) Dell’Erba, C.; Houman, A.; Morin, N.; Novi, M.; Petrillo, G.;
Pinson, J .; Rolando, C. J . Org. Chem. 1996, 61, 929.