Cascade radical reactions via α-(arylsulfanyl)imidoyl radicals: Competitive [4 + 2] and [4 + 1] radical annulations of alkynyl isothiocyanates with aryl radicals

Article abstract of DOI:10.1021/jo0267504

Aryl radicals react with 2-(2-phenylethynyl)phenyl isothiocyanate through a novel radical cascade reaction entailing formation of α-(arylsulfanyl)imidoyl radicals and affording a new class of compounds, i.e. thiochromeno[2,3-b]indoles. These derivatives are formed as mixtures of substituted analogues arising from competitive [4 + 2] and [4 + 1] radical annulations. The isomer ratio is strongly dependent on the aryl substituent and is correlated to its capability to delocalize spin density. The presence of a methylsulfanyl group in the ortho-position of the initial aryl radical results in complete regioselectivity and better yields, as the consequence of both strong spindelocalization effect, which promotes exclusive [4 + 1] annulation, and good radical leaving-group ability, which facilitates aromatization of the final cyclohexadienyl radical. Theoretical calculations support the hypothesis of competitive, independent [4 + 2] and [4 + 1] annulation pathways. They also suggest that rearrangement onto the sulfur atom of the [4 + 1] intermediate does not occur via a sulfuranyl radical but rather through either a transition state or a sulfur-centered (thioamidyl) radical; the latter is possibly the preferred route in the presence of an o-methylsulfanyl moiety that can act as a leaving group in the final ipso-cyclization process.

Full text of DOI:10.1021/jo0267504

Ca sca d e Ra d ica l Rea ction s via r-(Ar ylsu lfa n yl)im id oyl Ra d ica ls:
Com p etitive [4 + 2] a n d [4 + 1] Ra d ica l An n u la tion s of Alk yn yl
Isoth iocya n a tes w ith Ar yl Ra d ica ls§
Luisa Benati,† Gianluca Calestani,‡ Rino Leardini,† Matteo Minozzi,*,† Daniele Nanni,*,†
Piero Spagnolo,† Samantha Strazzari,† and Giuseppe Zanardi†
Dipartimento di Chimica Organica “A. Mangini”, Universita` di Bologna, Viale Risorgimento 4,
I-40136 Bologna, Italy, and Dipartimento di Chimica Generale ed Inorganica, Analitica e Chimica Fisica,
Universita` di Parma, Viale delle Scienze 78, I-43100 Parma, Italy
nanni@ms.fci.unibo.it
Received November 22, 2002
Aryl radicals react with 2-(2-phenylethynyl)phenyl isothiocyanate through a novel radical cascade
reaction entailing formation of R-(arylsulfanyl)imidoyl radicals and affording a new class of
compounds, i.e. thiochromeno[2,3-b]indoles. These derivatives are formed as mixtures of substituted
analogues arising from competitive [4 + 2] and [4 + 1] radical annulations. The isomer ratio is
strongly dependent on the aryl substituent and is correlated to its capability to delocalize spin
density. The presence of a methylsulfanyl group in the ortho-position of the initial aryl radical
results in complete regioselectivity and better yields, as the consequence of both strong spin-
delocalization effect, which promotes exclusive [4 + 1] annulation, and good radical leaving-group
ability, which facilitates aromatization of the final cyclohexadienyl radical. Theoretical calculations
support the hypothesis of competitive, independent [4 + 2] and [4 + 1] annulation pathways. They
also suggest that rearrangement onto the sulfur atom of the [4 + 1] intermediate does not occur
via a sulfuranyl radical but rather through either a transition state or a sulfur-centered (thioamidyl)
radical; the latter is possibly the preferred route in the presence of an o-methylsulfanyl moiety
that can act as a leaving group in the final ipso-cyclization process.
In tr od u ction
nitrogen-containing compounds, including phenanthri-
dines,2e quinolines,2d,f,g,4 indoles,5 benzotriazines,2h and
quinoxalines.4c,6
Within the great outbreak of kinetic, mechanistic, and
synthetic work carried out on radical reactions in the past
decades, our group has drawn a special attention to
imidoyl radicals. These species are very attractive inter-
mediates that can be readily produced by addition of
carbon- and heteroatom-centered radicals to isonitriles,1
by hydrogen atom abstraction from imines,2 and by
homolytic fragmentation of certain imidoylic precursors.3
As far as their synthetic potential is concerned, they have
been shown to perform smooth intra- and intermolecular
additions to double and triple carbon-carbon bonds, as
well as cyclizations onto aromatic rings, sulfur atoms, and
cyano groups. They have been therefore widely, efficiently
employed in cyclizations, annulations, and cascade reac-
tions leading to the construction of various heterocyclic
(2) (a) Ohta, H.; Tokumaru, K. J . Chem. Soc., Chem. Commun. 1970,
1601. (b) Danen, W. C.; West, C. T. J . Am. Chem. Soc. 1973, 95, 6872-
6874. (c) Davies, A. G.; Nedelec, J .-Y.; Sutcliffe, R. J . Chem. Soc., Perkin
Trans. 2 1983, 209-211. (d) Leardini, R.; Pedulli, G. F.; Tundo, A.;
Zanardi, G. J . Chem. Soc., Chem. Commun. 1984, 1320-1321. (e)
Leardini, R.; Tundo, A.; Zanardi, G.; Pedulli, G. F. Synthesis 1985,
107-110. (f) Leardini, R.; Nanni, D.; Pedulli, G. F.; Tundo, A.; Zanardi,
G. J . Chem. Soc., Perkin Trans. 1 1986, 1591-1594. (g) Leardini, R.;
Nanni, D.; Tundo, A.; Zanardi, G. Gazz. Chim. Ital. 1989, 119, 637-
641. (h) Leardini, R.; Nanni, D.; Tundo, A.; Zanardi, G. J . Chem. Soc.,
Chem. Commun. 1989, 757-758. (i) Leardini, R.; Nanni, D.; Santori,
M.; Zanardi, G. Tetrahedron 1992, 48, 3961-3970. (j) Guidotti, S.;
Leardini, R.; Nanni, D.; Pareschi, P.; Zanardi, G. Tetrahedron Lett.
1995, 36, 451-454. (k) Leardini, R.; McNab, H.; Nanni, D. Tetrahedron
1995, 51, 12143-12158. (l) Nanni, D.; Pareschi, P.; Tundo, A. Tetra-
hedron Lett. 1996, 37, 9337-9340.
(3) (a) Bachi, M. D.; Denenmark, D. J . Am. Chem. Soc. 1989, 111,
1886-1888. (b) Dan-oh, Y.; Matta, H.; Uemura, J .; Watanabe, H.;
Uneyama, K. Bull. Chem. Soc. J pn. 1995, 68, 1497-1507. (c) Fujiwara,
S.-i.; Matsuya, T.; Maeda, H.; Shin-ike, T.; Kambe, N.; Sonoda, N. J .
Org. Chem. 2001, 66, 2183-2185.
(4) (a) Curran, D. P.; Liu, H. J . Am. Chem. Soc. 1992, 114, 5863-
5864. (b) Curran, D. P.; J osien, H.; Ko, S.-B. Angew. Chem., Int. Ed.
Engl. 1995, 34, 2683-2684. (c) Curran, D. P.; Liu, H.; J osien, H.; Ko,
S.-B. Tetrahedron 1996, 52, 11385-11404. (d) Ryu, I.; Sonoda, N.;
Curran, D. P. Chem. Rev. 1996, 96, 177-194. (e) J osien, H.; Ko, S.-B.;
Bom, D.; Curran, D. P. Chem. Eur. J . 1998, 4, 67-83.
* Author to whom correspondence should be addressed. Fax: +39
051 2093654.
§ Dedicated to Prof. Antonio Tundo on the occasion of the 1st
anniversary of his death.
† Dipartimento di Chimica Organica “A. Mangini”, Bologna.
‡ Dipartimento di Chimica Generale ed Inorganica, Analitica
Chimica Fisica, Parma.
e
(1) (a) Leardini, R.; Nanni, D.; Zanardi, G. J . Org. Chem. 2000, 65,
2763-2772 and references cited therein. (b) For a comprehensive
review of radical isonitrile chemistry, see: Nanni, D. Isonitriles:
Useful Trap in Radical Chemistry. In Radicals in Organic Synthesis;
Renaud, P., Sibi, M. P., Eds.; Wiley-VCH: Weinheim, Germany, 2001;
Vol. 2, Chapter 1.3, pp 44-61.
(5) (a) Fukuyama, T.; Xiaoqi, C.; Peng, G. J . Am. Chem. Soc. 1994,
116, 3127-3128. (b) Kobayashi, Y.; Fukuyama, T. J . Heterocycl. Chem.
1998, 35, 1043-1055. (c) Tokuyama, H.; Yamashita, T.; Reding, M.
T.; Kaburagi, Y.; Fukuyama, T. J . Am. Chem. Soc. 1999, 121, 3791-
3792. (d) Rainier, J . D.; Kennedy, A. R.; Chase, E. Tetrahedron Lett.
1999, 40, 6325-6327.
a
10.1021/jo0267504 CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/04/2003
3454
J . Org. Chem. 2003, 68, 3454-3464

Annulations of Alkynyl Isothiocyanates with Aryl Radicals
SCHEME 1
A further way to generate imidoyl radicals has involved
the addition of carbon7 and, especially, tin or silicon8
radicals to the sulfur atom of isothiocyanates. Although
known for a few decades, the resulting R-sulfanyl-
substituted species have so far found only little use in
organic synthesis. Indeed, only since the late eighties
have they showed their synthetic appeal, first with the
conversion of glycosyl isothiocyanates into glycosyl iso-
nitriles and/or alditols,8d then with the Bachi synthesis
of thiolactams and thiopyroglutamates from alkenyl
isothiocyanates,8e-g and finally with the practicable alkyl-
radical-mediated isomerization of the isothiocyanate
moiety to the thiocyanate group reported by Barton in
1992.7 Until 1997, Bachi’s reaction has been the only
example concerning the synthesis of heterocyclic com-
pounds by radical addition to isothiocyanates, whereas
Barton’s work has been the only instance of addition of
carbon-centered radicals to the sulfur atom of isothio-
cyanates.
In 1997 we reported a novel radical cascade reaction
that involved carbon-centered radicals and aryl isothio-
cyanates and led to heterocycles containing both nitrogen
and sulfur.9 That reaction proceeded by 5-exo-dig cycliza-
tion onto the cyano group of R-[(2-cyanoaryl)sulfanyl]-
imidoyl radicals 3a , in turn obtained by addition of
2-cyanoaryl radicals 2a to aryl isothiocyanates 1; subse-
quent six-membered cyclization of the ensuing iminyl
radical 4a onto the aromatic ring of the original isothio-
cyanate allowed the one-pot synthesis of substituted
benzothienoquinoxalines 5a from commercially available
or easily accessible materials (Scheme 1, path a). That
paper was followed by the synthesis of benzothienoquino-
lines by a radical cascade reaction entailing, as in the
previous work,9 another rare [3 + 2] annulation10 initi-
ated by addition of 2-alkynylaryl radicals 2b to aryl
isothiocyanates.11 However, unlike before,9 this reaction
yielded two isomeric quinoline derivatives, viz. 5b and
6b, arising through competitive six- and five-membered
cyclizations of the final vinyl radical 4b onto the aromatic
ring of the isothiocyanate (Scheme 1, paths a and b).
The isomer ratio was found to depend on the nature
of the X-substituent, with the “rearranged” isomer 6b
becoming more and more favored by increasing the
electron-withdrawing capability of the substituent; this
has been explained in terms of a SOMO/LUMO-controlled
1,5-cyclization of the slightly nucleophilic vinyl radical
4b competing with a highly exothermic 1,6-cyclization
dominated by enthalpic factors.11
To further explore the synthetic usefulness of isothio-
cyanate radical reactions and to get some more informa-
tion on the stabilization and/or polar effects affecting the
tandem cyclizations of the intermediate R-sulfanyl-
substituted imidoyl radicals, we were next prompted to
study the reaction of aryl radicals with 2-alkynyl-
substituted aryl isothiocyanates. In particular, besides
the access to a totally new class of heterocyclic com-
pounds, we aimed at investigating the possible effects of
both the aryl radical substituents and the isothiocyanate
sulfur atom on the 1,5 and 1,6 competing cyclizations of
the intermediate vinyl radicals. Herein we therefore re-
port a new cascade radical reaction of 2-(phenylethynyl)-
phenyl isothiocyanate with some aryl radicals that allows
the one-pot synthesis of the polycondensed framework
of substituted thiochromeno[2,3-b]indoles.
Resu lts a n d Discu ssion
As before,9,11 aryl radicals 8 were generated from the
corresponding diazonium tetrafluoroborates 7 by reaction
with potassium acetate in ethyl acetate at room temper-
ature in the presence of [18-crown-6].12 When isothio-
cyanate 9 was allowed to react with 7, the 2-X- and 3-X-
substituted 11-phenylthiochromeno[2,3-b]indoles 10 and
11 were obtained in 40-50% overall yield (Scheme 2 and
Table 1).
(6) Nanni, D.; Pareschi, P.; Rizzoli, C.; Sgarabotto, P.; Tundo, A.
Tetrahedron 1995, 51, 9045-9062.
(7) Barton, D. H. R.; J aszberenyi, J . Cs.; Theodorakis, E. A.
Tetrahedron 1992, 48, 2613-2626.
(8) (a) J ohn, D. I.; Thomas, E. J .; Tyrrell, N. D. J . Chem. Soc., Chem.
Commun. 1979, 345-347. (b) Barton, D. H. R.; Bringmann, G.;
Lamotte, G.; Motherwell, W. B.; Hay Motherwell, R. S.; Porter, A. E.
A. J . Chem. Soc., Perkin Trans. 1 1980, 2657-2664. (c) J ohn, D. I.;
Tyrrell, N. D.; Thomas, E. J . J . Chem. Soc., Chem. Commun. 1981,
901-902. (d) Witczak, Z. J . Tetrahedron Lett. 1986, 27, 155-158. (e)
Bachi, M. D.; Denenmark, D. J . Org. Chem. 1990, 55, 3442-3444. (f)
Bachi, M. D.; Melman, A. J . Org. Chem. 1995, 60, 6242-6244. (g)
Bachi, M. D.; Bar-Ner, N.; Melman, A. J . Org. Chem. 1996, 61, 7116-
7124.
None of the compounds 10 and 11 were previously
reported in the literature and the general thiochromeno-
[2,3-b]indole framework is completely unprecedented as
well. The structures of 10 and 11 were confirmed by
X-ray crystallographic analysis of derivatives 10b,c and
11b-d , and then by spectral analogies with the other
congeners. In particular, a crucial support to structural
(9) Leardini, R.; Nanni, D.; Pareschi, P.; Tundo, T.; Zanardi, G. J .
Org. Chem. 1997, 62, 8394-8399.
1
assignment was brought by the peculiar H NMR signal
(10) For [3 + 2] radical annulations see: (a) Albertazzi, A.; Leardini,
R.; Pedulli, G. F.; Tundo, A.; Zanardi, G. J . Org. Chem., 1984, 49,
4482-4486. (b) Leardini, R.; Pedulli, G. F.; Tundo, A.; Zanardi, G. J .
Chem. Soc., Chem. Commun. 1985, 1390-1391. (c) Leardini, R.; Nanni,
D.; Tundo, A.; Zanardi, G. Tetrahedron Lett. 1998, 39, 2441-2442.
(11) Benati, L.; Leardini, R.; Minozzi, M.; Nanni, D.; Spagnolo, P.;
Zanardi, G. J . Org. Chem. 2000, 65, 8669-8674.
of the H-4 proton. Indeed, in almost all the examined
(12) Ru¨chardt, C.; Freudenberg, B. Tetrahedron Lett. 1964, 3623-
3628. See also: Rosenberg, D. E.; Beadle, J . R.; Korzeniowski, S. H.;
Gokel, G. W. Tetrahedron Lett. 1980, 21, 4141-4144.
J . Org. Chem, Vol. 68, No. 9, 2003 3455

Benati et al.
SCHEME 2
SCHEME 3
TABLE 1. Yield s of P r od u cts 10 a n d 11 in th e Rea ction s
of Dia zon iu m Sa lts 7 w ith Isoth iocya n a te 9
series
X
10 (%)
11 (%)
10/11
a
b
c
d
e
f
H
CF3
Cl
OMe
SMe
N3
50
27
27
20
8
23
23
20
32
40
43
1.17
1.17
1.00
0.25
0.10
0.05
SCHEME 4
4
2
g
C(O)Me
nil) yields (<5%) and presumably arose from competing
attack of the aryl radical 8 to the C-C triple bond of 9,
followed by cyclization of the ensuing vinyl radical onto
the nitrogen atom of the isothiocyanate moiety, loss of
CS, and eventual hydrogen atom abstraction (Scheme 4).
As can be seen in Table 1, the 10/11 yield ratio is
strongly dependent on the X-substituent. On the as-
sumption (suitably supported below) that indole 10 is
exclusively the result of the [4 + 2] annulation (Scheme
3, path a), whereas the isomer 11 totally arises from the
[4 + 1] route (Scheme 3, path b), the 1,5-cyclization of
vinyl radical 13 would become more and more favored
by moving rightward in Figure 2 from the trifluoromethyl
to the acetyl group. It is also evident that, unlike our
previous results,11 there is no correlation between the 10/
11 ratio and the electronic character of X, since two
substituents such as the trifluoromethyl and the acetyl
group, with very similar electron-withdrawing capabili-
ties, gave very different isomer ratios. It can therefore
be inferred that the sulfur atom linked to the aromatic
ring target of the cyclization exerts very different effects
with respect to the nitrogen atom of the related vinyl
radical 4b.
F IGURE 1. Structural assignment to 10 and 11 on the basis
of their 1H NMR spectra.
compounds, and mostly independently of the type of
substituent, H-4 appeared as a doublet at the downfield
limit of the spectrum: the magnitude of its coupling
constant with H-3 (ortho-coupling, compounds 10) or H-2
(meta-coupling, compounds 11) was therefore diagnostic
of the position of the X-substituent (Figure 1).
The reaction outcome can be accounted for through
initial addition of the aryl radical to the sulfur atom of
the isothiocyanate to give imidoyl radical 12; subsequent
cyclization of 12 onto the C-C triple bond leads to the
vinyl radical 13, which eventually undergoes two com-
petitive 1,5- and 1,6-cyclizations onto the aromatic ring
of the starting aryl radical (Scheme 3).
Ring strain could be a trivial explanation of the
considerable occurrence of rearranged products obtained
with radical 13. Indeed, with respect to the formation of
the pyrrole ring from radical 4b, in the case of 13 the
larger size of the sulfur atom could let the 1,5-cyclization
to the thiophene ring compete to a greater extent with
the 1,6-ring closure; this would in general favor the
formation of the rearranged isomers.13 However, this
possibility would not explain the observed dependence
of the 10/11 ratio on the X-substituent, thus we believe
On one hand, 1,6-ring closure leads directly to the
formation of the thiochromeno ring, whose sulfurated
part can thus be seen as arising from a [4 + 2] radical
annulation between radical 8 and isothiocyanate 9;
aromatization of cyclohexadienyl 14 eventually affords
compound 10. On the other hand, 1,5-cyclization produces
the spirocyclohexadienyl radical 15, whose thiophene ring
is hence the result of a [4 + 1] annulation between 8 and
9; ring expansion of radical 15 onto the sulfur atom
followed by aromatization yields the isomeric compound
11. The mechanism of such a ring expansion process will
be discussed below.
(13) Ring strain can probably be one of the reasons accounting for
the larger 1,6/1,5-cyclization ratios observed with radicals 4b with
respect to analogous vinyl radicals obtained by addition of R-phenyl-
N-arylimidoyl radicals to phenylacetylene (ref 2f).
The only identifiable reaction byproducts were indoles
17. These compounds were isolated in very low (or even
3456 J . Org. Chem., Vol. 68, No. 9, 2003

Annulations of Alkynyl Isothiocyanates with Aryl Radicals
F IGURE 3. Variation of the 10/11 ratio with the substituent
F IGURE 2. Effect of the substituents on the yields of 10 and
11 and the 10/11 ratio (the 10/11 ratio has been multiplied by
50 for better legibility).
radical effect (RE).15
that ring strain could only marginally influence the
reaction outcome.
Therefore, we turned our attention to the stabiliza-
tion role played by the substituent, believing that the
[4 + 2]/[4 + 1] competition should be mainly influenced
by the capability of the X-group to delocalize spin density
in radicals 15. Usually, in the absence of sizable steric
interactions, the description of the substituent effect on
carbon radical systems needs both polar and radical
stabilization parameters.14 To better understand the
nature of the substituent effect, a great deal of work has
been carried out in recent years trying to set up Ham-
mett-type radical spin delocalization σ• parameters and
to theoretically study the substituent effect on the C-H
BDE of suitable hydrocarbons. Very recently, Wu et al.
have reported a density functional study that attempts
to separate the polar contribution from the radical spin
delocalization effect.15 In that paper, the radical effect
(RE) on the BDE of the benzylic C-H bond of p-
substituted toluenes has been calculated and correlated
with spin density (to represent the spin delocalization
effect) and charge variations (to represent the polar
effect) at the benzylic radical center. Scattered points
were obtained by plotting RE against charge variations,
whereas good correlations were observed by plotting RE
against spin density and even when both spin density
and charge distribution were used as variables; this fact
denotes a very small contribution of the polar effect to
RE, which is therefore mainly the result of the spin
delocalization effect.
F IGURE 4. Correlation plot of the substituent radical effect
(RE)15 against the 10/11 ratio.
(RE): isomer 10 is strongly preferred over 11 with the
trifluoromethyl- and chloro-substituent, which have the
lowest RE values (0.26 and 0.28 kcal/mol, respectively),
whereas isomer 11 becomes dramatically favored with
the acetyl group, which is one of the best radical-
stabilizing (spin-density delocalizing) substituents (RE
) 1.64 kcal/mol). The RE value of the azido group is not
available, hence this substituent was not included in this
discussion; however, on the basis of the 10/11 ratio
obtained with X ) N3 (see Figure 2 and Table 1, entry
f), it can be assumed that its effect as a radical-stabilizing
substituent is quite strong, being comparable to that of
the acetyl group.
A good linear correlation (r2 ) 0.974) was observed by
plotting the 10/11 ratio against RE (Figure 4), thus
showing that the relationship between the isomer ratio
and RE is not simply established on a qualitative basis,
but it rather indicates a close interconnection between
the reaction outcome and the radical substituent effect.
Since the RE values are almost completely the result of
spin delocalization ability, we can reasonably assume
that, contrary to our previous results (Scheme 1),11 the
present reaction is only negligibly affected by polar
effects. Therefore, the [4 + 2]/[4 + 1] competing annu-
lations occur to an extent mostly dependent on the spin
delocalization degree of the resulting cyclohexadienyl
radicals.16
Since our intermediates 15 are basically cyclohexa-
dienyl radicals with the unpaired electron conjugated
with the X-group, hence essentially comparable to the
radical species studied by Wu et al., we thought that the
substituents could have played very similar roles. There-
fore, we tried to correlate the results of our reactions with
the literature theoretical data. At a first sight (Figure
3), one can see that the 10/11 ratio, which is a measure
of the relative occurrence of the [4 + 2] and [4 + 1]
pathways, is strongly dependent on Wu’s radical effect
(14) Viehe, H. G.; J anousek, Z.; Merenyi, R., Eds. Substituent Effects
in Radical Chemistry; Reidel: Dordrecht, The Netherlands, 1986.
(15) Wu, Y.-D.; Wong, C.-L.; Chan, K. W. K.; J i, G.-Z.; J iang, X.-K.
Remarkable results were obtained by inserting in the
ortho-position of the attacking aryl radical a substituent
that can function as both a radical-stabilizing moiety and
J . Org. Chem. 1996, 61, 746-750; see references therein for
a
comprehensive review of previous studies on the substituent effect on
carbon radical systems.
J . Org. Chem, Vol. 68, No. 9, 2003 3457

Benati et al.
SCHEME 5
relative positions of the methylsulfanyl and X groups.
This fact has an important consequence from a synthetic
point of view, since we can avoid lack of selectivity and
obtain either thiochromenoindole pure isomer simply by
using suitably o-sulfanyl-substituted aryl radicals. From
a mechanistic standpoint, the effect of the methylsulfanyl
group in pointing the reaction in a unique direction is
most likely the result of both its leaving group ability
and spin delocalization aptitude. Indeed, analogous reac-
tions carried out with phenyl radicals bearing an o-iodine
atom, which is another very good radical-leaving group,
gave, besides product 10a , minor amounts of a thio-
chromenoindole still containing the iodine atom (see
Experimental Section). The strong effect of the o-sulfanyl
group on both selectivity and yields could also suggest
some degree of reversibility in vinyl radical cyclizations,
at least when, in the absence of good leaving groups,
oxidation of cyclohexadienyl radicals may not be very
fast.
To verify the latter hypothesis and, in general, to get
some more insight into the overall reaction mechanism,
we carried out some theoretical calculations. Particularly,
we aimed at estimating the endo-/exothermicity and
activation barrier of the various steps, with the crucial
target to get some information about a possible re-
arrangement of the spirocyclohexadienyl radical onto the
C-C double bond and vis-a`-vis the actual mechanism of
its ring-expansion onto the sulfur atom.
As a matter of fact, spirocyclohexadienyl radical 15
could be in principle the unique reaction intermediate,
affording 16 by rearrangement onto the sulfur atom and
14 by ring-expansion onto the C-C double bond; the
latter route would occur via formation of intermediate
31 followed by fragmentation of the C-C intraannular
bond (Scheme 7, path a). This would also mean that, in
Scheme 3, path a should not operate, while route b should
be the only reaction pathway. Ring expansions of 5-exo
to 6-endo radicals, similar to the postulated conversion
of 15 into 14, are well documented in cyclizations of
â-multiply bonded alkyl or vinyl radicals.18 Unfortu-
nately, no experimental procedures can be envisaged to
establish, in Scheme 3, the possible occurrence of path a
instead of path (b + c).
As far as the conversion of 15 into 16 is concerned, at
least three distinct routes can be envisaged, viz. (i)
concerted ring enlargement with concomitant C-S bond-
making and bond-breaking through a transition state,
(ii) formation of the intermediate sulfuranyl radical 29
followed by scission of the C-S intraannular bond, and
(iii) fragmentation to the thioamidyl radical 30 followed
by recyclization (Scheme 7, paths b, c, and d, respec-
tively). Actually, SHi processes at sulfide moieties are
known to entail either transition states or sulfuranyl
radicals; the former were suggested to be involved with
alkyl sulfides,19 whereas the latter were assumed to be
true intermediates when electronegative atoms are linked
to the sulfur or intramolecular cyclizations onto diaryl-
sulfides occur.20 However, to the best of our knowledge,
no references to such strained intermediates as 29 have
been reported. On the other hand, the possibility that
a radical-leaving group. When 2-(methylsulfanyl)phenyl
radical 18 was allowed to react with isothiocyanate 9
under the usual conditions, compound 10a was obtained
in a yield significantly higher than that obtained in the
case of the unsubstituted radical 8a (70% instead of 50%)
(Scheme 5). No traces of any methylsulfanyl-substituted
thiochromenoindoles (i.e. 24 or 28, X ) H, Scheme 6)
were detected. At first sight, this result might be ex-
plained by assuming that vinyl radical 21 (X ) H) prefers
to undergo (reversible) 1,6-cyclization onto the aromatic
ring to give the cyclohexadienyl intermediates 22 and 23
(X ) H, Scheme 6, paths a and b).17 Due to the presence
of the good sulfanyl radical leaving group, under our non-
oxidative conditions the intermediate 22 (X ) H) might
be much proner to aromatize than the isomeric species
23 (X ) H), and thence able to afford product 10a in fairly
good yield.
However, the real mechanism is instead that depicted
in path d of Scheme 6 (framed structures), and this was
proved by carrying out the reaction with aryl radicals
19 and 20 (Scheme 5). With 19, the reaction did not afford
compound 10c, i.e. the expected product of direct ipso-
substitution of the sulfanyl moiety by vinyl radical 21
(X ) Cl), but it rather furnished the isomer 11c, ascrib-
able to spiro-cyclization of 21 to 25 (X ) Cl), followed by
ring expansion to cyclohexadienyl 27 (X ) Cl) and
eventual release of the sulfanyl group (Scheme 6).
Likewise, radical 20 exclusively afforded compound 10c
instead of 11c, with a mechanism wholly comparable to
that shown in Scheme 6 for radical 19.
We can conclude that, in the presence of the o-sulfide
moiety, the reaction proceeds by exclusive [4 + 1]
annulation via the spirocyclohexadienyl radical 25, fur-
nishing a single isomer whose structure depends on the
(16) Correlation of the RE values with the [14]/[15] ratio would be
more correct than that with the 10/11 yield ratio. However, since
products 10 and 11 are the only fate of intermediates 14 and 15,
respectively, we think that the 10/11 yield ratio should completely
mirror the [14]/[15] ratio, independent of the number of steps required
for the formation of the final products. We are considering the
possibility of performing ESR experiments to estimate the concentra-
tions of radicals 14 and 15.
(17) For a recent review of homolytic aromatic ipso-substitutions,
see: Studer, A.; Bossart, M. Homolytic Aromatic Substitutions. In
Radicals in Organic Synthesis; Renaud, P., Sibi, M. P., Eds.; Wiley-
VCH: Weinheim, Germany, 2001; Vol. 2, Chapter 1.4, pp 67-80.
(18) Capella, L.; Montevecchi, P. C.; Nanni, D. J . Org. Chem. 1994,
59, 3368-3374 and references therein.
(19) Ferris, K. F.; Franz, J . A.; Sosa, C.; Bartlett, R. J . J . Org. Chem.
1992, 57, 777-778.
3458 J . Org. Chem., Vol. 68, No. 9, 2003

Annulations of Alkynyl Isothiocyanates with Aryl Radicals
SCHEME 6
SCHEME 7
radical 15 could fragment to the fairly stable radical 30
cannot be in principle excluded. Indeed, ring openings
of sulfur-substituted spirocyclohexadienyl radicals have
been previously observed, although without concomitant
recyclization of the resulting sulfanyl radical onto the
aromatic ring.18 Homolytic aromatic substitution by
sulfur-centered radicals, although well-known (at least
with sulfanyl radicals), is in fact supposed to be a
reversible process,21 substitution products being formed
only in the presence of either good radical leaving groups,
e.g. halogen atoms22 or arylthio moieties,23 or in highly
oxidizing media able to oxidize the intermediate cyclo-
hexadienyl radicals.24 Very recently, the Mn(III)-medi-
ated oxidation of N,N′-diarylthioureas has been reported
to furnish, besides other major compounds, homolytic
substitution products of 30-like thioureidyl radicals,
whose yield increased in the presence of an o-methyl-
sulfanyl radical leaving group.25
First we examined the cyclization of vinyl radical 13.
DFT calculations found for 13 two conformational minima
separated by 4 kcal/mol; only the high-energy conformer
has the appropriate geometry for cyclization onto the
S-phenyl ring, therefore all the subsequent steps were
studied starting from this conformation (Figure 5 and
Figure S1, Supporting Information). Both the 1,5- ([4 +
1] annulation) and the 1,6-cyclization ([4 + 2] annulation)
were estimated to be exothermic (∆H ) -10.5 and -8.7
(20) Leardini, R.; Pedulli, G. F.; Tundo, A.; Zanardi, G. J . Chem.
Soc., Chem. Commun. 1985, 1390-1391 and references therein. See
also: Kampmeier, J . A.; Evans, T. R. J . Am. Chem. Soc. 1966, 88,
4096-4097. Beckwith, A. L. J .; Boate, D. R. J . Chem. Soc., Chem.
Commun. 1986, 189-190. Franz, J . A.; Roberts, D. H.; Ferris, K. F. J .
Org. Chem. 1987, 52, 2256-2262.
(21) Kice, J . L. Free Radicals; Kochi, J . K., Ed.; Wiley: New York,
1973; Vol. 2, pp 720-724.
(24) Gol’dfarb, Ya. A.; Pokhil, G. P.; Belen’kii, L. I. Dokl. Akad. Nauk
S.S.S.R. 1960, 167, 823-826; Chem. Abstr. 65, 12122. Benati, L.;
Camaggi, C. M.; Zanardi, G. J . Org. Chem. 1975, 40, 966-967.
(25) Calestani, G.; Capella, L.; Leardini, R.; Minozzi, M.; Nanni, D.;
Papa, R.; Zanardi, G. Tetrahedron 2001, 57, 7221-7233.
(22) Benati, L.; Camaggi, C. M.; Zanardi, G. J . Chem. Soc., Perkin
Trans. 1 1972, 2817-2819.
(23) Benati, L.; Montevecchi, P. C.; Tundo, A.; Zanardi, G. Gazz.
Chim. Ital. 1975, 105, 841-847.
J . Org. Chem, Vol. 68, No. 9, 2003 3459

Benati et al.
probably occur, especially in the absence of fast aroma-
tization pathways for 14, also due to the acceptable
activation barriers for the reverse reactions (13.8 and
16.4 kcal/mol for the ring-opening of 15 and 16, respec-
tively). This is consistent with the proposed stabilizing
substituent effect discussed above.
The evolution pathways of radical 15 are more uncer-
tain, due to the difficulties in optimizing both the
transition state (or intermediate 29) for ring-enlargement
(Scheme 7, path b or c) and the thioamidyl radical 30
(Scheme 7, path d). Route a of Scheme 7 was nevertheless
fully optimized; besides being of course altogether endo-
thermic by 1.8 kcal/mol, this path is characterized by
another quite strongly endothermic step, i.e. formation
of intermediate 31, which was located 14.6 kcal/mol above
15 (Figure 6 and Figure S2, Supporting Information). The
activation barrier for the conversion of 15 into 31 was
estimated at 16.6 kcal/mol, and the second barrier, that
for the transformation of 31 into 14, albeit quite low (3.3
kcal/mol), was however found to be higher than that of
the reconversion of 31 into 15 (2.0 kcal/mol).
F IGURE 5. Cyclization pathways of radical 13, with opti-
mized structures, ∆Hs, and activation barriers as obtained by
DFT calculations ([4 + 2] annulation, top path; [4 + 1]
annulation, bottom path).
To fully establish that this path is very unlikely we
needed to estimate the ∆H and activation barrier(s) for
the competitive one, i.e. the transformation of 15 into 16.
Although a full optimization of routes b, c, and d (Scheme
7) could not be obtained, nonetheless we think that
enough data were acquired at least to prove that 15
cannot rearrange into 14 and hence that the latter
radical, i.e. the [4 + 2] annulation intermediate, is the
result of a direct 1,6-cyclization of vinyl radical 13.
kcal/mol, respectively). The transition states for the two
cyclizations were optimized, giving activation barriers of
3.3 and 7.7 kcal/mol for the formation of 15 and 14,
respectively. The 1,5-cyclization seems therefore both
enthalpically and, above all, kinetically preferred over
the 1,6-ring closure. Nevertheless, the major distinction
between the activation barriers of 14 and 15 (∆Ea ) 4.4
kcal/mol), associated with the low enthalpy difference
(∆∆H ) 1.8 kcal/mol), suggests that the cyclization of 13
is not likely to be under complete kinetic control. Some
equilibration between the two cyclized species may
The rearrangement of 15 into 16 was calculated to be
exothermic by 9.5 kcal/mol (Figure 6 and Figure S2,
Supporting Information): this means that cyclohexa-
dienyl 16 is much more stable than its congener 14,
F IGURE 6. Rearrangement pathways of radical 15 into 14 and 16 according to routes a and b of Scheme 7, with optimized
structures, ∆Hs, and activation barriers as obtained by DFT calculations (* not fully optimized) (route a, top path; route b, bottom
path).
3460 J . Org. Chem., Vol. 68, No. 9, 2003

Annulations of Alkynyl Isothiocyanates with Aryl Radicals
energy and vibrational frequency calculation of the
structure found with PM3 methods for the maximum of
the 15-16 profile gave a structure with a single, physi-
cally consistent imaginary vibrational frequency that can
be assumed very close to the true transition state for the
process.
This structure entails an activation barrier of 10.0 kcal/
mol (Figure 6), much lower than that for the formation
of intermediate 31. Therefore, it can be reasonably
inferred that path a of Scheme 7 cannot efficiently
compete with path b, and hence spirocyclohexadienyl 15
is not on the reaction pathway leading to 14.
It remains to be established which path is actually
followed by spiro-radical 15 in its rearrangement into 16,
that is to distinguish between routes b, c, and d (Scheme
7). We already showed (Figures 6 and 7) that path b, i.e.
direct conversion through a transition state, is a viable
process with a reasonable activation barrier. The alter-
native formation of intermediate 29 (path c) should be
excluded. Indeed, both PM3 and DFT calculations did not
ever converge to any minima when trying to optimize
structure 29, affording instead either the starting radical
15 or the final cyclohexadienyl 16.
F IGURE 7. Energy profiles for the rearrangements of radical
15 into 31 and 16, with optimized structures, ∆Hs, and
activation barriers as obtained by PM3 calculations.
As for the possible formation of the sulfur-centered
radical 30, it could not be optimized by DFT calculations,
but its enthalpy and activation barriers for both 1,5-
(recyclization into 15) and 1,6-ring closure (cyclization
to 16) were estimated by PM3 methods (Figure 8 and
Figure S4, Supporting Information).
Although fragmentation of 15 into 30 is endothermic
by 4.4 kcal/mol, the activation energy for the scission is
quite low (8.2 kcal/mol) and the corresponding transition
state lies very close in energy to the transition state for
the direct conversion of 15 into 16 (∆Ea ) 1.6 kcal/mol).
Moreover, the barrier for the subsequent step, i.e. cy-
clization of 30 into 16, is very low (2.9 kcal/mol), and
however lower than the activation barrier for the reverse
probably as a result of direct conjugation of the unpaired
electron with the whole π-system of the chromenoindole
structure (in 14, substantial stabilization is probably
provided by the sulfur atom alone). The transition state
for the 15-16 conversion could not be fully optmized by
DFT methods; therefore, we tried to compare paths a and
b (Scheme 7) by calculating an energy profile, on a PM3
semiempirical basis, for the two competitive routes
(Figure 7 and Figure S3, Supporting Information). The
profiles clearly suggest that both enthalpic and kinetic
factors strongly favor ring expansion of 15 into 16 (Figure
7, right profile) rather than formation of the strained
intermediate 31 (Figure 7, left profile). DFT single-point
F IGURE 8. Rearrangements of radical 15 into 16 (routes b and d of Scheme 7), with optimized structures, ∆Hs, and activation
barriers as obtained by PM3 calculations (path b, bottom path; path d, top path).
J . Org. Chem, Vol. 68, No. 9, 2003 3461

Benati et al.
leum (40-70 °C) and light petroleum/diethyl ether/dichloro-
methane mixtures (up to 100% dichloromethane).
reaction (3.8 kcal/mol). These data do not allow an
unambiguous determination of the real rearrangement
pathway, but rather suggest that paths b and d of
Scheme 7 can be competitive, and both of them are
however clearly favored with respect to path a. On the
basis of the previous evidence that thioamidyl radicals
can actually perform homolytic aromatic substitution,25
it also might be concluded that path d may actually be
the main rearrangement route, at least when a proper
radical leaving group is present on the aromatic ring of
the original aryl radical, as in the case of the methyl-
sulfanyl-substituted radicals reported in Schemes 5 and
6.
Star tin g Mater ials. Aniline and 4-(trifluoromethyl)-, 4-chlo-
ro-, 4-methoxy-, 4-(methylsulfanyl)-, 4-acetyl-, 2-(methyl-
sulfanyl)-, and 2-iodoaniline were commercially available;
4-azidoaniline,26 4-chloro-2-(methylsulfanyl)aniline,27 5-chloro-
2-(methylsulfanyl)aniline,28 and 2-(2-phenylethynyl)aniline29
were synthesized according to the literature. The correspond-
ing diazonium tetrafluoroborates 7a -g were prepared follow-
ing usual procedures.30
2-(2-P h en ylet h yn yl)p h en yl isot h iocya n a t e (9). A di-
chloromethane (10 mL) solution of 2-(2-phenylethynyl)aniline
(7.64 g, 40 mmol) was added dropwise at 20 °C to a stirred
solution of thiophosgene (5.70 g, 50 mmol) in a dichloro-
methane (6 mL)/water (14 mL) mixture. After 3 h, the organic
layer was separated and dried (sodium sulfate). The solvent
was evaporated and the residue was crystallized from ligroin
(bp 80-95 °C) to give the title compound (60%), mp 50-51
°C: IR νmax 2217 (CtC), 2071 (NCS) cm-1; 1H NMR (300 MHz)
δ 7.10-7.15 (1 H, m), 7.18-7.29 (2 H, m), 7.33-7.39 (3 H, m),
7.49-7.54 (1 H, m), 7.62-7.69 (2 H, m); 13C NMR (75 MHz) δ
85.61 (C), 97.59 (C), 123.14 (C), 123.47 (C), 125.28, 127.57,
129.03, 129.55, 129.58, 132.33, 133.14 (C), 133.22, 138.62 (C,
NdCdS); MS m/z (rel intensity) 235 (M+, 100), 203 (6), 190
(15), 176 (6). Anal. Calcd for C15H9NS: C, 76.56; H, 3.86; N,
5.95. Found: C, 76.30; H, 3.87; N, 5.97.
Gen er a l P r oced u r e for th e Rea ction s of Isoth iocya n -
a te 9 w ith th e Tetr a flu or obor a tes 7.9,11,12 [18-Crown-6]
(0.25 mmol), potassium acetate (10 mmol), and the tetra-
fluoroborate (5 mmol) were added to a solution of isothio-
cyanate (15 mmol) in 20 mL of ethyl acetate. The reaction
mixture was kept at room temperature for ca. 1 h under
vigorous stirring, the solvent was then evaporated, and the
residue was chromatographed (silica gel, 3.5 cm φ × 40 cm h
column, unless otherwise stated) eluting with petroleum ether
to remove excess isothiocyanate and to separate indoles 17.
Subsequent elution with light petroleum/diethyl ether/di-
chloromethane mixtures furnished thiochromenoindoles 10
and 11. Yields of 10 and 11, as obtained by this method, are
shown in Table 1. The following reactions were carried out
according to this general procedure.
Rea ction of 9 w ith 1-P h en yld ia zon iu m Tetr a flu or o-
bor a te (7a ). Elution with light petroleum gave 2,3-diphenyl-
1H-indole (17a ) (5%), mp 113-116 °C (lit.31 mp 114-116 °C)
[IR νmax 3463 (NH), 1601, 1494, 1457 cm-1; 1H NMR (300 MHz)
δ 7.25-7.60 (12 H, m), 7.68 (1 H, br d, J ) 7.8 Hz), 7.76 (1 H,
br d, J ) 7.6 Hz), 8.36 (1 H, br s); MS m/z (rel intensity) 269
(M+, 100), 268 (17), 267 (25), 165 (26), 163 (11), 77 (20)]; further
elution with light petroleum/diethyl ether/dichloromethane 60:
20:20 v/v/v afforded 11-phenylthiochromeno[2,3-b]indole (10a )
(50%) as a red solid, mp 214-215 °C (from light petroleum/
benzene) [1H NMR (200 MHz) δ 6.40 (1 H, d, J ) 8.0 Hz), 6.93
(1 H, br t, J ) 8.0 Hz), 7.30-7.72 (10 H, m), 7.82 (1 H, d, J )
8.0 Hz); 13C NMR (50 MHz) δ 118.98, 122.74, 124.07, 126.54,
126.78 (C), 127.36, 128.91, 129.60, 129.68, 129.86, 130.06,
131.61, 133.93 (C), 136.86 (C), 146.53 (C), 155.69 (C), 162.09
(C) (2 quaternary signals not visible); MS m/z (rel intensity)
311 (M+, 100), 310 (24). Anal. Calcd for C21H13NS: C, 81.00;
H, 4.21; N, 4.50. Found: C, 81.30; H, 4.22; N, 4.49].
Con clu sion s
Aryl radicals 8 react with isothiocyanate 9 through a
novel radical cascade reaction affording a new class of
2-X- and 3-X-substituted thiochromeno[2,3-b]indoles (10
and 11). These isomeric compounds arise from competi-
tive [4 + 2] and [4 + 1] radical annulations; the [4 + 2]
pathway yields directly compound 10, whereas the [4 +
1] ring closure is followed by ring-enlargement onto the
sulfur atom to yield isomer 11.
The 10/11 ratio is strongly dependent on the X-
substituent. Contrary to analogous ring closures of vinyl
radicals onto nitrogen-substituted aromatic rings,11 the
ratio is not correlated to the electron-withdrawing (or
releasing) ability of the substituent, but it is instead
associated with its capability to delocalize spin density.
The presence of a methylsulfanyl group in the ortho-
position of the initial aryl radical results in complete
regioselectivity and better yields, as the consequence of
both strong spin-delocalization effect, which promotes
exclusive [4 + 1] annulation, and good radical leaving-
group ability, which facilitates aromatization of the final
cyclohexadienyl radical. It is therefore possible to obtain
a single thiochromenoindole derivative, bearing the X-
substituent in a specific position, starting from a properly
substituted aryl radical.
Semiempirical and DFT theoretical calculations sug-
gest that spirocyclohexadienyl radicals 15 do not re-
arrange by ring closure onto the C-C double bond but
by expansion onto the sulfur atom exclusively, and thence
isomeric indoles 10 and 11 are formed by two distinct,
competitive annulation routes. As far as the expansion
onto sulfur is concerned, formation of a sulfuranyl radical
seems very unlikely. The rearrangement through a
transition state and that via a sulfur-centered (thio-
amidyl) radical appear instead to be competitive; the
latter is probably favored in the presence of an o-
methylsulfanyl moiety that can act as a leaving group
in the final ipso-cyclization process.
R ea ct ion of 9 w it h 1-[4-(Tr iflu or om et h yl)p h en yl]d i-
a zon iu m Tetr a flu or obor a te (7b). Elution with light petro-
(26) Dyall, L. K.; L’abbe´, G.; Dehaen, W. J . Chem. Soc., Perkin Trans.
2 1997, 971-975.
Exp er im en ta l Section
(27) Ranken, P. F.; McKinnie, B. G. J . Org. Chem. 1989, 54, 2985-
Gen er a l P r oced u r es. 1H and 13C NMR spectra were
recorded in deuteriochloroform, unless otherwise stated, using
tetramethylsilane as an internal standard. Mass spectra (MS)
were performed by electron impact with a beam energy of 70
eV: relative intensities are given in parentheses. IR spectra
were recorded in CHCl3 solutions. Column chromatography
was carried out on silica gel (63-200, 60 Å) or basic aluminum
oxide (activity grade III) by gradual elution with light petro-
2988.
(28) Hodson, S. J .; Bishop, M. J .; Speake, J . D.; Navas, F., III;
Garrison, D. T.; Bigham, E. C.; Saussy, D. L., J r.; Liacos, J . A.; Irving,
P. E.; Gobel, M. J .; Sherman, B. W. J . Med. Chem. 2002, 45, 2229-
2239.
(29) Castro, C. E.; Gaughan, E. J .; Owsley, D. C. J . Org. Chem. 1966,
31, 4071-4078.
(30) Roe, A. Org. React. 1949, 5, 193-228.
(31) Nakamura, Y.; Ukita, T. Org. Lett. 2002, 4, 2317-2320.
3462 J . Org. Chem., Vol. 68, No. 9, 2003

Annulations of Alkynyl Isothiocyanates with Aryl Radicals
leum gave 2-phenyl-3-[4-(trifluoromethyl)phenyl]-1H-indole
C21H12ClNS: C, 72.93; H, 3.50; N, 4.05. Found: C, 73.15; H,
3.50; N, 4.06. The structure of 10c was confirmed by X-ray
diffraction (see Supporting Information)].
(17b) (3%) as an oil [IR νmax 3461 (NH), 1616, 1449, 1325 cm-1
;
1H NMR (300 MHz) δ 7.10-7.80 (13 H, m), 8.38 (1 H, br s);
13C NMR (75 MHz) δ 111.80, 119.94, 121.51, 123.68, 126.11
(q, J ) 3.6 Hz), 128.82, 129.03, 129.56, 130.84 (due to the small
amounts of this compound, only the CH signals are quoted);
MS m/z (rel intensity) 337 (M+, 100), 318 (6), 267 (19); HRMS
calcd for C21H14F3N 337.1078, found 337.1080]; further elution
with light petroleum/diethyl ether/dichloromethane 70:10:20
v/v/v yielded 11-phenyl-3-(trifluoromethyl)thiochromeno[2,3-
b]indole (11b) (23%) as a red solid, mp 250-251 °C (from
ethanol/benzene) [1H NMR (200 MHz) δ 6.42 (1 H, br d, J )
8.0 Hz), 6.98 (1 H, td, J t ) 8.0 Hz, J d ) 1.0 Hz), 7.39-7.75 (9
H, m), 8.14 (1 H, br s); 13C NMR (50 MHz) δ 119.10, 122.73 (q,
J ) 4.0 Hz), 123.16, 123.57 (C, q, J ) 280 Hz), 124.30 (q, J )
4.0 Hz), 124.41, 126.32 (C), 128.81, 129.90, 130.24, 130.38,
131.33 (C), 131.39 (C, q, J ) 36 Hz), 131.96, 133.97 (C), 136.11
(C), 144.92 (C), 155.70 (C), 161.15 (C) (2 quaternary signals
overlapped); MS m/z (rel intensity) 379 (M+, 100), 378 (25).
Anal. Calcd for C22H12F3NS: C, 69.65; H, 3.19; N, 3.69.
Found: C, 69.47; H, 3.19; N, 3.70. The structure of 11b was
confirmed by X-ray diffraction (see Supporting Information)],
and some impure 11-phenyl-2-(trifluoromethyl)thiochromeno-
[2,3-b]indole (10b), which was purified by chromatography on
aluminum oxide (1.5 cm φ × 20 cm h column) eluting with
light petroleum/diethyl ether/dichloromethane 80:10:10 v/v/v,
yield 27%, was also obtained as a red solid, mp 197-198 °C
(from ethanol/benzene) [1H NMR (200 MHz) δ 6.42 (1 H, br d,
J ) 7.7 Hz), 6.99 (1 H, td, J t ) 7.7 Hz, J d ) 1.0 Hz), 7.39-7.46
(2 H, m), 7.49 (1 H, dd, J 1 ) 7.8 Hz, J 2 ) 1.2 Hz), 7.67-7.83
(6 H, m), 7.99 (1 H, br d, J ) 8.5 Hz); 13C NMR (50 MHz) δ
119.21, 123.25, 124.20 (C, q, J ) 275 Hz), 124.32, 125.63 (q, J
) 4.0 Hz), 126.05 (C), 127.90, 128.02 (C), 128.70, 128.78 (q, J
) 4.0 Hz), 129.90, 130.08, 130.32 (C), 135.73 (C), 137.41 (C),
145.32 (C), 155.64 (C), 161.27 (C) (2 CH signals overlapped,
quaternary C-2 signal not visible); MS m/z (rel intensity) 379
(M+, 100), 378 (18). Anal. Calcd for C22H12F3NS: C, 69.65; H,
3.19; N, 3.69. Found: C, 69.45; H, 3.19; N, 3.70. The structure
of 10b was confirmed by X-ray diffraction (see Supporting
Information)].
Rea ction of 9 w ith 1-(4-Meth oxyp h en yl)d ia zon iu m
Tetr a flu or obor a te (7d ). Elution with light petroleum gave
3-[4-methoxyphenyl]-2-phenyl-1H-indole (17d)33 (trace amounts)
1
as an oil [IR νmax 3466 (NH), 1602, 1513 cm-1; H NMR (300
MHz) δ 3.80 (3 H, s), 6.85-7.65 (13 H, m), 8.20 (1 H, br s); MS
m/z (rel intensity) 299 (M+, 100), 284 (36); HRMS calcd for
C21H17NO 299.1310, found 299.1317]; further elution with light
petroleum/diethyl ether/dichloromethane 40:30:30 v/v/v yielded
2-methoxy-11-phenylthiochromeno[2,3-b]indole (10d ) (20%) as
a red solid, mp 156-157 °C (from ethanol/benzene) [1H NMR
(200 MHz) δ 3.70 (3 H, s), 6.40 (1 H, br d, J ) 7.8 Hz), 6.93 (1
H, td, J t ) 8.0 Hz, J d ) 1.0 Hz), 7.15 (1 H, d, J ) 2.6 Hz), 7.22
(1 H, dd, J 1 ) 2.6 Hz, J 2 ) 8.3 Hz), 7.37-7.47 (3 H, m), 7.62-
7.72 (4 H, m), 7.79 (1 H, d, J ) 8.3 Hz); 13C NMR (50 MHz) δ
56.09, 115.06, 118.36, 118.96, 122.64, 124.22, 125.69 (C),
126.76 (C), 128.39, 128.95, 129.75, 129.84, 130.22, 137.01 (C),
146.27 (C), 155.94 (C), 158.43 (C), 162.68 (C) (2 quaternary
signals not visible); MS m/z (rel intensity) 341 (M+, 100), 326
(20), 298 (13), 297 (14). Anal. Calcd for C22H15NOS: C, 77.39;
H, 4.43; N, 4.10. Found: C, 77.51; H, 4.45; N, 4.09], and
3-methoxy-11-phenylthiochromeno[2,3-b]indole (11d ) (20%) as
a yellow-orange solid, mp 234-236 °C (from ethanol/benzene)
[1H NMR (200 MHz) δ 3.93 (3 H, s), 6.38 (1 H, br d, J ) 7.8
Hz), 6.88-6.99 (2 H, m), 7.32-7.51 (5 H, m), 7.61-7.71 (4 H,
m); 13C NMR (50 MHz) δ 56.35, 110.14, 115.28, 118.86, 122.48,
123.41, 126.99 (C), 128.82, 128.87, 129.49, 129.96, 133.09,
136.63 (C), 137.10 (C), 146.97 (C), 155.23 (C), 160.84 (C),
161.82 (C) (2 quaternary signals not visible); MS m/z (rel
intensity) 341 (M+, 100), 326 (3), 298 (23), 297 (17). Anal. Calcd
for C22H15NOS: C, 77.39; H, 4.43; N, 4.10. Found: C, 77.53;
H, 4.44; N, 4.10. The structure of 11d was confirmed by X-ray
diffraction (see Supporting Information)].
R ea ct ion of 9 w it h 1-[4-(Met h ylsu lfa n yl)p h en yl]d i-
a zon iu m Tetr a flu or obor a te (7e). Elution with light petro-
leum/diethyl ether/dichloromethane 50:25:25 v/v/v yielded a
mixture of 2-(methylsulfanyl)-11-phenylthiochromeno[2,3-b]-
indole (10e) and 3-(methylsulfanyl)-11-phenylthiochromeno-
Reaction of 9 with 1-(4-Ch lor oph en yl)diazon iu m Tetr a-
flu or obor a te (7c). Elution with light petroleum gave 3-[4-
chlorophenyl]-2-phenyl-1H-indole (17c)32 (5%) as an oil [IR νmax
3461 (NH), 1599, 1486 cm-1; 1H NMR (200 MHz) δ 7.04-7.40
(12 H, m), 7.57 (1 H, bd, J ) 7.4 Hz), 8.20 (1 H, br s); MS m/z
(rel intensity) 305 (M+ + 2, 29), 303 (M+, 100), 267 (28). HRMS
calcd for C20H14ClN 303.0815, found 303.0819]; further elution
with light petroleum/diethyl ether/dichloromethane 70:15:15
v/v/v yielded 3-chloro-11-phenylthiochromeno[2,3-b]indole (11c)
(23%) as a red solid, mp 244-245 °C (from ethanol/benzene)
[1H NMR (200 MHz) δ 6.41 (1 H, d, J ) 8.0 Hz), 6.97 (1 H, td,
J t ) 8.0 Hz, J d ) 1.2 Hz), 7.31-7.53 (5 H, m), 7.64-7.74 (4 H,
m), 7.88 (1 H, d, J ) 1.8 Hz); 13C NMR (50 MHz) δ 118.53,
122.41, 123.46, 126.09, 126.45, 126.82 (C), 128.25, 128.91 (C),
129.19, 129.28, 129.57, 132.00, 134.73 (C), 135.68 (C), 135.83
(C), 155.03 (C), 160.90 (C) (2 quaternary signals not visible);
MS m/z (rel intensity) 347 (M+ + 2, 40), 345 (M+, 100), 344
(18), 309 (14). Anal. Calcd for C21H12ClNS: C, 72.93; H, 3.50;
N, 4.05. Found: C, 73.12; H, 3.49; N, 4.07. The structure of
11c was confirmed by X-ray diffraction (see Supporting
Information)], and 2-chloro-11-phenylthiochromeno[2,3-b]-
indole (10c) (27%) as a red solid, mp 160-162 °C (from ethanol/
benzene) [1H NMR (200 MHz) δ 6.39 (1 H, d, J ) 7.8 Hz), 6.97
(1 H, td, J t ) 7.8 Hz, J d ) 1.0 Hz), 7.36-7.50 (3 H, m), 7.51-
7.58 (2 H, m), 7.66-7.75 (4 H, m), 7.81 (1 H, br d, J ) 7.8 Hz);
13C NMR (50 MHz) δ 118.56, 122.43, 123.72, 125.95 (C), 127.90,
128.27, 129.36, 129.44, 129.59, 129.71, 130.12, 131.41 (C),
131.98 (C), 135.56 (C), 144.50 (C), 155.27 (C), 161.14 (C) (2
quaternary signals not visible); MS m/z (rel intensity) 347 (M+
+ 2, 37), 345 (M+, 100), 344 (12), 309 (8). Anal. Calcd for
1
[2,3-b]indole (11e) (40% overall yield, 1:4 10e/11e ratio by H
NMR analysis). This mixture could not be separated by
changing either the eluant or the stationary phase; however,
by washing the solid mixture with diethyl ether several times,
we obtained pure solid 11e together with a solution containing
10e and 11e in a 2:1 ratio. 11e: red solid, mp 227-230 °C
(from benzene/ligroin); 1H NMR (200 MHz) δ 2.55 (3 H, s), 6.38
(1 H, br d, J ) 7.7 Hz), 6.92 (1 H, br t, J ) 7.7 Hz), 7.18 (1 H,
dd, J 1 ) 2.0 Hz, J 2 ) 8.6 Hz), 7.33-7.46 (4 H, m), 7.57 (1 H,
d, J ) 2.0 Hz), 7.59-7.70 (4H, m); 13C NMR (50 MHz) δ 15.55,
119.01, 122.32, 122.72, 123.82, 124.25, 125.67 (C), 126.91 (C),
127.61 (C), 128.94, 129.36, 129.64, 130.08, 131.40, 135.19 (C),
136.87 (C), 143.24 (C), 146.60 (C), 155.58 (C), 161.53 (C); MS
m/z (rel intensity) 357 (M+, 100), 309 (20). Anal. Calcd for
C
22H15NS2: C, 73.91; H, 4.23; N, 3.92. Found: C, 74.12; H,
4.25; N, 3.92. 10e: 1H NMR (200 MHz, obtained by subtracting
the signals of 11e) δ 2.37 (3 H, s), 6.40 (1 H, br d, J ) 7.7 Hz),
6.95 (1 H, br t, J ) 7.7 Hz), 7.18-7.23 (1 H, m), 7.34-7.51 (5
H, m), 7.60-7.73 (3 H, m), 7.78 (1 H, d, J ) 8.4 Hz). Anal.
Calcd for C22H15NS2 (10e/11e mixture): C, 73.91; H, 4.23; N,
3.92. Found: C, 74.18; H, 4.26; N, 3.91.
Rea ction of 9 w ith 1-(4-Azid op h en yl)d ia zon iu m Tetr a -
flu or obor a te (7f). Elution with light petroleum/diethyl ether/
dichloromethane 34:33:33 v/v/v yielded 3-azido-11-phenyl-
thiochromeno[2,3-b]indole (11f) (40%) as a red solid that was
recrystallized from dichloromethane/diethyl ether and melted
between 210 and 250 °C with slow decomposition [IR νmax 2118
1
(N3) cm-1; H NMR (CD2Cl2, 300 MHz) δ 6.35 (1 H, br d, J )
7.7 Hz), 6.89 (1 H, td, J t ) 7.7 Hz, J d ) 1.0 Hz), 7.03 (1 H, dd,
(32) Arcadi, A.; Cacchi, S.; Marinelli, F. Tetrahedron Lett. 1992, 33,
3915-3918. Neither mp nor spectral data are reported.
(33) Mentzer, C.; Molho, D.; Berguer, Y. Bull. Chem. Soc. Fr. 1950,
555-561.
J . Org. Chem, Vol. 68, No. 9, 2003 3463

Benati et al.
J 1 ) 8.8 Hz, J 2 ) 2.4 Hz), 7.33-7.40 (3 H, m), 7.48-7.54 (2 H,
m), 7.59 (1 H, br d, J ) 7.9 Hz), 7.62-7.68 (3 H, m); MS m/z
(rel intensity) 352 (M+, 5), 324 (M+ - 28, 100)] and 2-azido-
11-phenylthiochromeno[2,3-b]indole (10f) (4%) as a red solid
that was recrystallized from light petroleum/diethyl ether and
melted between 160 and 190 °C with slow decomposition [IR
Rea ction of 9 w ith 1-[5-Ch lor o-2-(m eth ylsu lfa n yl)-
p h en yl]d ia zon iu m Tetr a flu or obor a te. Elution with light
petroleum/diethyl ether/dichloromethane 70:15:15 v/v/v yielded
10c exclusively (80%), with mp and spectral data identical to
those reported for 10c in the above reaction of 9 with 7c.
Rea ction of 9 w ith 1-(2-Iod op h en yl)d ia zon iu m Tetr a -
flu or obor a te. Elution with light petroleum/diethyl ether/
dichloromethane 70:15:15 v/v/v afforded 10a (30%), with mp
and spectral data identical to those reported for 10a in the
above reaction of 9 with 7a, and 4-iodo-11-phenylthiochromeno-
[2,3-b]indole (8%) as a red solid, mp 233-234 °C (from light
petroleum/benzene) [1H NMR (200 MHz) δ 6.37 (1 H, br d, J
) 8.1 Hz), 6.97 (1 H, td, J t ) 8.1 Hz, J d ) 1.2 Hz), 7.14 (1 H,
dd, J 1 ) 8.1 Hz, J 2 ) 7.6 Hz), 7.36-7.49 (3 H, m), 7.58-7.75
(5 H, m), 8.09 (1 H, dd, J 1 ) 7.6 Hz, J 2 ) 1.3 Hz); 13C NMR
(50 MHz) δ 96.50 (C), 118.65, 122.44, 123.53, 125.62 (C),
126.68, 128.33, 129.10, 129.40, 129.50, 130.79 (C), 131.26,
135.88 (C), 137.19 (C), 140.34, 145.38 (C), 155.28 (C), 163.31
(C) (1 quaternary signal not visible); MS m/z (rel intensity)
437 (M+, 100), 310 (20); HRMS calcd for C21H12INS 436.9735,
found 436.9740. The structure was confirmed by X-ray dif-
fraction (data not reported in the Supporting Information)].
ν
max 2120 (N3) cm-1; 1H NMR (200 MHz) δ 6.42 (1 H, br d, J )
8.0 Hz), 6.95 (1 H, td, J t ) 8.0 Hz, J d ) 1.0 Hz), 7.15 (1 H, d,
J ) 2.2 Hz), 7.26-7.50 (5 H, m), 7.64-7.72 (3 H, m), 7.85 (1
H, d, J ) 8.8 Hz); MS m/z (rel intensity) 352 (M+, 3), 324 (M+
- 28, 100)].
Rea ction of 9 w ith 1-(4-Acetylp h en yl)d ia zon iu m Tet-
r a flu or obor a te (7g). Elution with light petroleum/diethyl
ether/dichloromethane 34:33:33 v/v/v yielded 1-(11-phenyl-
thiochromeno[2,3-b]indol-3-yl)-1-ethanone (11g) (43%) as a red
solid, mp 239-241 °C (from ethanol/benzene) [IR νmax 1686
(CO) cm-1; 1H NMR (200 MHz) δ 2.62 (3 H, s), 6.37 (1 H, br d,
J ) 7.7 Hz), 6.90 (1 H, td, J t ) 7.7 Hz, J d ) 1.0 Hz), 7.32-
7.42 (3 H, m), 7.56-7.69 (5 H, m), 7.84 (1 H, dd, J 1 ) 2.0 Hz,
J 2 ) 8.7 Hz), 8.38 (1 H, d, J ) 2.0 Hz); 13C NMR (50 MHz) δ
27.52, 119.29, 123.26, 124.60, 125.63, 126.60 (C), 127.68,
128.97, 129.99, 130.36, 130.57, 131.81 (C), 131.89, 132.20 (C),
133.98 (C), 136.47 (C), 136.99 (C), 145.37 (C), 155.94 (C),
162.14 (C), 197.22 (CO); MS m/z (rel intensity) 353 (M+, 100),
310 (65), 43 (99). Anal. Calcd for C23H15NOS: C, 78.16; H, 4.28;
N, 3.96. Found: C, 78.31; H, 4.30; N, 3.97], and 1-(11-
phenylthiochromeno[2,3-b]indol-2-yl)-1-ethanone (10g) (2%) as
a red-orange solid, mp 201-203 °C (from ethanol) [IR νmax 1685
(CO) cm-1; 1H NMR (200 MHz) δ 2.71 (3 H, s), 6.33 (1 H, br d,
J ) 7.8 Hz), 6.89 (1 H, td, J t ) 7.8 Hz, J d ) 1.1 Hz), 7.32-
7.68 (8 H, m), 7.85 (1 H, br d, J ) 7.8 Hz), 8.22 (1 H, br d, J
) 8.8 Hz); MS m/z (rel intensity) 353 (M+, 100), 310 (24), 43
(66); HRMS calcd for C23H15NOS 353.0874, found 353.0877].
R ea ct ion of 9 w it h 1-[2-(Met h ylsu lfa n yl)p h en yl]d i-
a zon iu m Tetr a flu or obor a te. Elution with light petroleum/
diethyl ether/dichloromethane 60:20:20 v/v/v yielded 10a
(70%), with mp and spectral data identical to those reported
for 10a in the above reaction of 9 with 7a .
Ack n ow led gm en t. The authors gratefully acknowl-
edge financial support from MIUR (2000-2001 Funds
for “Radical Processes in Chemistry and Biology: Syn-
theses, Mechanisms and Applications”) and the Uni-
versity of Bologna (2001-2003 Funds for Selected
Research Topics).
Su p p or tin g In for m a tion Ava ila ble: Calculated reaction
pathways, calculation details, and Z-matrices of radicals 13,
14, 15, 16, 30, and 31 and the transition states connecting
13-14, 13-15, 15-31, 31-14, 15-16, 15-30, and 30-16 as
obtained by DFT and/or PM3 methods; X-ray molecular
structures, crystal data, and structure refinements of com-
pounds 10b,c and 11b,c,d ; CIF file for compounds 10b,c and
11b,c,d . This material is available free of charge via the
Internet at http://pubs.acs.org.
Rea ction of 9 w ith 1-[4-Ch lor o-2-(m eth ylsu lfa n yl)-
p h en yl]d ia zon iu m Tetr a flu or obor a te. Elution with light
petroleum/diethyl ether/dichloromethane 70:15:15 v/v/v yielded
11c exclusively (70%), with mp and spectral data identical to
those reported for 11c in the above reaction of 9 with 7c.
J O0267504
3464 J . Org. Chem., Vol. 68, No. 9, 2003