Alkanethioimidoyl Radicals: Evaluation of β-Scission Rates and of Cyclization onto S-Alkenyl Substituents

Article abstract of DOI:10.1021/jo0353637

Thioimidoyl radicals were generated by addition of alkylsulfanyl radicals to alkyl isonitriles and were characterized by electron paramagnetic resonance (EPR) spectroscopy. The β-scissions of their C·S-C bonds were studied by variable-temperature EPR spectroscopy and the fragmentation rate constants and activation energies were calculated. The scission rates depend on the stability of the released alkyl radicals but in any case, at room temperature, the processes were fast. Data collected on similar oxyimidoyls showed that their fragmentations are slightly slower compared to those of analogous thioimidoyls. The scission rates of selenoimidoyls could not be studied by EPR and were evaluated by theoretical calculations. EPR experiments also enabled both β-scission and 5-exo ring closure rate constants of two S-but-3-enyl-substituted imidoyl radicals to be determined, showing that cyclization prevails only at low temperatures. Density functional theory (DFT) theoretical calculations predicted that the fragmentation process preferentially occurs from the s-cis rotamers (X-C bond) of the imidoyl radicals. Thio- and seleno-imidoyls (but not oxyimidoyls) prefer s-trans conformations so that their fragmentations involve prior rotation about the X-C bond.

Full text of DOI:10.1021/jo0353637

Alk a n eth ioim id oyl Ra d ica ls: Eva lu a tion of â-Scission Ra tes a n d of
Cycliza tion on to S-Alk en yl Su bstitu en ts
†
,†
,‡
Matteo Minozzi, Daniele Nanni,* and J ohn C. Walton*
School of Chemistry, University of St. Andrews, Fife, KY16 9ST, United Kingdom, and Dipartimento di
Chimica Organica “A. Mangini”, Universit a` di Bologna, Viale Risorgimento 4, I-40136 Bologna, Italy
jcw@st-andrews.ac.uk
Received September 16, 2003
Thioimidoyl radicals were generated by addition of alkylsulfanyl radicals to alkyl isonitriles and
were characterized by electron paramagnetic resonance (EPR) spectroscopy. The â-scissions of their
•
C S-C bonds were studied by variable-temperature EPR spectroscopy and the fragmentation rate
constants and activation energies were calculated. The scission rates depend on the stability of the
released alkyl radicals but in any case, at room temperature, the processes were fast. Data collected
on similar oxyimidoyls showed that their fragmentations are slightly slower compared to those of
analogous thioimidoyls. The scission rates of selenoimidoyls could not be studied by EPR and were
evaluated by theoretical calculations. EPR experiments also enabled both â-scission and 5-exo ring
closure rate constants of two S-but-3-enyl-substituted imidoyl radicals to be determined, showing
that cyclization prevails only at low temperatures. Density functional theory (DFT) theoretical
calculations predicted that the fragmentation process preferentially occurs from the s-cis rotamers
(
X-C bond) of the imidoyl radicals. Thio- and seleno-imidoyls (but not oxyimidoyls) prefer s-trans
conformations so that their fragmentations involve prior rotation about the X-C bond.
In tr od u ction
SCHEME 1
1
•
2
Imidoyl radicals (R NdC R ) are finding increasing use
as synthetic intermediates, especially since the discovery
of several stylish annulations based around their forma-
tion from imines or isonitriles and subsequent cycliza-
tions to give quinolines, cyclopentaquinolines, and related
heterocycles.1 Thioimidoyl radicals (R NdC SR ) can be
-3
1
•
2
3
made from sulfanyl radical additions to isonitriles or by
aryl radical additions to isothiocyanates. The electron
paramagnetic resonance (EPR) spectra of three archetype
1
2
thioimidoyl radicals (R ) t-Bu; R ) Me, CF
3
, and t-Bu)
including kainic acid. Useful cascade sequences leading
were briefly described by Blum and Roberts4 in 1978.
Recently, important cyclization and cascade sequences
involving thioimidoyl radicals have been developed. For
6a
6b
to benzothienoquinoxalines and -quinolines and thio-
7
chromenoindoles have also been discovered.
Two â-scission modes are available to R-heteroatom-
substituted imidoyl radicals (Scheme 1). Scission of the
X-R2 bond (mode a) will give an isocyanate (isothio-
5
example, Bachi et al. exploited ring closures of thioimi-
doyl radicals onto unsaturated N-substituents (R1
)
2
•
alkenyl) for the preparation of a variety of heterocycles,
cyanate, isoselenocyanate) and C-centered radical R .
1
Alternatively, scission of the R -N bond will give C-
†
University of Bologna.
1•
centered radical R and a cyanate (thiocyanate, seleno-
‡
University of St. Andrews.
1) (a) Leardini, R.; Pedulli, G. F.; Tundo, A.; Zanardi, G. J . Chem.
cyanate) (mode b).
(
Soc., Chem. Commun. 1984, 1320. (b) Leardini, R.; Nanni, D.; Pedulli,
G. F.; Tundo, A.; Zanardi, G. J . Chem. Soc., Perkin Trans. 1 1986,
Knowledge of the mechanistic details of thioimidoyl
radical reactions and of the way â-scission rate constants
respond to variations in the attached groups would be
useful input for various aspects of synthetic design. For
example, ring closure of a thioimidoyl radical 3 (X ) S)
onto an S-alkenyl substituent had not been reported prior
to this work but appeared possible, provided the â-scis-
1
591.
2) (a) Curran, D. P.; Liu, H. J . Am. Chem. Soc. 1991, 113, 2127. (b)
(
Ryu, I.; Sonoda, N.; Curran, D. P. Chem. Rev. 1996, 96, 177. (c) J osien,
H.; Ko, S.-B.; Bom, D.; Curran, D. P. Chem. Eur. J . 1998, 4, 67.
(3) For a review on radical addition to isonitriles, see Nanni, D. In
Radicals in Organic Synthesis, Renaud, P., Sibi, M. P., Eds., Wiley-
VCH: Weinheim, Germany, 2001; Vol. 2, Chapt. 1.3.1, pp 44-61.
(4) (a) Blum, P. M.; Roberts, B. P. J . Chem. Soc., Chem. Commun.
1
1
976, 535. (b) Blum, P. M.; Roberts, B. P. J . Chem. Soc., Perkin Trans.
1978, 1313.
(6) (a) Leardini, R.; Nanni, D.; Pareschi, P.; Tundo, A.; Zanardi, G.
J . Org. Chem. 1997, 62, 8394. (b) Benati, L.; Leardini, R.; Minozzi,
M.; Nanni, D.; Spagnolo, P.; Zanardi, G. J . Org. Chem. 2000, 65, 8669.
(7) Benati, L.; Calestani, G.; Leardini, R.; Minozzi, M.; Nanni, D.;
Spagnolo, P.; Strazzari, S.; Zanardi, G. J . Org. Chem. 2003, 68, 3454.
(5) (a) Bachi, M. D.; Balanov, A.; Bar-Ner, N. J . Org. Chem. 1994,
5
9, 7752. (b) Bachi, M. D.; Melman, A. Synlett 1996, 60. (c) Bachi, M.
D.; Melman, A. J . Org. Chem. 1997, 62, 1896. (d) Bachi, M. D.; Melman,
A. Pure Appl. Chem. 1998, 70, 259.
1
0.1021/jo0353637 CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/19/2004
2
056
J . Org. Chem. 2004, 69, 2056-2069

Alkanethioimidoyl Radical â-Scission and Cyclization
SCHEME 2 a
a
For clearness’ sake, thioimidoyl radicals 6 are identified by both a capital letter (the N-substituent derived from the isonitrile) and
a small letter (the S-substituent derived from the disulfide).
sion reactions were not too fast. We report in this paper
our characterization of a representative range of thio-
imidoyl radicals and of their fragmentation rate con-
stants, mainly by EPR spectroscopy, supplemented by
end product analyses and by density functional theory
(
DFT) computations. For comparison purposes, several
model oxyimidoyl (3, X ) O) and selenoimidoyl radicals
3, X ) Se) were examined by similar methods. On the
(
basis of the measured fragmentation rates, it appeared
that ring closure onto an S-alkenyl group might be
observable and our study of the regiochemistry and rate
of this process is also reported. Part of this research was
published as a preliminary letter.8
Resu lts a n d Discu ssion
F IGURE 1. EPR spectrum of thioimidoyl radical 6Cc at 150
K in cyclopropane.
Thioimidoyl radicals have previously been generated
by addition of C-centered radicals to the S-atoms of
isothiocyanates.6 However, when a cyclopropane solu-
tion of acetyl peroxide and n-butyl isothiocyanate was
photolyzed in the resonant cavity of an EPR spectrom-
eter, only the methyl radical was observed in the whole
temperature range (150-250 K).
In subsequent spectroscopic work, thioimidoyl radicals
were generated exclusively by photolysis of a dialkyl
disulfide (5a -i) in the presence of an alkane isonitrile
,7,9
1
illustrates the striking difference in the appearance of
an archetype thioimidoyl radical spectrum as compared
with oxyimidoyl counterparts (see Supporting Informa-
tion for the latter). At higher temperatures the spectrum
of 6Cc weakened and the spectrum of the tert-butyl
radical appeared and became stronger as the tempera-
ture was raised. Above ca. 200 K only the tert-butyl
radical could be detected.
(
4A-D), and series of both substrates were prepared for
Similar spectra of alkane thioimidoyl radicals 6 were
observed for isonitriles 4A-D with disulfides 5a -d .
Attempts were also made to observe arenethioimidoyl
radicals by photolysis of diphenyl disulfide and other
diaryl disulfides, together with alkyl isonitriles. However,
no EPR spectra were obtained from these systems,
probably because of the insolubility of these disulfides
in cyclopropane. Hyperfine splittings (hfs) from H atoms
of the N-substituents of 6 were resolved (see Figure 1)
but hfs from H atoms of the S-substituents were not
resolved in any case. The measured g-factors were all in
the range 2.0008-2.0014 (Table 1) and are characteristic
this purpose by literature methods or adaptations thereof
(Scheme 2).
EP R Sp ectr a of Th ioim id oyl Ra d ica ls. When a
cyclopropane solution (ca. 0.2 mL) containing 4C (10 mg)
and 5c (10 mg) in a quartz tube was photolyzed at 150
K with light from a 500 W Hg arc lamp in the resonant
cavity of an EPR spectrometer, a three-line spectrum
with EPR parameters characteristic of the thioimidoyl
radical 6Cc was observed (Figure 1 and Table 1). Figure
(
8) Minozzi, M.; Nanni, D.; Walton, J . C. Org. Lett. 2003, 5, 901.
(9) Barton, D. H. R.; J aszberenyi, J . Cs.; Theodorakis, E. A.
4
Tetrahedron 1992, 48, 2613.
of oxyimidoyl and thioimidoyl radicals. The low g-factors
J . Org. Chem, Vol. 69, No. 6, 2004 2057

Minozzi et al.
TABLE 1. Exp er im en ta l EP R P a r a m eter s for
SCHEME 3
a
Th ioim id oyl a n d Rela ted Ra d ica ls
radical
T/K g-factor a(N)
a(Hâ)
•
a
6
6
6
6
6
6
6
6
Aa n-C12H25NdC S-n-Bu
255 2.0011
4.12 1.98 (2H)
4.43 1.96 (2H)
4.43 1.86 (2H)
4.64 1.86 (2H)
•
a
Ab n-C12H25NdC S-i-Pr
170 2.0008
•
a
Ad n-C12H25NdC S-c-C5H9 185 2.0008
•
a
Ac n-C12H25NdC S-t-Bu
160 2.0011
•
a
Ba c-C6H11NdC S-n-Bu
Bb c-C6H11NdC S-i-Pr
Bc c-C6H11NdC S-t-Bu
Ca t-BuNdC S-n-Bu
245 2.0010
4.4
2.20 (1H)
•
a
160 2.0008
4.64 2.20 (1H)
4.69 2.22 (1H)
•
a
160 2.0011
•
a
255 2.0012
4.58
4.70
4.58
4.83
4.74
5.0
5.1
4.58
4.5
b
2
10 2.0011
•
a
6
6
6
6
Ci t-BuNdC S-but-3-enyl 235 2.0011
Cb t-BuNdC S-i-Pr
Cd t-BuNdC S-c-C5H9
Cc t-BuNdC S-t-Bu
•
a
150 2.0009
R1 was confirmed by the unsymmetrical thioimidoyls,
•
•
a
215 2.0009
1
•
•
c
none of which showed any trace of R . GC-MS analyses
of the reaction products provided further evidence that
â-scission took place exclusively by fragmentation of the
188 2.0014
a
1
50 2.0011
•
a
6
6
6
6
6
Da AdNdC S-n-Bu
Di AdNdC S-but-3-enyl
Db AdNdC S-i-Pr
Dd AdNdC S-c-C5H9
230 2.0011
•
a
245 2.0012
2
13
S-R bond. In each case only the isothiocyanate
•
a
210 2.0009
4.80
4.85
5.10
1
-
•
a
R NdCdS was detected, not the alternative thiocyanate
NCS-R .
180 2.0008
2
•
a
Dc AdNdC S-t-Bu
175 2.0013
•
a
n-C12H25NdC O-t-Bu
205 2.0014 e0.7
1.36 (2H)
2.01 (1H)
Photolyses were also carried out with disulfides 5e-h
and tert-butyl isonitrile 4C. In each case the correspond-
ing thioimidoyl radical was not detected, even at the
lowest accessible temperature (ca. 120 K in n-propane
solvent). Instead the allyl, benzyl, pent-2-ynyl, or (meth-
oxycarbonyl)methyl radical was observed. Evidently, the
rate of the â-scission process was enhanced by the
resonance stabilization in the released radical.
•
a
c-C6H11NdC O-t-Bu
t-BuNdC O-t-Bu
220 2.0014 e0.7
•
c
2.0016 e0.3
a
All hyperfine splitting values are in gauss; solvent ) cyclo-
propane; Ad ) 1-adamantyl. Solvent ) tert-butylbenzene. c See
ref 4.
b
A search for selenoimidoyl radicals 8 was carried out
by photolysis of 4C with dibenzyl diselenide 7f (Scheme
3). EPR spectroscopy showed only the benzyl radical
down to 150 K in cyclopropane solvent.14 GC-MS analy-
sis of the reaction mixture established that the main
products were tert-butyl isoselenocyanate 9 and dibenzyl
(the latter to a major extent). This suggested that, at least
to some extent, benzylselenyl radicals had added to the
isonitrile and that â-scission had taken place. No spec-
troscopic evidence of the selenoimidoyl radical 8Cf was,
however, obtained.
F IGURE 2. Resonance stabilization in imidoyl radicals.
and other EPR parameters are consistent with the idea
that thioimidoyls are σ-radicals as suggested previ-
ously.4
,10,11
Imidoyl radicals normally exhibit low a(N) values (1.2-
.9 G) that have been accounted for by a spin polarization
1
mechanism that induces negative spin density at the
N-atom, somewhat balancing the positive spin density
resulting from resonance (Figure 2).
By analogy with thioimidoyls, selenoimidoyls with Se-
alkyl (non-resonance-stabilized) substituents should be
less susceptible to â-scission, and therefore the di-i-propyl
diselenide 7b was prepared and photolyzed with 4C. No
selenoimidoyl or i-propyl radicals were detected by EPR
spectroscopy in this case in the temperature range 150-
In thioimidoyl radicals, the ability of the S-atom to
stabilize an R-negative charge4 should increase the
contribution of the right-hand canonical structure of
Figure 2. This brings greater positive spin density onto
the N-atom and hence accounts for the somewhat higher
a(N) values (4.1-5.1 G) found for thioimidoyls (Table 1).
It should be noted that, for each series of 6, a(N) increases
as the extent of branching at the R-C-atom of the S-alkyl
substituent increases. Several oxyimidoyl radicals, gener-
ated by photolysis of di-tert-butyl peroxide (DTBP) and
an isonitrile, were also observed and their EPR param-
eters are included in Table 1.
2
00 K. However, GC-MS analysis of the product mixture
established again that tert-butyl isoselenocyanate 9 was
a major product. We tentatively conclude from this that
the selenoimidoyl radicals 8, although not observable by
EPR spectroscopy, are initially generated and undergo
a fast â-scission to 9 (see, however, below).
Kin etics of Th ioim id oyl â-Scission . The mechanism
of the process taking place under EPR conditions may
be represented by eqs 1-5:
For all the radicals 6Aa -6Dd the EPR spectra at
higher temperatures showed mainly the C-centered radi-
2
•
12
cal R with hfs identical to those given in the literature.
That R2 was released (by C-S bond scission) and not
•
1
ka
9
1
•
2
•
2
R -NC + X-R
8 R NC XR
(1)
4
6
(
10) Danen, W. C.; West, T. J . Am. Chem. Soc. 1973, 95, 6872-
kf
1
2•
6
874.
6
98 R NCX + R
(2)
(3)
(4)
(5)
(
11) Recent theoretical work has suggested that the bent σ-structure
is essentially independent of the electronic nature of the R-substitu-
ent: Guerra, M. J . Phys. Chem. 1996, 100, 19350.
2kt
2
•
2•
R
+ R
8 nonradical products
(
12) (a) Berndt, A. In Landolt-B o¨ rnstein; Magnetic Properties of Free
Radicals, Fischer, H., Hellwege, K., Eds.; Springer: Berlin, 1977; Vol.
b, p 452. (b) Berndt, A. In Landolt-B o¨ rnstein; Magnetic Properties of
kx
9
R2• + 6
8 nonradical products
9
Free Radicals, Fischer, H., Hellwege, K., Eds.; Springer: Berlin, 1987;
Vol. 17c, p 88.
2kt
6 + 6
8 nonradical products
2
058 J . Org. Chem., Vol. 69, No. 6, 2004

Alkanethioimidoyl Radical â-Scission and Cyclization
TABLE 2. Ra te a n d Ar r h en iu s P a r a m eter s for F r a gm en ta tion s of Th ioim id oyl a n d Oxyim id oyl Ra d ica ls
kf, s- (200 K)
1
log (Af, s-1
)
Ef (E13), kcal/mol
a
RSE, kcal/mol
b
radical â-scission
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Aa
Ab
Ad
Ac
Ba
Bb
Ca
Cb
Cd
Cc
Cg
Da
Db
Dc
n-C12H25NdC‚S-n-Bu
7.0
39
430
1600
12
14.0
11.0
12.4
10.2
13.2
12.1
14.0
13.2
12.8
10.0
[13]
13.0
12.6
12.6
10.7
11.0
12.2d
12.0 (10.9)
8.6 (10.6)
8.9 (9.5)
6.4 (8.7)
11.1 (10.9)
9.0 (9.8)
11.9 (10.7)
10.2 (10.0)
9.2 (9.3)
5.9 (8.4)
(e6.7)c
3.9
6.4
7.3
8.9
3.9
6.4
3.9
6.4
7.3
8.9
12.4
3.9
6.4
8.9
•
n-C12H25NdC S-i-Pr
•
n-C12H25NdC S-c-C5H9
•
n-C12H25NdC S-t-Bu
•
c-C6H11NdC S-n-Bu
•
c-C6H11NdC S-i-Pr
230
11
•
t-BuNdC S-n-Bu
•
t-BuNdC S-i-Pr
120
680
3500
•
t-BuNdC S-c-C5H9
•
t-BuNdC S-t-Bu
•
5
t-BuNdC SCH2CtCEt
g4 × 1 10 2
•
AdNdC S-n-Bu
10.9 (10.9)
9.5 (9.9)
8.5 (8.8)
7.6
•
AdNdC S-i-Pr
160
2400
100
18
•
AdNdC S-t-Bu
•
n-C12H25NdC O-t-Bu
•
c-C6H11NdC O-t-Bu
8.9
9.6
•
d
d
t-BuNdC O-t-Bu
50
a
E13 are the experimental â-scission activation energies standardized under the assumption that log (Af, s-1) ) 13. b Methane-based
stabilization energies of the released C-centered radicals. c Estimated (see text). d Data from Roberts et al.4
By application of the steady-state approximation, it can
easily be shown that
2
•
kx
2k [R ]
t
1
)
•
+
(6)
2
kf[6]
[
R ]
kf
Because the radicals involved are “small” molecules, the
termination rate constants will be close to the diffusion
rate and may all be considered the same15 (k
hence eq 6 simplifies to the usual expression:
x
t
) 2k );
2
• 2
kf
[R ]
2•
)
[R ] +
(7)
F IGURE 3. Arrhenius plot for â-scission of N-t-butyl-S-i-
propyl-thioimidoyl radical 6Cb.
2kt
[6]
For representative pairs of reactants, the correspond-
ing thioimidoyl â-scission (fragmentation) rate constants
of C-centered radical can be efficiently produced by this
route at T g 20 °C. The rate of S-C bond scission in
thioimidoyls is g16 times the rate of O-C bond scission
in oxyimidoyls at 200 K. This is in accord with expecta-
tion because of the longer and weaker C-S bonds.
Unfortunately, the â-scission rate constants for sele-
noimidoyl radicals could not be calculated by experimen-
tal data and an estimate will be reported in the theo-
(k
f
) were therefore obtained from EPR spectroscopic
measurements of the concentrations of the two radicals
1
6
by the usual method with DPPH as standard. The well-
established 2k
t
value of Schuh and Fischer,1 corrected
7
for the difference in viscosity between n-heptane and
cyclopropane as described previously,18 was introduced
to enable absolute values of k
f
to be derived. The
4
retical section (see below). Blum and Roberts reported
Arrhenius plot of the data obtained for the N-tert-butyl-
S-i-propyl-thioimidoyl radical 6Cb derived from tert-butyl
isonitrile and di-i-propyl disulfide (Figure 3) gives an
example of the quality of the data obtained. All of the
Arrhenius plots, with their linear regression equations,
and examples of the EPR spectra are in the Supporting
Information; the measured fragmentation rate constants
are given in Table 2.
the only previous kinetic datum for a thioimidoyl radical
5
-1
•
â-scission, i.e., k
t-Bu (6Cc). Extrapolation of our Arrhenius data (Table
f
(243 K) ) 3.1 × 10 s for t-BuNC S-
5
2
, E13) for this radical gives k
f
(243 K) ) 2.8 × 10 , in
good agreement.
Comparing the â-scission rate constants of the set of
thioimidoyl radicals with S-n-Bu groups (or the set with
S-t-Bu groups, Table 2) suggests that N-alkyl substitu-
ents have little influence on the rate. The rate constants
are more variable for other S-groups but, bearing in mind
that the error limits on the data are large, it is probable
that the effect of N-substituents is minor. S-Alkyl sub-
stituents, on the other hand, have a dramatic effect, and
going from S-n-alkyl to S-tert-butyl causes a rate increase
of ca. 2 orders of magnitude at 200 K. The rate increase
is even larger on going to a resonance-stabilized S-group
such as allyl or benzyl (see above). These trends make it
clear that two factors are important: (i) greater branch-
ing at the C-atom attached to sulfur dramatically in-
creases the rate of fragmentation, and (ii) resonance
The data reported in Table 2 show that â-scission of
the S-C bonds in thioimidoyl radicals is a fast process
5
6
-1
(
10 < k
f
< 10 s ) at 298 K. Thus, practically every kind
4
(
13) Note, however, that Blum and Roberts observed C-N â-scis-
•
•
sion of t-BuNdC OMe to afford the EPR spectrum of t-Bu radicals at
higher temperatures.
(14) An experiment with n-propane as solvent showed no interpret-
able spectra in the range 110-150 K.
(
15) Fischer, H.; Paul, H. Acc. Chem. Res. 1987, 20, 200.
(16) (a) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 193. (b)
Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317.
(
(
17) Schuh, H.; Fischer, H. Helv. Chim. Acta 1978, 61, 2130.
18) Maillard, B.; Walton, J . C. J . Chem. Soc., Perkin Trans. 2 1985,
4
43.
J . Org. Chem, Vol. 69, No. 6, 2004 2059

Minozzi et al.
R, which are isoelectronic with imidoyls but fragment by
scission of O-C bonds with concomitant formation of
2
2
CdO double bonds (elimination of carbon dioxide). At
room temperature, the fragmentations of O-tert-butyl
5
-1
oxyimidoyls (∼1 × 10 s at 298 K) are nearly as fast as
5
-1
those of the analogous O-tert-butyl carbonyl (1 × 10 s
at 293 K),2
2b,c
whereas scissions of S-tert-butyl thioimi-
doyls are 3-20 times faster, in agreement with the
weaker C-S bond. Interestingly, a comparison between
various heteroatom-alkyl-substituted radicals shows that
the fragmentation rate of alkoxycarbonyls is very sensi-
2
-1
tive to the released radical, passing from 2 × 10 s for
2
-1
primary alkyls to 5 × 10 s for secondary radicals and
5
-1
to as much as 1 × 10 s for tertiary and (cyclopropyl)-
2
2c
methyl radicals (1:2.5:550 rate ratio at 293 K). By way
of contrast, at room temperature, the fragmentations of
the corresponding thioimidoyls are always faster and
much less dependent on the extent of branching of the
f
F IGURE 4. Plot of experimental log (k ) values for thioimidoyl
radicals vs methane-based radical stabilization energies (RSE)
derived from thermochemical BDEs.
5
6
-1
stabilization in the released radical also strongly in-
creases the rate of fragmentation. Both these factors
contribute to the thermodynamic stabilization energy of
the released radical. In fact, a reasonably linear correla-
tion is obtained from the methane-based stabilization
energies of the C-centered radicals (RSE), derived from
the difference in the recommended C-H bond dissocia-
tion energy (BDE) of methane and those of the corre-
S-alkyl substituent, ranging from ∼10 to ∼10 s at 298
K. Thus, despite the formation of a very stable fragmen-
tation product (carbon dioxide), alkoxycarbonyls are quite
reluctant to release primary or secondary radicals, which
can be instead efficiently produced starting from thio-
imidoyls. Although the stabilization of the released
radical remains a key factor in controlling the rate of
fragmentation in thioimidoyls (see below), nevertheless
its influence on the rate seems to be less important
compared to alkoxycarbonyls; this is probably the result
of early (alkoxycarbonyls)2 or late (thioimidoyls) transi-
tion states in which the X-C bond is broken to very
different extents (see the section on theoretical calcula-
tion for further details).
2
19,20
sponding hydrocarbons [BDE(R H)],
f
with the log (k )
values (Figure 4). In deriving this correlation, a lower
limit for the â-scission rate constant of the alk-2-ynyl
thioimidoyl radical 6Cg was estimated from the fact that
only the pent-2-ynyl and not the corresponding thioimi-
doyl was observable in cyclopropane solution at 150 K.
From the measured concentration of the released pent-
2d
The photoinitiated reaction of 1-dodecaneisonitrile (4A)
with DTBP in cyclopropane showed spectra of a mixture
2
-ynyl radical, and assumption of the minimum thioimi-
doyl radical concentration consistent with the baseline
S/N, the k shown in Table 2 was derived. The good
of the cyclopropyl radical, derived from H-abstraction
•
f
from the solvent by t-BuO radicals, and the n-C12
H
25Nd
2
•
linearity of this correlation (r ) 0.967) indicates that the
stabilization of the released radical is a key factor in
controlling the rate of fragmentation. Of course, changes
in the strength of the bond being broken due to increased
frontal strain with increasing branching of the substitu-
ent is an interrelated factor.
C O-t-Bu radical in the temperature range 150-195 K.
At T > 230 K the tert-butyl radical started to appear.
For the competition below 195 K (eqs 8 and 9)
ka
9
t
•
n
•
t
BuO + 4A
8 C H NC O Bu
(8)
(9)
1
2
25
The literature provides very few examples of rate data
for radical fragmentation processes involving C-S bond
kH
t
•
c
t
c
•
BuO + C H
98 BuOH + C H
3 5
3
6
2
breaking. The exception is data for reactions of RS-CH -
•
•
CH R′ radicals undergoing â-scission with release of RS
radicals and an alkene. Several such fragmentations have
it can easily be shown that
rate constants in the range 107-10 s at 300 K and are
therefore faster than thioimidoyl fragmentation. Al-
8
-1
n
•
t
c
ka
[ C H NC O Bu][ C H ]
12
25
3
6
2
1
)
(10)
c
•
n
kH
[ C
H
][ C12
H
25NC]
though both reactions involve â-scission of initially
3
5
C-centered radicals via C-S bond breaking, the frag-
•
mentation of RS-CH
2
CH R′ radicals differs from that of
From EPR measurements of the radical concentrations
from two different c-C concentrations, a good, linear
Arrhenius plot of k /k was obtained (see Supporting
Information). Combining this with the known Arrhe-
thioimidoyls in that CdC bonds and sulfanyl radicals are
formed compared with isothiocyanate moieties and alkyl
radicals.
3 6
H
a
H
2
3
It is instead instructive to compare the behavior of
imidoyls with that of alkoxycarbonyl radicals (O)C O-
H
nius parameters for k gave
•
ka(M- s-1) ) 10.3-7.0 (kcal mol )/θ
1
-1
(11)
(19) (a) Seetula, J . A.; Russell, J . J .; Gutman, D. J . Am. Chem. Soc.
1
5
9
990, 112, 1347. (b) Seetula, J .; Gutman, D. J . Phys. Chem. 1992, 96,
401. (c) Berkowitz, J .; Ellison, G. B.; Gutman, D. J . Phys. Chem. 1994,
8, 2744.
•
and hence at 290 K the rate constant for t-BuO radical
addition to isonitrile 4A is 1.2 × 10 M
5
-1 -1
s . This is
(
20) Cohen, N. In General Aspects of the Chemistry of Radicals;
Alfassi, Z. B., Ed.; Wiley: Chichester, U.K., 1999; Chapt. 10, p 317.
21) Chatgilialoglu, C.; Altieri, A.; Fischer, H. J . Am. Chem. Soc.
slightly smaller than the only other known rate constant
5
-1
-1
of this type, i.e., k
a
(290 K) ) 2.6 × 10
M
s
for
(
•
2
002, 124, 12816.
addition of t-BuO to t-BuNC reported by Roberts and
2
060 J . Org. Chem., Vol. 69, No. 6, 2004

Alkanethioimidoyl Radical â-Scission and Cyclization
co-workers.4,24 These rate constants show that t-butoxyl
respectively,26 we tentatively estimated the variations in
the activation energies of fragmentations of thio- and
selenoimidoyls by assuming an average value of ca. 0.25
M for the concentration of the dichalcogenide. The
Arrhenius plots (see Figure 3) were recalculated by
means of eq 13, giving, for thioimidoyls, very minor
changes in the activation energies (e2%): this confirms
radical addition to NdC of isonitriles is somewhat slower
6
than addition to CdC of alkenes [k
a
(300 K) ∼ 10 ] but is
still a fast process. This is in good accord with the known
efficiency of this and related radical additions to isoni-
triles in a variety of synthetic procedures.
As far as the selenoimidoyls are concerned, the frag-
mentation rate constants could be neither calculated nor
estimated from the spectra of similar thioimidoyls.
Indeed, the photolyses of disulfides or diselenides in the
presence of an isonitrile are probably not fully compa-
rable, due to the presence of significant side reactions
with seleno compounds.
that the S 2 process does not influence the kinetic
H
outcome to significant extents and use of eq 7 instead of
4
eq 13 is a very good approximation. On the contrary,
the greater value of kHS for selenoimidoyls, in conjunction
with eq 13, gave activation-energy differences that can
-1
be as large as 1-1.5 kcal mol with respect to eq 7. This
Photolysis of dialkyl diselenides is known to afford
products deriving also from scission of the C-Se bond in
addition to those derived from Se-Se bond fragmenta-
tion.25 The presence of major amounts of dibenzyl in the
reaction of diselenide 7f with 4C, compared to the yield
of isoselenocyanate 9, suggests that great amounts of
benzyl radical could be formed by homolysis of the
supports the contention that the reactions of thio- and
selenoimidoyl radicals are not totally comparable and the
former cannot be used to acceptably estimate the â-scis-
sion rate constants for the latter. Indeed, small errors in
the estimation could result in activation-energy differ-
ences that do not allow us to establish whether sele-
noimidoyls fragment faster than thioimidoyls or vice
versa (see computational section).
PhCH -Se bond rather than fragmentation of selenoimi-
2
doyl 8Cf (which forms nonetheless, but probably to a
minor extent). With diselenide 7b the reaction afforded
greater amounts of isoselenocyanate 9, probably as a
result of the major formation of selenoimidoyl 8Cb, but
neither selenoimidoyl nor i-propyl radicals were detected.
The absence of any signal ascribable to the selenoimidoyl
could be due either to the effect of the selenium atom on
the relaxation time of the radical, with the effect of
broadening the lines beyond detection, or to a very fast
fragmentation process that might prevent buildup of
radical concentration. However, the lack of any signal of
the i-propyl radical suggested the presence in the reaction
medium of some very fast reactions that would be able
to scavenge those species as soon as they formed. A
possible scavenging route could be an intermolecular
homolytic substitution of the i-propyl radicals onto the
Se-Se bond of the starting diselenide, which is a known,
fast reaction that could also take place with disulfides,
although at a significantly lower rate.2
Rin g Closu r es of S-Alk -3-en yl Th ioim id oyl Ra d i-
1
•
f
ca ls. The k values for all the R NdC S-n-Bu radicals
5
-1
were close to 10 s at 300 K. That is marginally less
than the rate constants of many ring closure reactions.
For example, the archetype hex-5-enyl radical cyclizes
5
-1
28
with a rate constant of 2.3 × 10 s at 300 K. This
indicated that thioimidoyl radicals with primary S-
alkenyl (or S-alkynyl) substituents might be able to
cyclize before fragmentation under certain circumstances.
Ring closure onto the S-alkenyl group would produce
dihydrothiophen-2-ylidene amines (e.g., 14) that could
easily be hydrolyzed, thus affording γ-butyrothiolactones
by a novel radical route.
To test the feasibility of this idea, dibut-3-enyl disulfide
i was prepared and photolyzed with several isonitriles.
5
When dodecane-1-isonitrile 4A was used, broad unidenti-
fied spectra were obtained. However, when adamantane-
1-isonitrile 4D was employed, a strong EPR spectrum of
the expected thioimidoyl radical 6Di was observed in the
temperature range 170-220 K (Table 1). At higher
temperatures, two additional radicals appeared (Figure
5). One of these (marked B) was the but-3-enyl radical
10 with hfs identical to those in the literature.2 The other
radical (marked C) had g ) 2.0027, a(2H) ) 21.8, a(1H)
) 30.9, and a(1H) ) 0.9 G and can be positively identified
6,27
To verify whether this process could appreciably affect
the concentration of the alkyl radicals in the reactions
of both disulfides and diselenides, the kinetic law ob-
tained above (eq 7) was modified to take the inter-
9
H
molecular S 2 process (eq 12) into account; eq 13 was
thus obtained, where kHS is the rate constant for inter-
molecular substitution of R2 radicals on the dichalco-
genide:
•
(
23) Walton, J . C.; McCarroll, A. J .; Chen, Q.; Carboni, B.; Nziengui,
R. J . Am. Chem. Soc. 2000, 122, 5455.
24) The k value from ref 4 was recalculated with more recent
kHS
2
•
2
2
2
2
2
•
R
+ R XXR
8 R XR + R X
(12)
(
a
Arrhenius data for the reference reaction, i.e., the H-abstraction from
cyclopentane; see Baban, J . A.; Goddard, J . P.; Roberts, B. P. J . Chem.
Soc., Perkin Trans. 2 1986, 269.
2
• 2
2•
2
2
kf
[R ]
kHS[R ][R XXR ]
2•
)
[R ] +
+
(13)
2
kt
[6]
2k [6]
(25) See, for example, Franzi, R.; Geoffroy, M. J . Organomet. Chem.
t
1
981, 218, 321.
26) Curran, D. P.; Martin-Esker, A. A.; Ko, S.-B.; Newcomb, M. J .
(
From the reported data of 6 × 10 M s-1 and 3 × 10
4
-1
7
Org. Chem. 1993, 58, 4691.
(27) The feasibility of the S 2 process of alkyl radicals onto the
H
-
1
-1
M
s
(at 25 °C) for the addition of primary alkyl
dichalcogenide was also supported by the presence of dibutenyl sulfide
among the reaction products of disulfide 5i with isonitrile 4C (see the
section on cyclization). Although this product could also form by
coupling of butenyls with sulfanyl radicals, it is likely that the major
formation pathway is actually intermolecular attack of the butenyl
radicals onto the starting disulfide.
(28) Chatgilialoglu, C.; Ingold, K. U.; Scaiano, J . C. J . Am. Chem.
Soc. 1981, 103, 7739.
(29) Kochi, J . K.; Krusic, P. J .; Eaton, D. R. J . Am. Chem. Soc. 1969,
91, 1877.
radicals to dimethyl disulfide and diphenyl diselenide,
(
22) For some representative papers on alkoxycarbonyl radicals, see
with references therein) (a) Forbes, J . E.; Zard, S. Z. J . Am. Chem.
Soc. 1990, 112, 2034. (b) Beckwith, A. L. J .; Bowry, V. W. J . Am. Chem.
Soc. 1994, 116, 2710. (c) Simakov, P. A.; Martinez, F. N.; Horner, J .
H.; Newcomb, M. J . Org. Chem. 1998, 63, 1226. (d) Morihovitis, T.;
Schiesser, C. H.; Skidmore, M. A. J . Chem. Soc., Perkin Trans. 2 1999,
(
2
041.
J . Org. Chem, Vol. 69, No. 6, 2004 2061

Minozzi et al.
as the 2-substituted dihydrothiophenylmethyl radical
1D.
Very similar spectra containing signals from the three
1
analogous radicals were obtained when 5i was photolyzed
8
together with tert-butyl isonitrile 4C. These observations
confirmed that ring closure onto the S-butenyl substitu-
ent of a thioimidoyl radical could take place and that the
preferred mode was 5-exo. Above ca. 280 K all the
thioimidoyl radicals were converted to but-3-enyl and
dihydrothiophenylmethyl radicals with the former pre-
dominating.
Cyclopropane is a poor hydrogen-atom donor and
therefore, under EPR conditions, the main products were
1
expected to be the isothiocyanate (R NCS) together with
compounds from combination and disproportionation of
the three radicals. The main products identified by GC-
F IGURE 5. EPR spectrum resulting from photolysis of
isonitrile 1D and disulfide 5i at 250 K.
MS analyses of several photoinitiated reactions of 5i with
1
4
C (R ) t-Bu) are shown in Scheme 4 (see the Experi-
SCHEME 4
mental Section for minor products). Product analysis of
the reaction of 5i with 4D showed similar results but the
conversion was lower, so individual components were
more difficult to characterize.
Product analysis confirmed that disproportionation
(
(
e.g., 14, 15), combination (18), and cross-combination
e.g., 13) of the intermediate radicals 6, 10, and 11 were
the dominant processes. Butane, butene, and octa-1,7-
diene from the but-3-enyl radicals were not observed
because of their volatility. A small amount (0.4%) of a
+
second product having M ) 171 (C
9
H
17NS) was also
formed. This can probably be identified as the N-tert-
butyl-(tetrahydropyran-2-ylidene)amine 16. Comparing
the yield of this disproportionation product with that of
1
4 suggested the 5-exo/6-endo cyclization ratio was ca.
3
5:1 at 0 °C. Product analysis agreed with the reaction
mechanism shown in Scheme 4, except that minor
amounts of products derived from reactions of the bute-
nylsulfanyl radical, e.g., 19, were also implicated (<10%
total). If one neglects the minor amount of termination
involving butenylsulfanyl and assumes diffusion control,
it can easily be shown that eqs 14 and 15 hold:
2
kc
[11]
[6]
[10][11]
[6]
)
)
[11] +
[10] +
+
+
(14)
(15)
2
kt
2
kf
kt
[10]
[6]
[10][11]
[6]
2
The concentrations of radicals 6, 10, and 11 were
determined, by the EPR spectroscopic method described
above, for initial isonitriles 4C and 4D. Substitution into
eqs 14 and 15 led to the rate parameters listed in Table
3
.
Comparison of the â-scission rate data [k
f
(200 K) etc.]
1
of the S-but-3-enyl thioimidoyl radicals with the analo-
(Table 3); thus, as expected, the R -N substituent is too
1
•
gous data for R NdC S-n-alkyl radicals in Table 2 shows
far from the radical center to have much effect on the
rate. These kinetic data are the first available for ring
closure of any type of imidoyl radical. The regioselectivity
is strongly in favor of 5-exo cyclization, as is usual for
C-centered species.
Thioimidoyl radicals are isoelectronic with acyl radicals
such as 1-hexanoyl and the latter is also a σ-type species.
Acyl radicals show a preference for 6-endo ring closure
under certain conditions,30 but this is due to rearrange-
ment of the initial 2-oxocyclopentylmethyl radical via a
they are strikingly similar, and this lends confidence to
1
the reliability of the data, as does the fact that the R )
1
Ad and R ) t-Bu data are also nearly equal. At 200 K
the cyclization rates are higher than the â-scission rates,
but the reverse is true at 300 K. It follows that for these
systems â-scission will be the major reaction channel at
temperatures usual for synthetic procedures.
The measured cyclization rate parameters were very
similar for the two S-but-3-enyl thioimidoyl radicals
2
062 J . Org. Chem., Vol. 69, No. 6, 2004

Alkanethioimidoyl Radical â-Scission and Cyclization
TABLE 3. Ra te a n d Ar r h en iu s P a r a m eter s for Un im olecu la r Rea ction s of S-Bu t-3-en yl Th ioim id oyl Ra d ica ls
kc at 300 K [200 K], s-
1
log (Ac, s-1
)
Ec, kcal/mol
cyclization of
•
4
6
6
Ci
Di
t-BuNdC S(CH2)2CHdCH2
2.4 × 10 [45]
10.2
10.5
8.0
8.0
•
4
AdNdC S(CH2)2CHdCH2
3.8 × 10 [28]
-
1
log (Af, s-1
â-scission of
kf at 300 K [200 K], s
)
Ef, kcal/mol
•
5
6
6
Ci
Di
t-BuNdC S(CH2)2CHdCH2
2.0 × 10 [8]
13.7
13.7
11.7
11.6
•
5
AdNdC S(CH2)2CHdCH2
1.6 × 10 [13]
TABLE 4. Ca lcu la ted ∆H Va lu es a n d Activa tion En er gies for â-F r a gm en ta tion s of Som e Th ioim id oyl Ra d ica lsa
computed valuesb
1
2
â-scission N-R side
â-scission S-R side
1
2
exp data Eac (analogous radical)
radical (R NdC‚S-R )
∆H
Ea
∆H
Ea
•
6
6
6
F k
Gb
Cc
EtNdC SEt
9.9
7.6
4.5
21.2
19.3
15.4
-0.13
-2.6
-6.4
9.4
7.0
4.4
12.0 (6Aa )
9.0 (6Bb)
5.9 (6Cc)
•
i-PrNdC S-i-Pr
t-BuNdC S-t-Bu
•
a
All energy values (∆H and Ea) are in kilocalories per mole. b Computed by the UB3LYP method with a 6-31+G(2d,p) basis set.
c
Experimental activation energy for â-fragmentation at the S-R2 side of analogous radicals (see Table 2).
bicyclo[4.1.0]hexanoxyl radical. A similar rearrangement
of 6 could be envisaged via 20, but thermodynamic
control, which is a condition for this process to compete,
is most unlikely to prevail at the low temperatures of
the EPR experiments and no evidence of ring expansion
was obtained.
double bond (Ph, CO
of ring closure.
Com p u ta tion a l Stu d y of Im id oyl Ra d ica l â-Scis-
sion a n d Cycliza tion . To get a better insight into the
structure and reactivity of imidoyl radicals, some theo-
retical calculations were carried out on both the frag-
2
R), should tip the balance in favor
mentation and the cyclization reactions by the DFT
approach, UB3LYP variant.3
3,34
Attention was first fo-
cused on the â-scission modes of the thioimidoyl series.
Calculations were carried out on thioimidoyl radicals
bearing primary, secondary, and tertiary alkyl groups on
both nitrogen and sulfur (Table 4). For all species the
enthalpy variations and the activation barriers were
calculated for both â-fragmentation modes. Scission at
the nitrogen side was found to be always endothermic
and was characterized by quite high computed activation
energies (15.4-21.2 kcal/mol); on the contrary, fragmen-
tations at the sulfur side were always computed to be
Comparison with model species such as hex-5-enyl (see
above) shows that ring closure of 6Ci is about an order
of magnitude slower at 300 K. This is probably not a
consequence of the N) substitution at the radical center,
because analogous O) substitution did not depress the
5
rate of cyclization of hex-5-enoyl [k
c
(300 K) ) 2.2 × 10
-
1 31
s
]
compared to hex-5-enyl. Cyclization of 6Ci is also
slower than that of the isoelectronic (but-3-enyloxy)-
carbonyl radical2 (k
C).
Replacement of the R-methylene group of the hex-5-
enyl radical with a SiMe group led to a significant
2c
) 1.7 × 10 s vs 4 × 10 s at 80
5
-1
6
-1
(33) Calculations were carried out with the Gaussian 98W package.34
c
Density functional theory (DFT), UB3LYP variant, was employed. The
equilibrium geometries were fully optimized with respect to all
geometric variables with the 6-31+G(2d,p) (radicals 6F k , 6Gb, and
6Cc) and 6-311+G(d,p) (all other species) basis sets. For the calculation
of thermodynamic properties, total energies were adjusted for zero-
point vibrational energies and for thermal corrections to 298 K.
°
2
3
-1
32
reduction in k
c
(to 7.5 × 10 s at 298 K). This was
attributed to bond length effects, which made it difficult
for the radical center to approach the CCH end of the
double bond. Analogous bond length effects explain the
slower 5-exo cyclization rates of thioimidoyl radicals 6Ci
and 6Di and suggest that slower ring closure rates may
be general for radicals with second- (and subsequent) row
groups R- to C-centered hex-5-enyl type radicals. Ring
closure reactions of the archetype 6Xi will not be useful
for preparing dihydrothiophen-2-ylidene amines because
â-scission is too rapid. However, our results serve the
useful purpose of showing that only a small increase in
the rate of cyclization, for example, by inclusion of an
electron-withdrawing substituent at the terminus of the
2
Computed 〈S 〉 values were in the range 0.7530-0.7650 before an-
nihilation and 0.7500 after annihilation of higher multiplicity spin
states. The transition states for â-fragmentations (or cyclization) were
optimized by performing a progressive elongation of the bond involved
in the scission (or ring closure) and optimizing each single step (relaxed
scan) with a small basis set (UB3LYP/6-31G). Eventually, the energy
maximum was optimized with the usual basis set [6-31+G(2d,p)]. All
the transition states were characterized by a unique imaginary
frequency corresponding to the desired bond scission (see Supporting
Information). Isotropic EPR hfs were derived from the computed Fermi
contact integrals and were taken directly from the Gaussian output
file.
(34) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Salvador, P.; Dannenberg, J . J .; Malick, D. K.; Rabuck,
A. D.; Raghavachari, K.; Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.;
Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Keith, T.; Al-
Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill,
P. M. W.; J ohnson, B.; Chen, W.; Wong, M. W.; Andres, J . L.; Gonzalez,
C.; Head-Gordon, M.; Replogle, E. S.; Pople, J . A. Gaussian 98, revision
A.11 2/3; Gaussian, Inc.: Pittsburgh, PA, 2002.
(30) (a) Chatgilialoglu, C.; Crich, D.; Komatsu, M.; Ryu, I. Chem.
Rev. 1999, 99, 1991. (b) Chatgilialoglu, C.; Ferreri, C.; Lucarini, M.;
Venturini, A.; Zavitsas, A. A. Chem. Eur. J . 1997, 3, 376.
(
31) Brown, C. F.; Neville, A. G.; Rayner, D. M.; Ingold, K. U.;
Lusztyk, J . Aust. J . Chem. 1995, 48, 363.
32) (a) Wilt, J .; Lusztyk, J .; Peeran, M.; Ingold, K. U. J . Am. Chem.
Soc. 1988, 110, 281. (b) Wilt, J . Tetrahedron 1985, 41, 3979.
(
J . Org. Chem, Vol. 69, No. 6, 2004 2063

Minozzi et al.
TABLE 5. Th eor etica l a n d Exp er im en ta l Hyp er fin e Sp littin g Va lu es for Som e Th ioim id oyl Ra d ica lsa
computed spectrab
experimental spectrab
radical
a(N)
a(Hâ)
radical
a(N)
a(Hâ)
•
•
6
6
6
F k
Gb
Cc
EtNdC SEt
3.59
4.37
5.26
1.74 (2H)
6.51 (1H)
6Aa
6Bb
6Cc
n-C12H25NdC S-n-Bu
4.12
4.64
5.10
1.98 (2H)
2.20 (1H)
•
•
i-PrNdC S-i-Pr
t-BuNdC S-t-Bu
c-C6H11NdC S-i-Pr
t-BuNdC S-t-Bu
•
•
a
Computed by the UB3LYP method with a 6-31+G(2d,p) basis set. b All hyperfine splitting values are in gauss.
F IGURE 6. Theoretical fragmentation pathways for thioimi-
doyl radicals; for better comparison, all the energies were
shifted so as to make the energies of the starting thioimidoyls
equal to zero.
exothermic and had lower activation barriers (4.4-9.4
kcal/mol). Both modes became more and more favored,
from both a thermodynamic and a kinetic point of view,
by passing from primary to secondary and eventually to
tertiary alkyl substituents. Figure 6 clearly shows that
2
S-R fragmentation was predicted to be always favored
compared to the N-C scission, even when a primary
radical (E
with a tertiary one (E
a
) 9.4 kcal/mol) was released in competition
a
) 15.4 kcal/mol).
F IGURE 7. Optimized structures [UB3LYP, 6-31+G(2d,p)
basis set] of thioimidoyls 6F k , 6Gb, and 6Cc.
Table 4 also indicates that the calculated energies are
slightly underestimated, in comparison with experimen-
tal data for identical (6Cc) or analogous (6Bb) thioimi-
doyls, as expected for this theoretical method.35 On
in agreement with the recorded spectra. Good accord was
also observed between predicted and experimental a(H
values at the nitrogen side for radicals 6F k and 6Aa ,
whereas calculations failed in estimating the a(H ) hfs
â
)
passing from primary to secondary and from secondary
to tertiary alkyl groups, the percentage barrier variations
â
(34% and 58%, respectively) are strictly comparable to
for radicals bearing a secondary N-alkyl group (6Gb). In
the latter case, this is probably due not to modifications
in the radical core but rather to problems in optimizing
the conformation of the N-alkyl chain.36 Indeed, from the
conformational analysis of radical 6Gb, we could identify
two degenerate relative minima in which a methyl group,
instead of the hydrogen atom, is eclipsed with the orbital
containing the unpaired electron; for these species a value
the observed ones (34% and 52%, respectively).
Table 5 reports a comparison between calculated and
experimental hfs values for the same radical species
discussed in Table 4, and Figure 7 shows the optimized
structures of thioimidoyls 6F k , 6Gb, and 6Cc as obtained
by DFT calculations. As far as the a(N) data are con-
cerned, very good agreement between calculated and
experimental values was observed for all radicals (out-
standingly for radical 6Cc). Since the hfs are extremely
sensitive to the geometry of the radical, theoretical
results clearly support the EPR data in stating that
thioimidoyls are σ-radicals with a bent structure and a
trans configuration about the CdN double bond (Figure
â
of -0.98 was predicted for the a(H ) hfs. Taking into
account the Boltzmann distribution and the energy gap
between the two states, we could obtain a noteworthy
â
decrease in the average a(H ) value, which was neverthe-
less still significantly different from the experimental
one. Therefore, DFT calculations probably failed in
estimating both the real conformational minimum and
the energy gap between the possible conformers. It is
7
).
Again, for all radicals, the calculations predict negli-
gible spin densities at the â-protons on the sulfur side,
(
35) Foresman, J . B.; Frisch, Æ. Exploring chemistry with electronic
(36) Analogous differences were found by replacing the c-hexyl group
with other secondary chains.
structure methods, 2nd ed.; Gaussian Inc.: Pittsburgh, PA, 1996.
2
064 J . Org. Chem., Vol. 69, No. 6, 2004

Alkanethioimidoyl Radical â-Scission and Cyclization
SCHEME 5
TABLE 6. ∆H Va lu es a n d Activa tion Ba r r ier s for
a ,b
Isom er iza tion of Ra d ica ls 21 a n d 22
radicals
∆H
Ea
2
2
2
1a f 22a (X ) O)
1b f 22b (X ) S)
1c f 22c (X ) Se)
-0.86
0.44
0.96
5.1
6.2
5.4
a
Computed by the UB3LYP method with a 6-311+G(d,p) basis
set. b All values are in kilocalories per mole.
F IGURE 9. Fragmentation pathways for thioimidoyls 21a -
c/22a -c.
TABLE 7. Activa tion En er gies for â-F r a gm en ta tion s of
a ,b
Im id oyls 21a -c/22a -c
â-fragmentation N-side
â-fragmentation X-side
TS-W N
TS N
TS-W X
TS X
radical
(from 21)
(from 22)
(from 21)
(from 22)
a (X ) O)
b (X ) S)
c (X ) Se)
29.4
23.5
23.8
27.5
22.5
21.9
24.1
12.0
12.6
10.5
10.3
11.7
F IGURE 8. Energy profile for rotation of the C-S bond in
thioimidoyls 21b and 22b. φ is the dihedral angle formed by
the radical orbital and the C-S and S-Me bonds.
a
Computed by the UB3LYP method with a 6-311+G(d,p) basis
set. b All values are in kilocalories per mole.
intermediates 21 (TS-W N and TS-W X).37 Figure 9 and
Table 7 show all the evolution paths of radicals 21/22
and the related activation barriers.
worth noting that MM calculations on conformer distri-
bution for radical 6Gb predicted a global minimum with
the hydrogen atom not eclipsed with the radical center.
We conclude that thioimidoyl 6Gb probably stays longer
than predicted by DFT methods in a staggered rather
than an eclipsed conformation (with respect to the radical
center), thus giving rise to the low a(H ) hfs observed in
â
its EPR spectrum.
To probe the mechanism of the â-fragmentations, a
more detailed theoretical study was carried out on this
pathway for imidoyl radicals 21 with the 6-311+G(d,p)
basis set (Scheme 5).
For X ) O, the oxyimidoyl radical stays in the preferred
structure (22a ) already capable of fragmenting at the
O-Me side through the favored transition state (TS X
a ); the â-scission occurs with a ∆H of -16.9 kcal/mol and
an activation barrier of 10.5 kcal/mol. Instead, for X ) S
and X ) Se, taking into account that the imidoyl radicals
prefer conformation 21 but the rotation barriers around
•
the C -X bond are always lower than those for â-frag-
mentations (Table 6), calculations suggest that the real
evolution pathway should involve a two-step mechanism
requiring prior conversion of 21b,c into 22b,c, followed
by scission of the X-Me bond through the favored non-W
transition state (TS X b,c). Loss of the alkyl group at
the X-side seems therefore to resemble the anti elimina-
tion process typical of many ionic reactions. The frag-
mentations occur with ∆H values of 0.59 and 4.4 kcal/
For all imidoyl radicals, rotation around the single
C -X bond resulted in two conformational minima cor-
•
responding to the s-trans (21) and s-cis (22) rotamers;
for the oxyimidoyl (X ) O), the absolute minimum is the
s-cis structure 22a , whereas for thio- (X ) S) and
selenoimidoyl (X ) Se), the global minima correspond to
the s-trans structures 21b,c. Thermodynamic and kinetic
a
mol and E values of 10.3 and 11.7 kcal/mol for sulfur
•
data for rotation of the C -X bond (i.e., isomerization of
and selenium, respectively, so that fragmentations of
thioimidoyls appear to be noticeably favored compared
to those of analogous selenoimidoyls. As discussed above,
â-fragmentations at the N-side never compete with
â-scission at the X-side. The whole fragmentation mech-
anism is shown in Figure 10 for thioimidoyl 21b.
2
1 into 22) are reported in Table 6 and the energy profile
for the isomerization is shown in Figure 8 for thioimidoyls
1b and 22b. It is worth noting that rotamers 21 have
2
the W structure already postulated for analogous oxy-
imidoyls.4b
The activation energies for both fragmentation modes
It is worth pointing out that Table 7 shows that the
computed activation barrier for fragmentation of sele-
(
N- and X-side) were calculated by starting from both
radicals 21 and 22. Independently of X, the transition
states arising from structures 22 (TS N and TS X) were
lower in energy with respect to those derived from
(37) The activation energies reported in Table 4 are properly related
to the favored non-W (TS-X-like) transition states.
J . Org. Chem, Vol. 69, No. 6, 2004 2065

Minozzi et al.
F IGURE 10. Whole fragmentation mechanism for thioimidoyl 21b [UB3LYP method, 6-311+G(d,p) basis set, all values are in
kilocalories per mole].
noimidoyl 21c/22c is significantly higher than those of
the analogous oxy- and thioimidoyl radicals 21a ,b/22a ,b,
probably because scission of the weak Se-C bond is
associated with the formation of a much weaker isosele-
nocyanate moiety. Calculations therefore suggest that
selenoimidoyls should be less prone to fragmentation
compared to the analogous thioimidoyls; scission of the
latter species seems moreover to be slightly favored even
with respect to oxyimidoyls, in agreement with the
experimental data (see Table 2).
Spin density maps and hfs were calculated for radicals
2
1a -c and 22a -c (see Supporting Information). Calcu-
lations predict very low a(N) values for X ) O (-0.32 and
0.81 G for 21a and 22a , respectively), in agreement
with the experimental data (Table 1). For X ) S (3.68-
.39 G) and X ) Se (3.65-3.48 G), the a(N) values are
-
3
very close to each other, but selenium seems to be
marginally more able to delocalize spin density compared
to sulfur. Good agreement was again observed between
calculated and experimental nitrogen hfs of thioimidoyls
F IGURE 11. Spin density maps for radicals 21a -c and
2a -c [UB3LYP method, 6-311+G(d,p) basis set].
(3.68 G for 21b vs 4.12 G for 6Aa ), whereas, unfortu-
2
nately, no comparison with experimental data is possible
for selenoimidoyls, since no selenoimidoyl radical has
hitherto been detected.
a comparison between the two rotamers and the corre-
sponding transition states for thioimidoyls 21b/22b.
We also simulated the ring closure of S-butenyl-
substituted thioimidoyls by means of radical 6Ci [UB3-
LYP, 6-31+G(2d,p) basis set; see Supporting Information
for computed EPR hfs of radicals 6Ci and 11C]. The
cyclization of 6Ci was slightly exothermic (∆H ) -3.1
kcal/mol), with an activation barrier of 9.5 kcal/mol. In
terms of activation energy, there is therefore good agree-
ment with the experimental values (8.0 kcal/mol for both
radicals 6Ci and 6Di). Nevertheless, the underestimation
of the activation barrier for the associated fragmentation
[≈9-10 kcal/mol (see Table 4, radical 6F k , and Table 7,
radical 21b) compared with the experimental values of
As can be seen from Figure 11, the overall spin
distribution in radicals 22a -c is very different from that
calculated for radicals 21a -c. In particular, rotamers
2
2a -c are characterized by a significant spin density on
the X-methyl group, whereas isomers 21a -c do not seem
to distribute the spin on that moiety. This means that
radicals 22a -c show something akin to an embryonic
formation of the methyl radical and are therefore more
similar to the corresponding transition states (TS X a -c)
than are radicals 21a -c (TS-W X a -c): this could
account for the lower activation energies calculated for
all of the non-W TS X transition states. Figure 12 shows
2
066 J . Org. Chem., Vol. 69, No. 6, 2004

Alkanethioimidoyl Radical â-Scission and Cyclization
The rate of S-C bond scission in thioimidoyls is g16
times the rate of O-C bond scission in oxyimidoyls at
2
00 K. The â-scission rate constants for selenoimidoyl
radicals could not be estimated experimentally but they
were computed by DFT calculations, which suggested
that, within the imidoyl series, these radicals fragment
with the lowest rate.
The competition between fragmentation and 5-exo
cyclization was studied by EPR experiments on two
S-but-3-enyl-substituted thioimidoyl radicals. At 200 K
the cyclization rates are higher than the â-scission rates,
but the reverse is true at 300 K. It follows that for
unsubstituted systems ring closure can be the major
reaction channel only at low temperatures.
Computations on oxy-, thio-, and selenoimidoyls gave
further support to the postulated bent, trans σ-structure
of imidoyl radicals and reproduced well the observed
selectivity for S-alkyl fragmentation with respect to
N-alkyl scission, independent of the kind of the released
alkyl radical. Calculations also showed that the frag-
mentation process preferentially occurs from the s-cis
rotamer (X-C bond) of the imidoyl radical. Since thio-
and selenoimidoyls prefer an s-trans conformation, for
these radicals the actual scission pathway should there-
fore involve prior rotation of the X-C bond followed by
fragmentation. The kinetic preference for scission through
a non-W transition state was explained by comparison
of the spin density maps in the imidoyl rotamers and in
the corresponding transition states.
F IGURE 12. Spin density maps for thioimidoyls 21b/22b and
the corresponding transition states for â-fragmentation at the
S-side [UB3LYP method, 6-311+G(d,p) basis set].
1
1.6-11.7 kcal/mol (see Table 3)] makes the two pro-
38
cesses appear much more competitive than in reality.
Con clu sion s
Thioimidoyl radicals were generated at low tempera-
ture by photolysis of dialkyl disulfides in the presence of
alkyl isonitriles. Their EPR spectra showed hfs from
H-atoms of the N-substituents, whereas no hfs from those
of the S-substituents were resolved in any case. The low
measured g-factors (2.0008-2.0014) are consistent with
the idea that thioimidoyls are σ-radicals.
Exp er im en ta l Section
At higher temperatures the EPR spectra showed
mainly the C-centered radical originally linked to the
sulfur atom, as a result of â-fragmentation of the S-C
bond of the thioimidoyl. Experiments carried out on
unsymmetrical thioimidoyls showed that â-scission of the
alkyl-N bond did not compete in any case. Calculation
of the kinetic constants and Arrhenius parameters proved
that the fragmentation is a fast process and practically
every kind of C-centered radical can be efficiently pro-
Sta r tin g Ma ter ia ls. Cyclohexyl isonitrile, tert-butyl isoni-
trile, dibutyl disulfide, di-iso-propyl disulfide, di-tert-butyl
disulfide, diallyl disulfide, dibenzyl disulfide, dibenzyl dis-
elenide, cyclopentanethiol, methyl thioglycolate, 5-bromopent-
3
-yne, and iso-propyl bromide were obtained commercially.
But-3-ene-1-thiol, di-(but-3-enyl) disulfide, 1-adamantyl isoni-
trile, and dodecane-1-isonitrile were prepared as described
8
39
previously. Dicyclopentyl disulfide and dimethyl 2,2′-dithio-
diacetate40 were prepared from the corresponding thiols by
5
6
-1
oxidation with KMnO
4 4
/CuSO according to a reported proce-
duced by this route at T g 20 °C (10 < k
f
< 10 s at
dure.41
2
98 K). Thermodynamic stabilization of the released
EP R Sp ectr a . EPR spectra were obtained with a Bruker
EMX 10/12 spectrometer operating at 9.5 GHz with 100 kHz
modulation. For reactions performed in cyclopropane (up to
radical plays, however, an important role: going from
S-n-alkyl to S-tert-butyl causes a rate increase of ca. 2
orders of magnitude at 200 K, and release of resonance-
stabilized radicals is so fast that the corresponding
thioimidoyls could not be detected by EPR even at the
lowest accessible temperatures.
0
.5 mL), the solution was degassed on a vacuum line via the
freeze-pump-thaw technique, and the tube was flame-sealed.
In all cases where spectra were obtained, hfs were assigned
with the aid of computer simulations using the Bruker
Simfonia software package. For kinetic measurements, signals
were double-integrated by use of the Bruker WinEPR software
and radical concentrations were calculated by reference to a
known concentration of DPPH, as described previously.
Th ioa cetic Acid S-P en t-2-yn yl Ester . To a stirred solu-
tion of potassium thioacetate (2.86 g, 25.0 mmol) in acetone
(38) Computations on the cis-cyclized radical 11C (see Supporting
Information, page S32, right figure) indicated it is more stable by as
much as 7.4 kcal/mol than the trans isomer and its computed hfs are
in much better agreement with the experimental ones. Calculations
therefore suggest that the cyclized radical observed by EPR could be
cis- and not trans-11C. All the thioimidoyl radicals are more stable in
their trans configurations, and the CdN double bonds are not directly
involved in the ring closures. Hence, the formation of the cis-cyclized
radical implies a trans-cis equilibration of the starting thioimidoyls,
followed by thermodynamically controlled 5-exo cyclizations. This
explanation is also favored by the relatively high activation barrier
for reopening of cis-11C (19.7 kcal/mol compared to 12.7 kcal/mol for
reopening of trans-11C). Trans-cis isomerization of thioimidoyls can
be in principle a viable process, since they are isoelectronic with vinyl
radicals, which in turn are well-known to undergo fast E/ Z equilibra-
tion [see for instance (a) Singer, L. A. Selective Organic Transforma-
tions; Thyagarajan, B. S., Ed.; Wiley: New York, 1972; Vol. 11, p 239.
(250 mL) was added 1-bromo-pent-2-yne (3.68 g, 25.0 mmol).
The mixture was left stirring at room temperature for 12 h.
The solution was first evaporated in vacuo to reduce the
volume of the organic solvent and then partitioned between
H
2
O and CH
2
Cl
2
(3×). The combined organic layers were dried
over anhydrous magnesium sulfate and evaporated in vacuo.
(39) Balavoine, G.; Barton, D. H. R.; Gref, A.; Lellouche, I. Tetra-
hedron 1992, 48, 1883.
(40) Stoodley, R. J .; Watson, N. S. J . Chem. Soc., Perkin Trans. 1
1974, 252.
(b) Lui, M. S.; Soloway, S.; Wedegaetner, D. K.; Kampmeier, J . A. J .
Am. Chem. Soc. 1971, 93, 3809]. Further studies on this topic are
underway.
(41) Noureldin, N. A.; Caldwell, M.; Hendry, J .; Lee, D. G. Synthesis
1998, 1587.
J . Org. Chem, Vol. 69, No. 6, 2004 2067

Minozzi et al.
1
The crude product was purified by column chromatography
on silica gel with diethyl ether/hexane (1:9) as eluant to give
yellow oil. H NMR δ 1.44 (12 H, d, J ) 6.7 Hz), 3.24 (2 H,
heptet, J ) 6.7 Hz).
the target compound (2.60 g, 18.3 mmol, 73%) as a yellow oil.
P h otoch em ica l Rea ction of Diben zyl Diselen id e w ith
t-Bu tyl Ison itr ile. The diselenide 7f (ca. 15 mg) and tert-butyl
isonitrile (15 mg) were dissolved in benzene (0.5 mL) and
photolyzed at ambient temperature for 4 h with light from a
1
H NMR (CDCl
7.5 Hz, J ) 2.4 Hz), 2.35 (3 H, s), 3.64 (2 H, t, J ) 2.4 Hz);
) δ 13.11, 14.40, 18.89, 30.84, 74.28 (C), 85.43 (C),
95.14 (C).
P en t-2-yn e-1-th iol. According to the literature,42 the thio-
3
) δ 1.11 (3 H, t, J ) 7.5 Hz), 2.18 (2 H, qt, J
q
)
t
1
3
C (CDCl
3
1
4
00 W medium-pressure Hg lamp. The solvent was evaporated
and the mixture was analyzed by GC-MS. Peak 134, t-butyl
isoselenocyanate (9), m/z (%) 165 (6), 163 (39), 161 (18), 160
(5) 159 (6), 148 (3), 107 (3), 57 (100); peak 385, dibenzyl (main
product), m/z (%) 182 (M , 83), 92 (7), 91 (100), 65 (14), together
with minor unidentified products.
P h otoch em ica l Rea ction of Diisop r op yl Diselen id e
w ith t-Bu tyl Ison itr ile. The diselenide (ca. 15 mg) and tert-
butyl isonitrile (15 mg) were dissolved in benzene (0.5 mL)
and photolyzed at ambient temperature for 4 h with light from
a 400 W medium-pressure Hg lamp. The solvent was evapo-
rated and the mixture was analyzed by GC-MS. Peak 139,
tert-butyl isoselenocyanate (9).
P h otoch em ica l Rea ction of Dibu t-3-en yl Disu lfid e
w ith t-Bu tyl Ison itr ile. The disulfide (5i) (ca. 10 mg) and
tert-butyl isonitrile (ca. 10 mg) in a 4 mm o.d. quartz tube were
degassed on a vacuum line and dissolved in cyclopropane (ca.
acetic acid S-pent-2-ynyl ester (1.42 g, 10.0 mmol) was dis-
solved in 20 mL of anhydrous tetrahydrofuran and added
dropwise to a suspension of lithium aluminum hydride (0.38
g, 4 equiv) in 10 mL of anhydrous ether under a nitrogen
atmosphere at 0 °C. The reaction mixture was stirred at
ambient temperature for 30 min, and the excess lithium
aluminum hydride was destroyed by careful addition of 10 mL
of 1 N aqueous HCl. The ether layer was separated, dried over
anhydrous magnesium sulfate, and evaporated in vacuo. The
crude product was purified by column chromatography on
silica gel with ethyl ether/hexane (1:9) as eluant to give the
title thiol (0.88 g, 8.8 mmol, 88%) as a yellow oil. 1H NMR
+
(
2
CDCl
.20 (2 H, qt, J
Hz, J ) 2.3 Hz); C (CDCl
5.19 (C).
Bis(p en ty-2-yn yl) Disu lfid e (5g). According to the litera-
3
) δ 1.13 (3 H, t, J ) 7.5 Hz), 1.96 (1 H, t, J ) 7.2 Hz),
q
) 7.5 Hz, J
t
) 2.3 Hz), 3.27 (2 H, dt, J
d
) 7.2
1
3
t
3
) δ 13.13, 13.46, 14.49, 38.61 (C),
8
0
.5 mL). The tube was flame-sealed and photolyzed at 0 °C
with UV light from a 500 W super-pressure Hg arc lamp for 1
h. The tube was opened at low temperature, the cyclopropane
was carefully evaporated, and the contents were dissolved in
4
1
ture, the pent-2-yne-1-thiol (0.5 g, 5.0 mmol) was placed in
a 100 mL round-bottomed flask and dissolved in CH Cl (40
mL). A homogeneous mixture of KMnO (1.0 g) and CuSO
1.0 g) was added and the heterogeneous mixture was stirred
2
2
. The 1H NMR showed unreacted starting materials
4
4
CDCl
3
(
overlapping a complex sequence of product signals. GC-MS
analysis also showed starting materials and indicated the
following products (yields are given as % of total products):
peak 94, t-butyl isothiocyanate (17.3%), m/z (%) 115 (M , 92),
00 (23), 72 (10), 57 (100); peak 191, dibut-3-enyl sulfide
3.6%), m/z (%) 142 (M , 7), 114 (11), 101 (55), 88 (56), 85 (23),
3 (15), 67 (80), 59 (58), 55 (100); peak 253, tert-butyl-(3-
methyldihydrothiophen-2-ylidene)amine 14C (14.3%), m/z (%)
71 (M , 27), 156 (35), 143 (8), 116 (12), 87 (8), 68 (10), 57
100), 56 (19); peak 259, probably tert-butyl-(tetrahydrothi-
opyran-2-ylidene)amine 16C (0.4%), m/z (%) 171 (M , 8), 156
23), 142 (100), 127 (32), 113 (9), 108 (28), 99 (64), 79 (16), 67
(53), 57 (44); peak 280, tert-butyl-(3-methylenedihydrothiophen-
2-ylidene)amine 15C (11.3%), m/z (%) 169 (27), 154 (22), 113
(13), 67 (8), 57 (100); peak 356, probably N-tert-butyl-thio-
at room temperature. The progress of the reaction was
monitored by TLC until no thiol was detected. The contents
of the flask were then filtered through Celite and washed with
+
CH
2
Cl
2
(2 × 20 mL) and Et O (2 × 20 mL). The combined
2
1
(
organic layers were evaporated in vacuo and the crude product
was purified by column chromatography on aluminum oxide
+
7
with hexane as eluant to give the disulfide (0.29 g, 1.5 mmol,
1
5
)
9%) as a yellow sticky oil. H NMR (CDCl
3
) δ 1.14 (3 H, t, J
) 2.4 Hz), 3.60 (2 H,
) δ 13.26, 14.41, 29.01, 75.33 (C),
+
1
7.4 Hz), 2.22 (2 H, qt, J
q
) 7.4 Hz, J
t
(
1
3
t, J ) 2.4 Hz); C (CDCl
7.11 (C).
Isop r op yl Selen ocya n a te.43 A solution of potassium se-
3
+
8
(
lenocyanate (0.58 g, 4 mmol) in DMF (4 mL) under an argon-
filled balloon was heated to 75 °C, and i-propyl bromide (0.49
g, 4 mmol) was added. The resulting mixture was stirred for
+
formimidic acid but-3-enyl ester 17C (7.0%), m/z (%) 171 (M ,
3
h at 75 °C. Potassium carbonate [0.55 g, 4 mmol in water
7), 170 (9), 138 (14), 115 (4), 82 (13), 57 (100); peak 401, tert-
(
1.6 mL)] was slowly introduced by syringe. The resulting
butyl-(3-pent-4-enyldihydrothiophen-2-ylidene)amine
13C
+
mixture was stirred for 3 h at 75 °C and then hydrolyzed (10
mL of water) and extracted with ether (2 × 20 mL). The ether
fractions were combined, washed with water (2 × 20 mL), and
dried over anhydrous magnesium sulfate. The solvent was
evaporated in vacuo. The crude mixture was purified by
column chromatography on silica gel with pentane as eluant
to provide the target compound (0.41 g, 2.75 mmol, 69%) as a
(23.9%), m/z (%) 225 (M , 3), 210 (12), 170 (11), 157 (77), 114
(12), 101 (28), 57 (100); peak 494, 3-(but-3-enylsulfanylmeth-
yldihydrothiophen-2-ylidene)-tert-butylamine 19C (4.4%), m/z
+
(%) 257 (M , 9), 202 (27), 170 (11), 156 (29), 146 (60), 114 (33),
100 (12), 87 (17), 57 (100), 55 (31); peak 526, probably the
dimer 18C (2.6%), m/z (%) 170 (58), 114 (31), 57 (100), 55 (27);
peak 640, probably the product from cross-coupling radicals
6C and 11C (5.9%), m/z (%) 283 (8), 227 (9), 184 (30), 183 (100),
170 (29), 157 (55), 127 (19), 114 (15), 101 (22), 57 (96); several
minor unidentified components were also present (ca. 8%
total). Butene, butane, and octa-1,7-diene were not detected,
probably because their volatility caused their escape during
solvent removal.
1
colorless oil. H NMR δ 1.65 (6 H, d, J ) 6.9 Hz), 3.74 (1 H,
heptet, J ) 6.9 Hz); C NMR δ 25.5, 38.9, 102.1 (C).
1
3
Diisop r op yl Diselen id e (7b).43 To isopropyl selenocyanate
0.41 g, 2.75 mmol) and ethanol (7 mL) under an argon-filled
(
balloon was added sodium hydroxide (0.06 g, 1.38 mmol in 0.3
mL of water) by syringe. The resulting mixture was stirred
for 0.5 h at ambient temperature and then hydrolyzed (10 mL
of water) and extracted with ether (2 × 20 mL). The ether
fractions were combined, washed with water (2 × 20 mL), and
dried over anhydrous magnesium sulfate. The solvent was
evaporated in vacuo. The crude mixture was fractionated by
column chromatography on silica gel with pentane as eluant
to provide the diselenide (0.39 g, 1.60 mmol, 58%) as an intense
P h otoch em ica l Rea ction of Dibu t-3-en yl Disu lfid e
w ith Ad a m a n ta n e- 1-ison itr ile. The disulfide (5i) (ca. 10
mg) and adamantane-1-isonitrile (ca. 10 mg) in a 4 mm o.d.
quartz tube were degassed on a vacuum line and dissolved in
cyclopropane (ca. 0.5 mL). The tube was flame-sealed and
photolyzed at 0 °C with UV light from a 500 W super-pressure
Hg arc lamp for 1 h. The tube was opened at low temperature,
the cyclopropane was carefully evaporated, and the contents
were dissolved in CDCl
3
. The 1H NMR showed unreacted
(42) Yang, X. F.; Mague, J . T.; Li, C. J . J . Org. Chem. 2001, 66,
starting materials overlapping a complex sequence of products.
GC-MS analysis also showed starting materials and indicated
the following products: peak 433, adamantane-1-isothiocyan-
7
39.
(43) (a) Krief, A.; Dumont, W.; Delmotte, C. Angew. Chem., Int. Ed.
2
1
000, 39, 1669. (b) Krief, A.; Dumont, W.; Delmotte, C. Tetrahedron
997, 53, 12147.
+
ate, m/z (%) 193 (M , 7), 135 (100), 93 (11), 79 (8); peak 541,
2
068 J . Org. Chem., Vol. 69, No. 6, 2004

Alkanethioimidoyl Radical â-Scission and Cyclization
•
probably N-(1-adamantyl)-thioformimidic acid but-3-enyl ester
6 11
d, 6Ba-c, 6Ca-d,i, 6Da-d,i, n-C12H25NC O-t-Bu, and c-C H -
+
•
1
7D, m/z (%) 249 (M , 10), 192 (3), 135 (100); several additional
NC O-t-Bu; (iii) variable-temperature EPR spectra for frag-
mentations of imidoyl radicals 6Aa -d , 6Ba ,b, 6Ca -d , 6Da -
unidentified peaks were also present.
•
•
c, n-C12
6
H25NC O-t-Bu, and c-C H11NC O-t-Bu; (iv) Arrhenius
Ack n ow led gm en t. We thank the Royal Society of
Chemistry, St. Andrews University, MIUR (2002-2003
Funds for “Radical Processes in Chemistry and Biol-
ogy: Fundamental Aspects and Applications in Envi-
ronment and Material Science”), and the University of
Bologna (2001-2003 Funds for Selected Research Top-
ics) for financial support. We also thank Dr. B. Maillard
plots and equations for fragmentations of imidoyls 6Aa -d ,
•
6Ba ,b, 6Ca -d ,g, 6Da -c, n-C12
H
25NC O-t-Bu (both scission
11NC O-t-Bu; (v) spin densities and hfs
of imidoyl radicals 21a -c and 22a -c; (vi) computed structures
and hfs) of radicals 6Ci, 11C, and the connecting transition
state; (vii) Cartesian coordinates of radicals 6F k , 6Gb, 6Cc,
1a -c, 22a -c, 6Ci, 11C, and the transition states related to
•
a H 6
and k /k ), and c-C H
(
2
1
13
their reactions; and (viii) H and C NMR spectra of thioacetic
acid S-pent-2-ynyl ester, pent-2-yne-1-thiol, and disulfide 5g.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(Bordeaux) for providing samples of diallyl and di-i-
propyl disulfide and Dr. M. Guerra (ISOF-CNR, Bolo-
gna) for helpful discussions on theoretical calculations.
Su p p or tin g In for m a tion Ava ila ble: (i) General experi-
mental procedures; (ii) EPR spectra of imidoyl radicals 6Aa -
J O0353637
J . Org. Chem, Vol. 69, No. 6, 2004 2069