IV. In this case, 7 was obtained from 6 under the reaction con-
ditions. The intermediate IIId is more unstable than IIIb
because of the substitution of two butyl groups on the S-atom
rather than two methyl groups, so that the formation of IIIb
would be favoured over formation of IIId. Thus, the rate of
decomposition of 3d may be lower than that of 3b.
1465, 1320, 1290, 1260 (C᎐N), 1190, 1120, 1097, 1082, 1000,
900, 850 (C᎐NO2), 800 (para-substituted phenyl), 762, 707, 655,
626, 540, 502 and 480; δH(CDCl3) 0.98 (6H, t, J 6.0, Me), 1.0–
2.0 (8H, m, MeCH2CH2), 2.9 (4H, t, J 5.0, SCH2) and 7.38 (4H,
ABq, J 9.6, phenyl H); m/z 226 (Mϩ Ϫ Bu ϩ 1), 170 (Mϩ Ϫ
2Bu ϩ 2), 138 (Mϩ Ϫ 2BuS ϩ 2), 57, 45 and 41 (Found: C,
59.85; H, 7.85; N, 9.65. C14H22N2O2S requires C, 59.55; H, 7.8;
N, 9.95%).
The low ∆H‡ of 3d compared to that of 3b (Table 4) may
arise from the greater stability of IVd compared to IVb. The
lower ∆H‡ value of 3c compared to that of 3b (Table 4) can be
explained by the facile formation of a more stable 4-tolyl-
nitrenium ion rather than the 4-nitrophenylnitrenium ion.
From the above results, showing that the rate of decom-
position depends upon the ease of formation of both III and
IV, kobs might be represented as kK. The variation in which the
softer sulfonium ion 3 changes to the harder nitrenium ion
of III would mostly affect ∆S‡, leading to the large negative
∆S‡ due to an increase in the ion pair-association with the
counter-anion. The magnitudes of ∆H‡ and ∆S‡ and those of
the free energy are given in Table 4.
2,2,2-Trifluoroacetanilide derivatives 7a, 7b and 7c were syn-
thesized by reactions of trifluoroacetic anhydride with 6a, 6b
and 6c, respectively, and identified by the following data.
2,2,2-Trifluoroacetanilide 7a. The white crystalline com-
pound had mp 87–88 ЊC (lit.,11 87.6 ЊC); δH(CDCl3) 6.9–7.7
(5H, m, phenyl H) and 7.9–8.6 (1H, br, NH); m/z 189 (Mϩ), 120
(Mϩ Ϫ CF ), 92 (Mϩ Ϫ CF C᎐O), 77 (Phϩ), 65 and 51 (Found:
᎐
3
3
C, 50.55; H, 3.2; N, 7.5. Calc. for C8H6F3NO: C, 50.8; H, 3.2;
N, 7.4%).
2,2,2-Trifluoro-4Ј-nitroacetanilide 7b. The white crystalline
compound had mp 139–140 ЊC; νmax(nujol)/cmϪ1 3285 (NH),
1740 (C᎐O), 1695, 1618, 1565 (N᎐O), 1500, 1338 (N᎐O), 1288,
The thermal reaction of 11,2 or 23 with sulfides in TFA has
been proposed to produce 4 and 5 without giving 6 and 7 via
nitrenium species such as I and II, not by a bimolecular aro-
matic substitution of 1 or 2 (Scheme 1). However, the reaction
of 3 in TFA yields 6 and 7 without giving 4 and 5 via III
(Scheme 2). These facts imply that the nitrenium ion in III must
be different from the nitrenium ion in II. Thus, the nitrenium
ion in III may be in a lower energy state than II, and is interact-
ing strongly with sulfide as compared with the interaction of II
with sulfide. Because of such a strong interaction, III could give
6 and IV by the intramolecular hydride-shift from the α-C᎐H of
the sulfide. The details of the interaction between the nitrenium
ion and sulfide as shown in II will be reported in the future.
᎐
᎐
᎐
1253, 1203, 1148 (C᎐F), 1118, 903, 862 (C᎐NO2), 828 (para-
substituted phenyl), 755, 739, 693, 670, 655 and 502; δH(CDCl3–
[2H6]dimethyl sulfoxide) 5.6 (4H, ABq, J 9.0, phenyl H) and
8.3–8.8 (1H, br, NH); m/z 234 (Mϩ), 218, 204, 188, 168, 165
(Mϩ Ϫ CF3), 140, 134, 119, 109, 91, 69, 64 and 40 (Found: C,
41.3; H, 2.25; N, 11.95. Calc. for C8H5F3N2O3: C, 41.05; H,
2.15; N, 11.95%).
2,2,2-Trifluoro-4Ј-methylacetanilide 7c. The white crystalline
product had mp 102–102.5 ЊC; νmax(nujol)/cmϪ1 3280 (NH),
1700 (C᎐O), 1613, 1553, 1512, 1415, 1357, 1308, 1273, 1243,
᎐
1202, 1160 (C᎐F), 1132, 945, 820 (para-substituted phenyl),
728, 710 and 505; δH(CDCl3) 2.3 (3H, s, Me), 7.26 (4H, ABq,
J 9.0, phenyl H) and 7.6–8.4 (1H, br, NH); m/z 203 (Mϩ), 184
(Mϩ Ϫ CF ), 156 (Mϩ Ϫ CF C᎐O), 134, 106, 91, 65 and 40
᎐
3
3
Experimental
(Found: C, 53.4; H, 3.8; N, 6.85. Calc. for C9H8F3NO: C, 53.2;
H, 3.95; N, 6.9%).
IR spectra were obtained on a Hitachi EPI-G3 spectrometer.
1H NMR spectra were taken with a Nippondenshi PMX-60SI
instrument (J values are given in Hz). GLC–MS were recorded
with a Shimadzu QP-5000 spectrometer linked to a Shimadzu
GC-17A column (15 m × 0.25 mm) coated with DC-1 (0.25 µm
film thickness), made by J & W Scientific. GLC was performed
with a Shimadzu GC-8A instrument using a glass column (1
m × 3 mm) packed with 10% SE-30 on 60–80 mesh Chromo-
sorb W (AW-DMCS). Two runs agreed to within 3% error for
the yields of the products which are determined by replicated
GLC analyses.
TFA, CF3CH2OH, dimethyl sulfoxide 9a and aniline 6a were
purified by distillation of reagent grade materials (Nacalai
Tesque). The following compounds were reagent grade (Nacalai
Tesque) and were used without further purification; trifluoro-
methanesulfonic acid (TFSA), H2SO4, diphenylamine, 2-amino-
biphenyl, 4-nitroaniline 6b, 4-methylaniline 6c. Benzene,
AcOH, CH2Cl2 and MeOH were purified by standard methods
before use. HClO4 used in TFA was dehydrated by addition of
trifluoroacetic anhydride (TFAA); HClO4 and TFAA were
reagent grade (Nacalai Tesque). 4-Aminobiphenyl8 and di-n-
butyl sulfoxide 9d9 were prepared by the methods described in
the literature.
Reactions of S,S-dimethylanilinosulfonium salt 3a in TFA
The reaction of 3a in TFA was carried out with a variety of
acids as shown in Table 1. After the reaction, aq. Na2CO3 was
added until the solution reached pH > 7. The organic layer
was extracted with CH2Cl2 and the yields were determined by
GLC analysis. The products were isolated by thin layer chrom-
atography using silica gel (Merk 60F254) after column chrom-
atography (Fujisiriaru Chem. BW-127ZH). The structures of
the products 6a, 7a and 8a were characterised by comparison of
their mass spectra and retention times of GLC–MS with those
of authentic samples; we have already described the character-
isation of the structure of 8a.2 The effect of the counter ion is
summarised in Table 1.
The effect of solvent-nucleophilicity was investigated by
variation of solvent as shown in Table 2. The procedures used
for isolation and characterisation of the products and the
determination of their yields are indicated above.
The reaction of 3a (5.0 mmol) was carried out in a mixture
of TFA (7.5 cm3) and benzene (17.5 cm3) for 2 h at 25 ЊC or
under reflux. The reaction mixture was treated as described
above. According to the GLC analysis, diphenylamine and
2- and 4-aminobiphenyl were not detected, but 6a and 7a were
formed in 12% and trace or 25 and 46% yields for the reaction
at 25 ЊC or under reflux, respectively.
S,S-Dimethylanilinosulfonium picrate, S,S-dimethyl-N-(4-
nitrophenyl)imino-λ4-sulfane and S,S-dimethyl-4-methyl-
anilinosulfonium picrate were prepared by the literature
method,10 and had mp 129.6 ЊC (lit.,10 130–130.6 ЊC), 167 ЊC
(lit.,10 166–167 ЊC) and 167–169 ЊC (lit.,10 165–166 ЊC), respect-
ively. The following λ4-sulfane was synthesized by a similar
method to that described in the above literature, and confirmed
as indicated below.
Reactions of S,S-dialkylanilinosulfonium salts 3b–d in TFA
The detailed reaction conditions and the results are indicated in
Table 3. Isolation of the products 6b, 6c, 7b and 7c was carried
out as shown above. The products 9a and 9d were identified
by comparison of GC–MS and IR spectra with authentic
specimens. The reaction mixtures showed no peaks except those
of 6, 7 and 9 for the GLC analysis, and thus we thought that 8b
and 8c were not formed in the reactions.
S,S-Di-n-butyl-N-(4-nitrophenyl)imino-ë4-sulfane
This yellow crystalline compound had mp 68–69 ЊC; νmax
-
(KBr)/cmϪ1 3960, 3930, 2860, 1580 (N᎐S), 1570 (N᎐O), 1485,
᎐
᎐
J. Chem. Soc., Perkin Trans. 2, 1998
1745