TETRAHEDRON

LETTERS
Pergamon Tetrahedron Letters 44 (2003) 3459–3462

Catalyzed double Michael addition of anilines to vinyl sulfone

Jiong Jack Chen,* Cuong V. Lu and Rebecca N. Brockman†
Early Process Research and Development, Pharmacia Corporation, Kalamazoo, MI 49001, USA
Received 16 December 2002; revised 12 March 2003; accepted 13 March 2003

Abstract—Substituted anilines and vinyl sulfone undergo a facile double Michael addition to form substituted phenylthiomorpho-
line dioxide, catalyzed with AlCl3 or H3PO4. Scope and conditions were explored. © 2003 Elsevier Science Ltd. All rights reserved.

In the preparation of an active pharmaceutical interme- placement of the resulting alkyl chloride rather than a
diate, we needed a large quantity of compound 3. Since Michael addition? To test this hypothesis, 2,6-
2,6-difluoroaniline (1) and vinyl sulfone (2) are both difluoroaniline was heated with bis(2-chloroethyl)-
readily available, we sought to prepare compound 3 sulfone under neat conditions. This reaction gave no
using these two starting materials, based on a literature conversion, thus disproving this hypothesis. Interest-
procedure reported by Ford-Moore1 and later in a ingly, when AlCl3 was added, the reaction produced
patent by Straley and Fisher.2 Under the neat condi- trace amounts of desired product (3) along with some
tions reported by Ford-Moore, the reaction between vinyl sulfone. It was postulated that initially AlCl3
aniline and vinyl sulfone produced 56% of the desired catalyzed the formation of vinyl sulfone from bis(2-
double-Michael addition product, 4-phenylthiomorpho- chloroethyl)sulfone. Subsequently, the vinyl sulfone
line 1,1-dioxide.3 However, applying Ford-Moore’s underwent a catalyzed double Michael addition to gen-
conditions to the reaction between 2,6-difluoroaniline erate the desired product. This observation further sup-
(1) and vinyl sulfone (2) only resulted in mostly starting ported that the reaction proceeds via the Michael
materials and some uncyclized mono-Michael adduct addition mechanism (Scheme 1).
(4) after extensive heating. The double Michael addi-
tion did not proceed, presumably because 2,6- Various Lewis acids (i.e. FeCl3, BiCl3, ZnCl2, and
difluoroaniline is a poor Michael donor (two SnCl2) were then screened, and some were found to
electron-withdrawing fluorine atoms). Interestingly, effect the double Michael addition. However, only
when 1 equivalent of AlCl3 was added to the toluene AlCl3 gave a complete conversion within an adequate
solution of starting materials, the reaction proceeded timeframe (i.e. less than 24 h). By screening the various
smoothly to give desired compound 3. To our knowl- reaction conditions using AlCl3, we defined some key
edge, the use of a Lewis acid to activate vinyl sulfone as parameters. The reaction proceeds well within an ade-
a Michael acceptor is unprecedented.4 With this new quate timeframe at 90–110°C at a concentration
finding, we decided to explore the mechanism and between 0.5–1.0 M. By using the optimized conditions
define the scope of this interesting reaction.
(i.e. AlCl3 in chlorobenzene or toluene at 1 M concen-
tration at 110°C for 24 h) a yield of 68% was obtained
When a strong base, KOtBu, was added to the mixture
upon work-up. The low recovery was most likely the
of 2,6-difluoroaniline (1) and vinyl sulfone (2), the vinyl
result of a Lewis-acid promoted polymerization of vinyl
sulfone (2) polymerized instantaneously with a strong
exotherm. Thus, the activation of the Michael donor sulfone.
does not facilitate this reaction. As the chloride ion in
AlCl3 may add to the double bond in vinyl sulfone, is it To minimize the competing side reactions and to sim-
possible that the reaction mechanism involved the dis- plify the work-up, zeolites and protic acids were also
screened as the catalysts. Fourteen common zeolites
were first tried, none of them gave >5% of conversion
Keywords: Michael addition; vinyl sulfone; thiomorpholine; catalysis. based on GLC. The use of triflic acid was more promis-
* Corresponding author. Tel.: 1-269-833-9044; fax: 1-269-833-9282; ing. The reaction under neat conditions provided
e-mail: [email protected] mostly product within 40 h. However, the work-up was
†
Currently a graduate student at University of California, Berkerley. problematic because the reaction mixture solidified

0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)00702-0
3460 J. J. Chen et al. / Tetrahedron Letters 44 (2003) 3459–3462

Scheme 1.

upon cooling. If the reaction using triflic acid was tion was complete in 1 h with a 92% yield of com-
conducted in organic solvents, the reaction time pound 9. Under the phosphoric acid conditions, the
required for completion was beyond three days. reaction was complete after 5 h and provided an 84%
Weaker methanesulfonic acid and 98% sulfuric acid yield. 2,4-Dimethoxyaniline (10), an activated aniline,
were tried under the neat conditions, but no conver- was expected to yield similar results as compound 8.
sion was observed at all. To our surprise, heating the Interestingly, a major impurity was identified as the
starting materials in neat phosphoric acid at 140°C 4-demethoxy analog of compound 11. This impurity
provided a complete conversion in less than 24 h was formed even without any catalyst. Under the
along with some black residues. More surprisingly, phosphoric acid conditions, only 25% isolated yield of
the common 85% phosphoric acid proceeded cleaner the desired product 11 was obtained. For the deacti-
and faster than neat phosphoric acid at the same vated system such as 4-nitroaniline (12), the reaction
temperature. Under these conditions, the work-up and provided no conversion without catalyst. Under AlCl3
the yield were significantly improved. Upon comple- conditions, the reaction also gave no conversion due
tion, the reaction was simply cooled and poured into to the formation of an insoluble aniline–aluminum
water, stirred, and the product was then filtered to complex. After work-up, only starting material was
provide an 89% isolated yield. obtained. For the phosphoric acid conditions, the
reaction was complete after 17 h and provided a 76%
With optimized conditions found for AlCl3 (Method
yield of 13. For the 2,3,5,6-tetrafluroaniline (14), the
A)5 and 85% phosphoric acid (Method B),6 we were
reaction gave no conversion in the absence of a cata-
interested in the generality of this reaction toward
other substrates. Various anilines were employed to lyst. With AlCl3 as the catalyst, the reaction gave
study the scope of this reaction (Table 1). only a 65% conversion by GLC after 24 h of heating.
Attempts to work-up the reaction were unsuccessful
For the ‘neutral’ aniline (6), the reaction under the due to emulsion and gelling. Under the phosphoric
AlCl3 conditions provided a quantitative conversion acid conditions, the reaction provided compound 16
after only 45 min of heating at 110°C. After work-up, in a 66% yield.
a 73% isolated yield was obtained. Under the phos-
phoric acid conditions, an improved 78% isolated Using AlCl3 and phosphoric acid as catalysts, synthe-
yield was obtained. These unoptimized yields were ses of various substituted 4-phenylthiomorpholine 1,1-
significantly improved from the 56% yield reported by dioxide were demonstrated to provide good to
Ford-Moore.1,3 2,4-Dimethylaniline (8) should give excellent yields. The scope of this reaction appears to
better results since it is a better Michael donor. With be wide ranging to include both activated and de-acti-
no catalyst in toluene, the reaction provided a 90% vated anilines.8 This new finding should provide
conversion, 63% desired product, and 3% bis-Michael organic chemists with another synthetic tool, for the
adduct in 24 h based on GLC. With AlCl3, the reac- formation of thiomorpholine dioxides.
 J. J. Chen et al. / Tetrahedron Letters 44 (2003) 3459–3462 3461

Table 1.

References (1.8 kg). The column was eluted with methylene chloride
(16 L) until clear. The methylene chloride solution was
1. Ford-Moore, A. H. J. Chem. Soc. 1949, 2433–2440. concentrated to give 387 g of compound 3 as light brown
2. Straley, J. M.; Fisher, J. G. US Patent No: 3,585,182. solids (68% yield).
3. Using 1:1 ratio of aniline and vinyl sulfone, Ford-Moore 6. Method B typical procedure: 2,6-Difluoroaniline (1) (6.5
obtained 56% yield of desired 4-phenylthiomorpholine, mL, 60 mmol) and vinyl sulfone (2) (6.0 mL, 60 mmol)
1,1-dioxide and 10% yield of 2,2%-dianilinodiethyl sulfone. were charged to 30 mL of 85% H3PO4 at room tempera-
Using 2:1 ratio, 76% yield of the desired product was ture. The mixture was heated to 140°C for 16 h. The
obtained. However, for our purpose, neither starting mate- resulting suspension was cooled to 70°C and poured
rial was cheap, thus 2:1 ratio was not an option. Plus, into 100 mL of water. After cooling to room temperature,
purification became more challenging if excess aniline was the mixture was filtered. The beige solids were washed with
used. water and dried to give 13.2 g (89%) of compound 3.
4. Using Lewis acids to catalyze sulfone for other transfor- 7. Analytical data for previously unknown compounds 3, 4,
mations are known, such as Friedel–Crafts reaction: Trost, 5, 9, 11, 13, and 15.
B. M.; Matsuoka, R. T. Synlett 1992, 27–30. 4 - (2,6 - Difluorophenyl)thiomorpholine 1,1 - dioxide (3):
5. Method A typical procedure: Aluminum chloride (310 g, 1
H NMR (DMSO-d6) l 7.10 (m, 3H), 3.51 (m, 4H), 3.21
2.3 mol) was added to chlorobenzene (2.5 L, toluene works (m, 4H). HRMS calcd for [M+1]: 248.0557. Found:
also) to give a cloudy green suspension. Vinyl sulfone (2) 248.0552. Anal. calcd for C10H11F2NO2S: C, 48.58; H,
(230 mL, 2.3 mol) was added via funnel. 2,6-difluoroani- 4.48; N, 5.66. Found: C, 48.80; H, 4.37; N, 5.36.
line (1) (250 mL, 2.3mol) was added via funnel. The light (2-Ethenesulfonyl-ethyl)-2,6-difluorophenylamine (4): This
brown solution was heated to 110°C. Upon completion at compound was isolated in a non-catalyzed reaction. It
24 h, the heat was removed and the black solution was co-elutes with vinyl sulfone. Thus, a pure sample was not
cooled to 70°C. The reaction mixture was quenched into obtained. However, 1H NMR spectrum of the mixture
methylene chloride (4 L) and ice water (5 L). The aqueous indicated additional vinyl peaks. If this mixture was
phase was extracted with methylene chloride (4×2 L). The treated with Lewis acid, compound 3 was formed based on
combined organic layers were concentrated and added GLC.
branched octane (3 L), and then cooled to 0°C for 30 min. N - (2 - {[2 - (2,6 - Difluoroanilino)ethyl]sulfonyl}ethyl) - 2,6 - di-
The solids were filtered and washed with branched octane fluoroaniline (5): This compound was isolated in a run
(2×500 mL). The crude black solids were dissolved into without catalyst. NMRs were obtained. However, collec-
methylene chloride (3 L) and then loaded onto a SiO2 plug tion of other data was not possible in our hands due to its
3462 J. J. Chen et al. / Tetrahedron Letters 44 (2003) 3459–3462

partial conversion to compound 4. 1H NMR (CDCl3) l HRMS calcd for [M]: 256.0518. Found: 256.0514. Anal.
6.87–6.67 (m, 6H), 3.85 (m, 4H), 3.32 (m, 4H). 13C NMR calcd for C10H12N2O4S: C, 46.87; H, 4.72; N, 10.93; S,
(CDCl3) l 155.1, 151.9, 124.3 (t, JC–F=11.6 Hz), 118.8 (d, 12.51. Found: C, 47.67; H, 4.37; N, 10.74.
JC–F=8.7 Hz), 54.5, 40.0. 4-(2,3,5,6-Tetrafluorophenyl)thiomorpholine 1,1-dioxide
4 - (2,4 - Dimethylphenyl)thiomorpholine 1,1 - dioxide (9): (15):
1
H NMR (CDCl3): l 7.02 (s, 1H), 6.96 (m, 2H), 3.38 (m, 1
H NMR (CDCl3): l 6.83 (m, 1H), 3.74 (m, 4H), 3.35 (m,
4H), 3.19 (m, 4H), 2.28 (s, 3H), 2.27 (s, 3H). 13C NMR 4H). HRMS calcd for [M+1]: 284.0368. Found: 284.0923.
(CDCl3): l 148.1, 134.8, 132.9, 132.4, 127.8, 120.6, 53.0, Anal. calcd for C10H9F4NO2S: C, 42.40; H, 3.20; N, 4.95;
51.1, 21.1, 18.0. HRMS calcd for [M+1]: 240.1058. Found: S, 11.32; F, 26.83. Found: C, 41.88; H, 3.21; N, 4.69.
240.1054. Anal. calcd for C12H17NO2S: C, 60.22; H, 7.16;
8. Several monovinyl sulfones, such as phenyl vinyl sulfone
N, 5.85; S, 13.40. Found: C, 59.76; H, 7.13; N, 5.78.
and methyl vinyl sulfone, have been tried under the same
4 - (2,4 - Dimethoxyphenyl)thiomorpholine 1,1 - dioxide (11):
conditions. The Michael addition did proceed under cata-
1
H NMR (CDCl3): l 7.00 (m, 1H), 6.51 (m, 1H), 6.45 (m,
lyzed conditions. However, some Michael addition
1H), 3.90 (s, 3H), 3.80 (s, 3H), 3.54 (m, 4H), 3.24 (m, 4H).
13
C NMR (CDCl3): l 154.1, 121.6, 104.1, 100.4, 56.0, 55.9, product continued to the second Michael addition to give
52.6, 50.4. Anal. calcd for C12H17NO4S: C, 53.12; H, 6.32; a bis-adduct (an adduct derived from one aniline and two
N, 5.16; S, 11.82. Found: C, 52.95; H, 6.37; N, 5.10. monovinyl sulfone). A mixture of products was formed,
4 - (4 - Nitrophenyl)thiomorpholine 1,1 - dioxide (13): thus, makes this extended application less useful. Further-
1
H NMR (DMSO-d6): l 8.9 (d, 2H, J=9.3 Hz), 7.16 (d, more, the reaction mixture was very obnoxious, suggesting
2H, J=9.3 Hz), 4.02 (m, 4H), 3.18 (m, 4H). 13C NMR the decomposition of monovinyl sulfones under Lewis acid
(DMSO-d6): l 152.7, 138.2, 126.1. 113.8, 50.7, 45.8. conditions.