Dimethyl Sulfoxide-Trifluoroacetic Anhydride. A ... - ACS Publications

solid residue was washed well with water and crystallized from ether-petroleum ether (bp 30-60'), giving 66 mg (0.283 mmol,. 88%) of 4a: mp 66-67' dec...
0 downloads 0 Views 987KB Size
2758

J . Org. Chem., Vol. 40, No. 19, 1975

Sharma, Ku, Dawson, and Swern

5 ml of chloroform. After 10 min the solvent was removed, and the

solid residue was washed well with water and crystallized from ether-petroleum ether (bp 30-60'), giving 66 mg (0.283 mmol, (MeCN) (lom3e) 298 nm (2.35), 88%) of 4a: mp 66-67' dec; ,A,, 290 (2.46), and 239 (28.2); 1H NMR (CDC13) 8 7.95-7.8 (m, 2 H, aromatic), 7.3 (m, 6 H, aromatic), and 1.5 (s, 3 H, NH). The NH band was at 3200 cm-'. Anal. Calcd for C I ~ H S N O S " ~ ~ C,: 61.8; H, 4.75; N, 6.00; S, 13.7. Found: C, 61.9; H, 4.81; N, 5.86; S, 13.8. B.Compound 4a was also obtained by adding 10% sodium hydroxide solution to a suspension of 2 in ethanol and working up as before. Methylation of 4a. Formation of 5,5-Dihydro-5-(methylimino)phenoxathiin Iodide (5). To a solution of 110 mg (0.511 mmol) of 4a in ether was added 2 ml of methyl iodide. Pale yellow crystals of 5 deposited during 15 rnin of stirring, giving 149 mg (0.42 mmol, 81.5%), mp 121-122' dec. Conversion of 5 into 5,5-Dihydro-5-(methyIimino)phenoxathiin Perchlorate (6). An excess of silver perchlorate was added to a stirred solution of 100 mg (0.28 mmol) of 5 in acetonitrile. After 10 rnin the precipitated silver iodide was filtered, the solution was evaporated, and the residue was washed with water and crystallized from aqueous methanol, giving 88 mg (0.27 mmol, 96%) of 6, mp 160-162O dec, mmp with authentic 6 (see below) 159- 160' dec. Reaction of 1 with Methylamine. Formation of 6. A suspension of 1.02 g (3.41 mmol) of 1 in acetonitrile was stirred for 10 min and methylamine gas was bubbled in until the purple color disappeared. Reaction was slower than with ammonia. Work-up and chromatography, as earlier, gave 457 mg (2.28 mmol, 67%) of phenoxathiin, 31 mg (0.144 mmol, 4%) of phenoxathiin 5-oxide, and 143 mg (0.433 mmol, 13%)of 6, mp 158-159' dec, from aqueous ac(MeCN) e) 302 nm (5.431, 280 (3.54), and 233 etone: A,, (20.0); 'H NMR (MezSQ-dC)8 8.0-8.2 (m, 2 H, aromatic), 7.7 (m, 6 H, aromatic), and 2.2 ( s , 3 H, Me). The NH proton could not be detected in MezSO solvent, presumably because of exchange, but gave rise to a 3280-cm-' band in the infrared. Anal. Calcd for C I ~ H ~ ~ N S CC, ~O 47.3; S : H, 3.67; N, 4.25; S, 9.72; C1, 10.7. Found: C, 47.3; H, 3.42; N, 4.39; S, 9.92; C1, 10.7. Reaction of 4a with Tosyl Chloride. Formatian of N-Tosyl Phenoxathiin Sulfilimine (7). A suspension of 53 mg (0.17 mmol) of 2 in ether was deprotonated by addition of 1 ml of pyridine. Tosyl chloride (42 mg, 0.22 mmol) was added, and after 1 hr of stirring the solvent was removed. The residue was washed well with water and crystallized from aqueous ethanol t o give 26 mg (0.07 mmol, 41%) of 7, mp 166-168', infrared ident,ical with that of an authent,ic sample, mp 168-170°, made by reaction of phenoxathiin with chloramine-T according t o method R of Tsujihara et al.'S Reaction of 1 with 4a. Preparation of 3a. A suspension of 139 mg (0.463 mmol) of 1 in 10 ml of acetonitrile was stirred for 10 min

and a solution of 51 mg (0.237 mmol) of 4a in 3 ml of acetronitrile was added. The disappearance of the color of 1 was quite slow. After 20 rnin the pale purple color was discharged by adding 1 drop of water. The solution was stirred with a small amount of sodiuq carbonate (to neutralize perchloric acid) and evaporated. Column chromatography gave 47.6 mg (0.237 mmol, 51%) of phenoxathiin (benzene), 19 mg (0.09 mmol) of phenoxathiin 5-oxide (chloroform), and 101 mg (0.196 mmol, 42%) of 3a (acetone), mp 236-238O dec, from aqueous methanol. Reaction of 4a with Thianthrene Cation Radical Perchlorate (8). Formation of 5,5-Dihydro-5-(5-thianthreniumylimino)phenoxathiin Perchlorate (9). Reaction was carried out as with the reaction of 4a with 1, using 52 mg (0.165 mmol) of 8 and 19.5 mg (0.091 mmol) of 4a. After stirring with sodium carbonate the solution was poured into water and the precipitate was taken up in acetone and precipitated with ether. Crystallization from aqueous methanol gave 41 mg (0.077 mmol, 47%) of 9, mp 208210' dec. Anal. Calcd for C24H16NS3C105: C, 54.4; H, 3.04; N, 2.64; S, 18.2; C1,6.69. Found: C , 54.1; H, 3.02; N, 2.47; S, 18.4; C1,6.56. The acetone-ether filtrate from the precipitation of 9 was evaporated, and the residue was taken up in chloroform and subjected to TLC on silica gel with benzene development, giving 19.5 mg (0.09 mmol, 55%) of thianthrene and 7 mg (0.03 mmol) of thianthrene 5-oxide. Registry No.-1, 55975-63-8; 2, 55975-55-8; 3a, 55975-57-0; 4a, 54002-03-8; 5, 55975-58-1; 6, 55975-60-5; 7, 54462-91-8; 8, 3578771-4; 9, 55975-62-7; phenoxathiin, 262-20-4; methyl iodide, 74-884; silver perchlorate, 7783-93-9; methylamine, 74-89-5; tosyl chloride, 98-59-9.

References and Notes (1) Supported by Grant GP-25989X from the National Science Foundation, and Grant D-028 from the Robert A. Welch Foundation. (2) Part XXXiii: B. K. Bandiish, A. G. Padilia, and H. J. Shine, J. Org. Chem., 40, 2590 (1975). (3) B. Lamotte, A. Rassat, and P. Servoz-Gavin, C. R. Acad. Sci., 255, 1508 11962) - - - I

- - - I

(4) M. Tornita, S. Ueda, Y. Nakai, and Y. Deguchi, Tetrahedron Lett., 1189 ( 1963). (5) B. Lamotte and G. Berthier, J. Chim. Phys., 369 (1966). (6) U. Schmidt, K. Kabitzke, and K. Markau, Chem. Ber., 97, 498 (1983). (7) E. Volanshi and M. Hillebrand, Rev Room. Chim., 12, 751 (1967). (8) H.J. Shine and R. J. Srnaii, J. Org. Chem., 30, 2140 (1965). (9) C. Barry, G. Cauquis, and M. Maurey, Bo//.SOC.Chim. Fr., 2510 (1966). (IO) Y. Murata and H. J. Shine, J. Org. Chem., 34, 3368 (1969). (11) H. J. Shlne and J. J. Siiber, J. Am. Chem. Soc., 94, 1026 (1972). (12) P. Stoss and G. Satzinger, Tetrahedron Lett., 1973 (1974). (13) Y. Tamura, K. Sumoto, J. Minamikawa, and M. ikeda, Tetrahedron Lett., 4137 (1972). (14) H J. Shine and K. Kim, TehahedronLett., 99 (1974). (15) K. Tsujihara, N. Furukawa, K. Oae, and S. Oae, Bo//. Chem. SOC.Jpn., 42, 2631 (1969).

Dimethyl Sulfoxide-Trifluoroacetic Anhydride. A New and Efficient Reagent for the Preparation of Iminosulfuranes' Ashok K. Sharma, Thomas Ku, Arthur D. Dawson, and Daniel Swern* Fels Research Institute and Department of Chemistry, Temple University, Philadelphia, Pennsylvania 19122 Received March 10, 1975 The scope and limitations are described of the recently reported dimethyl sulfoxide-trifluoroacetic anhydride (DMSO-TFAA) reagent for the preparation of iminosulfuranes. Yields range from 40 to 90% with aryl amines, including ortho-substituted ones, aryl amides, aryl sulfonamides, and urea. Previously uncharacterized iminosulfuranes have been prepared from sulfanilamide (mono- and diylides) and sulfadiazine. Relatively basic amines (cyclohexylamine, benzylamine), o - and p-diaminobenzenes, anthranilamide, ansidine, and 2- and 4-aminopyridines failed to yield isolable iminosulfuranes.

?'his paper defines the scope and limitations of the recently reported dimethyl sulfoxide-trifluoroacetic anhydride (DMSQ-TFAA) reagent for the efficient preparation of iminosulfuranes (sulfilimines) and compares the re-

agent's utility with that of other "activated" DMSO reagents. In our preliminary communicationsla only a few iminosulfuranes were reported and no information was then available on the limitations of the new reagent.

J. Org. Chem., Vol. 40, No. 19, 1975 2759

Dimethyl Sulfoxide-Trifluoroacetic Anhydride The activation of DMSO by a number of electrophiles is well documented, typically chlorine,2 acetic anhydride,3 toluenesulfonic anhydride? methanesulfonic anhydride,4 alkyl chloroform ate^,^ toluenesulfonyl chloride: cyanuric ~hloricle,~ sulfur t r i ~ x i d e phosphorus , ~ ~ ~ ~ ~ pentox~ ide,6,8s9and dicyclohexy1carbodiimide.loJ1All of these electrophiles have been used to activate DMSO for the oxidation of alcohols but only the last three electrophiles (AE) and acetic anhydride have also been used for the preparation of iminosulfuranes (2) (eq 1).Preparation of 2 in good yield requires an intermediate (1) containing a good leaving group (OE-)readily displaced by the nucleophilic nitrogen compounds.12

+

+

AE

(CH,)$--O

--+

R”

[(CHJ,S-O-E]A1

+

[(CH,),S-NH-R]A-

---3

B

+

(CHJLS-fi-R

(1)

2

Results and Discussion Scope. Activation of DMSO with acyl halides and certain anhydrides at room temperature, particularly in the absence of a moderating solvent, can and does proceed explosively. TFAA falls into that category, but we correctly concluded that it should be possible to moderate its reaction with DMSO by working at low temperature in an unreactive solvent. Even under those conditions, DMSO and TFAA react almost instantly and exothermically a t -6OO in methylene chloride to produce a white precipitate which, for convenience, is written as 3 (eq 2). This precipitate is stable below -30’ but on warming the system it becomes homogeneous, and the Pummerer rearrangement product (4)forms; it is readily observed by NMR (6 5.35, 2 H, s; 2.28,3 H, s). DMSO -60”

I

_ . 9

CH,CI,

TFAA

0

+

1 aromatic ammes, amides, sulfonamides + 1,

II

[(CH,),S-O---C-CF,]-OCOCF,

2 (40-9Wo)

3

(2)

>-30’1

0

H,C---S-CHL-O-C-CF,

I1

4

We have no direct evidence for the structure of 3, as we have failed in all attempts to isolate it. We can intercept and trap 3, however, with a wide range of nitrogen-containing nucleophikes, such as certain aromatic amines, amides, and sulfonamides (eq 2). These nucleophiles react rapidly and cleanly. Crude 2, after basification (when required) with triethylamine or 5-10% aqueous sodium hydroxide, is obtained in 40-90% yields in almost analytical purity without further purification (Table I).17With many aromatic amines, the reactions are complete within a few minutes after all the amine has been added (TLC); with sulfonamides reaction takes about 30-60 min and with amides up to 240 min is required. In contrast, DMSO activated by sulfur trioxide or phosphorus pentoxide requires considerably longer reaction times (ca. 20 times longer).lb Table I lists the iminosulfuranes (2) (or their salts) prepared and also1 yields, melting points, NMR, and elemental analyses. In one case (61,both the ylide (y) and its picrate

(p) were isolated. In three other cases (5,7,8) only the picrate was stable enough for characterization. In one case (15) only the trifluoroacetate could be obtained. All compounds had the predicted NMR spectra. Compounds that had been previously reported by us and/or others (5, 7, 9, 11, 13, 14, 16, 17) were additionally characterized by melting point and ir comparison with authentic samples and, in several cases, by mixture melting point as well. New compounds were characterized spectrally and by elemental analyses. Thin layer chromatography was a useful monitor of purity. One noteworthy feature of the DMSO-TFAA reagent is is reactivity even with aromatic amines containing certain ortho substituents (CH3, NO& Such amines could not be converted to N-aryliminosulfuranes by our earlier procedure in which DMSO is activated by S03.1bFor completeness, o-fluoroaniline was also converted to the N-aryliminosulfurane (Table I, 8). Amides (benzamide, p-nitrobenzamide, and urea) readily form N-acyliminosulfuranes (Table I, 13, 14, 15) in excellent yield with the DMSO-TFAA reagent. Reaction times (90-240 min, as derived from TLC data) are longer than with aromatic amines and a small excess of the reagent (3) is required to obtain optimum yields. With benzamide, addition of base is essential to obtain the ylide (13) but with p-nitrobenzamide, the ylide 14 precipitates during the reaction and does not require basification to obtain a t least a 90% yield of almost pure product. The additional acidity of the NH proton resulting from the p-nitro group coupled with the low solubility of the product (14) shifts the equilibrium cleanly and almost quantitatively from salt to ylide. Urea, an amide with two identical nucleophilic sites, can in principle form a mono- and diylide as well as the corresponding salts. A pure dylide or disalt could not be obtained even with a large excess of the reagent (3); a complex mixture of products was shown by TLC. However, when equimolar quantities of urea and the reagent (3) were used, an excellent yield (80%)of the monotrifluoroacetate salt of the monoylide (Table I, 15) was obtained. The reaction of sulfonamides (benzene- and p-nitrophenylsulfonamide) with the reagent (3) turned out to be two to four times more rapid than that with amides, even though sulfonamides are poorer nucleophiles. Sulfonamides, however, are considerably stronger acids than amides and may exist in equilibrium with their conjugate base, the sulfonamido anions. Although only small quantities of these anions are likely to be present, they should be superior nucleophiles to free amides and, if the equilibrium is rapidly restored, the overall reaction rate should be higher with sulfonamides. This argument is supported by our work with p -aminobenzenesulfomamide (PARS) described below. With sulfonamides basification is not required as the NH proton is readily lost and the ylides precipitate from the reaction mixture. Sulfanilamide (18) poses an intriguing synthetic challenge as it can, in prinicple, form two monoiminosulfuranes, one with the ylide function on the amino side (20) and the other with the ylide function on the sulfonamido side (21), and one diiminosulfurane (22). Sulfadiazine (19) can yield only one iminosulfurane, the monoylide (23). The diylide 22 (eq 3) appeared to be the easiest of the group to prepare as we had already established that both the amino and sulfonamido functions react cleanly and rapidly with the reagent 3. When 18 was allowed to react with an excess of 3 in methylene chloride below -40° (usually -50 to -60’1 for about 2 hr followed by customary addition of triethylamine (TEA) to the reaction solution, a

2760

J . Org. Chem., Vol. 40, No. 19, 1975

18

19

21

22

23

Sharma, Ku, Dawson, and Swern in about 50% yield after addition of TEA to deprotonate the intermediate sulfonium salt. Compound 20, m p 135137' dec, gives a negative ninhydrin test and fails to form an azo dye with P-naphthol (free amino group absent) and it is soluble in aqueous base (free sulfonamido group present).18 Spectra and elemental analysis confirmed its structure. The direct preparation of 21 (iminosulfurane on the sulfonamido group) from 18 and 3 was expected to be difficult, if not impossible, because of the intrinsically lower nucleophilicity of the sulfonamido group. Even in cases where we attempted to reduce the nucleophilicity of the amino group markedly be making the trifluoroacetate salt of 18, we were not successful in the direct selective reaction of the sulfonamido group. Compound 21 was obtained unexpectedly, however, in about 30% yield when we attempted to convert 24, the monotrifluoroacetate salt (eq 3), to the diiminosulfurane (22) by treatment with basic ion-exchange resins in an aqueous system (eq 5). Treatment of 24 with Amberlite

75-95% yield of 24, mp 174-176' dec, the monotrifluoroacetate salt of 22 rather than 22, was unexpectedly formed. To obtain 22 it was necessary to react 24 with virtually neat TEA (some methylene chloride was needed to enhance the solubility of 24) in a separate step. DMSO-TFAA 3 (excess)

+

18

I cH,CI. 14 0

s

II -C-NH,

Calcd for C5H,F,N203S: C , 25.6; H, 3.87; F, 24.4; N, 12.0; S, 13.7. Found: C , 26.0; H, 3.84; F , 24.2; N, 11.7; S, 13.4. NMR 3.2, 6 H, s; 7.0, 3 H, s

15

(broad) -so&$ 16

80

124-131

128-1 30.5' 13116

85

184-185

183-185Ib 186'O

NMR 2.7, 6 H, s; 7.35, 3 H, m ; 7.85, 2 H, m NMR 2.8, 6 H, s; 8.0, 2 H, m; 8.4, 2 H ,

m

CDCla or DMSO-& solution; XL-100 NMR spectrometer; 6 values (Mersi = 0). * Isolated as picrate only. c Both ylide and picrate isolated; y = ylide and p = picrate. The free ylide is unstable; an elemental analysis was not obtained. e New compounds. f This compound was not completely characterized and its structure is still uncertain. 8 Isolated as trifluoroacetate only. a

The sequence that best explains the formation of 25 from 24 via 25a and 25b in the presence of basic ion-exchange resins is shown in eq 6.

Sulfadiazine (19) readily forms an iminosulfurane (231, sodium salt (dihydrate), mp 265-268O dec, in 7096 yield from reagent 3 and 19 (eq 7). In contrast to 19,23 is readily

2762 J.Org. Chem., Vol. 40,No. 19,1975

Sharma, Ku, Dawson, and Swern

I. CH-CI..