|
1. |
Synthesis and pyrolytic behaviour of thiazolidin-2-one 1,1-dioxides |
|
Journal of the Chemical Society, Perkin Transactions 1,
Volume 1,
Issue 14,
1997,
Page 2139-2146
R. Alan Aitken,
Preview
|
|
摘要:
J. Chem. Soc. Perkin Trans. 1 1997 2139 Synthesis and pyrolytic behaviour of thiazolidin-2-one 1,1-dioxides R. Alan Aitken,a,* David P. Armstrong,a Ronald H. B. Galt b and Shaun T. E. Mesher a a School of Chemistry University of St. Andrews North Haugh St. Andrews Fife UK KY16 9ST b Zeneca Pharmaceuticals Alderley Park Macclesfield Cheshire UK SK10 4TG Four examples of the chiral thiazolidin-2-one 1,1-dioxides 5 have been prepared by reaction of the appropriate amino alcohols 11 with CS2 in aqueous sodium hydroxide to give the thiazolidine-2-thiones 12 followed by oxidation with KMnO4 under phase-transfer conditions in the presence of benzoic acid either directly or via the thiazolidin-2-ones 13. Upon flash vacuum pyrolysis (FVP) at 650 8C 5a–c decompose mainly by loss of SO2 to give an alkene and benzyl isocyanate together with other products from fragmentation of the N-benzyl group.A significant minor pathway involves net loss of CO2 and water to give the 2-phenyl-4,5-dihydrothiazoles 21 together with their aromatisation products 22 and 23. A mechanism for this new heterocyclic transformation is proposed involving initial expansion to a cyclic carbamic–sulfinic anhydride (2,1,4-oxathiazin-3-one 1-oxide). The fully assigned 13C NMR spectra are presented for 5 12 and 13 and the 33S NMR spectrum has been obtained for 5c. Thermal and photochemical extrusion of SO2 from suitable ring systems has recently been used to achieve a wide variety of synthetic transformations.1 Among the most interesting targets have been b-lactams and these have been obtained in several cases by SO2 extrusion from appropriate thiazolidin-4-one 1,1- dioxides 1.2 In one case stereoselectivity was achieved with the cis compound 1 giving mainly the cis product 2 photochemically but mainly the trans isomer on pyrolysis (Scheme 1).3 The corresponding reaction of the thiazolidine-2,4-dione 1,1- dioxides 3 to give malonic acid imides 4 has also been reported.4 We were interested to examine the isomers of 1 the thiazolidin-2-one 1,1-dioxides 5 as possible precursors of the b-lactams 6. These have the advantage of being readily accessible in enantiomerically pure form from amino acid-derived amino alcohols and if the extrusion were possible it would be of interest to examine the reaction of their 5-anions with electrophiles in which the expected diastereoselectivity might lead to an efficient overall asymmetric synthesis of b-lactams.In this paper we describe the synthesis of representative thiazolidin-2- one 1,1-dioxides 5 in enantiomerically pure form and the first study of their pyrolytic behaviour.5 Scheme 1 N S O2 O N R3 R2 R3 R2 O N S O2 O N O O O R1 R R R1 N S O2 O R2 R3 R1 N O R1 R2 R3 – SO2 3 4 2 1 – SO2 heat or hn – SO2 5 6 heat or hn Results and discussion The target thiazolidin-2-one 1,1-dioxides are a very little known class of compounds and the only previous examples were obtained by Gaul and Fremuth in 1961 by oxidation of the corresponding thiazolidine-2-thiones using peracetic acid.6 The synthesis of the latter by reaction of b-amino alcohols with CS2 in aqueous sodium hydroxide is well known but as recently examined in detail by Delaunay et al.,7 it may also lead to the corresponding oxazolidine-2-thiones or to mixtures of both depending on the conditions used.We first subjected (S)- phenylalaninol 7c to reaction with CS2 in aqueous sodium hydroxide using the conditions reported by Roth and Schlump.8 The product consisted of a mixture of the oxazolidine-2-thione 8c and the thiazolidine-2-thione 9c but the former could be converted entirely into the latter by treatment with P2S5 (Scheme 2). Although this was obtained in low yield it had properties including the 13C NMR data (Table 1) in good agreement with those reported.7 When 9c was subjected to oxidation using a wide variety of oxidants and conditions complex mixtures of intractable products were obtained most likely owing to oxidative dimerisation through the 2- mercaptothiazoline tautomeric form a reaction previously reported for these compounds,9 followed by ring-opening and further degradation.Attention was therefore turned to N-substituted examples since it was these that were used in the successful oxidation method of Gaul and Fremuth;6 to facilitate later deprotection we chose the N-benzyl compounds. The required N-benzylamino alcohols 11a–c were readily obtained by condensation of 7a–c with benzaldehyde to give 10 followed by catalytic hydrogenation. When these were subjected to reaction with CS2 in aqueous sodium hydroxide under the same conditions as before the desired thiazolidinethiones 12a–c were obtained in moderate to good yield and no oxazolidinethiones were formed. This is in agreement with the work of Delaunay et al.7 where N-methylamino alcohols also gave exclusively the thiazolidinethiones.The colourless crystalline compounds 12a–c gave analytical and spectroscopic data in good agreement with expectations and the fully assigned 13C NMR spectra presented in Table 1 formed a highly consistent pattern. Attempted oxidation of 12 using peracetic acid under a variety of conditions did give the desired sulfones 5 but in disap- 2140 J. Chem. Soc. Perkin Trans. 1 1997 pointing yield and always accompanied by some of the thiazolidinones 13. The reagent of choice was found to be KMnO4 in a mixed phase system of CH2Cl2–water with both 0.1 equiv. benzyltriethylammonium chloride and 1 equiv. of benzoic acid as additives a system which has also recently proved valuable for the oxidation of 4,5-dihydrothiazoles to their 1,1-dioxides.10 Using this system the oxidation could be completely controlled with 3 equiv.of KMnO4 converting 12 cleanly into 13 and either 12 or 13 being converted into 5 using 5 and 2 equiv. KMnO4 respectively the yields in all reactions exceeding 60%. In view of problems later encountered in the pyrolysis (vide Scheme 2 Reagents i CS2 aq. NaOH; ii P2S5; iii PhCHO cat. TsOH; iv H2 Pd/C; v KMnO4 PhCO2H phase transfer catalyst CH2Cl2–H2O; vi MeI acetone; vii NaOMe MeOH; viii AcOOH R2 R3 HO H2N R2 R3 HO R1HN N S S R1 R2 R3 N N S O2 O R1 R2 R3 S O R1 R2 R3 HN S S R2 R3 HN O S R2 R3 R2 R3 HO N PhHC N+ S MeS H HN H HO3S 13 11 10 9 5 iii i 12 v v 8 v + 7 ii iv i I– vi vii viii (12d) 15 R1 R2 R3 a PhCH2 H Et b PhCH2 Pri H c PhCH2 PhCH2 H d —(CH2)3— H 14 (13d) infra) it was desirable to have an example of 5 in which R1 was linked to either R2 or R3 and 5d was therefore prepared by starting from (S)-prolinol 11d which reacted with CS2 in aqueous sodium hydroxide under the standard conditions to give 12d.In this case it was found to be preferable to obtain 5 via oxidation of 13 rather than directly from 12; 13d was obtained in good overall yield by reaction of 12d with methyl iodide to give 14 followed by treatment with sodium methoxide to give 13d. This unusual method for converting a thiazolidinethione to a thiazolidinone is based on a recently reported method for the thiazolinethione to thiazolinone transformation.11 Attempted oxidation of 13d using peracetic acid did give 5d but in low yield and this was accompanied by a new product which proved to be the sulfonic acid 15 resulting from hydrolytic ringopening of 5 decarboxylation and further oxidation.Again the permanganate–benzoic acid method provided the method of choice for conversion of 13d into 5d. Although the compounds 5a–d were perfectly stable under dry conditions and gave good analytical and spectroscopic data some hydrolysis as hinted at by the formation of 15 could be observed on prolonged storage. This is not surprising since a-oxo sulfones are notoriously elusive and in cases where they have been obtained they are readily hydrolysed.12 Acyclic carboxamido sulfones have been obtained before and are somewhat more resistant to hydrolysis.13 As shown in Table 1 the 13C NMR data for 12a–d 13a–d and 5a–d form a consistent pattern and the trends on going from 12 to 13 to 5 are somewhat surprising.The fall of ca. 25 ppm in the value for C-2 on going from 12 to 13 is as expected but the reason for the further fall of ca. 12 ppm for C-2 on going from 13 to 5 is not clear particularly when at the same time the values for C-5 increase by 17–20 ppm upon S-oxidation. The signals for the remaining ring carbon C-4 are also affected to a surprising degree by oxidation with falls of ca. 8 ppm associated with each oxidation step. Further confirmation of the five-membered ring structure of the compounds 5 as opposed to the isomeric sixmembered cyclic carbamic–sulfinic anhydride structure 19 was obtained by 33S NMR spectroscopy. The use of this technique to clarify a similar structural ambiguity has been described by Farrar et al.,14 and relies on the fact that the line widths in 33S NMR spectra are highly dependant on the degree of symmetry around the sulfur atom.Thus 5 would be expected to give a Table 1 13C NMR spectra of heterocycles 9 12 13 and 5 N Y X R1 R2 R3 5 4 2 dC 9c 12a 12b 12c 12d 13a 13b 13c 13d 5a 5b 5c 5d R1 H CH2Ph CH2Ph CH2Ph ](CH CH2Ph CH2Ph CH2Ph ](CH CH2Ph CH2Ph CH2Ph ](CH R2 CH2Ph H Pri CH2Ph 2)3] H Pri CH2Ph 2)3] H Pri CH2Ph 2)3] R3 H Et HH H Et HH H Et HH H X S SSS S OOO OOOO O Y S SSS SSSS S SO2 SO2 SO2 SO2 C-2a 200.5 197.0 197.4 196.7 191.1 172.0 172.9 171.8 169.8 159.8 160.6 159.5 157.7 C-4 65.1 67.7 71.0 67.5 71.9 59.1 62.0 59.5 63.0 51.3 54.5 51.7 52.6 C-5 39.7 31.7 26.9 32.2 35.8 29.9 24.9 30.4 33.2 47.2 42.7 47.9 53.8 R1 signals — 135.2 (4ry) 128.8 (2C) 127.9 127.7 (2C) 50.1 135.1 (4ry) 128.7 (2C) 127.8 (3C) 50.0 135.4 (4ry) 128.9 (2C) 128.2 128.0 (2C) 50.7 46.3 31.4 28.8 136.2 (4ry) 128.6 (2C) 127.7 (2C) 127.6 46.3 135.9 (4ry) 128.7 (2C) 128.0 (2C) 127.7 46.6 136.3 (4ry) 128.6 (2C) 128.0 (2C) 127.9 46.7 43.3 30.8 27.2 133.4 (4ry) 129.3 (2C) 128.8 128.2 (2C) 47.2 133.4 (4ry) 129.2 (2C) 128.7 128.2 (2C) 47.1 133.2 (4ry) 129.2 (2C) 129.0 128.5 (2C) 47.4 43.7 32.4 23.4 R2 and R3 signals 135.7 (4ry) 129.1 (2C) 129.0 (2C) 127.3 37.9 24.1 9.2 28.9 18.6 14.7 135.9 (4ry) 129.1 (2C) 129.0 (2C) 127.2 36.3 — 24.3 8.6 28.1 18.2 14.5 136.4 (4ry) 129.2 (2C) 128.7 (2C) 127.1 37.3 — 24.6 8.7 27.4 18.2 13.9 134.7 (4ry) 129.4 (2C) 129.3 (2C) 127.8 38.1 — a dC Values are given with reference to Me4Si as the internal standard.J. Chem. Soc. Perkin Trans. 1 1997 2141 relatively sharp signal while 19 would give a signal too broad to be observed.In the event the spectrum of 5c was readily obtained at natural abundance and consisted of a single signal of w1/2 130 Hz. The chemical shift of dS 26.5 with respect to aqueous Na2SO4 is in the expected range for cyclic sulfones,15 although no a-oxo sulfone has previously been observed. The sulfones 5a–c were subjected to flash vacuum pyrolysis (FVP) using a conventional flow system with a horizontal furnace tube operating at 1023 Torr and involving contact times of ª1–10 ms. Under these conditions all three compounds underwent complete reaction at the relatively mild temperature of 650 8C to give rather complex mixtures of products as shown in Table 2. It is disappointing to note that the desired extrusion of SO2 does occur but is accompanied by complete fragmentation to give the alkene 16 together with benzyl isocyanate 17 obtained largely in the form of its hydrolysis product 18 (Scheme 3).Pyrolysis of an authentic sample of 17 under the same conditions confirmed both that it does not undergo any further thermal reactions and that using our normal techniques it underwent substantial hydrolysis owing to adventi- Scheme 3 N S O2 O PhCH2 R2 R3 N O R2 R3 S O O R2 R3 PhCH2 N N• S O H R2 R3 PhCH2 Ph •O S R2 R3 H N PhCH HO S N R2 R3 S Ph R2 R3 N S Ph R3 N S Ph R2 O HO HN R2 R3 Ph PhCH2NH NHCH2Ph O 5a–c PhCH2NCO – SO2 FVP FVP 23 20 19 22 21 • – CO2 16 • 17 24 PhCOCl 7 H2O – H2O 18 P2S5 + – R3H – R2H Table 2 Products from FVP of thiazolidin-2-one 1,1-dioxides 5 at 650 8C (%) Starting material Product 16 17 18 21 22 23 PhC]] ] N PhCH2CH2Ph PhCH]] NCH2Ph PhCHO PhMe PriCHO EtCHO 5a 13 2 24 5 —3 19 10 524 —2 5b 12 — 10 444 16 67215 — 5c 18 — 15 3827 12 —45 —— tious moisture in the cold trap to give 18.It appears that the extrusion from 5a–c requires more forcing conditions as compared to 1 such that the b-lactam cannot survive intact. The formation of benzonitrile bibenzyl toluene N-benzylidenebenzylamine and benzaldehyde in all cases is probably associated with fragmentation of the N-benzyl group. The origin of the aliphatic aldehydes corresponding to R2/R3CHO is unclear. Most interesting however is the formation of small but signifi- cant quantities of the 2-phenyl-4,5-dihydrothiazoles 21 and their aromatization products 22 and 23. The identity of these unexpected products was demonstrated by comparison with authentic samples prepared by reaction of 7a–c with benzoyl chloride to give 24 followed by cyclisation with P4S10.10 Heating 21b,c with sulfur at 200–210 8C afforded samples of 23b,c while 2-phenylthiazole 23a (=22b,c) was prepared by a literature method16 and these were identical with the pyrolysis products.The mechanism of this unprecedented heterocyclic transformation is believed to involve the sequence of steps shown in Scheme 3 resulting in the required net loss of CO2 and H2O. Ring expansion to the cyclic sulfinic–carbamic anhydride 19 a process well known in the pyrolysis of cyclic sulfones,1 allows ready loss of CO2. Rearrangement of the resulting diradical and intramolecular abstraction of the benzylic CH gives the imino sulfenic acid 20 which can then lose water to afford 21.Overall the process is somewhat reminiscent of the pyrolysis of benzothiophene 1,1-dioxide to give benzothiete,17 which also involves loss of CO2 and initial ring expansion. In an attempt to prevent the fragmentation to alkene and isocyanate we then examined the pyrolysis of 5d in which the routes leading to 21–23 are also impossible. This underwent complete reaction at the lower temperature of 600 8C but the product consisted of a complex mixture of products which could not be identified. The presence of alkene signals in the NMR spectra pointed to ring-opening and this might be expected as shown in Scheme 4 since the diradical resulting from loss of SO2 can readily open to give the pentenyl isocyanate 25 while additional loss of CO can lead to pentenylnitrene 26.Synthesis of an authentic sample of 27 an alternative possible product from the diradical shown confirmed that it was not present. Both 25 and 26 are expected to be highly reactive and can undergo a variety of secondary reactions either in the furnace or in the cold trap so the complex mixture produced is not surprising. In a final attempt to obtain a b-lactam 5a–c were subjected to photolysis in a variety of solvents. In contrast to the isomeric compounds 1,3 they were found to be photochemically inert and the only new product obtained in low yield from 5c was the amino sulfonic acid 28 resulting from hydrolysis by adventitious moisture decarboxylation and oxidation. In conclusion it is clear that the thiazolidin-2-one 1,1-dioxides 5 are not suitable precursors for the thermal or photochemical generation of Scheme 4 N S O2 O H OCN N N H Me :N – SO2 • • – SO2 – CO 5d 27 25 26 HO3S PhCH2NH CH2Ph H 28 2142 J.Chem. Soc. Perkin Trans. 1 1997 b-lactams in contrast to the isomeric thiazolidin-4-one 1,1- dioxides 1. Due to subtle differences between the two ring systems the more severe conditions required to achieve SO2 extrusion in the former case lead to complete fragmentation to an alkene and isocyanate. The unexpected formation of 21–23 is however of some mechanistic interest. Experimental Melting points were determined on a Reichert hot-stage microscope and are uncorrected. Infrared spectra were recorded for solids as Nujol mulls and for liquids as thin films on a Perkin- Elmer 1420 spectrophotometer. NMR spectra were recorded for 1H at 80 MHz on a Bruker WP80 instrument or at 300 MHz on a Bruker AM300 instrument for 13C at 20 MHz on a Varian CFT 20 or at 75 MHz on a Bruker AM300 instrument and for 33S at 38 MHz on a Bruker MSL500 spectrometer.Spectra were obtained for solutions in CDCl3 unless otherwise indicated with Me4Si as internal reference for 1H and 13C and aqueous Na2SO4 as external reference for 33S. Chemical shifts are reported in ppm relative to the reference and coupling constants J are given in Hz. In the assignments for the 13C NMR data 4ry refers to quaternary carbon. Mass spectra were obtained on an A.E.I. MS902 instrument using electron impact at 70 eV. GC–MS was performed with a Hewlett Packard 5890A chromatograph coupled to a Finnigan Incos 50 mass spectrometer. Optical rotations were measured on an Optical Activity AA1000 polarimeter and are given in units of 1021 deg cm2 g21.The amino alcohols 7a–c and 11d were prepared by reduction of the corresponding amino acids or were commercially available. Preparation of 2-benzylideneamino alcohols 10 Benzaldehyde (24.4 g 230 mmol) was added to a stirred solution of the appropriate amino alcohol 7 (220 mmol) in toluene (250 cm3) and the mixture heated under reflux for 1 h using a Dean–Stark separator. Evaporation yielded the product which was recrystallised from hexane. Using this method the following compounds were prepared. (2R)-2-Benzylideneaminobutan-1-ol 10a. (2R)-2-Aminobutan- 1-ol 7a gave 10a as colourless needles (77%) mp 57– 58 8C (Found C 74.6; H 8.8; N 7.9. C11H15NO requires C 74.5; H 8.5; N 7.9%); [a]D 20 137.8 (c 1.0 in CH2Cl2); nmax/cm21 3280 (OH) 1645 (CN) 1060 1000 780 and 705; dH 8.20 (1 H s) 7.65 (2 H m) 7.35 (3 H m) 3.78 (1 H half AB pattern of d J 12 10) 3.72 (1 H half AB pattern of d J 12 4) 3.18 (1 H m) 2.86 (1 H br s) 1.60 (2 H m) and 0.85 (3 H t J 7); dC 162.0 (CH) 135.8 (4ry) 130.7 (CH) 128.5 (2 CH) 128.3 (2 CH) 74.7 (CH) 66.0 (CH2) 25.0 (CH2) and 10.7 (CH3); m/z 177 (M1 15%) 176 (50) 146 (100) 132 (25) 118 (30) 104 (50) 91 (85) 77 (35) and 41 (60).(2S)-2-Benzylideneamino-3-methylbutan-1-ol 10b. (2S)-2- Amino-3-methylbutan-1-ol 7b gave 10b as colourless crystals (77%) mp 70–71 8C (Found C 75.2; H 9.0; N 7.3. C12H17NO requires C 75.3; H 9.0; N 7.3%); [a]D 25 283.3 (c 0.3 in CHCl3); nmax/cm21 3700–2400 (br OH) 1640 1470 1450 1380 1260 1220 1060 and 1020; dH 8.29 (1 H s) 7.85–7.6 (2 H m) 7.6–7.3 (3 H m) 3.80 (2 H m) 3.2–2.8 (1 H m) 1.90 (1 H octet J 7) 0.95 (3 H d J 7) and 0.90 (3 H d J 7); dC 161.7 (C]] N) 136.0 (4ry) 130.4 (CH) 128.4 (4 CH) 79.2 (CH) 64.1 (CH2) 30.0 (CH) 19.7 (CH3) and 19.2 (CH3); m/z 190 (M 2 H1 5%) 189 (2) 160 (100) 148 (70) 130 (25) and 118 (35).(2S)-2-Benzylideneamino-3-phenylpropan-1-ol 10c. (2S)-2- Amino-3-phenylpropan-1-ol 7c gave 10c as colourless prisms (64%) mp 78–80 8C (Found C 80.1; H 7.2; N 5.8. C16H17NO requires C 80.0; H 7.1; N 5.8%); [a]D 25 2215.6 (c 2.0 in CHCl3); nmax/cm21 3600–2700 (br OH) 1640 1490 1450 1380 1220 1030 and 700; dH 7.98 (1 H s) 7.7–7.55 (2 H m) 7.45–7.3 (3 H m) 7.25–7.1 (5 H m) 3.85 (1 H half of AB pattern of d J 10 6) 3.70 (1 H half of AB pattern of d J 10 4) 3.7–3.4 (1 H m) 3.00 (1 H half of AB pattern of d J 14 5) 2.80 (1 H half of AB pattern of d J 14 8) and 2.26 (1 H br s); dC 162.4 (C]] N) 138.6 (4ry) 135.6 (4ry) 130.6 (CH) 129.6 (2 CH) 128.4 (2 CH) 128.2 (4 CH) 126.0 (CH) 74.4 (CH) 65.6 (CH2) and 38.9 (CH2); m/z 208 (M1 2 CH2OH 8%) 148 (M1 2 CH2Ph 50) 130 (12) 128 (32) 127 (35) and 91 (100).Preparation of 2-benzylamino alcohols 11 A solution of the appropriate benzylideneamino alcohol 10 (0.52 mol) and 5% palladium/charcoal catalyst (3.0 g) in ethyl acetate (500 cm3) was stirred vigorously in the presence of hydrogen gas (12 dm3 0.54 mol) at room temp. for 24 h. The solution was then filtered through Celite and evaporated to afford the product. Using this method the following compounds were prepared. (2R)-2-Benzylaminobutan-1-ol 11a. (2R)-2-Benzylideneaminobutan- 1-ol 10a gave 11a following recrystallisation from hexane as a colourless solid (77%) mp 74–75 8C (Found C 73.4; H 9.6; N 7.7.C11H17NO requires C 73.7; H 9.6; N 7.8%); [a]D 20 228.5 (c 1.0 in CH2Cl2); nmax/cm21 3400–3000 (OH) 3280 (NH) 1060 865 745 and 700; dH 7.30 (5 H m) 3.80 and 3.72 (2 H AB pattern J 14) 3.62 (1 H half of AB pattern of d J 10 4) 3.35 (1 H half AB pattern of d J 10 6) 2.62 (1 H m) 2.40 (2 H br s) 1.6–1.4 (2 H m) and 0.90 (3 H t J 7); dC 140.3 (4ry) 128.5 (2 CH) 128.1 (2 CH) 127.1 (CH) 62.6 (CH2) 59.8 (CH) 51.0 (CH2) 24.2 (CH2) and 10.4 (CH3); m/z 179 (M1 1%) 148 (100) 106 (55) 91 (100) 77 (50) 65 (75) and 56 (70). (2S)-2-Benzylamino-3-methylbutan-1-ol 11b. (2S)-2-Benzylideneamino- 3-methylbutan-1-ol 10b gave 11b following Kugelrohr distillation as a colourless oil (80%) bp (oven temp.) 106– 108 8C at 0.4 Torr (lit.,18 103–107 8C at 0.2 Torr).(2S)-2-Benzylamino-3-phenylpropan-1-ol 11c. (2S)-2-Benzylideneamino- 3-phenylpropan-1-ol 10c gave 11c following recrystallisation from hexane–ethyl acetate (5 1) as colourless prisms (71%) mp 124–126 8C (Found C 79.6; H 8.0; N 5.7. C16H19NO requires C 79.6; H 7.9; N 5.8%); [a]D 25 249.8 (c 2.0 in CHCl3); nmax/cm21 3700–2400 (br OH) 1640 1490 1450 1400 1220 1110 1030 910 and 700; dH 7.25 (11 H m) 3.9–3.6 (1 H m) 3.80 (2 H s) 3.60 (1 H half AB pattern of d J 10 4) 3.40 (1 H half AB pattern of d J 10 5) 2.90 and 2.70 (2 H AB pattern of d J 8 4) and 2.80 (1 H br s); dC 139.9 (4ry) 138.8 (4ry) 129.2 (2 CH) 128.4 (4 CH) 128.0 (2 CH) 126.9 (CH) 126.2 (CH) 62.6 (CH2) 59.7 (CH) 51.1 (CH2) and 37.8 (CH2); m/z 242 (M 1 H1 1.2%) 241 (M1 1) 210 (8) 150 (20) and 91 (100).Preparation of thiazolidine-2-thiones 9 and 12 A mixture of the appropriate amino alcohol (45 mmol) 2 M sodium hydroxide (150 cm3) and carbon disulfide (9.8 cm3 12.4 g 163 mmol) was stirred at room temp. for 20 h. A further portion of carbon disulfide (5.0 cm3 6.3 g 83 mmol) was added and the solution stirred for an additional 4 h. The mixture was extracted with CH2Cl2 and the organic layer washed with water dried and evaporated to afford the product. Using this method the following compounds were prepared. (4S)-4-Benzylthiazolidine-2-thione 9c. (2S)-2-Amino-3- phenylpropan-1-ol 7c (8.0 g 52 mmol) gave a mixture of the desired product 9c and the corresponding oxazolidine-2-thione 8c. This was dissolved in toluene (200 cm3) and heated under reflux with P2S5 (20 g 90 mmol) for 48 h.Filtration and evaporation followed by column chromatography [SiO2 diethyl ether– petroleum (bp 40–60 8C) 1 1] gave a red solid which was recrystallised from hexane–ethyl acetate (5 1) to give the product as red needles (18%) mp 79–80 8C (lit.,7 84–85 8C) (Found C 57.4; H 5.3; N 6.6. C10H11NS2 requires C 57.4; H 5.3; N 6.7%); [a]D 25 2112.2 (c 1.7 in CHCl3); nmax/cm21 3480 1470 1290 1250 1220 1140 1040 1010 960 and 700; dH 8.40 (1 H br s) 7.4–7.2 (3 H m) 7.2–7.1 (2 H m) 4.46 (1 H quintet J 10) 3.50 and 3.26 (2 H AB pattern of d J 14 10) 3.05 and 2.93 (2 H AB pattern of d J 12 10); dC see Table 1; m/z 209 (M1 40%) 182 (12) 167 (15) 146 (93) 132 (12) 118 (27) 117 (20) and 91 (100). J. Chem. Soc. Perkin Trans. 1 1997 2143 (4R)-3-Benzyl-4-ethylthiazolidine-2-thione 12a.(2R)-2-Benzylaminobutan- 1-ol 11a gave 12a following recrystallisation from hexane–ethyl acetate (2 1) as colourless crystals (72%) mp 61–62 8C (Found C 60.7; H 6.1; N 5.9. C12H15NS2 requires C 60.7; H 6.4; N 5.9%); [a]D 20 191.3 (c 1.0 in CH2Cl2); nmax/cm21 3060 3040 1475–1425 1225 1175 1025 (CS) 760 and 700; dH 7.30 (5 H m) 5.75 and 4.25 (2 H AB pattern J 17) 4.00 (1 H m) 3.35 (1 H half AB pattern of d J 10 8) 2.96 (1 H half AB pattern of d J 10 5) 1.77 (2 H m) and 0.92 (3 H t J 7); dC see Table 1; m/z 237 (M1 15%) 148 (100) 132 (5) 121 (10) 104 (5) 91 (70) and 65 (25). (4S)-3-Benzyl-4-isopropylthiazolidine-2-thione 12b. (2S)- Benzylamino-3-methylbutan-1-ol 11b gave 12b following recrystallisation from hexane–ethyl acetate (5 1) with cooling (220 8C) as colourless crystals (36%) mp 77–78 8C (Found C 62.2; H 6.9; N 5.6.C13H17NS2 requires C 62.1; H 6.8; N 5.6%); [a]D 25 2143.1 (c 0.5 in CHCl3); nmax/cm21 1460 1450 1330 1240 1220 1200 1180 1130 1040 990 and 960; dH 7.40 (5 H s) 6.00 and 4.14 (2 H AB pattern J 16) 4.05 (1 H m) 3.20 (1 H half AB pattern of d J 11 9) 3.05 (1 H half AB pattern of d J 11 6) 2.34 (1 H septet of d J 7 4) 0.95 (3 H d J 7) and 0.90 (3 H d J 7); dC see Table 1; m/z 251 (M1 100%) 208 (15) 187 (24) 148 (82) 144 (24) and 91 (35). (4S)-3,4-Dibenzylthiazolidine-2-thione 12c. (2S)-2-Benzylamino- 3-phenylpropan-1-ol 11c gave 12c following recrystallisation from hexane–ethyl acetate (3 1) as colourless needles (53%) mp 137–139 8C (Found C 68.4; H 5.55; N 4.65. C17H17NS2 requires C 68.2; H 5.7; N 4.7%); [a]D 25 225.8 (c 1.6 in CH2Cl2); nmax/cm21 1490 1450 1420 1350 1300 1220 1170 1080 1030 920 and 700; dH 7.4–7.2 (8 H m) 7.1–7.0 (2 H m) 5.82 and 4.20 (2 H AB pattern J 16) 4.30–4.15 (1 H m) 3.20 (1 H half AB pattern of d J 12 8) 3.15 (1 H half AB pattern of d J 14 5) 2.86 (1 H half AB pattern of d J 12 10) and 2.83 (1 H half AB pattern of d J 14 10); dC see Table 1; m/z 299 (M1 42%) 277 (20) 238 (10) 208 (100) 148 (92) and 117 (31).(5S)-3-Thia-1-azabicyclo[3.3.0]octane-2-thione 12d. (2S)-2- Hydroxymethylpyrrolidine 11d gave 12d following recrystallisation from ethanol as colourless crystals (49%) mp 130– 131 8C (lit.,19 132–133 8C) (Found C 45.1; H 5.5; N 8.78. C6H9NS2 requires C 45.2; H 5.7; N 8.8%); [a]D 20 2159.8 (c 1.0 in CH2Cl2); nmax/cm21 1360 1340 1245 1210 1180 1055 1030 (CS) 940 and 850; dH 4.63 (1 H m) 3.60 (1 H m) 3.48 (1 H m) 3.32 (2 H dd J 7 2) 2.5–2.3 (2 H m) 2.20 (1 H m) and 1.80 (1 H m); dC see Table 1; m/z 159 (M1 70%) 126 (5) 118 (10) 85 (30) 72 (25) 67 (50) 45 (35) and 41 (100).Preparation of thiazolidin-2-ones 13a–c A solution of the appropriate thiazolidinethione 12 (5 mmol) benzoic acid (0.62 g 5 mmol) and benzyltriethylammonium chloride (0.11 g 0.5 mmol) in dichloromethane (50 cm3) was stirred vigorously with a solution of potassium permanganate (2.37 g 15 mmol) in water (100 cm3) for 3 h. Sufficient solid sodium metabisulfite was added to decolourise the mixture which was then filtered through Celite the organic layer was separated and the aqueous layer washed with dichloromethane (3 × 50 cm3). The combined organic extracts were washed with 1 M hydrazine dihydrochloride followed by aqueous sodium carbonate dried with anhydrous magnesium sulfate and evaporated to give the product.Using this method the following compounds were prepared. (4R)-3-Benzyl-4-ethylthiazolidin-2-one 13a. (4R)-3-Benzyl-4- ethylthiazolidine-2-thione 12a gave 13a following Kugelrohr distillation as a pale green oil (76%) bp (oven temp.) 215 8C at 0.7 Torr (Found C 65.6; H 7.0; N 6.6%; M 221.0859. C12H15NOS requires C 65.1; H 6.8; N 6.3%; M 221.0874); [a]D 20 226.1 (c 1.07 in CH2Cl2); nmax/cm21 2970–2940 1670 (CO) 1460 1410 1230 and 710; dH 7.35–7.2 (5 H m) 4.96 and 4.00 (2 H AB J 15) 3.55 (1 H m) 3.26 (1 H half AB pattern of d J 11 8) 2.93 (1 H half AB pattern of d J 11 6) 1.75–1.5 (2 H m) and 0.86 (3 H t J 7); dC see Table 1; m/z 221 (M1 90%) 192 (85) 165 (20) 122 (25) 104 (70) 91 (100) and 65 (80).(4S)-3-Benzyl-4-isopropylthiazolidin-2-one 13b. (4S)-3- Benzyl-4-isopropylthiazolidine-2-thione 12b gave 13b following Kugelrohr distillation as a pale yellow solid (43%) mp 33– 35 8C bp (oven temp.) 185 8C at 0.7 Torr (Found C 66.5; H 7.7; N 6.1%; M 235.1026. C13H17NOS requires C 66.3; H 7.3; N 6.0%; M 235.1031); [a]D 20 134.0 (c 1.02 in CH2Cl2); nmax/cm21 3025 2964 1723 1664 (CO) 1455 1435 1260 1215 and 705; dH 7.35–7.2 (5 H m) 5.10 and 3.90 (2 H AB pattern J 17) 3.57 (1 H m) 3.10 (1 H half AB pattern of d J 13 9) 3.03 (1 H half of AB pattern of d J 13 7) 2.20 (1 H m) 0.87 (3 H d J 9) and 0.85 (3 H d J 9); dC see Table 1; m/z 235 (M1 15%) 192 (45) 176 (5) 133 (10) 105 (5) 91 (100) and 77 (5). (4S)-3,4-Dibenzylthiazolidin-2-one 13c.(4S)-3,4-Dibenzylthiazolidine- 2-thione 12c gave 13c following Kugelrohr distillation as a colourless oil which formed colourless prisms with time (45%) bp (oven temp.) 225 8C at 0.3 Torr; mp 70–71 8C (Found C 72.2; H 6.3; N 4.8. C17H17NOS requires C 72.0; H 6.1; N 4.9%); [a]D 25 111.8 (c 0.9 in CHCl3); nmax/cm21 1650 1490 1450 1440 1400 1350 1200 1080 1030 970 and 930; dH 7.4– 7.2 (8 H m) 7.08 (2 H m) 5.08 and 4.00 (2 H AB pattern J 16) 3.80 (1 H m) 3.15–3.05 (2 H m) 2.92 (1 H half AB pattern of d J 12 4) and 2.77 (1 H half AB pattern of d J 12 8); dC see Table 1; m/z (CI) 284 (M 1 H1 100%) 192 (46) 108 (7) 91 (65) and 65 (7). (5S)-2-Methylthio-3-thia-1-azabicyclo[3.3.0]oct-1-en-1-ium iodide 14 A solution of (5S)-3-thia-1-azabicyclo[3.3.0]octane-2-thione 12d (4.0 g 25 mmol) and methyl iodide (15.6 cm3 35.5 g 250 mmol) in acetone (110 cm3) was stirred for 16 h at room temp.The resulting precipitate was filtered off and washed with diethyl ether. The filtrate was concentrated and a second crop of the product filtered off and washed with diethyl ether. The solids were combined to yield the product (6.74 g 90%) as a pale yellow powder mp 111–112 8C (Found C 27.8; H 3.9; N 4.6. C7H12INS2 requires C 27.9; H 4.0; N 4.7%); [a]D 20 2256.5 (c 1.66 in CH2Cl2); nmax/cm21 1555 1300 1200 1170 and 950; dH 5.20 (1 H m) 3.87 (2 H m) 3.68 (1 H m) 3.58 (1 H m) 2.77 (3 H s) 2.46 (2 H m) and 2.25–2.10 (2 H m); dC 186.7 (4ry) 77.5 (CH) 49.3 (CH2) 38.5 (CH2) 29.5 (CH2) 29.4 (CH2) and 19.8 (CH3); m/z 159 (M1 2 MeI 30%) 126 (5) 118 (10) 85 (30) 82 (10) and 67 (50). (5S)-3-Thia-1-azabicyclo[3.3.0]octan-2-one 13d (5S)-2-Methylthio-3-thia-1-azabicyclo[3.3.0]oct-1-en-1-ium iodide 14 (22.58 g 75 mmol) was added to a solution of sodium methoxide (75 mmol) in methanol (200 cm3) and the mixture stirred for 16 h at room temp.Water (400 cm3) was added and the mixture extracted with CH2Cl2. The combined organic layers were washed with water dried and evaporated to yield a yellow solid. Recrystallisation of this from diethyl ether–ethyl acetate with cooling (220 8C) afforded the product (8.37 g 78%) as colourless crystals mp 70–71 8C (Found C 50.1; H 6.3; N 9.6. C6H9NOS requires C 50.3; H 6.3; N 9.8%); [a]D 20 235.4 (c 1.0 in CH2Cl2); nmax/cm21 3320 1700 (CO) 1385 930 and 890; dH 4.22 (1 H m) 3.55 (1 H m) 3.38 (1 H half AB pattern of d J 12 9) 3.22 (1 H half AB pattern of d J 12 10) 3.17 (1 H m) 2.3–2.0 (3 H m) and 1.62 (1 H m); dC see Table 1; m/z 143 (M1 30%) 114 (5) 85 (5) 80 (5) 74 (20) 70 (30) and 55 (100).Preparation of thiazolidin-2-one 1,1-dioxides 5 Exactly the same method was used as described for the thiazolidin-2-ones above except that the quantity of potassium permanganate was increased to 3.95 g (25 mmol) and the products were recrystallised from diethyl ether–CH2Cl2 (1 1). Using this method the following compounds were prepared. (4R)-3-Benzyl-4-ethylthiazolidin-2-one 1,1-dioxide 5a. (4R)-3- Benzyl-4-ethylthiazolidine-2-thione 12a gave 5a as colourless 2144 J. Chem. Soc. Perkin Trans. 1 1997 crystals (72%) mp 102–103 8C (Found C 56.8; H 6.0; N 5.5. C12H15NO3S requires C 56.9; H 6.0; N 5.5%); [a]D 20 147.0 (c 0.1 in CH2Cl2); nmax/cm21 3420 1710 (CO) 1320 1140 940 850 755 and 700; dH 7.4–7.3 (3 H m) 7.25–7.2 (2 H m) 5.10 and 4.22 (2 H AB pattern J 15) 3.70 (1 H m) 3.35 (1 H half AB pattern of d J 14 8) 3.15 (1 H half AB pattern of d J 14 4) 1.95 (1 H m) 1.76 (1 H m) and 0.94 (3 H t J 8); dC see Table 1; m/z 189 (M1 2 SO2 2%) 161 (2) 133 (50) 105 (30) 91 (100) and 77 (10).This product could alternatively be prepared from (4R)-3- benzyl-4-ethylthiazolidin-2-one 13a using 2 equiv. of KMnO4. (4S)-3-Benzyl-4-isopropylthiazolidin-2-one 1,1-dioxide 5b. (4S)-3-Benzyl-4-isopropylthiazolidine-2-thione 12b gave 5b as pale yellow needles (33%) mp 114–115 8C (Found C 58.4; H 6.4; N 5.2. C13H17NO3S requires C 58.4; H 6.4; N 5.2%); [a]D 20 239.6 (c 1.02 in CH2Cl2); nmax/cm21 3420 1720 (CO) 1325 and 1135 (SO2) 760 740 and 700; dH 7.4–7.3 (3 H m) 7.3–7.2 (2 H m) 5.10 and 4.18 (2 H AB pattern J 15) 3.77 (1 H m) 3.26 (1 H half AB pattern of d J 14 8) 3.12 (1 H half AB pattern of d J 14 6) 2.38 (1 H m) 0.89 (3 H d J 7) and 0.85 (3 H d J 7); dC see Table 1; m/z 203 (M1 2 SO2 15%) 160 (10) 133 (90) 105 (30) 91 (100) and 77 (5).(4S)-3,4-Dibenzylthiazolidin-2-one 1,1-dioxide 5c. (4S)-3,4- Dibenzylthiazolidine-2-thione 12c gave 5c as colourless needles (67%) mp 143–144 8C (Found C 65.0; H 5.4; N 4.4. C17H17NO3S requires C 64.7; H 5.4; N 4.4%); [a]D 25 222.6 (c 0.7 in CHCl3); nmax/cm21 1730 (CO) 1490 1450 1420 1330 1220 1140 and 770; dH 7.45–7.35 (3 H m) 7.3–7.2 (5 H m) 7.07 (2 H m) 5.16 and 4.20 (2 H AB pattern J 16) 3.90 (1 H m) 3.34 (1 H half AB pattern of d J 16 6) 3.20 (1 H half AB pattern of d J 12 4) 3.03 (1 H half AB pattern of d J 12 8) and 2.88 (1 H half AB pattern of d J 16 10); dC see Table 1; dS 26.5 (w1/2 130 Hz); m/z 316 (M 1 H1 1%) 251 (M1 2 SO2 7) 192 (8) 176 (19) 160 (28) 134 (12) 118 (38) and 91 (100).(5S)-3-Thia-1-azabicyclo[3.3.0]octan-2-one 3,3-dioxide 5d. (5S)-3-Thia-1-azabicyclo[3.3.0]octan-2-one 13d gave 5d as colourless crystals (70%) mp 175–176 8C (Found C 41.1; H 5.2; N 7.95. C6H9NO3S requires C 41.1; H 5.2; N 8.0%); [a]D 20 130.7 (c 0.104 Me2SO); nmax/cm21 3440 1740 (CO) 1320 and 1130 (SO2) and 1160; dH(CD2Cl2) 3.95 (1 H m) 3.80 (1 H dd J 13 6) 3.55 (2 H m) 3.02 (1 H dd J 13 9) 2.39 (1 H m) 2.20 (1 H m) 2.02 (1 H m) and 1.56 (1 H qd J 12 8); dC see Table 1; m/z 111 (M1 2 SO2 25%) 82 (10) 68 (80) 67 (100) 55 (70) and 53 (55). An attempt to prepare 5d by oxidation of 13d using 32% peroxyacetic acid in acetic acid,6 gave the desired product in 37% yield but this was now accompanied by (2S)-pyrrolidine- 2-methanesulfonic acid 15 (24%) as a colourless powder mp 260 8C (decomp.) (Found C 36.0; H 6.6; N 8.2.C5H11NO3S requires C 36.4; H 6.7; N 8.5%); [a]D 20 131.2 (c 1.0 in H2O); nmax/cm21 2900–2400 1377 1172 and 1045; dH(CD3SOCD3) 8.90 (1 H br s) 8.40 (1 H br s) 3.75 (1 H m) 3.15 (2 H m) 2.90 (1 H m) 2.87 (1 H d J 2) 2.10 (1 H m) 1.90 (1 H m) 1.80 (1 H m) and 1.60 (1 H m); dC(CD3SOCD3) 56.4 (CH) 51.8 (CH2) 44.8 (CH2) 29.9 (CH2) and 22.7 (CH2); m/z 165 (M1 5%) 157 (5) 122 (5) 111 (10) 97 (10) 84 (55) and 44 (90). FVP of thiazolidin-2-one 1,1-dioxides 5 The apparatus used is similar to one which has been illustrated and described recently.20 The sample was volatilised from a tube in a Büchi Kugelrohr oven through a 30 × 2.5 cm horizontal fused quartz tube.This was heated externally by a Carbolite Eurotherm tube furnace MTF-12/38A to a temperature of 600– 650 8C the temperature being monitored by a Pt/Pt–13%Rh thermocouple situated at the centre of the furnace. The products were collected in a U-shaped trap cooled in liquid nitrogen. The whole system was maintained at a pressure of 1–2 × 1023 Torr by an Edwards Model E2M5 high capacity rotary oil pump the pressure being measured by a Pirani gauge situated between the cold trap and the pump. Under these conditions the contact time in the hot zone was estimated to be ª10 ms. After the material had all sublimed the products were recovered directly from the cold trap and analysed by 1H and 13C NMR spectroscopy and GC–MS the identity of the products being determined by comparison with authentic samples.Yields were determined by calibration of the 1H NMR spectra by adding an accurately weighed quantity of a solvent such as CH2Cl2 and comparing integrals a procedure estimated to be accurate to ±10% or for products such as benzonitrile which did not show a distinctive NMR signal from the GC integrals. Pyrolysis of 5a. 5a (0.10 g 650 8C) gave a yellow oil. Careful analysis of the 13C and 1H NMR spectra and GC–MS showed eleven compounds to be present but-1-ene 16a (13%) benzyl isocyanate 17 (2%) dibenzylurea 18 (24%) 4-ethyl-2-phenyl- 4,5-dihydrothiazole 21a (5%) 2-phenylthiazole 23a (3%) benzonitrile (19%) bibenzyl (10%) N-benzylidenebenzylamine (5%) benzaldehyde (2%) toluene (4%) and propanal (2%).Pyrolysis of 5b. 5b (103 mg 650 8C) afforded a yellow oil in the cold trap. Analysis of the 13C and 1H NMR spectra and GC–MS showed eleven compounds to be present 3-methylbut- 1-ene 16b (12%) dibenzylurea 18 (10%) 4-isopropyl-2-phenyl- 4,5-dihydrothiazole 21b (4%) 2-phenylthiazole 22b (4%) 4- isopropyl-2-phenylthiazole 23b (4%) benzonitrile (16%) bibenzyl (6%) N-benzylidenebenzylamine (7%) benzaldehyde (2%) toluene (1%) and 2-methylpropanal (5%). Pyrolysis of 5c. 5c (118 mg 650 8C) afforded a yellow oil at the furnace exit and in the cold trap. Analysis of the 13C and 1H NMR spectra and GC–MS showed the oil to contain nine products allylbenzene 16c (18%) dibenzylurea 18 (15%) 4- benzyl-2-phenyl-4,5-dihydrothiazole 21c (3%) 2-phenylthiazole 22c (8%) 4-benzyl-2-phenylthiazole 23c (2%) bibenzyl (12%) benzonitrile (7%) benzaldehyde (4%) and toluene (5%).Pyrolysis of 5d. 5d (120 mg 600 8C) afforded a yellow oil in the cold trap. The 13C and 1H NMR spectra showed a large number of compounds to be present but identification proved inconclusive. GC–MS analysis showed a major product with m/z 83 (C5H9N) but examination of the 13C NMR spectrum showed that the signals for the likely product 3,4-dihydro-5- methyl-2H-pyrrole 27 were absent. Synthesis of authentic samples of FVP products Preparation of 3-methylbut-1-ene 16b. The FVP of isoamyl acetate (2.5 g 19 mmol 750 8C 7.0 × 1023 Torr) afforded 3- methylbut-1-ene (0.2 g 15%); dH 5.8–5.75 (1 H m) 5.0–4.85 (2 H m) 2.28 (1 H m) and 0.98 (6 H d J 8); dC 146.0 (CH) 111.1 (CH2) 32.0 (CH) and 22.0 (2 CH3).Flash vacuum pyrolysis of benzyl isocyanate 17. The FVP of benzyl isocyanate (0.20 g 650 8C 1.0 × 1023 Torr) produced no change in the starting compound. Upon standing overnight the liquid solidified to afford dibenzylurea 18 (0.18 g 99%); dH 7.4–7.1 (10 H m) 6.5 (2 H br s) and 4.25 (4 H d J 4); dC(CD3SOCD3) 158.1 (CO) 140.7 (2 4ry) 128.1 (4 CH) 126.9 (4 CH) 126.5 (2 CH) and 42.9 (2 CH2). 4,5-Dihydrothiazoles 21 and thiazoles 22 and 23. The 4,5- dihydrothiazoles 21a–c were prepared as previously described 10 by acylation of the appropriate amino alcohol 7 with benzoyl chloride to give 24 followed by treatment with P2S5. These were then used to obtain the 2,4-disubstituted thiazoles 22a 23b and 23c by treatment with sulfur.10 2-Phenylthiazole 23a was prepared by the method of Lawson and Searle,16 as a colourless oil (25%) bp (oven temp.) 160 8C at 1.0 Torr (lit.,16 267–279 8C at 760 Torr); dH 7.95–7.90 (2 H m) 7.81 (1 H d J 3) 7.35–7.30 (3 H m) and 7.20 (1 H d J 3); dC 168.2 (4ry) 143.6 (CH) 133.5 (4ry) 129.9 (CH) 128.9 (2 CH) 126.5 (2 CH) and 118.7 (CH).Preparation of N-benzylidenebenzylamine. Benzaldehyde (5.43 g 51.2 mmol) was added to a stirred solution of benzylamine (5.48 g 51.2 mmol) in toluene (150 cm3). Heating under reflux for 1 h using a Dean–Stark separator followed by evaporation J. Chem. Soc. Perkin Trans. 1 1997 2145 of the solution afforded a yellow oil which was Kugelrohr distilled to yield N-benzylidenebenzylamine (9.0 g 90%) as a colourless oil bp (oven temp.) 175 8C at 0.5 Torr (lit.,21 200–202 8C at 10–20 Torr); dH 8.25 (1 H s) 7.7 (2 H m) 7.35–7.1 (8 H m) and 4.7 (2 H s); dC 162.6 (CH) 140.1 (4ry) 136.9 (4ry) 131.4 (CH) 129.3 (2 CH) 129.2 (2 CH) 129.0 (2 CH) 128.7 (2 CH) 127.7 (CH) and 65.6 (CH2).Preparation of 3,4-dihydro-5-methyl-2H-pyrrole 27. An ethereal solution of methyllithium (1.4 M; 37.5 cm3 50 mmol) was cooled to 220 8C and a solution of N-vinylpyrrolidin-2-one (5.0 g 45 mmol) dissolved in diethyl ether (50 cm3) was added dropwise over a period of 2 min. The mixture was stirred for a further 2 min at 220 8C and then 1 M hydrochloric acid (70 cm3) was added and the mixture stirred for an additional 2 min. The organic layer was separated and extracted with dilute hydrochloric acid the combined aqueous layers were washed with diethyl ether and then treated with aqueous sodium hydroxide until pH 10 was reached.The imine was extracted with CH2Cl2 and the extracts combined dried evaporated and Kugelrohr distilled to afford 3,4-dihydro-5-methyl-2H-pyrrole (1.83 g 49%) as a colourless oil bp (oven temp.) 50 8C at 14 Torr (lit.,22 103–105 8C at 760 Torr); dH 3.38 (2 H m) 2.10 (2 H t J 9) 1.66 (3 H s) and 1.50 (2 H quintet J 9); dC 174.7 (4ry) 61.1 (CH2) 38.7 (CH2) 23.0 (CH2) and 19.7 (CH3). Photolysis of 5c A solution of 5c (20 mg) in [2H6]acetone (0.5 cm3) in a dry NMR tube was irradiated with a 100 W medium pressure mercury lamp. After 10 days the NMR spectra showed the dissolved material to be completely unchanged but a small crystal (ª2 mg) had been deposited which was found to be 2- benzylamino-3-phenylpropane-1-sulfonic acid 28 (Found C 61.7; H 6.4; N 4.5.C16H19NO3S?0.4H2O requires C 61.5; H 6.4; N 4.5%); dH(CD3SOCD3) 9.4–9.2 (1 H br s) 9.2–9.0 (1 H br s) 7.6–7.4 (5 H m) 7.4–7.2 (5 H m) 4.5–4.3 (2 H m) 3.65 (1 H m) 3.35 (1 H half AB pattern of d J 20 15) 2.90 (1 H half AB pattern of d J 20 10) 2.70 (1 H half AB pattern of d J 15 10) and 2.65 (1 H half AB pattern of d J 15 4) dC(CD3SOCD3) 136.0 (4ry) 132.0 (4ry) 129.5 (2 CH) 129.3 (2 CH) 129.0 (CH) 128.8 (2 CH) 128.6 (2 CH) 127.0 (CH) 56.5 (CH) 48.8 (CH2) 47.5 (CH2) and 35.2 (CH2). Acknowledgements We thank Zeneca Pharmaceuticals and SERC for a Case studentship (D. P. A.) and the Royal Society for a Warren Research Fellowship (R. A. A.). References 1 R. A. Aitken I. Gosney and J. I. G. Cadogan Prog. Heterocycl. Chem. 1992 4 1; 1993 5 1. 2 J. M. Decazes J. L. Luche H.B. Kagan R. Parthasarthy and J. Ohrt Tetrahedron Lett. 1972 3633; D. Bellus Helv. Chim. Acta 1975 58 2509. 3 M. R. Johnson M. J. Fazio D. L. Ward and L. R. Sousa J. Org. Chem. 1983 48 494. 4 J. M. Bohen and M. M. Joullié J. Org. Chem. 1973 38 2652; W. Hanefeld and M. A. Jalili Liebigs Ann. Chem. 1986 1787. 5 Preliminary communication R. A. Aitken D. P. Armstrong S. T. E. Mesher and R. H. B. Galt Tetrahedron Lett. 1994 35 6143. 6 R. J. Gaul and W. J. Fremuth US Pat. 3 006 919 1961; R. J. Gaul and W. J. Fremuth J. Org. Chem. 1961 26 5103. 7 D. Delaunay L. Toupet and M. Le Corre J. Org. Chem. 1995 60 6604. 8 H. J. Roth and H. Schlump Arch. Pharm. (Weinheim Ger.) 1963 296 213. 9 A. G. M. Barrett D. H. R. Barton and R. Colle J. Chem. Soc. Perkin Trans. 1 1980 665. For a general discussion of oxidation of these ring systems see R.A. Aitken D. P. Armstrong and S. T. E. Mesher Prog. Heterocycl. Chem. 1990 2 1. 10 R. A. Aitken D. P. Armstrong R. H. B. Galt and S. T. E. Mesher J. Chem. Soc. Perkin Trans. 1 1997 935. 11 C. Roussel J.-L. Stein and F. Beauvais New J. Chem. 1990 14 169. 12 K. Schank and F. Werner Liebigs Ann. Chem. 1979 1977. 13 D. H. R. Barton D. P. Manly and D. A. Widdowson J. Chem. Soc. Perkin Trans. 1 1975 1568. 14 T. C. Farrar B. M. Trost S. L. Tang and S. E. Springer-Wilson J. Am. Chem. Soc. 1985 107 262. 15 G. Barbarella Prog. Nucl. Magn. Reson. Spectrosc. 1993 25 317; S. Berger S. Braun and H.-O. Kalinowski NMR Spektroskopie von Nichtmetallen Thieme Stuttgart 1992 vol. 1 p. 119. 16 A. Lawson and C. E. Searle J. Am. Chem. Soc. 1957 79 1556. 17 W.J. M. van Tilborg and R. Plomp J. Chem. Soc. Chem. Commun. 1977 130. 18 S. Itsuno K. Ito A. Hirao and S. Nakahama J. Chem. Soc. Perkin Trans. 1 1984 2887. 19 J. R. Piper and T. P. Johnston J. Org. Chem. 1963 28 981. 20 J. T. Sharp I. Gosney and A. G. Rowley Practical Organic Chemistry Chapman and Hall London 1989 p. 51. 21 M. Freifelder M. B. Moore M. R. Vernstein and G. R. Stone J. Am. Chem. Soc. 1958 80 4320. 22 J. Bielawski S. Brandage and L. Lindblom J. Heterocycl. Chem. 1978 15 97. Paper 7/00521K Received 22nd January 1997 Accepted 25th March 1997 J. Chem. Soc. Perkin Trans. 1 1997 2139 Synthesis and pyrolytic behaviour of thiazolidin-2-one 1,1-dioxides R. Alan Aitken,a,* David P. Armstrong,a Ronald H. B. Galt b and Shaun T. E. Mesher a a School of Chemistry University of St.Andrews North Haugh St. Andrews Fife UK KY16 9ST b Zeneca Pharmaceuticals Alderley Park Macclesfield Cheshire UK SK10 4TG Four examples of the chiral thiazolidin-2-one 1,1-dioxides 5 have been prepared by reaction of the appropriate amino alcohols 11 with CS2 in aqueous sodium hydroxide to give the thiazolidine-2-thiones 12 followed by oxidation with KMnO4 under phase-transfer conditions in the presence of benzoic acid either directly or via the thiazolidin-2-ones 13. Upon flash vacuum pyrolysis (FVP) at 650 8C 5a–c decompose mainly by loss of SO2 to give an alkene and benzyl isocyanate together with other products from fragmentation of the N-benzyl group. A significant minor pathway involves net loss of CO2 and water to give the 2-phenyl-4,5-dihydrothiazoles 21 together with their aromatisation products 22 and 23.A mechanism for this new heterocyclic transformation is proposed involving initial expansion to a cyclic carbamic–sulfinic anhydride (2,1,4-oxathiazin-3-one 1-oxide). The fully assigned 13C NMR spectra are presented for 5 12 and 13 and the 33S NMR spectrum has been obtained for 5c. Thermal and photochemical extrusion of SO2 from suitable ring systems has recently been used to achieve a wide variety of synthetic transformations.1 Among the most interesting targets have been b-lactams and these have been obtained in several cases by SO2 extrusion from appropriate thiazolidin-4-one 1,1- dioxides 1.2 In one case stereoselectivity was achieved with the cis compound 1 giving mainly the cis product 2 photochemically but mainly the trans isomer on pyrolysis (Scheme 1).3 The corresponding reaction of the thiazolidine-2,4-dione 1,1- dioxides 3 to give malonic acid imides 4 has also been reported.4 We were interested to examine the isomers of 1 the thiazolidin-2-one 1,1-dioxides 5 as possible precursors of the b-lactams 6. These have the advantage of being readily accessible in enantiomerically pure form from amino acid-derived amino alcohols and if the extrusion were possible it would be of interest to examine the reaction of their 5-anions with electrophiles in which the expected diastereoselectivity might lead to an efficient overall asymmetric synthesis of b-lactams. In this paper we describe the synthesis of representative thiazolidin-2- one 1,1-dioxides 5 in enantiomerically pure form and the first study of their pyrolytic behaviour.5 Scheme 1 N S O2 O N R3 R2 R3 R2 O N S O2 O N O O O R1 R R R1 N S O2 O R2 R3 R1 N O R1 R2 R3 – SO2 3 4 2 1 – SO2 heat or hn – SO2 5 6 heat or hn Results and discussion The target thiazolidin-2-one 1,1-dioxides are a very little known class of compounds and the only previous examples were obtained by Gaul and Fremuth in 1961 by oxidation of the corresponding thiazolidine-2-thiones using peracetic acid.6 The synthesis of the latter by reaction of b-amino alcohols with CS2 in aqueous sodium hydroxide is well known but as recently examined in detail by Delaunay et al.,7 it may also lead to the corresponding oxazolidine-2-thiones or to mixtures of both depending on the conditions used.We first subjected (S)- phenylalaninol 7c to reaction with CS2 in aqueous sodium hydroxide using the conditions reported by Roth and Schlump.8 The product consisted of a mixture of the oxazolidine-2-thione 8c and the thiazolidine-2-thione 9c but the former could be converted entirely into the latter by treatment with P2S5 (Scheme 2).Although this was obtained in low yield it had properties including the 13C NMR data (Table 1) in good agreement with those reported.7 When 9c was subjected to oxidation using a wide variety of oxidants and conditions complex mixtures of intractable products were obtained most likely owing to oxidative dimerisation through the 2- mercaptothiazoline tautomeric form a reaction previously reported for these compounds,9 followed by ring-opening and further degradation. Attention was therefore turned to N-substituted examples since it was these that were used in the successful oxidation method of Gaul and Fremuth;6 to facilitate later deprotection we chose the N-benzyl compounds.The required N-benzylamino alcohols 11a–c were readily obtained by condensation of 7a–c with benzaldehyde to give 10 followed by catalytic hydrogenation. When these were subjected to reaction with CS2 in aqueous sodium hydroxide under the same conditions as before the desired thiazolidinethiones 12a–c were obtained in moderate to good yield and no oxazolidinethiones were formed. This is in agreement with the work of Delaunay et al.7 where N-methylamino alcohols also gave exclusively the thiazolidinethiones. The colourless crystalline compounds 12a–c gave analytical and spectroscopic data in good agreement with expectations and the fully assigned 13C NMR spectra presented in Table 1 formed a highly consistent pattern.Attempted oxidation of 12 using peracetic acid under a variety of conditions did give the desired sulfones 5 but in disap- 2140 J. Chem. Soc. Perkin Trans. 1 1997 pointing yield and always accompanied by some of the thiazolidinones 13. The reagent of choice was found to be KMnO4 in a mixed phase system of CH2Cl2–water with both 0.1 equiv. benzyltriethylammonium chloride and 1 equiv. of benzoic acid as additives a system which has also recently proved valuable for the oxidation of 4,5-dihydrothiazoles to their 1,1-dioxides.10 Using this system the oxidation could be completely controlled with 3 equiv. of KMnO4 converting 12 cleanly into 13 and either 12 or 13 being converted into 5 using 5 and 2 equiv.KMnO4 respectively the yields in all reactions exceeding 60%. In view of problems later encountered in the pyrolysis (vide Scheme 2 Reagents i CS2 aq. NaOH; ii P2S5; iii PhCHO cat. TsOH; iv H2 Pd/C; v KMnO4 PhCO2H phase transfer catalyst CH2Cl2–H2O; vi MeI acetone; vii NaOMe MeOH; viii AcOOH R2 R3 HO H2N R2 R3 HO R1HN N S S R1 R2 R3 N N S O2 O R1 R2 R3 S O R1 R2 R3 HN S S R2 R3 HN O S R2 R3 R2 R3 HO N PhHC N+ S MeS H HN H HO3S 13 11 10 9 5 iii i 12 v v 8 v + 7 ii iv i I– vi vii viii (12d) 15 R1 R2 R3 a PhCH2 H Et b PhCH2 Pri H c PhCH2 PhCH2 H d —(CH2)3— H 14 (13d) infra) it was desirable to have an example of 5 in which R1 was linked to either R2 or R3 and 5d was therefore prepared by starting from (S)-prolinol 11d which reacted with CS2 in aqueous sodium hydroxide under the standard conditions to give 12d.In this case it was found to be preferable to obtain 5 via oxidation of 13 rather than directly from 12; 13d was obtained in good overall yield by reaction of 12d with methyl iodide to give 14 followed by treatment with sodium methoxide to give 13d. This unusual method for converting a thiazolidinethione to a thiazolidinone is based on a recently reported method for the thiazolinethione to thiazolinone transformation.11 Attempted oxidation of 13d using peracetic acid did give 5d but in low yield and this was accompanied by a new product which proved to be the sulfonic acid 15 resulting from hydrolytic ringopening of 5 decarboxylation and further oxidation. Again the permanganate–benzoic acid method provided the method of choice for conversion of 13d into 5d.Although the compounds 5a–d were perfectly stable under dry conditions and gave good analytical and spectroscopic data some hydrolysis as hinted at by the formation of 15 could be observed on prolonged storage. This is not surprising since a-oxo sulfones are notoriously elusive and in cases where they have been obtained they are readily hydrolysed.12 Acyclic carboxamido sulfones have been obtained before and are somewhat more resistant to hydrolysis.13 As shown in Table 1 the 13C NMR data for 12a–d 13a–d and 5a–d form a consistent pattern and the trends on going from 12 to 13 to 5 are somewhat surprising. The fall of ca. 25 ppm in the value for C-2 on going from 12 to 13 is as expected but the reason for the further fall of ca. 12 ppm for C-2 on going from 13 to 5 is not clear particularly when at the same time the values for C-5 increase by 17–20 ppm upon S-oxidation.The signals for the remaining ring carbon C-4 are also affected to a surprising degree by oxidation with falls of ca. 8 ppm associated with each oxidation step. Further confirmation of the five-membered ring structure of the compounds 5 as opposed to the isomeric sixmembered cyclic carbamic–sulfinic anhydride structure 19 was obtained by 33S NMR spectroscopy. The use of this technique to clarify a similar structural ambiguity has been described by Farrar et al.,14 and relies on the fact that the line widths in 33S NMR spectra are highly dependant on the degree of symmetry around the sulfur atom. Thus 5 would be expected to give a Table 1 13C NMR spectra of heterocycles 9 12 13 and 5 N Y X R1 R2 R3 5 4 2 dC 9c 12a 12b 12c 12d 13a 13b 13c 13d 5a 5b 5c 5d R1 H CH2Ph CH2Ph CH2Ph ](CH CH2Ph CH2Ph CH2Ph ](CH CH2Ph CH2Ph CH2Ph ](CH R2 CH2Ph H Pri CH2Ph 2)3] H Pri CH2Ph 2)3] H Pri CH2Ph 2)3] R3 H Et HH H Et HH H Et HH H X S SSS S OOO OOOO O Y S SSS SSSS S SO2 SO2 SO2 SO2 C-2a 200.5 197.0 197.4 196.7 191.1 172.0 172.9 171.8 169.8 159.8 160.6 159.5 157.7 C-4 65.1 67.7 71.0 67.5 71.9 59.1 62.0 59.5 63.0 51.3 54.5 51.7 52.6 C-5 39.7 31.7 26.9 32.2 35.8 29.9 24.9 30.4 33.2 47.2 42.7 47.9 53.8 R1 signals — 135.2 (4ry) 128.8 (2C) 127.9 127.7 (2C) 50.1 135.1 (4ry) 128.7 (2C) 127.8 (3C) 50.0 135.4 (4ry) 128.9 (2C) 128.2 128.0 (2C) 50.7 46.3 31.4 28.8 136.2 (4ry) 128.6 (2C) 127.7 (2C) 127.6 46.3 135.9 (4ry) 128.7 (2C) 128.0 (2C) 127.7 46.6 136.3 (4ry) 128.6 (2C) 128.0 (2C) 127.9 46.7 43.3 30.8 27.2 133.4 (4ry) 129.3 (2C) 128.8 128.2 (2C) 47.2 133.4 (4ry) 129.2 (2C) 128.7 128.2 (2C) 47.1 133.2 (4ry) 129.2 (2C) 129.0 128.5 (2C) 47.4 43.7 32.4 23.4 R2 and R3 signals 135.7 (4ry) 129.1 (2C) 129.0 (2C) 127.3 37.9 24.1 9.2 28.9 18.6 14.7 135.9 (4ry) 129.1 (2C) 129.0 (2C) 127.2 36.3 — 24.3 8.6 28.1 18.2 14.5 136.4 (4ry) 129.2 (2C) 128.7 (2C) 127.1 37.3 — 24.6 8.7 27.4 18.2 13.9 134.7 (4ry) 129.4 (2C) 129.3 (2C) 127.8 38.1 — a dC Values are given with reference to Me4Si as the internal standard.J. Chem. Soc. Perkin Trans. 1 1997 2141 relatively sharp signal while 19 would give a signal too broad to be observed. In the event the spectrum of 5c was readily obtained at natural abundance and consisted of a single signal of w1/2 130 Hz. The chemical shift of dS 26.5 with respect to aqueous Na2SO4 is in the expected range for cyclic sulfones,15 although no a-oxo sulfone has previously been observed.The sulfones 5a–c were subjected to flash vacuum pyrolysis (FVP) using a conventional flow system with a horizontal furnace tube operating at 1023 Torr and involving contact times of ª1–10 ms. Under these conditions all three compounds underwent complete reaction at the relatively mild temperature of 650 8C to give rather complex mixtures of products as shown in Table 2. It is disappointing to note that the desired extrusion of SO2 does occur but is accompanied by complete fragmentation to give the alkene 16 together with benzyl isocyanate 17 obtained largely in the form of its hydrolysis product 18 (Scheme 3). Pyrolysis of an authentic sample of 17 under the same conditions confirmed both that it does not undergo any further thermal reactions and that using our normal techniques it underwent substantial hydrolysis owing to adventi- Scheme 3 N S O2 O PhCH2 R2 R3 N O R2 R3 S O O R2 R3 PhCH2 N N• S O H R2 R3 PhCH2 Ph •O S R2 R3 H N PhCH HO S N R2 R3 S Ph R2 R3 N S Ph R3 N S Ph R2 O HO HN R2 R3 Ph PhCH2NH NHCH2Ph O 5a–c PhCH2NCO – SO2 FVP FVP 23 20 19 22 21 • – CO2 16 • 17 24 PhCOCl 7 H2O – H2O 18 P2S5 + – R3H – R2H Table 2 Products from FVP of thiazolidin-2-one 1,1-dioxides 5 at 650 8C (%) Starting material Product 16 17 18 21 22 23 PhC]] ] N PhCH2CH2Ph PhCH]] NCH2Ph PhCHO PhMe PriCHO EtCHO 5a 13 2 24 5 —3 19 10 524 —2 5b 12 — 10 444 16 67215 — 5c 18 — 15 3827 12 —45 —— tious moisture in the cold trap to give 18.It appears that the extrusion from 5a–c requires more forcing conditions as compared to 1 such that the b-lactam cannot survive intact.The formation of benzonitrile bibenzyl toluene N-benzylidenebenzylamine and benzaldehyde in all cases is probably associated with fragmentation of the N-benzyl group. The origin of the aliphatic aldehydes corresponding to R2/R3CHO is unclear. Most interesting however is the formation of small but signifi- cant quantities of the 2-phenyl-4,5-dihydrothiazoles 21 and their aromatization products 22 and 23. The identity of these unexpected products was demonstrated by comparison with authentic samples prepared by reaction of 7a–c with benzoyl chloride to give 24 followed by cyclisation with P4S10.10 Heating 21b,c with sulfur at 200–210 8C afforded samples of 23b,c while 2-phenylthiazole 23a (=22b,c) was prepared by a literature method16 and these were identical with the pyrolysis products.The mechanism of this unprecedented heterocyclic transformation is believed to involve the sequence of steps shown in Scheme 3 resulting in the required net loss of CO2 and H2O. Ring expansion to the cyclic sulfinic–carbamic anhydride 19 a process well known in the pyrolysis of cyclic sulfones,1 allows ready loss of CO2. Rearrangement of the resulting diradical and intramolecular abstraction of the benzylic CH gives the imino sulfenic acid 20 which can then lose water to afford 21. Overall the process is somewhat reminiscent of the pyrolysis of benzothiophene 1,1-dioxide to give benzothiete,17 which also involves loss of CO2 and initial ring expansion. In an attempt to prevent the fragmentation to alkene and isocyanate we then examined the pyrolysis of 5d in which the routes leading to 21–23 are also impossible.This underwent complete reaction at the lower temperature of 600 8C but the product consisted of a complex mixture of products which could not be identified. The presence of alkene signals in the NMR spectra pointed to ring-opening and this might be expected as shown in Scheme 4 since the diradical resulting from loss of SO2 can readily open to give the pentenyl isocyanate 25 while additional loss of CO can lead to pentenylnitrene 26. Synthesis of an authentic sample of 27 an alternative possible product from the diradical shown confirmed that it was not present. Both 25 and 26 are expected to be highly reactive and can undergo a variety of secondary reactions either in the furnace or in the cold trap so the complex mixture produced is not surprising.In a final attempt to obtain a b-lactam 5a–c were subjected to photolysis in a variety of solvents. In contrast to the isomeric compounds 1,3 they were found to be photochemically inert and the only new product obtained in low yield from 5c was the amino sulfonic acid 28 resulting from hydrolysis by adventitious moisture decarboxylation and oxidation. In conclusion it is clear that the thiazolidin-2-one 1,1-dioxides 5 are not suitable precursors for the thermal or photochemical generation of Scheme 4 N S O2 O H OCN N N H Me :N – SO2 • • – SO2 – CO 5d 27 25 26 HO3S PhCH2NH CH2Ph H 28 2142 J. Chem. Soc. Perkin Trans. 1 1997 b-lactams in contrast to the isomeric thiazolidin-4-one 1,1- dioxides 1.Due to subtle differences between the two ring systems the more severe conditions required to achieve SO2 extrusion in the former case lead to complete fragmentation to an alkene and isocyanate. The unexpected formation of 21–23 is however of some mechanistic interest. Experimental Melting points were determined on a Reichert hot-stage microscope and are uncorrected. Infrared spectra were recorded for solids as Nujol mulls and for liquids as thin films on a Perkin- Elmer 1420 spectrophotometer. NMR spectra were recorded for 1H at 80 MHz on a Bruker WP80 instrument or at 300 MHz on a Bruker AM300 instrument for 13C at 20 MHz on a Varian CFT 20 or at 75 MHz on a Bruker AM300 instrument and for 33S at 38 MHz on a Bruker MSL500 spectrometer. Spectra were obtained for solutions in CDCl3 unless otherwise indicated with Me4Si as internal reference for 1H and 13C and aqueous Na2SO4 as external reference for 33S.Chemical shifts are reported in ppm relative to the reference and coupling constants J are given in Hz. In the assignments for the 13C NMR data 4ry refers to quaternary carbon. Mass spectra were obtained on an A.E.I. MS902 instrument using electron impact at 70 eV. GC–MS was performed with a Hewlett Packard 5890A chromatograph coupled to a Finnigan Incos 50 mass spectrometer. Optical rotations were measured on an Optical Activity AA1000 polarimeter and are given in units of 1021 deg cm2 g21. The amino alcohols 7a–c and 11d were prepared by reduction of the corresponding amino acids or were commercially available. Preparation of 2-benzylideneamino alcohols 10 Benzaldehyde (24.4 g 230 mmol) was added to a stirred solution of the appropriate amino alcohol 7 (220 mmol) in toluene (250 cm3) and the mixture heated under reflux for 1 h using a Dean–Stark separator.Evaporation yielded the product which was recrystallised from hexane. Using this method the following compounds were prepared. (2R)-2-Benzylideneaminobutan-1-ol 10a. (2R)-2-Aminobutan- 1-ol 7a gave 10a as colourless needles (77%) mp 57– 58 8C (Found C 74.6; H 8.8; N 7.9. C11H15NO requires C 74.5; H 8.5; N 7.9%); [a]D 20 137.8 (c 1.0 in CH2Cl2); nmax/cm21 3280 (OH) 1645 (CN) 1060 1000 780 and 705; dH 8.20 (1 H s) 7.65 (2 H m) 7.35 (3 H m) 3.78 (1 H half AB pattern of d J 12 10) 3.72 (1 H half AB pattern of d J 12 4) 3.18 (1 H m) 2.86 (1 H br s) 1.60 (2 H m) and 0.85 (3 H t J 7); dC 162.0 (CH) 135.8 (4ry) 130.7 (CH) 128.5 (2 CH) 128.3 (2 CH) 74.7 (CH) 66.0 (CH2) 25.0 (CH2) and 10.7 (CH3); m/z 177 (M1 15%) 176 (50) 146 (100) 132 (25) 118 (30) 104 (50) 91 (85) 77 (35) and 41 (60).(2S)-2-Benzylideneamino-3-methylbutan-1-ol 10b. (2S)-2- Amino-3-methylbutan-1-ol 7b gave 10b as colourless crystals (77%) mp 70–71 8C (Found C 75.2; H 9.0; N 7.3. C12H17NO requires C 75.3; H 9.0; N 7.3%); [a]D 25 283.3 (c 0.3 in CHCl3); nmax/cm21 3700–2400 (br OH) 1640 1470 1450 1380 1260 1220 1060 and 1020; dH 8.29 (1 H s) 7.85–7.6 (2 H m) 7.6–7.3 (3 H m) 3.80 (2 H m) 3.2–2.8 (1 H m) 1.90 (1 H octet J 7) 0.95 (3 H d J 7) and 0.90 (3 H d J 7); dC 161.7 (C]] N) 136.0 (4ry) 130.4 (CH) 128.4 (4 CH) 79.2 (CH) 64.1 (CH2) 30.0 (CH) 19.7 (CH3) and 19.2 (CH3); m/z 190 (M 2 H1 5%) 189 (2) 160 (100) 148 (70) 130 (25) and 118 (35).(2S)-2-Benzylideneamino-3-phenylpropan-1-ol 10c. (2S)-2- Amino-3-phenylpropan-1-ol 7c gave 10c as colourless prisms (64%) mp 78–80 8C (Found C 80.1; H 7.2; N 5.8. C16H17NO requires C 80.0; H 7.1; N 5.8%); [a]D 25 2215.6 (c 2.0 in CHCl3); nmax/cm21 3600–2700 (br OH) 1640 1490 1450 1380 1220 1030 and 700; dH 7.98 (1 H s) 7.7–7.55 (2 H m) 7.45–7.3 (3 H m) 7.25–7.1 (5 H m) 3.85 (1 H half of AB pattern of d J 10 6) 3.70 (1 H half of AB pattern of d J 10 4) 3.7–3.4 (1 H m) 3.00 (1 H half of AB pattern of d J 14 5) 2.80 (1 H half of AB pattern of d J 14 8) and 2.26 (1 H br s); dC 162.4 (C]] N) 138.6 (4ry) 135.6 (4ry) 130.6 (CH) 129.6 (2 CH) 128.4 (2 CH) 128.2 (4 CH) 126.0 (CH) 74.4 (CH) 65.6 (CH2) and 38.9 (CH2); m/z 208 (M1 2 CH2OH 8%) 148 (M1 2 CH2Ph 50) 130 (12) 128 (32) 127 (35) and 91 (100).Preparation of 2-benzylamino alcohols 11 A solution of the appropriate benzylideneamino alcohol 10 (0.52 mol) and 5% palladium/charcoal catalyst (3.0 g) in ethyl acetate (500 cm3) was stirred vigorously in the presence of hydrogen gas (12 dm3 0.54 mol) at room temp. for 24 h. The solution was then filtered through Celite and evaporated to afford the product. Using this method the following compounds were prepared. (2R)-2-Benzylaminobutan-1-ol 11a. (2R)-2-Benzylideneaminobutan- 1-ol 10a gave 11a following recrystallisation from hexane as a colourless solid (77%) mp 74–75 8C (Found C 73.4; H 9.6; N 7.7. C11H17NO requires C 73.7; H 9.6; N 7.8%); [a]D 20 228.5 (c 1.0 in CH2Cl2); nmax/cm21 3400–3000 (OH) 3280 (NH) 1060 865 745 and 700; dH 7.30 (5 H m) 3.80 and 3.72 (2 H AB pattern J 14) 3.62 (1 H half of AB pattern of d J 10 4) 3.35 (1 H half AB pattern of d J 10 6) 2.62 (1 H m) 2.40 (2 H br s) 1.6–1.4 (2 H m) and 0.90 (3 H t J 7); dC 140.3 (4ry) 128.5 (2 CH) 128.1 (2 CH) 127.1 (CH) 62.6 (CH2) 59.8 (CH) 51.0 (CH2) 24.2 (CH2) and 10.4 (CH3); m/z 179 (M1 1%) 148 (100) 106 (55) 91 (100) 77 (50) 65 (75) and 56 (70).(2S)-2-Benzylamino-3-methylbutan-1-ol 11b. (2S)-2-Benzylideneamino- 3-methylbutan-1-ol 10b gave 11b following Kugelrohr distillation as a colourless oil (80%) bp (oven temp.) 106– 108 8C at 0.4 Torr (lit.,18 103–107 8C at 0.2 Torr). (2S)-2-Benzylamino-3-phenylpropan-1-ol 11c. (2S)-2-Benzylideneamino- 3-phenylpropan-1-ol 10c gave 11c following recrystallisation from hexane–ethyl acetate (5 1) as colourless prisms (71%) mp 124–126 8C (Found C 79.6; H 8.0; N 5.7.C16H19NO requires C 79.6; H 7.9; N 5.8%); [a]D 25 249.8 (c 2.0 in CHCl3); nmax/cm21 3700–2400 (br OH) 1640 1490 1450 1400 1220 1110 1030 910 and 700; dH 7.25 (11 H m) 3.9–3.6 (1 H m) 3.80 (2 H s) 3.60 (1 H half AB pattern of d J 10 4) 3.40 (1 H half AB pattern of d J 10 5) 2.90 and 2.70 (2 H AB pattern of d J 8 4) and 2.80 (1 H br s); dC 139.9 (4ry) 138.8 (4ry) 129.2 (2 CH) 128.4 (4 CH) 128.0 (2 CH) 126.9 (CH) 126.2 (CH) 62.6 (CH2) 59.7 (CH) 51.1 (CH2) and 37.8 (CH2); m/z 242 (M 1 H1 1.2%) 241 (M1 1) 210 (8) 150 (20) and 91 (100). Preparation of thiazolidine-2-thiones 9 and 12 A mixture of the appropriate amino alcohol (45 mmol) 2 M sodium hydroxide (150 cm3) and carbon disulfide (9.8 cm3 12.4 g 163 mmol) was stirred at room temp.for 20 h. A further portion of carbon disulfide (5.0 cm3 6.3 g 83 mmol) was added and the solution stirred for an additional 4 h. The mixture was extracted with CH2Cl2 and the organic layer washed with water dried and evaporated to afford the product. Using this method the following compounds were prepared. (4S)-4-Benzylthiazolidine-2-thione 9c. (2S)-2-Amino-3- phenylpropan-1-ol 7c (8.0 g 52 mmol) gave a mixture of the desired product 9c and the corresponding oxazolidine-2-thione 8c. This was dissolved in toluene (200 cm3) and heated under reflux with P2S5 (20 g 90 mmol) for 48 h. Filtration and evaporation followed by column chromatography [SiO2 diethyl ether– petroleum (bp 40–60 8C) 1 1] gave a red solid which was recrystallised from hexane–ethyl acetate (5 1) to give the product as red needles (18%) mp 79–80 8C (lit.,7 84–85 8C) (Found C 57.4; H 5.3; N 6.6.C10H11NS2 requires C 57.4; H 5.3; N 6.7%); [a]D 25 2112.2 (c 1.7 in CHCl3); nmax/cm21 3480 1470 1290 1250 1220 1140 1040 1010 960 and 700; dH 8.40 (1 H br s) 7.4–7.2 (3 H m) 7.2–7.1 (2 H m) 4.46 (1 H quintet J 10) 3.50 and 3.26 (2 H AB pattern of d J 14 10) 3.05 and 2.93 (2 H AB pattern of d J 12 10); dC see Table 1; m/z 209 (M1 40%) 182 (12) 167 (15) 146 (93) 132 (12) 118 (27) 117 (20) and 91 (100). J. Chem. Soc. Perkin Trans. 1 1997 2143 (4R)-3-Benzyl-4-ethylthiazolidine-2-thione 12a. (2R)-2-Benzylaminobutan- 1-ol 11a gave 12a following recrystallisation from hexane–ethyl acetate (2 1) as colourless crystals (72%) mp 61–62 8C (Found C 60.7; H 6.1; N 5.9. C12H15NS2 requires C 60.7; H 6.4; N 5.9%); [a]D 20 191.3 (c 1.0 in CH2Cl2); nmax/cm21 3060 3040 1475–1425 1225 1175 1025 (CS) 760 and 700; dH 7.30 (5 H m) 5.75 and 4.25 (2 H AB pattern J 17) 4.00 (1 H m) 3.35 (1 H half AB pattern of d J 10 8) 2.96 (1 H half AB pattern of d J 10 5) 1.77 (2 H m) and 0.92 (3 H t J 7); dC see Table 1; m/z 237 (M1 15%) 148 (100) 132 (5) 121 (10) 104 (5) 91 (70) and 65 (25).(4S)-3-Benzyl-4-isopropylthiazolidine-2-thione 12b. (2S)- Benzylamino-3-methylbutan-1-ol 11b gave 12b following recrystallisation from hexane–ethyl acetate (5 1) with cooling (220 8C) as colourless crystals (36%) mp 77–78 8C (Found C 62.2; H 6.9; N 5.6. C13H17NS2 requires C 62.1; H 6.8; N 5.6%); [a]D 25 2143.1 (c 0.5 in CHCl3); nmax/cm21 1460 1450 1330 1240 1220 1200 1180 1130 1040 990 and 960; dH 7.40 (5 H s) 6.00 and 4.14 (2 H AB pattern J 16) 4.05 (1 H m) 3.20 (1 H half AB pattern of d J 11 9) 3.05 (1 H half AB pattern of d J 11 6) 2.34 (1 H septet of d J 7 4) 0.95 (3 H d J 7) and 0.90 (3 H d J 7); dC see Table 1; m/z 251 (M1 100%) 208 (15) 187 (24) 148 (82) 144 (24) and 91 (35).(4S)-3,4-Dibenzylthiazolidine-2-thione 12c. (2S)-2-Benzylamino- 3-phenylpropan-1-ol 11c gave 12c following recrystallisation from hexane–ethyl acetate (3 1) as colourless needles (53%) mp 137–139 8C (Found C 68.4; H 5.55; N 4.65. C17H17NS2 requires C 68.2; H 5.7; N 4.7%); [a]D 25 225.8 (c 1.6 in CH2Cl2); nmax/cm21 1490 1450 1420 1350 1300 1220 1170 1080 1030 920 and 700; dH 7.4–7.2 (8 H m) 7.1–7.0 (2 H m) 5.82 and 4.20 (2 H AB pattern J 16) 4.30–4.15 (1 H m) 3.20 (1 H half AB pattern of d J 12 8) 3.15 (1 H half AB pattern of d J 14 5) 2.86 (1 H half AB pattern of d J 12 10) and 2.83 (1 H half AB pattern of d J 14 10); dC see Table 1; m/z 299 (M1 42%) 277 (20) 238 (10) 208 (100) 148 (92) and 117 (31).(5S)-3-Thia-1-azabicyclo[3.3.0]octane-2-thione 12d. (2S)-2- Hydroxymethylpyrrolidine 11d gave 12d following recrystallisation from ethanol as colourless crystals (49%) mp 130– 131 8C (lit.,19 132–133 8C) (Found C 45.1; H 5.5; N 8.78. C6H9NS2 requires C 45.2; H 5.7; N 8.8%); [a]D 20 2159.8 (c 1.0 in CH2Cl2); nmax/cm21 1360 1340 1245 1210 1180 1055 1030 (CS) 940 and 850; dH 4.63 (1 H m) 3.60 (1 H m) 3.48 (1 H m) 3.32 (2 H dd J 7 2) 2.5–2.3 (2 H m) 2.20 (1 H m) and 1.80 (1 H m); dC see Table 1; m/z 159 (M1 70%) 126 (5) 118 (10) 85 (30) 72 (25) 67 (50) 45 (35) and 41 (100).Preparation of thiazolidin-2-ones 13a–c A solution of the appropriate thiazolidinethione 12 (5 mmol) benzoic acid (0.62 g 5 mmol) and benzyltriethylammonium chloride (0.11 g 0.5 mmol) in dichloromethane (50 cm3) was stirred vigorously with a solution of potassium permanganate (2.37 g 15 mmol) in water (100 cm3) for 3 h. Sufficient solid sodium metabisulfite was added to decolourise the mixture which was then filtered through Celite the organic layer was separated and the aqueous layer washed with dichloromethane (3 × 50 cm3). The combined organic extracts were washed with 1 M hydrazine dihydrochloride followed by aqueous sodium carbonate dried with anhydrous magnesium sulfate and evaporated to give the product. Using this method the following compounds were prepared. (4R)-3-Benzyl-4-ethylthiazolidin-2-one 13a.(4R)-3-Benzyl-4- ethylthiazolidine-2-thione 12a gave 13a following Kugelrohr distillation as a pale green oil (76%) bp (oven temp.) 215 8C at 0.7 Torr (Found C 65.6; H 7.0; N 6.6%; M 221.0859. C12H15NOS requires C 65.1; H 6.8; N 6.3%; M 221.0874); [a]D 20 226.1 (c 1.07 in CH2Cl2); nmax/cm21 2970–2940 1670 (CO) 1460 1410 1230 and 710; dH 7.35–7.2 (5 H m) 4.96 and 4.00 (2 H AB J 15) 3.55 (1 H m) 3.26 (1 H half AB pattern of d J 11 8) 2.93 (1 H half AB pattern of d J 11 6) 1.75–1.5 (2 H m) and 0.86 (3 H t J 7); dC see Table 1; m/z 221 (M1 90%) 192 (85) 165 (20) 122 (25) 104 (70) 91 (100) and 65 (80). (4S)-3-Benzyl-4-isopropylthiazolidin-2-one 13b. (4S)-3- Benzyl-4-isopropylthiazolidine-2-thione 12b gave 13b following Kugelrohr distillation as a pale yellow solid (43%) mp 33– 35 8C bp (oven temp.) 185 8C at 0.7 Torr (Found C 66.5; H 7.7; N 6.1%; M 235.1026.C13H17NOS requires C 66.3; H 7.3; N 6.0%; M 235.1031); [a]D 20 134.0 (c 1.02 in CH2Cl2); nmax/cm21 3025 2964 1723 1664 (CO) 1455 1435 1260 1215 and 705; dH 7.35–7.2 (5 H m) 5.10 and 3.90 (2 H AB pattern J 17) 3.57 (1 H m) 3.10 (1 H half AB pattern of d J 13 9) 3.03 (1 H half of AB pattern of d J 13 7) 2.20 (1 H m) 0.87 (3 H d J 9) and 0.85 (3 H d J 9); dC see Table 1; m/z 235 (M1 15%) 192 (45) 176 (5) 133 (10) 105 (5) 91 (100) and 77 (5). (4S)-3,4-Dibenzylthiazolidin-2-one 13c. (4S)-3,4-Dibenzylthiazolidine- 2-thione 12c gave 13c following Kugelrohr distillation as a colourless oil which formed colourless prisms with time (45%) bp (oven temp.) 225 8C at 0.3 Torr; mp 70–71 8C (Found C 72.2; H 6.3; N 4.8.C17H17NOS requires C 72.0; H 6.1; N 4.9%); [a]D 25 111.8 (c 0.9 in CHCl3); nmax/cm21 1650 1490 1450 1440 1400 1350 1200 1080 1030 970 and 930; dH 7.4– 7.2 (8 H m) 7.08 (2 H m) 5.08 and 4.00 (2 H AB pattern J 16) 3.80 (1 H m) 3.15–3.05 (2 H m) 2.92 (1 H half AB pattern of d J 12 4) and 2.77 (1 H half AB pattern of d J 12 8); dC see Table 1; m/z (CI) 284 (M 1 H1 100%) 192 (46) 108 (7) 91 (65) and 65 (7). (5S)-2-Methylthio-3-thia-1-azabicyclo[3.3.0]oct-1-en-1-ium iodide 14 A solution of (5S)-3-thia-1-azabicyclo[3.3.0]octane-2-thione 12d (4.0 g 25 mmol) and methyl iodide (15.6 cm3 35.5 g 250 mmol) in acetone (110 cm3) was stirred for 16 h at room temp. The resulting precipitate was filtered off and washed with diethyl ether. The filtrate was concentrated and a second crop of the product filtered off and washed with diethyl ether.The solids were combined to yield the product (6.74 g 90%) as a pale yellow powder mp 111–112 8C (Found C 27.8; H 3.9; N 4.6. C7H12INS2 requires C 27.9; H 4.0; N 4.7%); [a]D 20 2256.5 (c 1.66 in CH2Cl2); nmax/cm21 1555 1300 1200 1170 and 950; dH 5.20 (1 H m) 3.87 (2 H m) 3.68 (1 H m) 3.58 (1 H m) 2.77 (3 H s) 2.46 (2 H m) and 2.25–2.10 (2 H m); dC 186.7 (4ry) 77.5 (CH) 49.3 (CH2) 38.5 (CH2) 29.5 (CH2) 29.4 (CH2) and 19.8 (CH3); m/z 159 (M1 2 MeI 30%) 126 (5) 118 (10) 85 (30) 82 (10) and 67 (50). (5S)-3-Thia-1-azabicyclo[3.3.0]octan-2-one 13d (5S)-2-Methylthio-3-thia-1-azabicyclo[3.3.0]oct-1-en-1-ium iodide 14 (22.58 g 75 mmol) was added to a solution of sodium methoxide (75 mmol) in methanol (200 cm3) and the mixture stirred for 16 h at room temp.Water (400 cm3) was added and the mixture extracted with CH2Cl2. The combined organic layers were washed with water dried and evaporated to yield a yellow solid. Recrystallisation of this from diethyl ether–ethyl acetate with cooling (220 8C) afforded the product (8.37 g 78%) as colourless crystals mp 70–71 8C (Found C 50.1; H 6.3; N 9.6. C6H9NOS requires C 50.3; H 6.3; N 9.8%); [a]D 20 235.4 (c 1.0 in CH2Cl2); nmax/cm21 3320 1700 (CO) 1385 930 and 890; dH 4.22 (1 H m) 3.55 (1 H m) 3.38 (1 H half AB pattern of d J 12 9) 3.22 (1 H half AB pattern of d J 12 10) 3.17 (1 H m) 2.3–2.0 (3 H m) and 1.62 (1 H m); dC see Table 1; m/z 143 (M1 30%) 114 (5) 85 (5) 80 (5) 74 (20) 70 (30) and 55 (100). Preparation of thiazolidin-2-one 1,1-dioxides 5 Exactly the same method was used as described for the thiazolidin-2-ones above except that the quantity of potassium permanganate was increased to 3.95 g (25 mmol) and the products were recrystallised from diethyl ether–CH2Cl2 (1 1).Using this method the following compounds were prepared. (4R)-3-Benzyl-4-ethylthiazolidin-2-one 1,1-dioxide 5a. (4R)-3- Benzyl-4-ethylthiazolidine-2-thione 12a gave 5a as colourless 2144 J. Chem. Soc. Perkin Trans. 1 1997 crystals (72%) mp 102–103 8C (Found C 56.8; H 6.0; N 5.5. C12H15NO3S requires C 56.9; H 6.0; N 5.5%); [a]D 20 147.0 (c 0.1 in CH2Cl2); nmax/cm21 3420 1710 (CO) 1320 1140 940 850 755 and 700; dH 7.4–7.3 (3 H m) 7.25–7.2 (2 H m) 5.10 and 4.22 (2 H AB pattern J 15) 3.70 (1 H m) 3.35 (1 H half AB pattern of d J 14 8) 3.15 (1 H half AB pattern of d J 14 4) 1.95 (1 H m) 1.76 (1 H m) and 0.94 (3 H t J 8); dC see Table 1; m/z 189 (M1 2 SO2 2%) 161 (2) 133 (50) 105 (30) 91 (100) and 77 (10).This product could alternatively be prepared from (4R)-3- benzyl-4-ethylthiazolidin-2-one 13a using 2 equiv. of KMnO4. (4S)-3-Benzyl-4-isopropylthiazolidin-2-one 1,1-dioxide 5b. (4S)-3-Benzyl-4-isopropylthiazolidine-2-thione 12b gave 5b as pale yellow needles (33%) mp 114–115 8C (Found C 58.4; H 6.4; N 5.2. C13H17NO3S requires C 58.4; H 6.4; N 5.2%); [a]D 20 239.6 (c 1.02 in CH2Cl2); nmax/cm21 3420 1720 (CO) 1325 and 1135 (SO2) 760 740 and 700; dH 7.4–7.3 (3 H m) 7.3–7.2 (2 H m) 5.10 and 4.18 (2 H AB pattern J 15) 3.77 (1 H m) 3.26 (1 H half AB pattern of d J 14 8) 3.12 (1 H half AB pattern of d J 14 6) 2.38 (1 H m) 0.89 (3 H d J 7) and 0.85 (3 H d J 7); dC see Table 1; m/z 203 (M1 2 SO2 15%) 160 (10) 133 (90) 105 (30) 91 (100) and 77 (5).(4S)-3,4-Dibenzylthiazolidin-2-one 1,1-dioxide 5c. (4S)-3,4- Dibenzylthiazolidine-2-thione 12c gave 5c as colourless needles (67%) mp 143–144 8C (Found C 65.0; H 5.4; N 4.4. C17H17NO3S requires C 64.7; H 5.4; N 4.4%); [a]D 25 222.6 (c 0.7 in CHCl3); nmax/cm21 1730 (CO) 1490 1450 1420 1330 1220 1140 and 770; dH 7.45–7.35 (3 H m) 7.3–7.2 (5 H m) 7.07 (2 H m) 5.16 and 4.20 (2 H AB pattern J 16) 3.90 (1 H m) 3.34 (1 H half AB pattern of d J 16 6) 3.20 (1 H half AB pattern of d J 12 4) 3.03 (1 H half AB pattern of d J 12 8) and 2.88 (1 H half AB pattern of d J 16 10); dC see Table 1; dS 26.5 (w1/2 130 Hz); m/z 316 (M 1 H1 1%) 251 (M1 2 SO2 7) 192 (8) 176 (19) 160 (28) 134 (12) 118 (38) and 91 (100).(5S)-3-Thia-1-azabicyclo[3.3.0]octan-2-one 3,3-dioxide 5d. (5S)-3-Thia-1-azabicyclo[3.3.0]octan-2-one 13d gave 5d as colourless crystals (70%) mp 175–176 8C (Found C 41.1; H 5.2; N 7.95. C6H9NO3S requires C 41.1; H 5.2; N 8.0%); [a]D 20 130.7 (c 0.104 Me2SO); nmax/cm21 3440 1740 (CO) 1320 and 1130 (SO2) and 1160; dH(CD2Cl2) 3.95 (1 H m) 3.80 (1 H dd J 13 6) 3.55 (2 H m) 3.02 (1 H dd J 13 9) 2.39 (1 H m) 2.20 (1 H m) 2.02 (1 H m) and 1.56 (1 H qd J 12 8); dC see Table 1; m/z 111 (M1 2 SO2 25%) 82 (10) 68 (80) 67 (100) 55 (70) and 53 (55). An attempt to prepare 5d by oxidation of 13d using 32% peroxyacetic acid in acetic acid,6 gave the desired product in 37% yield but this was now accompanied by (2S)-pyrrolidine- 2-methanesulfonic acid 15 (24%) as a colourless powder mp 260 8C (decomp.) (Found C 36.0; H 6.6; N 8.2.C5H11NO3S requires C 36.4; H 6.7; N 8.5%); [a]D 20 131.2 (c 1.0 in H2O); nmax/cm21 2900–2400 1377 1172 and 1045; dH(CD3SOCD3) 8.90 (1 H br s) 8.40 (1 H br s) 3.75 (1 H m) 3.15 (2 H m) 2.90 (1 H m) 2.87 (1 H d J 2) 2.10 (1 H m) 1.90 (1 H m) 1.80 (1 H m) and 1.60 (1 H m); dC(CD3SOCD3) 56.4 (CH) 51.8 (CH2) 44.8 (CH2) 29.9 (CH2) and 22.7 (CH2); m/z 165 (M1 5%) 157 (5) 122 (5) 111 (10) 97 (10) 84 (55) and 44 (90). FVP of thiazolidin-2-one 1,1-dioxides 5 The apparatus used is similar to one which has been illustrated and described recently.20 The sample was volatilised from a tube in a Büchi Kugelrohr oven through a 30 × 2.5 cm horizontal fused quartz tube. This was heated externally by a Carbolite Eurotherm tube furnace MTF-12/38A to a temperature of 600– 650 8C the temperature being monitored by a Pt/Pt–13%Rh thermocouple situated at the centre of the furnace.The products were collected in a U-shaped trap cooled in liquid nitrogen. The whole system was maintained at a pressure of 1–2 × 1023 Torr by an Edwards Model E2M5 high capacity rotary oil pump the pressure being measured by a Pirani gauge situated between the cold trap and the pump. Under these conditions the contact time in the hot zone was estimated to be ª10 ms. After the material had all sublimed the products were recovered directly from the cold trap and analysed by 1H and 13C NMR spectroscopy and GC–MS the identity of the products being determined by comparison with authentic samples. Yields were determined by calibration of the 1H NMR spectra by adding an accurately weighed quantity of a solvent such as CH2Cl2 and comparing integrals a procedure estimated to be accurate to ±10% or for products such as benzonitrile which did not show a distinctive NMR signal from the GC integrals.Pyrolysis of 5a. 5a (0.10 g 650 8C) gave a yellow oil. Careful analysis of the 13C and 1H NMR spectra and GC–MS showed eleven compounds to be present but-1-ene 16a (13%) benzyl isocyanate 17 (2%) dibenzylurea 18 (24%) 4-ethyl-2-phenyl- 4,5-dihydrothiazole 21a (5%) 2-phenylthiazole 23a (3%) benzonitrile (19%) bibenzyl (10%) N-benzylidenebenzylamine (5%) benzaldehyde (2%) toluene (4%) and propanal (2%). Pyrolysis of 5b. 5b (103 mg 650 8C) afforded a yellow oil in the cold trap. Analysis of the 13C and 1H NMR spectra and GC–MS showed eleven compounds to be present 3-methylbut- 1-ene 16b (12%) dibenzylurea 18 (10%) 4-isopropyl-2-phenyl- 4,5-dihydrothiazole 21b (4%) 2-phenylthiazole 22b (4%) 4- isopropyl-2-phenylthiazole 23b (4%) benzonitrile (16%) bibenzyl (6%) N-benzylidenebenzylamine (7%) benzaldehyde (2%) toluene (1%) and 2-methylpropanal (5%).Pyrolysis of 5c. 5c (118 mg 650 8C) afforded a yellow oil at the furnace exit and in the cold trap. Analysis of the 13C and 1H NMR spectra and GC–MS showed the oil to contain nine products allylbenzene 16c (18%) dibenzylurea 18 (15%) 4- benzyl-2-phenyl-4,5-dihydrothiazole 21c (3%) 2-phenylthiazole 22c (8%) 4-benzyl-2-phenylthiazole 23c (2%) bibenzyl (12%) benzonitrile (7%) benzaldehyde (4%) and toluene (5%). Pyrolysis of 5d. 5d (120 mg 600 8C) afforded a yellow oil in the cold trap. The 13C and 1H NMR spectra showed a large number of compounds to be present but identification proved inconclusive.GC–MS analysis showed a major product with m/z 83 (C5H9N) but examination of the 13C NMR spectrum showed that the signals for the likely product 3,4-dihydro-5- methyl-2H-pyrrole 27 were absent. Synthesis of authentic samples of FVP products Preparation of 3-methylbut-1-ene 16b. The FVP of isoamyl acetate (2.5 g 19 mmol 750 8C 7.0 × 1023 Torr) afforded 3- methylbut-1-ene (0.2 g 15%); dH 5.8–5.75 (1 H m) 5.0–4.85 (2 H m) 2.28 (1 H m) and 0.98 (6 H d J 8); dC 146.0 (CH) 111.1 (CH2) 32.0 (CH) and 22.0 (2 CH3). Flash vacuum pyrolysis of benzyl isocyanate 17. The FVP of benzyl isocyanate (0.20 g 650 8C 1.0 × 1023 Torr) produced no change in the starting compound. Upon standing overnight the liquid solidified to afford dibenzylurea 18 (0.18 g 99%); dH 7.4–7.1 (10 H m) 6.5 (2 H br s) and 4.25 (4 H d J 4); dC(CD3SOCD3) 158.1 (CO) 140.7 (2 4ry) 128.1 (4 CH) 126.9 (4 CH) 126.5 (2 CH) and 42.9 (2 CH2).4,5-Dihydrothiazoles 21 and thiazoles 22 and 23. The 4,5- dihydrothiazoles 21a–c were prepared as previously described 10 by acylation of the appropriate amino alcohol 7 with benzoyl chloride to give 24 followed by treatment with P2S5. These were then used to obtain the 2,4-disubstituted thiazoles 22a 23b and 23c by treatment with sulfur.10 2-Phenylthiazole 23a was prepared by the method of Lawson and Searle,16 as a colourless oil (25%) bp (oven temp.) 160 8C at 1.0 Torr (lit.,16 267–279 8C at 760 Torr); dH 7.95–7.90 (2 H m) 7.81 (1 H d J 3) 7.35–7.30 (3 H m) and 7.20 (1 H d J 3); dC 168.2 (4ry) 143.6 (CH) 133.5 (4ry) 129.9 (CH) 128.9 (2 CH) 126.5 (2 CH) and 118.7 (CH).Preparation of N-benzylidenebenzylamine. Benzaldehyde (5.43 g 51.2 mmol) was added to a stirred solution of benzylamine (5.48 g 51.2 mmol) in toluene (150 cm3). Heating under reflux for 1 h using a Dean–Stark separator followed by evaporation J. Chem. Soc. Perkin Trans. 1 1997 2145 of the solution afforded a yellow oil which was Kugelrohr distilled to yield N-benzylidenebenzylamine (9.0 g 90%) as a colourless oil bp (oven temp.) 175 8C at 0.5 Torr (lit.,21 200–202 8C at 10–20 Torr); dH 8.25 (1 H s) 7.7 (2 H m) 7.35–7.1 (8 H m) and 4.7 (2 H s); dC 162.6 (CH) 140.1 (4ry) 136.9 (4ry) 131.4 (CH) 129.3 (2 CH) 129.2 (2 CH) 129.0 (2 CH) 128.7 (2 CH) 127.7 (CH) and 65.6 (CH2).Preparation of 3,4-dihydro-5-methyl-2H-pyrrole 27. An ethereal solution of methyllithium (1.4 M; 37.5 cm3 50 mmol) was cooled to 220 8C and a solution of N-vinylpyrrolidin-2-one (5.0 g 45 mmol) dissolved in diethyl ether (50 cm3) was added dropwise over a period of 2 min. The mixture was stirred for a further 2 min at 220 8C and then 1 M hydrochloric acid (70 cm3) was added and the mixture stirred for an additional 2 min. The organic layer was separated and extracted with dilute hydrochloric acid the combined aqueous layers were washed with diethyl ether and then treated with aqueous sodium hydroxide until pH 10 was reached. The imine was extracted with CH2Cl2 and the extracts combined dried evaporated and Kugelrohr distilled to afford 3,4-dihydro-5-methyl-2H-pyrrole (1.83 g 49%) as a colourless oil bp (oven temp.) 50 8C at 14 Torr (lit.,22 103–105 8C at 760 Torr); dH 3.38 (2 H m) 2.10 (2 H t J 9) 1.66 (3 H s) and 1.50 (2 H quintet J 9); dC 174.7 (4ry) 61.1 (CH2) 38.7 (CH2) 23.0 (CH2) and 19.7 (CH3).Photolysis of 5c A solution of 5c (20 mg) in [2H6]acetone (0.5 cm3) in a dry NMR tube was irradiated with a 100 W medium pressure mercury lamp. After 10 days the NMR spectra showed the dissolved material to be completely unchanged but a small crystal (ª2 mg) had been deposited which was found to be 2- benzylamino-3-phenylpropane-1-sulfonic acid 28 (Found C 61.7; H 6.4; N 4.5. C16H19NO3S?0.4H2O requires C 61.5; H 6.4; N 4.5%); dH(CD3SOCD3) 9.4–9.2 (1 H br s) 9.2–9.0 (1 H br s) 7.6–7.4 (5 H m) 7.4–7.2 (5 H m) 4.5–4.3 (2 H m) 3.65 (1 H m) 3.35 (1 H half AB pattern of d J 20 15) 2.90 (1 H half AB pattern of d J 20 10) 2.70 (1 H half AB pattern of d J 15 10) and 2.65 (1 H half AB pattern of d J 15 4) dC(CD3SOCD3) 136.0 (4ry) 132.0 (4ry) 129.5 (2 CH) 129.3 (2 CH) 129.0 (CH) 128.8 (2 CH) 128.6 (2 CH) 127.0 (CH) 56.5 (CH) 48.8 (CH2) 47.5 (CH2) and 35.2 (CH2).Acknowledgements We thank Zeneca Pharmaceuticals and SERC for a Case studentship (D. P. A.) and the Royal Society for a Warren Research Fellowship (R. A. A.). References 1 R. A. Aitken I. Gosney and J. I. G. Cadogan Prog. Heterocycl. Chem. 1992 4 1; 1993 5 1. 2 J. M. Decazes J. L. Luche H. B. Kagan R. Parthasarthy and J. Ohrt Tetrahedron Lett. 1972 3633; D. Bellus Helv. Chim. Acta 1975 58 2509. 3 M. R. Johnson M. J. Fazio D. L. Ward and L. R. Sousa J. Org. Chem. 1983 48 494. 4 J. M. Bohen and M.M. Joullié J. Org. Chem. 1973 38 2652; W. Hanefeld and M. A. Jalili Liebigs Ann. Chem. 1986 1787. 5 Preliminary communication R. A. Aitken D. P. Armstrong S. T. E. Mesher and R. H. B. Galt Tetrahedron Lett. 1994 35 6143. 6 R. J. Gaul and W. J. Fremuth US Pat. 3 006 919 1961; R. J. Gaul and W. J. Fremuth J. Org. Chem. 1961 26 5103. 7 D. Delaunay L. Toupet and M. Le Corre J. Org. Chem. 1995 60 6604. 8 H. J. Roth and H. Schlump Arch. Pharm. (Weinheim Ger.) 1963 296 213. 9 A. G. M. Barrett D. H. R. Barton and R. Colle J. Chem. Soc. Perkin Trans. 1 1980 665. For a general discussion of oxidation of these ring systems see R. A. Aitken D. P. Armstrong and S. T. E. Mesher Prog. Heterocycl. Chem. 1990 2 1. 10 R. A. Aitken D. P. Armstrong R. H. B. Galt and S. T. E. Mesher J. Chem. Soc. Perkin Trans.1 1997 935. 11 C. Roussel J.-L. Stein and F. Beauvais New J. Chem. 1990 14 169. 12 K. Schank and F. Werner Liebigs Ann. Chem. 1979 1977. 13 D. H. R. Barton D. P. Manly and D. A. Widdowson J. Chem. Soc. Perkin Trans. 1 1975 1568. 14 T. C. Farrar B. M. Trost S. L. Tang and S. E. Springer-Wilson J. Am. Chem. Soc. 1985 107 262. 15 G. Barbarella Prog. Nucl. Magn. Reson. Spectrosc. 1993 25 317; S. Berger S. Braun and H.-O. Kalinowski NMR Spektroskopie von Nichtmetallen Thieme Stuttgart 1992 vol. 1 p. 119. 16 A. Lawson and C. E. Searle J. Am. Chem. Soc. 1957 79 1556. 17 W. J. M. van Tilborg and R. Plomp J. Chem. Soc. Chem. Commun. 1977 130. 18 S. Itsuno K. Ito A. Hirao and S. Nakahama J. Chem. Soc. Perkin Trans. 1 1984 2887. 19 J. R. Piper and T. P. Johnston J. Org. Chem. 1963 28 981.20 J. T. Sharp I. Gosney and A. G. Rowley Practical Organic Chemistry Chapman and Hall London 1989 p. 51. 21 M. Freifelder M. B. Moore M. R. Vernstein and G. R. Stone J. Am. Chem. Soc. 1958 80 4320. 22 J. Bielawski S. Brandage and L. Lindblom J. Heterocycl. Chem. 1978 15 97. Paper 7/00521K Received 22nd January 1997 Accepted 25th March 1997
ISSN:1472-7781
DOI:10.1039/a700521k
出版商:RSC
年代:1997
数据来源: RSC
|
|