Organic & Biomolecular Chemistry

Kinetic and thermodynamic stereocontrol in the atroposelective formation of sulfoxides by oxidation of 2-sulfanyl-1-naphthamides

Mark S. Betson, Jonathan Clayden*, Madeleine Helliwell and David Mitjans
School of Chemistry, University of Manchester, Oxford Rd., Manchester, UK M13 9PL. E-mail: clayden@man.ac.uk; Fax: +44 161 275 4939

Atropisomeric aromatic amides bearing 2-sulfanyl groups are oxidised by m-CPBA to the corresponding sulfoxides apparently with very high diastereoselectivity. NMR studies and oxidations of chiral benzamides however indicate that the kinetic selectivity of the oxidation is in fact relatively poor, and that the final diastereoisomeric ratio (typically >991) is under thermodynamic control, with relatively unhindered Ar–CO rotation readily converting the less stable to the more stable product diastereoisomer. Molecular mechanics indicates that the thermodynamic diastereoselectivity results principally from electrostatic repulsion between the CO and S–O dipoles.

Amide groups bonded to aromatic rings have the potential to display atropisomerism,¹ both when the amide is attached to the ring via the carbonyl group (e.g. 1)^2–4 and by the nitrogen atom (e.g. 2). We (for the former, benzamide class of amide)⁵ and others (mainly for the latter, anilide or maleimide, class)⁶ have demonstrated that these atropisomeric amides, with their perpendicular arrangement of the planar amide group and ring, undergo stereoselective reactions in which the amide is able to exert kinetic control over the formation of new stereogenic centres. We have reported in full our observations of the way in which an amide group may control addition of an ortholithiated amide to an aldehyde,³ addition of a nucleophile to formyl and acyl groups in the 2- or 8-positions of a naphthamide,⁷ and addition of electrophiles to a laterally lithiated amide.⁸ In this publication, we continue this series of publications with a report in full⁹ of the results of an investigation of the role of an adjacent conformationally constrained amide group in the stereoselectivity of oxidation of some sulfides to sulfoxides. We show that very high stereoselectivity may be a feature of this reaction, but only when the reaction is under thermodynamic control. The kinetically controlled sulfoxidation is by contrast relatively unselective.

Three sulfides were prepared by ortholithiation of N,N-diisopropyl-1-naphthamide 3, treating with s-BuLi in THF (the addition of TMEDA is unnecessary with diisopropylnaphthamides)¹⁰ and quenching the yellow organolithium with diphenyl, di-tert-butyl or dimethyl disulfide (Scheme 1).

	Scheme 1 Synthesis of 2-sulfanylnaphthamides.

The three sulfides were treated with m-CPBA at –15 °C in CH₂Cl₂ (Scheme 2) and the product mixtures were analysed by NMR spectroscopy both before and after purification. In principle, two diastereoisomers, anti- and syn-5, could be formed from each sulfide, but in the event >98% of a single diastereoisomer was observed by NMR in each case (for a more accurate quantification for 5a and 5b see below). In the case of the phenylsulfinyl-substituted amide 5c the product was crystalline and an X-ray crystal structure (Fig. 1) proved the anti relative stereochemistry; we assume the other isolated sulfoxides 5b and 5c are also the anti diastereoisomer.

	Scheme 2 Oxidation of 2-sulfanylnaphthamides.

	Fig. 1 X-Ray crystal structure of anti-5c.

This seemed to us a remarkable result, and in order to obtain and characterise the syn diastereoisomer we took the three sulfoxides anti-5a–c and heated them in toluene at 80 °C for 3 h. For anti-5b some decomposition was observed during this time, but no new peaks suggestive of another diastereoisomer appeared in any of the three samples.

This meant that, unless the syn and anti diastereoisomers in all three cases had identical NMR spectra, which we ruled out as unlikely, either 80 °C was an insufficiently high temperature to overcome the kinetic barrier to epimerisation of the sulfoxides, or the sulfoxide atropisomers were already at thermodynamic equilibrium and that one (the anti) diastereoisomer was much more stable than the other. Given that the second row elements P, Si and S typically present poor barriers to bond rotation in related amides,^4,11 probably due to the length of their bonds to C, we judged the latter, though remarkable, to be the more likely explanation for the high selectivity.

If the diastereoisomeric sulfoxides were indeed already in thermodynamic equilibrium then again there are two possible ways in which this ratio might have arisen: either the kinetic stereoselectivity of the sulfide oxidation happens to coincide with the thermodynamic stability of the products, or the initial kinetically controlled ratio of sulfoxides is, even at room temperature, rapidly overturned by easy equilibration to a thermodynamically controlled preponderance of the anti diastereoisomer.

To distinguish between these two possibilities we oxidised a sulfoxide already containing a chiral centre as well as the potentially atropisomeric amide axis. Sulfide 6 was available from previous work on the synthesis of potential atropisomeric chiral ligands.¹¹ If the kinetic selectivity of the amide-directed oxidation of 4 is high, then it is reasonable to expect the oxidation of 6 also to provide a sulfoxide with a high degree of stereoselectivity. However, the alternative mechanism, by which the oxidation of 4 gives a single diastereoisomer of 5 by virtue of epimerisation about the Ar–CO axis after oxidation, is not possible for 6 since, although Ar–CO epimerisation may still take place, there is no available mechanism for inversion of the relative stereochemistry between the sulfoxide and silicon-bearing stereogenic centres.

Oxidation of 6 with m-CPBA gave two separable diastereoisomers of the sulfoxide 7 (Scheme 3) in a 31 ratio. Both aspects of the relative stereochemistry of anti-7 are assigned on the basis of the discussion presented below; in syn-7 the axial stereochemistry is unknown.

	Scheme 3 Oxidation of a chiral sulfide.

From the poor selectivity in this reaction we concluded that the selective formation of anti-5 was not due to a simple stereoselective oxidation of 4 under kinetic control. A more reasonable explanation of the selective formation of anti-5 consists of a kinetically controlled oxidation of 4, of unknown selectivity, to yield a mixture of syn- and anti-5, followed by rapid (at room temperature at least) equilibration about the Ar–CO bond which converts syn-5 to the thermodynamically more stable anti-5 (Scheme 4).

	Scheme 4 Thermodynamic control over stereoselectivity.

In order to discover more about both the kinetic stereoselectivity of the oxidation and the rate of equilibration of the diastereoisomers we followed oxidation of 5a by NMR at temperatures below room temperature. We carried out three such experiments. In the first, 25 mg of sulfide 4a in CDCl₃ was treated, at –70 °C, with a slight excess of m-CPBA and the resulting mixture lowered into the NMR probe at 0 °C. Spectra were acquired every 2.7 min for 89 min, and the evolution of the spectrum during this time is shown in Fig. 2. The second experiment was identical except the temperature of the probe was –15 °C; spectra were acquired every 2.7 min for 84 min, and the evolution of the spectrum during this time is shown in Fig. 3. In the third experiment, a deficit (ca. 0.9 equiv.) of m-CPBA was added to 25 mg of sulfide 4a in CDCl₃ at –70 °C and the mixture was lowered into the probe at –55 °C. Spectra were acquired every 2.7 min for 73 min, and the evolution of the spectrum during this time is shown in Fig. 4.

	Fig. 2 Oxidation of 4a at 0 °C.

	Fig. 3 Oxidation of 4a at –15 °C.

	Fig. 4 Oxidation of 4a at –55 °C.

A total of four separate methyl singlets were discernible during these experiments: that of the starting sulfide 4a (at = 2.60), that of an over-oxidation product, the sulfone 8 (at = 3.25), that of the anti-sulfoxide anti-5a at = 2.92 and finally a further peak which we assume, for reasons given below, is the syn-sulfoxide syn-5a at = 2.98.

At 0 °C, oxidation of the sulfide is complete even before the first spectrum is acquired, and a mixture of sulfoxides and sulfone is immediately evident (Fig. 2). The ratio of anti to syn sulfoxides in the first spectrum is 928; after 10 spectra have been acquired (27 min) it is 973, and by the end of the experiment (89 min) it is a more or less constant 991. The proportion of sulfone does appear slowly to increase during the experiment, but assuming that any differential rate of oxidation of the two sulfoxides to the sulfone is negligible, the change in ratio confirms our rationalisation of the source of stereoselectivity. The conversion of syn- to anti-5a follows a first order rate law with k₁ (the rate constant for Ar–CO rotation) = 0.0005 s^–1 at 0 °C, or a half-life of 20 min.

At –15 °C, oxidation of the sulfide is incomplete in the first few spectra (Fig. 3), but an initial ratio of 7525 antisyn-5a is evident, which slowly decreases to 919 during the course of the reaction, as the amount of sulfone 8 also increases to about 15% of the total. However, we were able to establish that the change in the antisyn ratio is not due to selective oxidation of the minor isomer to the sulfone by warming this sample to 0 °C. After 3.5 h, the ratio of antisyn was 991, with no further increase in the proportion of sulfone. The conversion of syn- to anti-5a follows a first order rate law with k₁ = 0.0002 s^–1 at –15 °C, or a half-life of 50 min.

At –55 °C, with a deficit of m-CPBA, sulfide is never fully consumed, and no sulfone is formed. The ratio of antisyn-5a remains more or less constant between 6535 and 6733 (Fig. 4) This ratio presumably represents the true kinetic selectivity of the oxidation at –55 °C, and probably represents a good estimate of the kinetic selectivity at higher temperatures too.

The rate of equilibration of the diastereoisomers of 5a at 0 and at –15 °C extrapolates (assuming S = 0) to a half-life at 20 °C of less than 1 min, meaning that isolation of the unstable syn diastereoisomer of syn-5a, and presumably of sulfoxides syn-5 in general, is impossible. 2-Substituted naphthamides are typically atropisomeric¹² at room temperature,⁴ but comparably short half-lives for isomerisation about this bond are shown for example by 2-trialkylsilylnaphthamides;^4,13 presumably, like the C–Si bond, the C–S bond is relatively long and flexible.

Identifying the syn diastereoisomer in the ¹H NMR spectrum of 5a also allows us to quantify accurately the ratio of sulfoxides returned by work up of the oxidation at room temperature (Scheme 2). The syn-sulfoxide's methyl singlet at = 2.98 has a peak height less than that of the ¹³C satellites of the anti-sulfoxide's methyl singlet (Fig. 5a) at = 2.92, and we therefore have confidence in quantifying the ratio as >99.50.5. Given the fact that the rate of interconversion of the diastereoisomers at this temperature is fast, we also ran the NMR spectrum in CD₃OD to investigate whether the ratio of diastereoisomers was solvent-dependent The ratio of peak heights changed from 99.50.5 in CDCl₃ to 946 in CD₃OD (Fig. 5b). For the tert-butyl sulfoxide 5b a similar change in peak ratio was seen from >982 to 973, although in this case we cannot be certain that the minor peak is indeed the syn-sulfoxide.

	Fig. 5 (a) Portion of the 1H spectrum of 5a in CDCl3. (b) Portion of the 1H NMR spectrum of 5a in CD3OD.

Taken together these results point to a general mechanism for the oxidation of 2-sulfanyl-1-naphthamides to 2-sulfinyl-1-naphthamides in which an initial rather unselective (6535 at –55 °C, presumably even less selective at 20 °C) kinetically controlled oxidation is followed by equilibration of the mixture to a product ratio which is entirely under thermodynamic control because of the rapid interconversion of the sulfoxides by rotation about the Ar–CO bond, as shown in Scheme 4. A few other examples of the diastereoselective synthesis of atropisomers under thermodynamic control have been reported by ourselves^14,15 and by others.^16–18

The origin of the very strong thermodynamic preference for anti-5 over syn-5 was investigated by a molecular mechanics study. Two Monte Carlo conformational searches on 5a (5000 steps for each search) were carried out using Macromodel,¹⁹ one starting from a structure resembling anti-5a, with an Ar–CO dihedral angle of +90°, the other starting from a structure resembling syn-5a, with an Ar–CO dihedral angle of –90°. Energy minimisations were performed using MM2*, and in both searches the Ar–CO dihedral angle was constrained to within ±90° of its starting value, ensuring that only one of the two diastereoisomeric conformers about Ar–CO was minimised in each case.

The global minimum found from the first search was the conformer shown as ac-anti-5, with the sulfur lone pair occupying the most hindered inside position, and the CO and S–O dipoles aligned almost exactly antiparallel, giving an anticlinal (ac) relationship between Ar–CO and S–Me or S–O bonds. The second search found two minima, each shown in Scheme 5. The more stable conformer is shown as sc-syn-5, in which the CO and S–O dipoles are opposed but which also suffers a synclinal interaction between the O and Me substituents and the amide Ar–CO bond. The less stable conformer ac-syn-5 avoids this synclinal interaction by placing the S lone pair inside but pays the price by aligning the CO and S–O dipoles more parallel. The individual energy terms bear out this interpretation: ac-syn-5 has an electrostatic potential some 15–20 kJ mol^–1 higher than either of the other conformers, but is of comparable stability to ac-anti-5 in its steric (Stretch + Bend + Torsion + van der Waals) terms (Table 1). The synclinal conformer sc-syn-5a on the other hand has the lowest electrostatic potential, but incurs bending and torsional steric penalties. Only in anti-5a can both steric and electrostatic potential be minimised. The difference in energy between the lowest energy conformer of anti-5a and the lowest energy conformer of syn-5a is 10.4 kJ mol^–1, which generates, by Boltzmann distribution, a 98.51.5 ratio at 298 K, a figure within an order of magnitude of that observed spectroscopically. The slightly lowered conformational selectivity in the more polar solvent CD₃OD furthermore argues in favour of an electronic, rather than a purely steric, explanation for the selectivity.

	Scheme 5 Conformational preference in sulfoxides 5.

A pair of dihedral drive studies (Fig. 6) allowed us to locate the three significant conformers shown in Scheme 5 on the potential energy curve created by rotating (a) the anti and (b) the syn diastereoisomer of 5a about their ArC–S bond. Starting from the minimised conformation of each diastereoisomer the dihedral angle (Scheme 5) in anti-5a (solid circles) and syn-5a (open circles) was varied from 0 to 360° in 3° steps and the total energy was minimised at each step. For anti-5a by far the most stable conformer is ac-anti-5a ( = 240°), with a very shallow and inconsequential local minimum as the lone pair passes through the plane of the ring ( = 60°). For syn-5a, sc-syn-5a is a broad global minimum ( = 60 ± 30°), with another significant minimum resembling ac-anti-5a ( = 230°) and a shallow higher energy local minimum as the S–O bond passes through the plane of the ring ( = 180°). Whereas anti-5a lies happily in a low energy conformational well, syn-5a just can't get comfortable, in vain exchanging steric for dipole repulsions as the ArC–S bond rotates.

	Fig. 6 Energy profile for rotation about Ar–S in (a) anti-5a (solid circles) and (b) syn-5a (open circles).

We are currently exploiting the fact that sulfoxides may be formed in enantiomerically pure form as a means of controlling the absolute orientation of functional groups, enabling the asymmetric synthesis of certain classes of atropisomers.

General descriptions of spectrometers etc. have been provided before.¹⁵ Amides 3³ and 6¹¹ were prepared as previously described. The X-ray crystal structure of anti-5c has been deposited with the Cambridge Crystallographic Database (187714).

sec-Butyllithium (1.3 mol dm^–3 solution in hexane; 3.0 cm³, 4.0 mmol) was added dropwise to a stirred solution of N,N-diisopropyl-1-naphthamide 3 (1.003 g, 3.9 mmol) in THF (18 cm³) at –70 °C under nitrogen. After 35 minutes, dimethyl disulfide (0.5 cm³, 5 mmol) was added dropwise to the yellow solution, which was warmed to room temperature. The mixture was treated with water (50 cm³) and extracted with CH₂Cl₂ (3 × 10 cm³). The combined extracts were washed with water (2 × 10 cm³), dried (MgSO₄) and evaporated to give a crude product (1.076 g, 91%) which was recrystallised twice from petroleum ether to yield the sulfide 4a (0.615 g, 52%). R_f(petroleum ether–EtOAc, 41) 0.47; _max (film)/cm^–1 1620 (CO); _H (300 MHz; CDCl₃) 8.0–7.3 (6 H, m, ArH), 3.65 (1 H, septet, J 7 Hz, NCH), 3.55 (1 H, septet, J 7 Hz, NCH), 2.60 (3 H, s, SCH₃), 1.80 (3 H, d, J 7 Hz), 1.75 (3 H, d, J 7 Hz), 1.25 (3 H, d, J 7 Hz) and 1.05 (3 H, d, J 7 Hz) (CH₃ × 4); _C (75 MHz; CDCl₃) 167.9 (CO), 136.0, 131.7, 130.9, 129.9, 128.3, 127.9, 127.0, 125.9, 125.8, 124.5 (Ar), 51.3, 46.1 (NCH × 2), 21.2, 20.8, 20.7, 20.3 (CH₃ × 4), 17.2 (SCH₃); m/z (CI) 302 (100%, M + H), m/z (EI) 201 (100%, M – N{CH[CH₃]₂}₂) (Found: M⁺, 301.1505. C₁₈H₂₃NOS requires M, 301.15003).

In a similar way, sec-butyllithium (1.1 mol dm^–3 solution in hexane; 3.7 cm³, 4.1 mmol), N,N-diisopropyl-1-naphthamide 3 (1.004 g, 3.9 mmol), and di-tert-butyl disulfide (1.0 cm³, 5 mmol) in THF (18 cm³) gave a crude product (1.299 g, 96%) which was purified by flash chromatography (eluting with petroleum ether–EtOAc, 141) to yield the sulfide 4b (0.870 g, 65%) as a solid. R_f(petroleum ether–EtOAc, 41) 0.80; _max (film)/cm^–1 1632 (CO); _H (300 MHz; CDCl₃) 7.9–7.4 (6 H, m, ArH), 3.63 (1 H, septet, J 7 Hz, NCH), 3.43 (1 H, septet, J 7 Hz, NCH), 1.79 (3 H, d, J 7 Hz), 1.74 (3 H, d, J 7 Hz) (CH[CH₃]₂), 1.43 (9 H, s, SC[CH₃]₃), 1.22 (3 H, d, J 7 Hz) and 0.96 (3 H, d, J 7 Hz) (CH[CH₃]₂); _C (75 MHz; CDCl₃) 167.9 (CO), 142.5, 134.1, 133.1, 130.2, 127.8, 127.2, 126.78, 126.76, 126.6, 125.6 (Ar), 51.0 (NCH), 48.0 (C[CH₃]₃), 46.0 (NCH), 31.8 (C[CH₃]₃), 21.3, 20.7, 20.5, 20.0 (CH₃ × 4); m/z (CI) 344 (100%, M + H) (Found: M⁺, 343.1972. C₂₁H₂₉NOS requires M, 343.19697).

In a similar way, sec-butyllithium (1.1 mol dm^–3 solution in hexane; 3.6 cm³, 4 mmol), N,N-diisopropyl-1-naphthamide 3 (1.004 g, 3.9 mmol) and diphenyl disulfide (1.099 g, 5 mmol) gave a crude product (1.429 g, 99%) which was recrystallised with petroleum ether and ethyl acetate to yield the sulfide 4c (1.022 g, 72%) as a solid. R_f(petroleum ether–EtOAc, 41) 0.58; _max (film)/cm^–1 1630 (CO); _H (300 MHz; CDCl₃) 7.9–7.2 (11 H, m, ArH), 3.66 (2 H, m, NCH × 2), 1.81 (3 H, d, J 7 Hz), 1.74 (3 H, d, J 7 Hz), 1.23 (3 H, d, J 7 Hz) and 1.07 (3 H, d, J 7 Hz) (CH₃ × 4); _C (75 MHz; CDCl₃) 167.7 (CO), 137.5, 135.5, 132.2, 131.2, 130.0, 129.1, 128.8, 128.4, 128.0, 127.4, 127.12, 127.06, 126.4, 124.9 (Ar), 51.3, 46.3 (NCH × 2), 21.2, 20.9, 20.7, 20.3 (CH₃ × 4); m/z (CI) 364 (100%, M + H), m/z (EI) 363 (50%, M), 263 (100%, M – N{CH[CH₃]₂}₂) (Found: M⁺, 363.1666. C₂₃H₂₅NOS requires M, 363.16568).

m-CPBA (70%, 0.415 g, 1.7 mmol) was added to a stirred solution of the sulfide 4a (0.543 g, 1.8 mmol) in CH₂Cl₂ (9 cm³) at –15 °C. The mixture was stirred for 2 h 45 min. and poured into water. The CH₂Cl₂ layer was separated and washed with aqueous NaHSO₃, aqueous NaHCO₃ and water. The CH₂Cl₂ layer was dried (MgSO₄) and concentrated to give a solid which was purified by flash chromatography (eluting with EtOAc–petroleum ether 2.51) to yield the sulfoxide 5a (0.497 g, 78%) as a solid. Mp 157–157.5 °C; R_f(petroleum ether–EtOAc, 12) 0.13; _max (film)/cm^–1 1625 (CO), 1059 (SO); _H (300 MHz; CDCl₃) 8.2–7.5 (6 H, m, ArH), 3.69 (1 H, septet, J 7 Hz, NCH), 3.58 (1 H, septet, J 7 Hz, NCH), 2.89 (3 H, s, SO[CH₃]), 1.76 (3 H, d, J 7 Hz), 1.70 (3 H, d, J 7 Hz), 1.21 (3 H, d, J 7 Hz) and 1.08 (3 H, d, J 7 Hz) (CH₃ × 4); _C (75 MHz; CDCl₃) 166.1 (CO), 138.7, 134.6, 133.8, 130.0, 128.6, 128.0, 127.83, 127.82, 125.0, 118.5 (Ar), 51.6, 46.5 (NCH × 2), 43.4 (SCH₃), 21.0, 20.7, 20.5 and 20.2 (CH₃ × 4); m/z (CI) 318 (80%, M + H), 217 (100%, M – N{CH[CH₃]₂}₂), m/z (EI) 217 (100%, M – N{CH[CH₃]₂}₂) (Found: M⁺, 317.1452. C₁₈H₂₃NO₂S requires M, 317.14494).

In a similar way, m-CPBA (0.577 g, 2.3 mmol) and sulfide 4b (0.846 g, 2.46 mmol) in CH₂Cl₂ (14 cm³) gave a crude product which was purified by flash chromatography (eluting with petroleum ether–EtOAc, 11) to yield the sulfoxide 5b (0.657 g, 74%) as a solid. Mp 132.5–133.5 °C; R_f (petroleum ether–EtOAc, 41) 0.09; _max (film)/cm^–1 1632 (CO), 1043 (SO); _H (300 MHz; CDCl₃) 8.1–7.5 (6 H, m, ArH), 3.67 (1 H, septet, J 7 Hz, NCH), 3.47 (1 H, septet, J 7 Hz, NCH), 1.78 (3 H, d, J 7 Hz), 1.70 (3 H, d, J 7 Hz) (CH[CH₃]₂), 1.34 (9 H, s, SC[CH₃]₃), 1.28 (3 H, d, J 7 Hz) and 0.97 (3 H, d, J 7 Hz) (CH[CH₃]₂); _C (75 MHz; CDCl₃) 166.3 (CO), 137.5, 134.5, 134.0, 129.0, 128.5, 128.3, 128.2, 127.6, 125.7, 121.5 (Ar), 58.0 (C[CH₃]₃), 51.3, 46.4 (NCH × 2), 23.7 (C[CH₃]₃) 21.1, 20.6, 20.4 and 19.9 (CH₃ × 4); m/z (CI) 360 (80%, M + H), 259 (70%, M – N{CH[CH₃]₂}₂) (Found: M + H, 360.1996. C₂₁H₂₉NO₂S requires M + H, 360.1997).

In a similar way, m-CPBA (0.666 g, 2.7 mmol) and sulfide 4c (1.017 g, 2.8 mmol) in CH₂Cl₂ (15 cm³) gave a crude product which was purified by flash chromatography (eluting with petroleum ether–EtOAc, 51) to yield the sulfoxide 5c (0.758 g, 74%) as a solid. Mp 202–203 °C; R_f (petroleum ether–EtOAc, 41) 0.16; _max (film)/cm^–1 1627 (CO), 1046 (SO); _H (300 MHz; CDCl₃) 8.0–7.2 (11 H, m, ArH), 3.75 (2 H, m, NCH × 2), 1.83 (3 H, d, J 7 Hz), 1.78 (3 H, d, J 7 Hz), 1.36 (3 H, d, J 7 Hz) and 1.12 (3 H, d, J 7 Hz) (CH₃ × 4); _C (75 MHz; CDCl₃) 166.3 (CO), 144.7, 139.0, 136.0, 134.5, 130.3, 130.2, 129.0, 128.6, 128.5, 128.4, 127.8, 125.5, 124.3, 120.4 (Ar), 51.8, 46.6 (NCH × 2), 21.1, 20.9, 20.6 and 20.2 (CH₃ × 4); m/z (CI) 380 (100%, M + H), m/z (EI) 379 (20%, M) 279 (60%, M – N{CH[CH₃]₂}₂) (Found: M⁺, 379.1604. C₂₃H₂₅NO₂S requires M, 379.16059).

In a similar way, m-CPBA (70%, 0.025 g, 0.10 mmol) and amide 6 (0.047 g, 0.11 mmol) in CH₂Cl₂ (1 cm³) gave a crude product which was purified by flash chromatography (eluting with petroleum ether–EtOAc, 61) to yield 7a (0.012 g) and 7b (0.006 g). 7a: _max (film)/cm^–1 1627 (CO), 1046 (SO); _H (300 MHz; CDCl₃) 8.0–7.1 (8 H, m, ArH), 3.79 (1 H, septet, J 7 Hz, NCH), 3.66 (1 H, septet, J 7 Hz, NCH), 2.27 (1 H, q, J 7 Hz, CHCH₃Si), 1.68 (3 H, d, J 7 Hz, CHCH₃), 1.64 (3 H, d, J 7 Hz, CHCH₃), 1.43 (3 H, d, J 7 Hz, SiCHCH₃), 1.32 (3 H, d, J 7 Hz, CHCH₃), 1.23 (3 H, d, J 7 Hz, CHCH₃) and 0.2–0.0 (9 H, m, Si[CH₃]₃); m/z (CI) 430 (100%, M + H) (Found: M + H, 430.2246. C₂₃H₂₅NO₂S requires M + H, 430.2236).

Footnote

Crystal data C23H25NO2S; M = 379.50; orthorhombic P212121; a = 10.712 Å; b = 13.782 Å; c = 14.260 Å; V = 2105.2 Å3; T = 293(2) K; Z = 4; µ = 0.170 mm–1; 4202 reflections; Rint = 0.042; R1 = 0.0314 and wR2 = 0.0877 [I > 2(I)]. This data was reported in reference 9. CCDC reference number 187714. See http://dx.doi.org/10.1039/b511452g for crystallographic data in CIF or other electronic format.