Improved methods for the oxidation of primary and secondary alcohols

摘要

1614 J.C.S. Perkin IImproved Methods for the Oxidation of Primary and Secondary AlcoholsBy Derek H. R. Barton and Craig P. Forbes, Department of Chemistry, Imperial College, London SW7 2AYThe two-stage procedure for the oxidation of secondary alcohol chloroformates by treatment with dimethylsulphoxide and then with triethylamine has been much improved by the addition of an acid scavenger, conveniently1.2-epoxypropane, in the first stage of the sequence.In agreement with mechanistic considerations, treatment of cholestanol, or of its nitrite, with nitrosyl cationsin the presence of hexamethyldisiloxane as acid scavenger gave an excellent yield of cholestanone.SOME years ago we described a new procedure for theoxidation of alcohols (Scheme 1). The alcohol was con-verted into its chloroformate and the latter treated withdimethyl sulphoxide.After loss of carbon dioxide theintermediate alkoxydimethylsulphonium salt wastreated with triethylamine to give the carbonylcompound and dimethyl sulphide. This procedure gavebsol; COCl2 bsol;/ /CH-OH 4 CH-O-COCl-SCHEME 1good yields in the oxidation of primary alcohols toaldehydes, as further confirmed by a recent appli~ation.~The method has the advantage that the reagents arecheap and that all operations are carried out at roomtemperature or lower under near neutral conditions.Large-scale operations would be no problem.The mechanism by which the alkoxydimethyl-sulphonium salt eliminates dimethyl sulphide hassubsequently been clarified by isotope labelling experi-m e n t ~ .~ It is known that first the triethylamineabstracts a proton from one of the methyl groupsbonded to positive sulphur and then an intramolecularhydride transfer takes place.we found that a secondaryalcohol cholestanol (I) was oxidised to the ketone (11)in poor yield (20). We have now discovered that theyields of ketone can be improved remarkably by theaddition of a non-basic acid scavenger, suitably 1,2-epoxypropane, in the first stage of the reaction.When the oxidation of cholestanol (I) was attemptedin dry tetrahydrofuran in the usual way,l no cholestanone(11) was formed. The product was a mixture of un-changed alcohol (I) and 3p,3rsquo;p-methylenedioxydichole-stane (III).5 The structure of this acetal was confirmedIn our original workD.H. R. Barton, B. J. Garner, and R. H. Wightman, J .Chem. SOC., 1964, 1865.N. Kornblum, J. W. Powers, G. J. Anderson, W. J. Jones,H. 0. Larson, 0. Levand, and W. M. Weaver, J . Anzer. Ckem.SOG., 1957, 79, 6662; N. Kornblum, W. J. Jones, and G. J.Anderson, ibid., 1959, 81, 4113.N. Finch, J.J. Fitt, andI. H. S. Hsu, J . Org. Chem., 1975, 40,K. Torssell, Tetrahedrort Letters, 1966, 4445.J. E. Herz, J. Lucero, Y. Santoyo, andE. S. Waight, Canad.206.J . Chem., 1971, 49, 24181975 1615by n.m.r. spectroscopy, elemental analysis, and its readyhydrolysis to cholestanol. Its formation was favouredif the reaction was carried out in dichloromethane inthe presence of dissolved hydrogen chloride or phosgene.However, in the presence of a proton scavenger like(Yl1 ,2-epoxypropane, l-chloro-2,3-epoxypropane, or epoxy-ethylbenzene, the oxidation proceeded smoothly toproduce a reasonable yield of the desired ketone.Foursteroidal chloroformates were oxidised in this way byusing epoxypropane. The results are given in theTable. The low yield of ll-oxoprogesterone mayindicate that even our modified procedure is not applic-able to the formation of hindered ketones.Oxidation of steroidal chloroformates by dimethylsulphoxideChloroformate of Product ()Pregnenolone Progesterone 73YieldCholestanol Cholestanone 80Cholesterol Cholest-4-enone 711 la-Hydroxyprogesterone 1 l-Oxoprogesterone 24The other products of the oxidation were the alcoholfrom which the chloroformate was derived and a lowyield of a sulphur-containing steroid, probably themethylthiomethyl ether.The initial oxidation productsfrom cholesterol and pregnenolone chloroformates werethe amp;-unsaturated ketones. Isomerisation to theconjugated ketones by oxalic acid and ethanol wascarried out with scrupulous exclusion of oxygen prior towork-up. In one case, concentration and crystallisationof the crude product from cholesterol chloroformateafforded cholest-5-en-3-one in 47 yield, but generallythe product mixtures contained too much alcohol forpurification by crystallisation and the oxidation productswere isolated by column or preparative thin-layerchromatography.Despite variation of solvent, relative concentrations ofreactants, reaction times, and the absence or presence ofmoisture, some alcohol was always re-formed.Thisalcohol formation does not appear to be occurring in thefinal step of the reaction because the decomposition ofthe intermediate alkoxydimethylsulphonium salt (IV)was found to proceed in high yield without any alcoholformation. Furthermore, the presence of alcohol in thereaction mixtures was shown by t.1.c. in all cases priorto addition of triethylamine. It seems that alcoholformation is occurring primarily in the rearrangementof the intermediate (V).The salt (IV) formed a crystalline precipitate whichcould be isolated in analytically pure state by filtrationin a sealed system. It was unstable in air and meltedwith decomposition over the range 90-120 ldquo;C.Re-duction with sodium borohydride in bis- (2-methoxy-ethyl) ether gave 3@-cholestanol, with no detectable3a-cholestanol. This excludes the possibility that theintermediate (V) suffers an SN2 inversion by dimethylsulphoxide to give the salt (VI) prior to elimination.Such a mechanism could have explained the loss ofcarbon dioxide from the intermediate (V) and the pooryield in the formation of 1 l-oxoprogesterone.H + Y bsol;/ . / +/ + /CH-O-N=O lsquo;CH-O-N=Ovlsquo;CH-O-N-o -+ lsquo;CH-OH + NO+y=0lsquo;CH-O-N=O + NO+ bsol;CH---O-y=O/ /lsquo;C=O t H++ 2NO/SCHEME 2In a previous paper6 we proposed a mechanism(Scheme 2) for the acid-catalysed decomposition ofnitrites. If this mechanism is correct, treatment of analcohol or nitrite with a source of nitrosyl cations in thepresence of a proton scavenger should favour theformation of a ketonic product. This was found to bethe case.Treatment of cholestanol with nitrosyl tetra-fluoroborate in methylene chloride-acetonitrile in thepresence of hexamethyldisiloxane as scavanger produceda nearly quantitative yield of cholestanone. Benzylalcohol was similarly oxidised to benzaldehyde in 90yield. This is an interesting result because the chloro-formate method of oxidation of benzyl alcohol gave onlybenzyl chloride. The formation of ketone by nitrosationof nitrite supports our Scheme 2.We also examined the alkylation of nitrites as acomparable source of carbonyl compounds. Treatmentof cholest any1 nitrite with triethyloxonium tetrafluoro-borate gave cholestanone (50).EXPERIMENTALM.p.s were carried out on a Kofler hot-stage apparatus.Optical rotations were measured for solutions in chloroform.1.r.spectra were obtained on a Unicam SP 200 spectro-photometer. N.m.r. spectra (solutions in CDCI,) wereobtained on a Varian T60 spectrometer. Qualitative andpreparative t.1.c. was carried out on silica gel (G254)D. H. R. Barton, G. C. Ramsey, and D. Wege, J . Chem. SOC.(C), 1967, 19151616 J.C.S. Perkin Ideveloped with 20 ethyl acetate in light petroleum orbenzene. Column chromatography was carried out withgrade I11 alumina eluted with increasing proportions ofmethylene chloride in light petroleum. All solvents werepurified by standard procedures. Light petroleum refersto the fraction of b.p.60-80". All chloroformates wereprepared as in our previous publicati0n.lOxidation of Cholestan-3P-yZ Chloroformate.-(i) Cholestan-3p-yl chloroformate (2.15 g, 5 mmol) in dry tetrahydrofuran(10 ml) was treated with dimethyl sulphoxide (1.56 g,20 mmol) and stirred a t room temperature for 20 h.Triethylamine (0.8 g, 8 mmol) was added and the mixturewas stirred for 30 min, diluted with methylene chloride,washed with water, and dried (Na,SO,). T.1.c. indicatedthe presence of a mixture of cholestanol and a compoundof higher RF values.(ii) The reaction was repeated in methylene chloride(10 ml) containing dissolved hydrogen chloride (1 mmol) .After work-up t.1.c. indicated the presence of the same twocompounds but a greater amount of the high RF com-ponent.P.1.c. produced 3~,3'~-methylenedioxycholestane(111) (1.1 g), m.p. 203-206" (from acetone) (lit.,6 206-207"), 6 (CDCl,) 3.6 (2 H, s), and 4.9 (2 H, s) (Found: C,83.55; H, 12.0. Calc. for C,,H,,O,: C, 83.7; H, 12.25).(iii) The reaction was repeated in tetrahydrofuran in thepresence of 1,2-epoxypropane (480 mg, 8.3 mmol). T.1.c.indicated the presence of cholestanol and cholestanone (bycomparison with authentic samples). A minor high l i psulphur-containing steroid (palladium chloride spray re-agent) was also present but none of the acetal (111).Column chromatography gave pure cholestan-3-one ( 1.55 g,4 mmol), m.p. 130-132', mixed m.p. 130-132', a,3-44.5" (c 1.86).(iv) Cholestan-3P-yl chloroformate (670 mg, 1.5 mmol)in methylene chloride (4 ml) was treated with 1,2-epoxy-propane (140 mg, 2.5 mmol) and dimethyl sulphoxide(1 ml) and the solution was left a t room temperature for20 h.The crystals of choZestan-3~-yloxydimethylsul~honiumchloride (V) were filtered off in a Craig tube, washed fivetimes with small volumes of dry ether and dried undervacuum for 4 h; m.p. 90-120" (decomp.) (Found: C,71.6; H, 10.85; C1, 7.0; S, 6.3. Camp;amp;1OS requires C,71.75; H, 11.0; C1, 7.3; S, 6.6).The reaction was repeated and without removing thesalt the suspension was cooled to 0 "C, treated with sodiumborohydride (250 mg, 6.6 mmol) in bis-(2-methoxyethyl)ether (5 ml), and left for 3 h to warm to room temperature.It was diluted with 0.5~-hydrochloric acid and extractedwith ether.The organic phase was washed, dried, andconcentrated to leave a crystalline solid. T.1.c. showed thepresence of only cholestan-3P-01. No cholestan-3a-01 waspresent (comparison with an authentic specimen).Oxidation of Cholesterol ChZmoformate.-(i) Cholesterolchloroformate (10 g, 22.3 mmol) in tetrahydrofuran (60 ml)was treated with 1,2-epoxypropane (2.5 g, 44 mmol) anddimethyl sulphoxide (16 g, 200 mmol) and the mixture wasstirred under nitrogen for 20 h. Triethylamine (3 g, 30No ketone was formed.mmol) was added and stirring was continued for 30 min.The mixture was concentrated to 20 ml, treated withethanol (25 ml) and oxalic acid (2.0 g) and refluxed undernitrogen for 2 h. The solution was diluted with waterand extracted with methylene chloride, and the organicphase was dried and concentrated to a solid.Chromato-graphy gave pure cholest-4-en-3-one (6.0 g, 15.6 mmol).(ii) The reaction was repeated with cholesteryl chloro-formate (4.52 g, 10 mmol). The solution was concentratedto dryness without work-up. The residue in ether (30 ml)was treated with methanol (15 ml), concentrated to halfits volume, and cooled, to give cholest-5-en-3-one (180 mg,4.7 mmol), m.p. 119-124' (from ethanol), a, -4.4'(c 1.96).Oxidation of Pregnenolone Chloroformate.-The oxidationwas carried out with pregnenolone chloroformate (6.3 g,16.7 mmol) exactly as for cholesterol chloroformate. Afterisomerisation and work-up, chromatography gave pro-gesterone (3.76 g, 12.4 mmol), m.p.129-132' (fromethanol), aID +216" (c 2.4). An authentic sample crystal-lised in the same way had the same m.p. and a, f214"(c 1.08).Oxidation of 1 lu-Hydroxyprogesterone ChZo7oformate.-The oxidation was carried out with 1 la-hydroxyprogesteronechloroformate (490 mg, 1.25 mmol) exactly as for cholestan-3p-yl chloroformate. After work-up and concentration,p.1.c. gave 11-oxoprogesterone (100 mg, 0.32 mmol),identical with an authentic specimen.Oxidations with NitrosyZ Tetvufluorobwate.-Cholestan-3p-01 (883 mg, 1.5 mmol) in acetonitrile (12 ml) andmethylene chloride (3 ml) was treated with hexamethyl-disiloxane (2.9 g, 18 mmol) and nitrosyl tetrafluoroborate(1.10 g, 9.0 mmol) and stirred a t 0 "C for 2 h. The solutionwas washed with aqueous 5 sodium disulphite and water,dried, and concentrated to a crystalline solid (540 mg).Crystallisation from acetone gave pure cholestan-3-one,m.p. 128-130", a, +44" (c 1.8).(ii) Benzyl alcohol (326 mg, 3.0 mmol) in methylenechloride ( 15 ml) was treated with hexamethyldisiloxane(900 mg, 5.6 mmol) and nitrosyl tetrafluoroborate (670 m,5.5 mmol). The suspension was stirred for 15 min a t roomtemperature, worked up as above, and added to methanolic2,4-dinitrophenylhydrazine. The precipitated benzalde-hyde 2,4-dinitrophenylhydrazone (820 mg, 2.7 mmol),crystallised from benzene-ethanol, was identical with anauthentic specimen.AlkyZation of Cholestan-3P-yZ Nitrite.-Cholestan-3P-y1nitrite (417 mg, 1 mmol) in methylene chloride (5 ml) wastreated with triethyloxonium tetrafluoroborate (569 mg,3 mmol) and stirred under nitrogen a t room temperaturefor 50 min. The solution was washed with water, dried,and concentrated to leave a solid. Column chromato-graphy gave cholestan-amp;one (187 mg, 0.5 mmol).support and for gifts of chemicals.We thank the Schering-Plough Corporation for financial6/494 Received, 13th March, 1976