Non-rearrangement Reactions of the Neopentyl-Oxygen Bond. New Syntheses of Neopentyl Halides

Feb. 5, 1954 NON-REARRANGEMENT REACTIONS OF t o react with 6 g. of anhydrous aluminum chloride over a period of two hours. The gas evolved during the reaction was passed into bromine and identified as ethylene by the formation of ethylene bromide, b.p. 128-130°, n Z o D 1.5387, in 64% yield. Fractionation of the liquid reaction product gave 19 g. (0.12 mole) of diethyldichlorosilane, b.p. 127128' (734 mm.), nZ0D 1.4311, d*o 1.0507, neut. equiv. 79.0 (calcd. 78.6), M R D38.72 (calcd. 38.8), a yield of 75%. Chloromethyldimethylchlorosilane and Aluminum Chloride.-Chloromethyldimethylchlorosilane,'2 35.0 g. (0.25 mole) was refluxed for three hours with 2.0 g. of anhydrous aluminum chloride. During this time the originally waterwhite liquid became almost black. Distillation, however, gave 34.0 g. (0.24 mole) of chloromethyldimethylchlorosilane, b.p. 112-115' at 735 mm., neOD 1.4360, a 96%recovery of the original compound. Dichloromethyldimethylchlorosilane and Aluminum Chloride .-Dichloromethyldimethylchlorosilane, 62 .O g . (0.35 mole), was refluxed for two hours with 3.0 g. of anhydrous aluminum chloride. The same black coloration noted in the above experiment with chloromethyldimethylchlorosilane developed almost a t once. After decanting the liquid product from the catalyst, distillation gave 43.5 g. (0.25 mole) of dichloromethyldimethylchlorosilane, b.p. 146-148", 70% recovery of the starting material. (12) R.H. Krieble and J. R. Elliott, THIS JOURNAL, 67, 1810 (1945). THE NEOPENTYL-OXYGEN BOND 80.7 Intramolecular Rearrangement of Dichloromethyltrimethy1silane.-The violence and exothermal nature of the action of anhydrous aluminum chloride on dichloromethyltrimethylsilane is such as to necessitate a warning by the authors that unless care is exercised this reaction may lead to explosions. The procedure described above for the rearrangement of a-chloroethyltrimethylsilane was found to be satisfactory when adapted for dichloromethyltrimethylsilane. It is important (for reasons of safety) to note that the organosilicon compound is added to the aluminum chloride in both cases. During six hours, 125.0 g. (0.80 mole) of dichloromethyltrimethylsilanels reacted exothermally with 4.5 g. of anhydrous aluminum chloride. Ethylene was evolved and was converted by liquid bromine into ethylene bromide, 115 g. (0.61 mole), b.p. 129-130" (735 mm.), ~ Z O D 1.5380, in 76% yield. The crude product in the reaction flask, 92 g., was combined with 10 g. of liquid which was collected in the Dry Ice-acetone trap and distilled to give 91 g. (0.71 mole) of dimethyldichlorosilane,14 b.p. 69-70" (737 mm.), % C1 54.4 (calcd. 54.9), in 88% yield. (13) J. L. Speier and B. F. Daubert, ibid., 70,1400 (1948). (14) W. F. Gilliam, H. A. Liebhafsky and A. F. Winslow, ibid., 63, 801 (1941). STATE COLLEGE, PA. [ CONTRIBUTION FROM THEWHITMORE LABORATORY, PENNSYLVANIA COLLEGE THE STATE ] Non-rearrangement Reactions of the Neopentyl-Oxygen Bond. New Syntheses of Neopentyl Halides1 BY LEO H. SOMMER, HERBERT BLANKMAN PAUL MILLER D. AND C. RECEIVED APRIL 18, 1953 New syntheses for neopentyl bromide and neopentyl chloride are reported which make these interesting aliphatic halides readily available in contrast to the previous methods used. These syntheses also provide unequivocal examples of reactions of the neopentyl-oxygen bond proceeding without rearrangement. Anionoid substitution reactions of neopentyl compounds have long been known to proceed with rearrangement of the carbon skeleton.2 C c-LC-x I C Y I I +c-c-c-c C Indeed, these changes are often cited in textbooks as classical simple examples of the Wagner-Meerwein rearrangement. While these changes are of considerable interest for the theory of intramolecular rearrangements, neopentyl compounds, especially the neopentyl halides, were previously rather difficult to prepare, since the chloride was available only from vaporphase chlorination of neopentane, s and the bromide was made from the chloride by a multi-step procedure involving conversion to the Grignard reagent, treatment of the latter with mercuric chloride and, finally, reaction of neopentylmercuric chloride with bromine.2c I n a recent elegant series of papers, however, Dostrovsky, Hughes and Ingold4 have demon(1) For a preliminary Communication see L. H. Sommer, H. D. Blankman and P. C. Miller, THIS JOURNAL, 73,3542 (1951). (2) (a) F. C. Whitmore, ibid., 64, 3274 (1932); (b) F. C. Whitmore and H. S. Rothrock, ibid., 54, 3431 (1932); (c) F. C.Whitmore, E. L. Wittle and B. R. Harriman, ibid., 61, 1586 (1939). (3) Whitmore and Fleming, ibid., 66, 4161 (1933). (4) I. Dostrovsky, E. D. Hughes and C. K. Ingold, J . Chem. Soc., 157 (1946). strated rearrangement for neopentyl bromide in SN~ reactions and non-rearrangement in S N ~ reactions. This suggested that similar non-rearrangement reactions of the neopentyl-oxygen bond might be isolated by appropriate structural variations in neopentyl alcohol (which left intact the neopentyl-oxygen configuration) or by reaction conditions unfavorable to reaction of the alcohol by an S Nmechanism. ~ I n this article we report unequivocal examples of reactions of the neopentyl-oxygen bond proceeding without rearrangement, which provide new and convenient syntheses of neopentyl bromide and neopentyl chloride. Me3CCH20SiEt8 + PBr3 CoH79,HX + EtsSiBr (1) + SOCIz Me3CCH2Cl + EtaSiCl MeXCH20H + PBrj + CoH7N--+ MesCCHBBr + CSH;?;.HBr (2) MesCCHzBr MesCCHzOSiEts CoH7X.HX (3) An important feature of the above reactions is the choice of reactants which will not give hydrogen chloride or hydrogen bromide as reaction products (5) Subsequent to the completion of this work, F. G. Bordwell, E.M. Pitt and M. Knell, THISJOURNAL, 73, 5004 (1951), reported nonrearrangement reactions of neopentyl fi-toluenesulfonate. Evidence for non-rearrangement of d-pinacolyl hydrogen sulfate in sulfuric acid during racemization has been presented by N. C. Deno and M. S. Newman. ibid.. 73, 1920 (1951). 804 L. H. SOMMER, n. L A N K M A N H. B (in 3 an excess of quinoline is available for imrnediate combination with hydrogen bromide). Hydrogen bromide gives extensive rearrangement with neopentyl alcohol.2b Experimental Neopentyl Alcohol.-To 27 moles of t-butylmagnesium chloride in 10 liters of ether contained in a copper-lined stirred Grignard reactor there was added (12 hours) 13.5 moles of methyl formate. After stirring the reactants for an additional 12 hours, the product was hydrolyzed in the usual way. Fractionation gave 9.75 moles (72y0 yield) of neopentyl alcohol, b.p. 111" (730 mm.), m . p . 51-52". This convenient preparation of neopentyl alcohol is made possible by Grignard reduction of the intermediate trirnethylacetaldehyde. Triethylneopentoxysilane.-In a 12-liter, three-necked, round-bottomed flask equipped with a condenser, mercurysealed anchor-type glass stirrer, and protected from moisture by a sulfuric acid trap were mixed 514 g. (5.84 moles), of neopentyl alcohol, 770 g. (5.95 moles) of quinoline and 2700 cc. of dry benzene. T o this mixture there was added, with rapid stirring, over a period of 2.5 hours, 981 g. (6.50 moles) of triethylchlorosilane.6 After the addition, the reactants were heated on the steam-bath with stirring for an additional 20 hours. Fractionation followed by treatment of the triethylneopentoxysilane with silica gel t o remove quinoline and quinoline hydrochloride gave 4.97 moles (85% yield), h.p. 63"(7 mm.), ? z ~ O D 1.4189. Anal. Calcd. for CllH260Si: Si, 13.84. Found: Si, 13.78. Neopentyl Bromide from Triethylneopentoxysi1ane.-In a two-liter glass-jointed, round-bottomed flask were placed 411 g. (2.03 moles), of triethylneopentoxysilane, 1078 g. (3.97 moles) of phosphorus tribromide, and 3 . 0 g. of quinoline hydrochloride. T h e flask was attached to a n efficient multiple-bulb condenser, which was open to the atmosphere through a Dry Ice-ethanol trap and a concd. sulfuric acid trap. Mixing of the reactants was accompanied by a temperature rise of 45'. T h e reaction mixture was then heated a t reflux temperature for 16 hours. During t h a t period the reaction temperature dropped from 173 to 148'. After cooling, fractional distillation gave 264 g. (1.75 moles) of pure neopentyl bromide, h.p. 104" (733 m m . ) , n 2 0 1.4371, d201.199; lit.2a ~ h.p. 105', n Z o ~ 1.4370, d20 1.199; anilide, m.p. and mixed 130'; less than 0.5% reaction with S a O E t in EtOH at reflux for four hours; (calcd. for C5H::Br: Br, 52.9. Found: Br, 52.9). The yield of pure neopentyl bromide was 86%. Hydrolysis of the higher-boiling material gave hexaethyldisiloxane, from hydrolysis of EttSiBr, in 94% yield. Neopentyl Chloride from Triethy1neopentoxysiane.-The apparatus used here was the same as t h a t used above in the neopentyl bromide preparation. In the reaction flask were placed 413 g. (2.05 moles) of triethylneopentoxysilane, 3.73 moles of thionyl chloride and 2.9 g. of quinoline hydrochloride. IVhen the contents of the flask were well mixed, (6) For a convenient preparation see P. A. Di Giorgio, W. . Strong, A L. H. Sommer and F. C. Whitmore, THIS JOURNAL, 68, 1380 (1946). AND P. c. MILLER 1701. 76 a temperature rise of 40" was noted, causing the contents t o start t o boil. The flask was quickly charged t o t h e condenser and heated; the time of heating was 23 hours; the temperature of the reactants in the flask during heating at reflux dropped from 115 t o 103". After cooling t o room temperature, the reaction products were added to cracked ice to remove excess thionyl chloride. The organic layer was separated, washed with water five times, and then dried over Drierite. Fractional distillation gave 132 g. (1.24 moles, 60% yield) of pure neopentyl chlo~ ride, b . p . 83.5' (725mm.), n 2 0 1.4043, d200.8659, lit.3b.p. 83.5 (740 mm.), T Z ~ O D 1.4043, d20 0.865; completely inert t o NaOEt in EtOH; anilide, m.p. and mixed 130" (calcd. for COH11Cl: C1, 33.27. Found: Cl, 32.58). Neopentyl Bromide from Neopentyl Alcohol.-In a twoliter, three-necked, round-bottomed flask containing a thermometer well and fitted with a dropping funnel and mercurysealed stirrer and protected from moisture by a sulfuric acid trap, were placed the following: 176 g. (2.0 moles) of neopentyl alcohol, 313 g. (2.43 moles) of quinoline and 740 cc. of dry bromobenzene. The flask was placed in an ice-salt-bath and the contents Then, over a period of six hours, 402 g. cooled t o -,5'. (1.48 moles) of phosphorus tribromide was added with stirring. During the addition the temperature was allowed to rise gradually until i t reached 15'. The material in the flask was transferred, with the aid of a n additional 150 cc. of bromohenzene, t o a 2-liter glass-jointed flask which was attached t o a n efficient multiple-bulb condenser; heat was applied t o the flask. The time of heating was 24 hours, the temperature of the mixture dropping from 181 to 162' during this period. The reaction mixture was then allowed to cool, giving a clear reddish liquid over a red, viscous residue. The liquid was decanted and the residue broken up and extracted with three 65-cc. portions of bromobenzene. The combined bromobenzene solution was poured onto ice t o remove excess phosphorus tribromide; the resulting organic layer was washed with water and dried, first over calcium chloride and then over Drierite. Fractional distillation gave 143 g. (0.93 mole) of pure neopentyl bromide, b.p. 105", %*OD 1.4370, inert t o NaOEt in EtOH, anilide m.p. and mixed m.p. 130°, a yield of 47%. Neopentyl Alcohol and Phosphorus Tribromide.-Reaction of 38 g. (0.433 mole) of neopentyl alcohol with 134 g. (1.49 moles) of phosphorus tribromide in 180 cc. of dry bromobenzene was carried out by a procedure similar to that used above. After heating for 19 hours at 131-136" the product was poured onto ice, separated and dried. Frartional distillation gave a mixture of amyl bromides, b . p . 104-114', ~ * O D 1.4380-1.4433; reaction to the extent of 485S70 with NaOEt in EtOH; 28.6 g., 44% yield. The Uncatalyzed Reaction of Triethylneopentoxysiane and Phosphorus Tribromide.-In the absence of quinoline hydrochloride, reaction of 52.1 g. (0.257 mole) of triethylneopentoxysilane with 107 g. (0.395 mole) of phosphorus tribromide, under conditions identical with those used in the quinoline hydrochloride-catalyzed reaction, gave a mix~ ture of amvl bromides. b.D. 106-112'. n Z o1.4376-1.4407. reaction with alcoholic pdtassium hydroxide to the evtenf of 20-40%; a yield of only 26%. STATECOLLEGE, PA.