March, 1948
PREPARATION AND PROPERTIES OF TRIMETHYLSILYLMETHANOL1117
distilled at 99.6" (750 mm.); %*OD 1.3910. It was assumed t o be an azeotrope between MZ (b. p. 100.4') and toluene (b. p. 110.8'). After the excess toluene was removed, the next fraction (10 g., 0.036 mole) was distilled a t about 250", leaving as a residue a low viscosity oil, which was assumed t o consist of linear compounds of the type MzD:. The 250" fraction was redistilled; b. p. 95O (5.0 mm.), ~ Z O D 1.4468, dzo4 0.9080, RD 0.2941, R D (calcd.) 0.2929. Anal. Calcd. for C13H2~Si30z(MzD'): C, 52.30; €1, 8.78. Found: C, 52.8, 52.6; H, 8.9, 8.7. Acknowledgment.-The author is indebted to Miss Mary L. Caldwell, who carried out the microanalyses for carbon and hydrogen.
[CONTRIBUTION FROM MELLON INSTITUTE
AND
Summary 1. Two stereoisomeric cyclic trimers and a cyclic tetramer have been obtained by the hydrolysis of m e ~ y l p h e ~ y ~ d i c ~ o r o s i ~ a n e . 2* Linear methylphenylpolysiloxanes have been prepared from the cyclic ones by means of hexamethyldisiloxane or tetramethyl-1,3-diphenyldisiloxane and an alkaline catalyst. 3. Evidence for an equilibrium between siloxanes, silano~sand silyl ethers has been found. SCHENECTADY, XEW
YORK
RECEIVED OCTOBER 3, 1947
DEPARTMENT OF CHEMISTRY, UNIVERSITY O F PITTSBURGH]
Preparation and Properties of Trimethylsilylmethanol BY JOHN L. SPEIER,B. F. DAUBERT AND R. R. MCGREGOR Organo-silicon compounds containing organic functional groups such as the carbinol group are not readily prepared by the usual methods of synthesis of organo-silicon compounds. The literature records the preparation of only two cornpounds in which this group is contained. The first is triethylsilylethanol of uncertain structure' obtained by the chlorination of tetraethylsilane, conversion of the chloride to the acetate and hydrolysis of the acetate. The second is l-trimethylsilyl-2-propanol prepared from trimethylsilylmethylmagnesium chloride and acetaldehyde. Since it was desired t o investigate the properties of an a-hydroxyalkylsilane, trimethylsilylmethano1 was synthesized in good yield through the sequence of reactions :
sodium carbonate, was slightly amber in color and weighed 416.5 g. (92%). The mixture was distilled through a three-foot Stedman column and was found to boil entirely between the temperatures of 136.2-136.8' a t 748 mm., all but 5 ml. distilling at the higher temperature. The ester has the following constants: b. p. 136.8" at 748 mm., [ n l Z 5 1.4060, ~ [dlz64 0.8667. Molar refraction: Calcd.6 for Me3SiCHfOAc, 41.31. Found: 41.37. Sapn. equiv.: Calcd., 146.2. Found: 146.5, 146.6. Anal. Calcd. for CeH140zSi: Si, 19.2. Found: Si, 19.3, 19.1. Preparation of Trimethylsilylmethanol .-Trimethylsilylmethyl acetate (420 ml., 2.5 moles) was dissolved in absolute methanol (9 moles) and acidified with 10 drops of concentrated sulfuric acid. After standing a t room temperature for two days, 165 ml. of the azeotropic mixture of methanol and methyl acetate was removed by distillation. The charge was then diluted with methanol (4 moles) and permitted t o stand four days a t room temperature before fractionation t o yield 268 ml. (80%) of constant boiling trimethylsilylmethanol, b. p. 121.6" KOAc a t 729 mm. The alcohol redistilled from lime had the HOAc following properties: b. p. 121.7-121.9' a t 751 mm., Me3SiCH&!lS [nIz6D1.4169, [d]Z6, 0.8261. Molar refraction: C a l ~ d . ~ 190"C. for MeaSiCHzOH, 31.83. Found, 31.71. The alcohol MeOH a strong odor resembling that of menthol. MeaSiCHzOAc --+ Me3SiCH20H has Preparation of Trimethylsilylmethyl 3,s-DinitrobenzoHzSQ ate.-Trimethylsilylmethanol (1.1g.) was added dropwise The trimethylsilylmethanol so prepared was found to 3,5-dinitrobenzoyl chloride (2.5 9.) and heated t o 90" no more hydrogen chloride was evolved. The mixto be approximately six times as reactive in form- until ture was cooled, and ground into a fine powder in the ing a phenylurethan as its carbon analog, neo- presence of two portions of warm 54% sodium carbonate pentyl alcohol, and approximately three times as solution. I t was heated with a third portion of sodium carbonate solution until it fused. On stirring vigorously reactive as methanol in the same reaction. as it cooled, solidification occurred. The solids were washed with water in a suction filter, dissolved in boiling Experimental 95% ethanol, and filtered while hot. On cooling t o Preparation of Trimethylsilylmethyl Acetate .-Chlororoom temperature, 2.3 g. of crystalline product was obmethyltrimethyl~ilane~(3.1 moles), potassium acetate tained which melted at 69.5-70". The mother liquor was (3.8 moles), and glacial acetic acid (420 ml.) were charged heated t o boiling and diluted with water t o the point of into a stainless steel high-pressure reactor and shaken a t incipient turbidity. Again on cooling, 0.9 g. of crystalline 190-192' for eighteen hours. The contents of the reactor product formed, m. p . 68.5-70". The two crops of cryswere then washed thoroughly with distilled water. The tals were combined and recrystallized once from 95y0 water-insoluble liquid, after drying over anhydrous ethanol. The recrystallized product melted a t 70-70.5". Further recrystallizations did not change this melting (1) Friedel and Crafts, Compl. rend., 61, 792 (1865). point. (2) Whitmore, Sommet, Gold and Van Strien, THIS JOURNAL, 69, Ana2. Calcd. for CI1Hl4O~NzSi:C, 44.26; €1, 4.73; 1551 (1947). N, 9.33; Si,9.41. Found: C,44.04; H, 4.41; N, 9.33; (3) Whitmore and Sommer, ibid., 68, 481 (1946), first prepared Si, 9.47,9.45. chloromethyltrimethylsilane and showed that it reacted with potassium acetate in glacial acetic acid. Preparation of (Trimethylsilylmethoxy) -trimethylsilane . (4) Chloromethyltrimethylsilane was prepared for use in this -Trimethylsilylrnethanol (19.7 g.) dissolved in dry
-
work by the method of Whitmore, Sommer and Gold, ibid., 69, 1976 (1947).
( 5 ) Warrick, ibid., 68, 2455 (1946).
1118
JOHN
L. SPEIER, B. F.DAUBERT AND R. R.MCGREGOR
Vol. 70
chloroform-quinoline solution was treated with an equiv- mole fraction of the silicon derivative present was calcualent amount of trimethylchlorosilane. The mixture, lated (Table I). which was kept cool during the addition of the trimethylTABLE I chlorosilane, separated into two layers. After the adCompeting dition was complete, the mixture was shaken vigorously alcohol Trial % Si in products of reaction KdKROH for fifteen minutes and then diluted with absolute ether Neopentyl 1 10.13 10.04 10.09 10.04 6.56 (280 ml.). The quinoline hydrochloride was removed by 2 10.12 10.10 filtration. The filtrate was distilled t o yield a fraction (36 ml., 84%) boiling at 129.8' at 738 mm. which was 3 10.07 found t o be (trimethylsilylmethoxy) -trimethylsilane of Methyl 1 9.53 9.56 2.92 the following properties: b. p. 129.8' at 738 mm., [ n l a S ~ 2 9.53 9.54 1.3971, [dIz64 0.7781. Molar refraction: Calcd.6 for 3 MepSiCHzOSiMer, 54.00. Found, 53.91. Anal. Calcd. 9.55 for CrHpoOSi2: Si, 31.81; C, 47.71; H, 11.42. Found: 4 9.53 Si, 31.4, 31.4; C, 47.86; H, 11.61. Ethyl 1 9.68 3.57 Reactivity of Trimethylsilylmethanol with Phenyl 9.62 2 Isocyanate Relative to Other Alcohols.-The method of Davis and Famums for comparing the relative reaction The relative reaction rate constants were calculated rate constants of alcohols by means of the competitive reaction of one equivalent of each of two alcohols for one according t o the formula given by Davis and Parnum,6 equivalent of phenyl isocyanate in benzene solution was where K s l / K a o ~is the ratio of t h e reaction rate constants adapted for use in studying the reactivity of trimethyl- of the silicon-containing alcohol and the other alcohol silylmethanol. These authors showed that the reaction concerned. Method of Analysis for Organo-silicon Compounds.of alcohols with phenyl isocyanate t o form the phenylurethans is rapid, essentially irreversible, complete, and All determinations of silicon content in the compounds free from side reactions. Also, the products can be in this communication were carried out by the following isolated quantitatively free from solvent and excess procedure: A "Vycor" brand Kjeldahl flask (3O-ml; capacity) was brought t o constant weight at about 800 alcohol. in a furnace. The flask was cooled in a desiccator, stopThe methyl, ethyl, neopentyl and trimethylsilylmethyl alcohols used in this study were all carefully dried and pered tightly t o keep moisture from the interior, and purified. commercial absolute methanol was refluxed permitted t o stand in the balance case for from twenty over magnesium shavings and distilled. Commercial t o thirty minutes so that the exterior surfaces reached absolute ethanol was refluxed two days over amalgamated equilibrium with atmospheric moisture. The weight of aluminum foil and distilled. The neopentyl alcohol' the flask and stopper was determined accurately. A sample was prepared from t-butylmagnesium chloride and formal- yielding between 500 and 1500 mg. of silica on oxidation dehyde. Only the portion of the product distilled from was added to the flask and weighed carefully. About lime boiling at 111.1' a t 737 mm., m. p. 50-53' was used. one and one-half ml. of 30% fuming sulfuric acid was The phenyl isocyanate was freshly distilled in vacuo im- added t o the flask t o dissolve the sample. Fuming nitric mediately before use t o ensure its being free of carbanilide. acid was cautiously added dropwise with gentle heating An equivalent of phenyl isocyanate (ca. 1 9.) re- between additions until the oxidized material in the flask acted with two equivalents of trimethylsilylmethanol in turned white. The flask was slowly brought t o a high dry benzene (10 ml.) in a vial closed with a well-fitted temperature to drive off all volatile material, and the flask ground-glass stopper. The vial was permitted to stand and contents brought t o constant weight a t about 800'. twenty-four hours at approximately 25'. The benzene The weighing procedure described above was employed and excess alcohol were then evaporated in a stream of again, the gain in weight of the flask being taken as due filtered air with final drying under high vacuum for t o silica formed by oxidation. The silica was quite twenty-four hours. The product so obtained had a melt- hygroscopic. Therefore, the flask during weighing was inrr Doint of 80-80.5'. Reueated recrvstallization from stoppered to exclude moisture. All organo-silicon comseGeral solvents caused no dhange in this melting point. pounds may be analyzed in this manner except those that Anal. Calcd. for Cl1HI1O2NSi: Si, 12.57; C, 59.23; are volatile or not readily soluble in fuming sulfuric acid. H, 7.68; 37, 6.28. Found: Si, 12.59, 12.61; C, 58.89, Discussion 59.10; H. 7.37, 7.46; N, 6.15. 6.07. The unusual reactivity of trimethylsilylmethaThe same technique was appfied t o a sample of neopentyl alcohol t o produce a derivative melting sharply at 113'. no1 in forming the phenylurethan indicates a proRecrystallization from several solvents did not change nounced activating influence by the silicon atom this melting point. The phenylurethan of neopentyl upon the 0-H bond. Primary alcohols branched alcohol is reported as melting a t 114°.880 Trimethylsilylmethanol was made t o compete with the in the ,8 position are generally less reactive than alcohols mentioned essentially by the method of Davis their normal homologs. The results of several and Farnum.6 A dry benzene solution containing exactly different methods of determining the relative reequimolar amounts of trimethylsilylmethanol and neopentyl alcohol was prepared. From a buret there was activities of alcohols agree quite well in that they added a n amount of this solution calculated t o contain usually place the alcohols in the same relative exactly enough of each alcohol t o react with 1.096 g. of order. Table I1 shows some of the results of variphenyl isocyanate delivered from a carefully calibrated ous workers, and the results of this study. pipet into a glass-stoppered vial. The solution was From the data in Table I1 it can be seen that the quickly and thoroughly mixed and permitted t o stand a t least twenty-four hours at approximately 25'. In the d e e t of branching in the B position is most apparsame fashion solutions of trimethylsilylmethanol and ent in neopentyl alcohol. The same effects of methanol, trimethylsilylmethanol, and ethanol were pre- branching either are not present in the silicon anapared and reacted. The amounts of silicon in the non- log of neopentyl alcohol or they are completely volatile products of the reaction were determined and the (6) Davis and Farnum, THIS JOURNAL, 66, 883 (1934). (7) Rice, Jenkins and Hardin, J . A m . Phorm. Assoc.. 47, 303 (1938). (8) Whitmore and Rothrock, THISJOURNAL, 64, 3431 (1932). (9) Richard, Ann. Chemi0, [e] 91, 337 (1910).
masked by a pronounced activating influence of the silicon atom upoa the 0-H bond. The large covalent radius 1.17 A. for the silicon atom compared to 0.77 A. for carbon causes less packing and a u s less interaction between the methyl groups on . .
March, 1948
TERMOLECULAR MECHANISM OF TRIPHENYLMETHYL HALIDEDISPLACEMENTS1119
method of Kharasch and Marker.14 This refationship of reactivity of alcohol to electronegativity of the radical seems to hold true for the normal alcohols but glaring exceptions are noticeable where branching in the alcohol occurs. Thus (CHd aSiCHz0H ..... 2.92 ... * . .. CHaOH 1.000 1 .oo 1 .oo 1.oo 56.6 benzyl, neopentyl and t-butyl radicals are about C:HiH 0.961 0.82 0.46 0.81 46.9 equal in electronegativity, but the corresponding n-OHtOH .782 .79 46.9 .36 alcohols are quite different in relative reactivities. tt-C4HsOH .40 .972 .80 46.8 The reactivity of an alcohol probably is dependent n-CrHuOH .908 ., .43 .98 .. 4-GHoOH .. .I7 .66 .693 upon the electronegativity of the radical as well as (C:HI)(CHdCHCHeOH ..... ' 19 .. upon the steric effects of structure and the sepa(CHdrCCHIOH ..... .45 ... .. .. ration of the two effects is not possible. .30R ( CHa)nCHOH .054 .55 26.5 Trimethylsilylmethanol dissolves sodium slowly .. .53 22.6 ( C H I ) ~ ( C Z H ~ ) C H O H .321 .04 ,00318 .. ,015 .. a t room temperature with the evolution of hydro(CHalaCOH 1.4 00728 ,014 (CH3)n(CzHs)COH .8 gen. A sample (22 g.) containing a small amount the silicon atom than would be the case for similar of dissolved sodium was found to be unchanged by groups on a carbon atom, as has been pointed out heating a t 175' in a sealed tube for four days. by Whitmore and S ~ m m e r . The ~ interaction of The same sample diluted with a mixture of methathe closely packed groups on a carbon atom prob- nol (4ml.) and water (0.5 ml.) was heated a t 180' ably is partly responsible for the inactivity of neo- for sixteen hours followed by heating a t 210' for pentyl alcohol. The lack of this effect, however, two hours. No appreciable change was observed. The alcohol dissolved slowly in boiling 75% can scarcely explain the enhanced activity of its silicon analog which appears to be more active potassium hydroxide solution over a period of than any other alcohol. The explanation for the several hours to form a clear solution from which enhanced activity of this compound might be re- an infusible combustible gel was precipitated by lated to the highly electronegative nature of the the addition of acid. Apparently cleavage of Me3SiCHz-radical as found by Whitmore, Som- silicon-carbon bonds occurred resulting in the promer and B e r n ~ t e i n , ~who ~ ' ~found that order of duction of a polymeric siloxane material upon certain radicals was as follows: phenyl > Measi- acidification of the solution. Trimethylsilylmethanol dissolved clean amalCH2- > Me- > Et- > n-Pr- > (n-Bu-, %-Hex-)> Benzyl > (MeaCCH2-,t-Bu-) as determined by the gamated aluminum foil very quickly with the evo(10) Norris. Ashdown and Cortese, THISJOURNAL, 47, 837 (1925); lution of much heat when a trace of carbon tetrachloride was added to catalyze the reaction. 49, 2640 (1927); found these rate constants for esterification of alcohols with p-nitrobenzoyl chloride recalculated on basis of the rate Summary for methanol equal to unity. (11) Fehlandt and Adkins, ibid., 67, 193 (1935); Hatch and The synthesis of trimethylsilylmethanol and a Adkins, i b i d . , I S , 1694 (1937), found relative reactivities from study of its chemical reactivity is reported. TABLE I1 RELATIVEREACTION RATES OF ALIPHATICALCOHOLS K (this % reAlcohol KO work) K 10 K 1' actedl'
....
methanolysis studies. (12) Menschutkin. Ann. chim. phys., [ e ] 80, 81 (1883), reacted alcohols with acetic acid one hour at 155'. 60, 2626 (1938). (13) Whitmore and Bernstein, THISJOURNAL,
[CONTRIBUTION 8 0 . 1115 FROM THE GATESAND CRELLIN FROM THE DEPARTMENT OF CHEMISTRY,
(14) Kharasch and Marker, ibid., 48, 3140 (1926).
PITTSBURGH, PENNSYLVANIA RECEIVED SEPTEMBER 2, 1947
LABORATORIES OF CALIFORNIA INSTITUTE OF TECMNOLOGY A N D MASSACHUSETTS INSTITUTE OF TECHNOLOGY]
Kinetic Evidence for a Termolecular Mechanism in Displacement Reactions of Triphenylmethyl Halides in Benzene Solution BY C . GARDNER SWAIN' According to current theory there are only two commonly occurring mechanisms for nucleophilic displacement reactions. One is the bimolecular, kinetically second order (sN2) substitution or "Walden In this mechanism the (1) National Research Fellow 1945-1946; American Chemical Society Fellow 1946-1947. Paper presented a t the N e w York meeting of the American Chemical Society, September 17,1947. (2) (a) Walden, Ber., 88, 1287 (1895); (b) Lewis, "Valence and the Structure of Atoms and Molecules," 1923, p. 113; (c) Hughes. Juliusberger, Masterman, Topley and Weiss, J . Chsm. Soc.. 1526 (1935); (d) A.G . Evms, "Reactions of Organic Halides in Solution," Manche ter University Press (1946); Trans. F a r a d a y SOC.,48, 719 ( 1 946).
driving force is exclusively a "push" or nzlcleophilic attack by the entering atom or group, which has an unshared pair of electrons that it is eager to donate in bond formation. The attack is entirely from the back side. An example is the reaction of primary halides, such as methyl bromide, with pyridine in benzene solution.
