Silane Reductions in Acidic Media
(Rated as: excellent)
(Requested ref Post 433308 (not existing))
Journal of Organometallic Chemistry, 117 (1976) 129-140
Silane Reductions in Acidic Media
VIII. Boron Trifluoride Catalysed Organosilane Reductions. Selectivity and Mechanism
M. P. Doyle, C. T. West, S. J. Donnelly and C. C. McOsker
Reductions of organic compounds by organosilanes require activation of the carbon center by a suitable acid catalyst. The nature of the acid employed in these reactions influences the rate of reduction and, more importantly, the extent of reactions subsequent to the hydride transfer step . In trifluoroacetic acid media, for example, hydrosilylation of the carbonyl group by hindered organosilanes yields relatively stable alkyl silyl ethers ; subsequent reactions of the alkyl silyl ether with trifluoroacetic acid give alcohol and silyl trifluoroacetate products. With less sterically encumbered organosilanes such as triethylsilane, however, reactions subsequent to hydrosilylation are rapid and the observed reaction products from aldehyde and ketone reductions are alkyl trifluoroacetates and silyl trifluoroacetates.
Although Lewis acid catalysed reductions of carbonyl compounds by organosilanes have been less thoroughly investigated than those employing Bronsted acids, hydrosilylation is the principle result and subsequent reactions of alkyl silyl ethers with common Lewis acids are not observed [4-6]. Ketone reductions by triethylsilane employing catayltic amounts of anhydrous zinc chloride or aluminum chloride, for example, yield alkyl silyl ethers and, through a parallel competing process, symmetrical ethers. Rhodium(I) catalysts have more recently been investigated as hydrosilylation catalysts, principally to effect asymmetric reductions of prochiral ketones [7-10]; although the mechanism of these reductions is not known with certainty, the rhodium(I) catalyst does appear to act as a Lewis acid template for the reducible substrate.
In reductions by organosilanes the nature of the Lewis acid is expected to influence reaction selectivity as well as the stability of the hydrosilylation product. Metal fluoride catalysts, in particular, are of interest because of their pronounced ability to effect fluoride transfer to silicon derivates [11-13] and their potential influence on the geometry of the transition state during the transfer of hydride from silicon to carbon in hydrosilylation reactions . In this paper we report the stereochemistry and mechanism of organosilane reductions involving boron trifluoride.
Results and discussion
Organosilane reductions in boron trifluoride etherate
In contrast to hydrosilylation reactions catalysed by metal chlorides, aldehydes and ketones are rapidly reduced at room temperature by triethylsilane in boron trifluoride etherate primarily to borate esters and symmetrical ethers.
Triethylsilyl fluoride is the oxidation product. Product yields from reductions of representitive carbonyl compounds after basic hydrolysis are presented in Table 1.
Symmetrical ethers are formed preferentially in reactions with aldehydes and are less evident in ketone reductions; similar results were observed for triethylsilane reductions in trifluoroacetic acid . Increasing the concentration of the acid also leads to a decrease in ether formation. However, increasing the amount of ethyl ether leads to an increase in the relative yield of symmetrical ether; for example, employing BF3-2Et2O in a molar amount equivalent to that of cyclohexanone results in a 39% yield of cyclohexyl ether, nearly three-times more than that obtained with BF3-Et2O under similar conditions.
Boron trifluoride is consumed during the reaction of triethylsilane with carbonyl compounds. When less than one-third molar equivalent of boron trifluoride etherate is employed, reduction is incomplete; the extent of reaction corresponds to a stoichiometric requirement of one BF3 to three R2C=O for both borate ester (eq. 1) and ether (eq. 2) formation
a Reactions were run at room temperature. A 10% molar excess of triethylsilane to carbonyl compound was used.
b Relative yield of products.Borate esters were inferred from 1H NMR spectra of the reaction mixtures prior to basic hydrolysis but were not further identified. Attempts to isolate the borate reaction products from the reation of cyclohexanone with triethylsilane in BF3-Et2O by distillation led, instead, to cyclohexene which was recovered in 68% yield. The formation of cyclohexene by this procedure represents an unusual but facile one-pot reductive elimination (eq. 3).
In scheme 3 fluoride transfer occurs from a boron fluoride not associated with the carbonyl compound that is being reduced; the stereochemical course of the process may parallel that observed in the reduction of trityl halides by R3Si*H . in methylene chloride, racemization, or the high inversion of configuration in the reaction of R3Si*H with silver tetrafluoroborate in ether . Additional information is necessary to clarify the mechanism of carbonyl group reduction by unhindered organosilanes.
Stereoselectivity in cyclic ketone reductions
Table 2 presents the relative yields of the less stable alcohol isomer from reductions of alkyl-substituted cyclohexanones by triethylsilane in BF3-Et2O and compares these data to similar results from reductions in trifluoroacetic acid and aqueous sulfuric acid/ethyl ether media .
Table 2: Stereoselectivities of triethylsilane reductions of alkyl-substituted cyclohexanones
a Reactions were run at room temperature. Alcohols were the only products observed after basic hydrolysis.
b Data taken from ref. 22.
Use of a 3-fold molar excess of BF3-Et2O over ketone, completely suppresses symmetrical ether formation in these reductions; by comparison, with a [R2C=O]/BF3-Et2O molar ratio of 3.0 for the triethylsilane reduction of 4-tert-butylcychexanone, 4-tert-butylcyclohexyl ether is produced in 42% yield (44% cis,cis-, 46% cis,trans-, and 10% trans,trans-). Elimination to cycloalkenes is not observed even from reactions of triethylsilane with 2-methylcyclohexanone in BF3-Et2O.
The yield of the less stable alcohol isomer is consistently greater from reduction by triethylsilane in BF3-Et2O than from those in Bronsted acids. This increase in selectivity for hydride transfer from the equatorial direction is caused by the charge in the Lewis acid. The effects of Lewis acids (or metal ions) on reactivity and selectivity in hydride transfer reactions have been generally overlooked. These results and those from other laboratories [23,24] provide an increasing amount of evidence that in hydride transfer reactions Lewis acid complexation with the carbonyl oxygen plays a major roles in determining product selectivity.
Table 3 presents the stereochemical results from reductions of 4-tert-butylcyclohexanone by triethylsilane which emply selected catalysts. Compared with protonic acids, the use of Lewis acids generally leads to significantly higher relative yields of the less stable alcohol derivative. Indeed, by comparing the data in table 3 it is evident that changing the steric bulk of the acid required to activate the carbonyl group in the silane reduction dramatically affects the stereoselectivity of hydride transfer.
Table 3: Acid catalyzed reductions of 4-tert-butylcyclohexanone by triethylsilanea
a All reactions were run at room temperature.
b Observed redution product.
c Data taken from ref. 22.
d 27% cis, cis-, 53% cis, trans-, and 20% trans, trans-4-tert-butylcyclohexyl ether.
e 43% cis, cis-, 45% cis, trans-, and 12% i]trans, trans[/i]-4-tert-butylcyclohexyl ether.
f Redution of R2C=OCHR2 (R2CH = cis- or trans-4-tert-butylcyclohexyl, R2C=O = 4-tert-butylcyclohexanone) leading to symmetrical ethers during triethylsilane reduction of 4-tert-butylcyclohexanone in trifluoroacetic acid.
Because of competing symmetrical ether formation anhydrous aluminum chloride and stannous chloride were generally less satisfactory than either zinc chloride or boron trifluoride etherate for carbonyl group reductions by triethylsilane. In reductions catalyzed by zinc chloride, reation stereoselectivity was remarkably sensitive to the amount of zinc chloride employed and to the reaction temperature. Reductions at room temperature, catalyzed by 10 mol% of zinc chloride with respect to 4-tert-butylcyclohexanone, led to 67% of the less stable alcohol derivative; in contrast, when equal amounts of zinc chloride and ketone were used reation times were significantly shorter but the yield of cis-4-tert-butylcyclohexyl triethylsilyl ether was only 32%. Zinc chloride catalyzed triethylsilane reductions at 25° C were much slower than those run at 100° C but yielded a significantly higher relative yield of the less stable cis-isomer: 67% at 25° C, 33% at 100°C. These results are compatible with the coordination of zinc chloride with simple ketones to form momo- and di-carbonyl zinc chloride complexes (eq. 9, 10). [26,27]. Dicarbonyl complexes, offering greater steric restrictions for hydride transfer, promote
reductions that yield predominantly the less stable alkyl silyl ether isomer. Similar coordination is not possible with boron trifluoride and, thereofre, stereoselectivity in cyclic ketone reductions is not dependent on the molar ratio of ketone to boron trifluoride etherate.
methods and materials
Instrumentation has previously been described . A varian Model 485 digital integrator was used to determine peak areas in GLC analysis. Commercial boron trifluoride etherate was purified by distilation from calcium hydride through a 10 cm Vigreux column under a slow flow of nitrogen and was stored over calcium hydride under nitrogen in a refrigerator at 5° C. Purified boron trifluoride etherate could be stored and used for more than six weeks without noticeable deterioration. Othe anhydrous Lewis acids were commercial reagents and were not further purified. (+)-alpha-naphthylphenylmethylsilane having an optical rotation of +31° (ether, 91% optical purity) and racemic alpha-naphthylphenylmethylmethoxysilane were prepared by the method of Sommer .
General reduction procedure in boron trifluoride etherate
Boron trifluoride etherate was added dropwise to an ice-bath cooled and rapidly stirred solution of the reduceable substrate and organosilane which were contained in a round bottom flask fitted with a condenser and drying tube. After addition was complete the homogenous solution was allowed to warm to room temperature. In reactions with carbonyl compounds, addition of the boron trifluoride etherate initiated a mildly exothemic reaction. Generalyy a white precipitate formed as the reaction progressed. The process of the reaction was followed by 1H NMR analysis. Reaction times for aliphatic ketone and aldehyde reductions were less than one hour when one or more molar equivalents of BF3-Et2O to carbonyl compounds were used. Reductions of acetophenone and benzophenone were performed with a molar ratio of BF3-Et2O to ketone between 1.0 and 2.5; these reactions were complete within two hours. Benzoic acid, benzamide, ethyl phenylacetate, nitrobenzene, and 1-methylcyclohexene were not reduced by triethylsilane even after reaction times as long as six days. Upon complete reduction, an excess of 3N sodium hydroxide was slowly added to the reaction mixture which was then stirred at room temperature for four hours. After approximately 5 min of hydrolysis the white solid had completely dissipated. The hydrolyzed mixture was then extracted three times with ether, the combined ether extract was dried over and filtered from anhydrous magnesium sulfate, and the magnesium sulfate filter cake was rinsed several times with smallportions of ether.
The combined ether washes and extract were concentrated under reduced pressure, and the products were subjected to 1H NMR and GLC analyses. Triethylsilyl fluoride (>95%) and hexaethyldisiloxane (<5%) were the only silicon containing products from reactions employing triethylsilane following hydrolysis, as determined by spectral and chromatographic analyses. Mass spectrum (70 eV) Et3SiF: 136 (P + 2, 0.015), 135 (P + 1, 0.42), 134 (P, 3.20), 115 (45), 87 (12), 77 (100), 59 (30), 49 (29), 47 (33), 31 (10), 29 (11). Alcohols, symmetrical ethers, and alkenes were analyzed after hydrolysis; these compounds were the sole products from reductions of aliphatic ketones and aldehydes. Prior to hydrolysis borate ester products (RO)nBF3-n, were inferred from 1H NMR analyses: R = benzyl (s, delta 4.92 ppm), R = 1-octyl (t, delta 3.93 ppm[/i], R = cyclohexyl (m, delta 4.35-3.85 ppm). These 1H NMR spectra were identical to those of the tribenzyl, tri-1-octyl, and tricyclohexyl borate esters that were produced from boric acid and the corresponding alcohols by standard procedures.
Synthesis of cyclohexen from cyclohexanone
Following complete reduction of cyclohexanone (5.0g, 50 mmol) by triethylsilane (6.4g, 55 mmol) in BF3-Et2O (3.08g, 22 mmol) the reaction mixture was directly distilled at atmospheric pressure. Two fractions were collected and analysed by GLC and 1H NMR spectroscopy: fraction 1, b.p. 35-83° C and fraction 2, b.p. 83-140° C. Fraction 2 was composed of 10.1g of a mixture of triethylsilane (0.6g 5.0 mmol), triethylfluorosilane (6.7g, 50 mmol), and cyclohexene (2.8g, 34 mmol, 68% yield). Analysis of the pot residue after distillation showed cyclohexen, cyclohexyl ether, and unidentified materials.
Diphenylmethane from acylation-reduction
Boron trifluoride etherate (5.69g, 40.2 mmol) was added to a stirred solution of benzoyl chloride (1.43g, 10.2 mmol), benzene (3.16g 40.5 mmol), and triethylsilane (3.58g, 30.7 mmol) in a round bottom flask fitted with a reflux condenser and drying tube. After heating at 95° C for 18 h the reaction solution was cooled, quenched with 10% aqueous sodium hydroxide, and extracted as previously described. GLC analysis using an internal standard identified diphenylmethane in 30% yield. Identical yields of diphenylmethane were obtained when triethylsilane was added to the reaction solution subsequent to acylation of benzene by benzoyl chloride and when only 1.6 equivalent of BF3-Et2O, based on benzoyl chloride, was employed for acylation-reduction. Benzyl alcohol was not produced in these reactions. The relatively low yield of diphenylmethane was, therefore, a consequence of the acylation process rather than the reduction step and is consistent with yields from boron trifluoride etherate catyzed acylation reactions that employ acid chlorides rather than acid fluoride .
Reactions with (+)-alpha-naphthylphenylmethylsilane
To a mixture of (+)-alpha-naphthylphenylmethylsilane (1.24g, 5.0 mmol, 91% optical purity) and acetone (2.90g, 50 mmol) in a round bottom flask was added 1.30g of BF3-Et2O (9.2 mmol). Extreme care was used to avoid the introduction of water into the reaction mixture: oven dried glassware was used, BF3-Et2O was transferred by syringe, and reagents were added under a slow flow of nitrogen.
After reduction was complete (2 h), as evidenced by the disappearance of the delta 5.39 (q, Si-H, J = 4 Hz) and 0.65 ppm (d, Si-CH3, J = 4 Hz) signals and the appearance of the fluorosilane methyl doublet at delta 0.75 ppm (J = 7.5 Hz), the reaction solution was quenched with 15 ml of saturated sodium bicarbonate. Pentane was added, the resulting mixture was extracted, and the aqueous solution was washed twice with pentane. The combined pentane solution was washed twice with water and dried over anhydrous magnesium sulfate. Removal of pentane under reduced pressure gave an oil which was identified by 1H NMR sectroscopy as the fluorosilane: (CCl4, internal TMS) delta 0.75 (d, J = 7.5 Hz, 3H) and 7.2-82. ppm (m, 13H). Distillation of the oil at 0.3 Torr gave alphanaphthylphenylmethylfluorosilane (b.p. 129-132° C) in 78% isolated yield: [alpha]D = -4.2° (ether, ca. 0.81). Recrystallization of the fluorosilane had no effect on the specific rotation. In an alternative workup procedure, employed for the majority of the reductions with (+)-alpha-naphthylphenylmethylsilane, the reaction solution was evacuated to dryness after complete reduction. The remaining solid was recrystallized twice from pentane to give the same stereochemical result. Net inversion (in parentheses) was observed for acetone reductions with relative silane: BF3-Et2O: acetone concentrations given: 1: 1.80 : 10 (10%), 1 : 2.0 : 2.9 (13%), 1 : 1.9 : 2.3 (11%), 1 : 0.5 : 1.8 (9%). In a control experiment (+)-alpha-naphthylphenylmethylsilane was treated with a five-fold molar excess of BF3-Et2O. No change in the specific rotation of the silane was observed after three days at room temperature.
Reductions of menthone-isomenthone were performed as previously described. Racemic alpha-naphthylphenylmethylsilane, isolated by a procedure identical to that used by Sommer [11a], was the only silicon product from these reactions. The olefinic proton from menthene was evident in 1H NMR analyses of the reaction solutions (m, delta 5.2-5.3).
Attempted reductions of hexachloroacetone and di-tert-butyl ketone by (+)-alpha-naphthylphenylmethylsilane gave evidence for hydrolytic cleavage of naphthalene. In reactions eith hexachloroacetone phenylmethylfluorosilane was formed and dissipated during a 16 day period at 50° C:1H NMR (BF3-Et2O, internal TMS) delta 0.40 (doublet of doublets, J (F-CH3) = 7.5 Hz, J (H-CH3) = 2.5 HZ, 3 H) and 5.22 ppm (doublet of quartets, J (F-H) = 54 Hz, J (CH3-H) = 2.5 Hz, 1H). Naphthalene was isolated and identified by spectral and GLC methods. Similar observations were made in reductions of di-tert-butyl ketone, but reactiontimes were three-times longer. No evidence for the reduction of these ketones by alpha-naphthylphenylmethylsilane was obtained. In a separate experiment naphthalene was formed from alpha-naphthylphenylmethylsilane and BF3-Et2O when trace amounts of water were introduced into the reaction solution.
Redution of alkyl-substituted cyclohexanones by triethylsilane in BF3-Et2O
Reactions were performed as previously described. Product yields were determind by GLC analyses. Isomeric alcohols from 2-, 3-, and 4-methylcyclohexanone reductions were separated and analyzed on 5ft., 25% glycerol columns at 100° C. Isomeric alcohols from 4-tert-butylcyclohexanone reductions were separated and analyzed on a 5 ft., 20% Carbowax 20M column programmed from 135 to 180° C at 4°C/min. Isomeric alcohols from 3,3,5-trimethylcyclohexanone reductions were separated and analyzed on a 10 ft., 20% Carbowax 20M column at 180° C. The individual thermal conductivities of alcohol and symmetrical ether products were determined and used to obtain absolute yields. The thermal conductivities of the geometrical isomers of each alcohol were assumed to be identical ; those of the symmetrical ethers were identical within experimental error.
Triethylsilane reductions of 4-tert-butylcyclohexanone catalyzed by zinc chloride
Anhydrous zinc chloride was added to the solution of triethylsilane and 4-tert-butylcyclohexanone and the resulting mixture was stirred at 25° C or at 100° C. Reactions employing an equivalent amount of zinc chloride relative to ketone were complete within four hours at 25° C; however, as the reaction progressed a thick gel formed.
Methylene chloride was used successfully to break up the gel to the point where good mixing of reagents occured. Reactions employing 10 mol % of zinc chloride relative to ketone were slow at room temperature; after two days less than 20% reduction had occured. At 100° C reactions catalzed by zinc chloride were complete within 24 h. Reaction solutions were quenched with excess saturated sodium bicarbonate and extracted with ether. The isomeric 4-tert-butylcyclohexyl triethylsilyl ethers were separated and analyzed on a 5' 15% SE-30 column at 200° C.
 M.P. Doyle, C.C. McOsker and C.T. West, J. Org. Chem., 41 (1976) 1393.
 M.P. Doyle, D.J. DeBruyn, S.J. Donnelly, D.A. Kooistra, A.A. Odubela, C.T. West and S.M. Zonnebelt, J. Org. Chem, 39 (1974) 2740.
 M.P. Doyle and C.T. West, J. Org. Chem. 40 (1975) 3835
 R. Calas, E. Frainnet and J. Bonastre, Compt. Rend., 251 (1960) 2987.
 I.I. Lapkin, T.N. Povarnitsyna and L.A. Kostareva, Zh. Obshch. Khim., 38 (1968) 1578.
 N.E. Glushkova and N.P. Kharitonova, Izvest. Akad. Nauk SSSR Ser. Khim., (1976) 88.
 I. Ojima, J. Syn. Org. Chem. Jap., 32 (1974) 687.
 K. Yamamoto, T. Hayashi and M. Kumada, J. Organometal. Chem., 54 (1973) C45.
 W. Dumont. J. -C. Poulin, T.-P. Dang, and H.P Kagan, J. Amer. Chem. Soc., 95 (1973) 8295.
 R.J.P. Corriu and J.J.E. Moreau, J. Organometal. Chem., 64(1974) C51.
 (a) L.H. Sommer, C.L. Frye, G.A. Parker and K.W. Micheal, J. Amer. Chem. Soc., 86 (1964) 3271;
(b) L.H. Sommer, J.D. Citron and G.A. Parker, ibid., 91 (1969) 4729.
 S.C.Pace, J.C. Elkaim and J.G. Reiss J. Organometal. Chem., 56, (1973) 141
 E.J. Corey and A. Venkateswarlu, J. Amer. Chem. Soc., 94 (1972)6190
 C.T. West, S.J. Donnelly, D.A. Kooistra and M.P. Doyle, J. Org. Chem., 38 (1973) 2675.
 D.N. Kursanov, Z.N. Parnes and N.M. Loim, Synthesis, (1974) 633.
 G.A. Olah, Friedel-Crafts and related reactions, Interscience Publishers, Inc., New York, N.Y., 1963-1964.
 A.J. Vanderweilen and M.A. Ring, Inorg. Nucl. Chem. Lett., 8(1972) 421.
 L.H. Sommer, Stereochemistry, Mechanism and Silicon, McGraw-Hill, New York, N.Y., 1965 chapter 4.
 M.P. Doyle and C.T. West, J. Org. Chem., 40 (1975) 3829.
 L.H. Sommer and D.L. Bauman, J. Amer. Chem. Soc. 91 (1969) 7076.
 L.H. Sommer, W.D. Korte, and C.L. Frye, J. Amer. Chem. Soc. 94(1972) 3463.
 M.P. Doyle and C.T. West, J. Org. Chem. 40 (1975) 3821.
 D.N. Kirk and A. Mudd, J. Chem. Soc., C, (1969) 804.
 E.C. Ashby, J.R. Boone and J.P. Oliver, J. Amer. Chem. Soc. 95(1973) 5427
 (a) H.C. Brown and S. Krishnamurthy, J. Amer. Chem. Soc., 94 (1972) 7159; (b) C.A. Brown, ibid, 95 (1973) 4100.
 R.J. Berni, R.R. Benerito, W.M. Ayres and H.B. Jonassen, J. Inorg. and Nucl. Chem., 25 (1963) 807.
 M. Nardelli, G. Fava and G. Giraldi, Acta Cryst., 16 (1963) 343.
 E.L. Eliel, S.J. Schroeter, T.J. Brett, F.J. Biros and J.-C.Richer, J. Amer. Chem. Soc., 88 (1966) 3327.